clinical waste handling and obstacles in malaysia - Centro de ...
clinical waste handling and obstacles in malaysia - Centro de ...
clinical waste handling and obstacles in malaysia - Centro de ...
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
EDITORIAL TEAM<br />
Editors<br />
Celso Augusto Guimarães Santos, Fe<strong>de</strong>ral University of Paraíba, Brazil<br />
Masuo Kashiwadani, Ehime University, Japan<br />
Dragan Savic, University of Exeter, United K<strong>in</strong>gdom<br />
Vicente L. Lopes, Texas State University, United States<br />
Associate Editors<br />
Koichi Suzuki, Ehime University, Japan<br />
Hafzullah Aksoy, Istanbul Technical University, Turkey<br />
António Pais Antunes, University of Coimbra, Portugal<br />
Roberto Leal Pimentel, Fe<strong>de</strong>ral University of Paraíba, Brazil<br />
Max Billib, Hannover University, Germany<br />
Bernardo Arantes do Nascimento Teixeira, Fe<strong>de</strong>ral University of São Carlos, Brazil<br />
Generoso <strong>de</strong> Angelis Neto, State University of Mar<strong>in</strong>gá, Brazil<br />
FOCUS <strong>and</strong> SCOPE<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE) provi<strong>de</strong>s a forum for orig<strong>in</strong>al papers <strong>and</strong> for the exchange of<br />
<strong>in</strong>formation <strong>and</strong> views on significant <strong>de</strong>velopments <strong>in</strong> urban <strong>and</strong> environmental eng<strong>in</strong>eer<strong>in</strong>g worldwi<strong>de</strong>. The scope of the<br />
journal <strong>in</strong>clu<strong>de</strong>s:<br />
(a) Water Resources <strong>and</strong> Waste Management: This topic <strong>in</strong>clu<strong>de</strong>s (i) <strong>waste</strong> <strong>and</strong> sanitation; (ii) environmental<br />
issues; (iii) the hydrological cycle on the Earth; (iv) surface water, groundwater, snow <strong>and</strong> ice, <strong>in</strong> all their physical,<br />
chemical <strong>and</strong> biological processes, their <strong>in</strong>terrelationships, <strong>and</strong> their relationships to geographical factors, atmospheric<br />
processes <strong>and</strong> climate, <strong>and</strong> Earth processes <strong>in</strong>clud<strong>in</strong>g erosion <strong>and</strong> sedimentation; (v) hydrological extremes <strong>and</strong> their<br />
impacts; (vi) measurement, mathematical representation <strong>and</strong> computational aspects of hydrological processes; (vii)<br />
hydrological aspects of the use <strong>and</strong> management of water resources <strong>and</strong> their change un<strong>de</strong>r the <strong>in</strong>fluence of human<br />
activity; (viii) water resources systems, <strong>in</strong>clud<strong>in</strong>g the plann<strong>in</strong>g, eng<strong>in</strong>eer<strong>in</strong>g, management <strong>and</strong> economic aspects of<br />
applied hydrology.<br />
(b) Constructions <strong>and</strong> Environment: Build<strong>in</strong>gs <strong>and</strong> <strong>in</strong>frastructure constructions (bridges/footbridges, pipel<strong>in</strong>es etc)<br />
are part of every urban area. In recent years there is a grow<strong>in</strong>g <strong>in</strong>terest <strong>in</strong> seek<strong>in</strong>g rationality of construction systems, <strong>in</strong><br />
balance with environmental a<strong>de</strong>quacy <strong>and</strong> harmony <strong>in</strong> an urban area. This <strong>in</strong>volves, among others, a<strong>de</strong>quacy of structural<br />
systems (shapes, functionality, rational <strong>de</strong>sign etc), use of alternative materials for construction (recycled,<br />
environmentally friendly materials etc) <strong>and</strong> solutions seek<strong>in</strong>g energy efficiency.<br />
(c) Urban Design: This topic covers the arrangement, appearance <strong>and</strong> functionality of towns <strong>and</strong> cities, <strong>and</strong> <strong>in</strong><br />
particular the shap<strong>in</strong>g <strong>and</strong> uses of urban public space (e.g. streets, plazas, parks <strong>and</strong> public <strong>in</strong>frastructure), <strong>in</strong>clud<strong>in</strong>g also<br />
urban plann<strong>in</strong>g, l<strong>and</strong>scape architecture, or architecture issues (e.g. thermic <strong>and</strong> acoustic comfort).<br />
(d) Transportation Eng<strong>in</strong>eer<strong>in</strong>g: This topic covers such area as Traffic & Transport Management, Rail Transport, Air<br />
Transport, International Transport, Logistics/Physical Distribution/Supply Cha<strong>in</strong> Management, Management Information<br />
Systems & Computer Applications, Motor Transport, Regulation/Law, Transport Policy, <strong>and</strong> Water Transport.
SUMMARY<br />
CLINICAL WASTE HANDLING AND OBSTACLES IN MALAYSIA……………………………………….. 47-54<br />
Shaidatul Shida Razali<br />
Multi-criteria evaluation of the expansion of natural gas distribution network by the urban<br />
dynamics………………………………………..………………………………………..……………………………… 55-62<br />
Vanessa Meloni Massara, Miguel Edgar Morales Udaeta<br />
ANALYSIS OF NEIGHBORHOOD IMPACTS ARISING FROM IMPLEMENTATION OF<br />
SUPERMARKETS IN CITY OF SÃO CARLOS………………………………………………………………….. 63-73<br />
Pedro Silveira Gonçalves Neto, José Augusto <strong>de</strong> Lollo<br />
DETERMINING INDICATORS OF URBAN HOUSEHOLD WATER CONSUMPTION THROUGH<br />
MULTIVARIATE STATISTICAL TECHNIQUES……………………………………………………………….. 74-80<br />
Gledsneli Maria Lima L<strong>in</strong>s, Walter Santa Cruz, Zédna Mara Castro Lucena Vieira,<br />
Francisco <strong>de</strong> Assis Costa Neto, Érico Alberto Albuquerque Mir<strong>and</strong>a<br />
COMPARATIVE ANALYSIS OF EROSION MODELING TECHNIQUES IN A BASIN OF<br />
VENEZUELA………………………………………………..……………………………………………………………. 81-104<br />
Adriana Marquez Romance, Edilberto Guevara Perez
Razali <strong>and</strong> Ishak<br />
47<br />
J U E E<br />
Journal of Urban <strong>and</strong> Environmental<br />
Eng<strong>in</strong>eer<strong>in</strong>g, v.4, n.2, p.47-54<br />
ISSN 1982-3932<br />
doi: 10.4090/juee.2010.v4n2.047054<br />
Journal of Urban <strong>and</strong><br />
Environmental Eng<strong>in</strong>eer<strong>in</strong>g<br />
www.journal-uee.org<br />
CLINICAL WASTE HANDLING AND OBSTACLES<br />
IN MALAYSIA<br />
Shaidatul Shida Razali ∗ <strong>and</strong> Mohd Bakri Ishak<br />
Department of Environmental Management, Faculty of Environmental Studies, Universiti Putra Malaysia,<br />
Serdang, Selangor, Malaysia.<br />
Received 22 March 2010; received <strong>in</strong> revised form 22 April 2010; accepted 27 December 2010<br />
Abstract:<br />
Keywords:<br />
As <strong>in</strong> many other <strong>de</strong>velop<strong>in</strong>g countries, the generation of <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> <strong>in</strong> Malaysia has<br />
<strong>in</strong>creased significantly over the last few <strong>de</strong>ca<strong>de</strong>s. Even though the serious impact of the<br />
<strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> on human be<strong>in</strong>gs <strong>and</strong> the environment is significant, only m<strong>in</strong>or attention<br />
is directed to its proper <strong>h<strong>and</strong>l<strong>in</strong>g</strong> <strong>and</strong> legal aspects. This study seeks to exam<strong>in</strong>e the<br />
management of <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> <strong>in</strong> Selangor’s government hospitals as well as problems<br />
that arise from the current practice of <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> management. A <strong>de</strong>pth <strong>in</strong>terview<br />
with the responsible concession who h<strong>and</strong>les the <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> management <strong>in</strong> those<br />
hospitals also has been taken. In general, it was found that the consortium’s<br />
adm<strong>in</strong>istration was reasonably aware of the importance of <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> management.<br />
However, significant voids were presented that need to be addressed <strong>in</strong> future <strong>in</strong>clud<strong>in</strong>g<br />
efficient segregation, better <strong>h<strong>and</strong>l<strong>in</strong>g</strong> <strong>and</strong> transfer means, as well as the need for<br />
tra<strong>in</strong><strong>in</strong>g <strong>and</strong> awareness programs for the personnel. Other <strong>obstacles</strong> faced by<br />
consortiums were to h<strong>and</strong>le the <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> <strong>in</strong>clud<strong>in</strong>g the operational costs. Waste<br />
m<strong>in</strong>imiz<strong>in</strong>g <strong>and</strong> recycl<strong>in</strong>g, as well as the alternative treatment methods for <strong>in</strong>c<strong>in</strong>eration<br />
are regar<strong>de</strong>d to be major challenges <strong>in</strong> the future.<br />
<strong>cl<strong>in</strong>ical</strong> <strong>waste</strong>, management, <strong>obstacles</strong><br />
© 2010 Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE). All rights reserved.<br />
∗ Correspon<strong>de</strong>nce to: Shaidatul Shida Razali, Tel.: +60102177483.<br />
E-mail: shida_upm@yahoo.com<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.47-54, 2010
Razali <strong>and</strong> Ishak<br />
48<br />
INTRODUCTION<br />
Medic<strong>in</strong>e is one of the important sectors show<strong>in</strong>g<br />
enhancement throughout the recent <strong>de</strong>ca<strong>de</strong>s (Birp<strong>in</strong>ar et<br />
al., 2008). A hospital is midpo<strong>in</strong>t that supplies various<br />
healthcare services to the community. Its activity may<br />
<strong>in</strong>clu<strong>de</strong> curative, rehabilitation, preventive, <strong>and</strong><br />
promot<strong>in</strong>g of health education. In un<strong>de</strong>rtak<strong>in</strong>g the<br />
activities, the hospital may also generate <strong>waste</strong> (Chaerul<br />
et al., 2008). However, the fraction of <strong>waste</strong> generated<br />
at medical <strong>in</strong>stitutions has not attracted the level of<br />
attention as other types of <strong>waste</strong>s, especially <strong>in</strong><br />
<strong>de</strong>velop<strong>in</strong>g countries (Birp<strong>in</strong>ar et al., 2008). Even<br />
though concrete steps like propos<strong>in</strong>g a policy <strong>and</strong> laws<br />
to regulate <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> management, there is still<br />
problems presence due to several factors. Lack of<br />
tra<strong>in</strong><strong>in</strong>g, awareness, <strong>and</strong> f<strong>in</strong>ancial resources had been<br />
seemed as factors contribute to the mismanagement of<br />
<strong>cl<strong>in</strong>ical</strong> <strong>waste</strong>.<br />
Cl<strong>in</strong>ical <strong>waste</strong>s differ from any other <strong>waste</strong>s that<br />
be<strong>in</strong>g produced <strong>in</strong> hospitals. Sharps, human tissues or<br />
body parts <strong>and</strong> other <strong>in</strong>fectious materials conta<strong>in</strong> <strong>in</strong><br />
<strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> poses potential health <strong>and</strong> environmental<br />
risks (Bareja et al., 2000). By weight, approximately 15<br />
– 25% of <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> is consi<strong>de</strong>red <strong>in</strong>fectious (Sh<strong>in</strong>ee<br />
et al., 2008). Even though the current practices of<br />
<strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> management are different from hospital to<br />
hospital, the problematic are similar for all healthcare<br />
<strong>in</strong>stitutional from segregation, collection, packag<strong>in</strong>g,<br />
storage, transport, treatment <strong>and</strong> disposal (Tsanoka et<br />
al., 2007).<br />
Environmental pollution, unpleasant odors<br />
encourages <strong>in</strong>sects, ro<strong>de</strong>nts <strong>and</strong> worms to breed that<br />
may lead to transmission of disease like cholera,<br />
hepatitis or typhoid through <strong>in</strong>juries from contam<strong>in</strong>ated<br />
sharps (Abdulla et al., 2008).<br />
Proper manner of <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> management is<br />
greatly important to avoid health risks <strong>and</strong> damage to<br />
the flora, fauna <strong>and</strong> the environment (Yong, et al.,<br />
2009).<br />
Cl<strong>in</strong>ical Waste Management <strong>in</strong> Malaysia<br />
Up until 1980s, Malaysia has no proper system for the<br />
management of <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong>. With the emergence of<br />
HIV, M<strong>in</strong>istry of Health <strong>in</strong> collaboration with<br />
Department of Environment took an <strong>in</strong>itiative to revise<br />
policies <strong>and</strong> gui<strong>de</strong>l<strong>in</strong>es for prevention <strong>and</strong> control of<br />
<strong>in</strong>fectious disease <strong>and</strong> <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> <strong>h<strong>and</strong>l<strong>in</strong>g</strong>.<br />
It is estimated that total bed’s strength is about 47<br />
000 <strong>and</strong> 35 000 came from government’s hospital with<br />
the occupancy rate 65%. Cl<strong>in</strong>ical <strong>waste</strong> <strong>in</strong> Malaysia<br />
may be <strong>de</strong>f<strong>in</strong>ed as:<br />
i. Any <strong>waste</strong> which consist wholly or partly of human<br />
animal tissue, blood or other body fluids, excretions,<br />
drugs or other pharmaceutical products, swabs or<br />
dress<strong>in</strong>gs, syr<strong>in</strong>ges, needles or other sharp <strong>in</strong>struments,<br />
be<strong>in</strong>g <strong>waste</strong> which unless ren<strong>de</strong>red safe may cause<br />
hazardous to any person com<strong>in</strong>g <strong>in</strong>to contact with it.<br />
ii. Any other <strong>waste</strong> aris<strong>in</strong>g from medical, nurs<strong>in</strong>g,<br />
<strong>de</strong>ntal, veter<strong>in</strong>ary, pharmaceutical or similar practice,<br />
<strong>in</strong>vestigation, treatment, care, teach<strong>in</strong>g or research, or<br />
the collection of blood for transfusion, be<strong>in</strong>g <strong>waste</strong><br />
which may cause <strong>in</strong>fection to any person com<strong>in</strong>g <strong>in</strong>to<br />
contact with it.<br />
Currently, <strong>cl<strong>in</strong>ical</strong> is reported together with<br />
pharmaceutical <strong>waste</strong>. The total <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong><br />
generated is about 8000 tonnes per year <strong>and</strong> DOE<br />
(2005) estimated that the generation rate for <strong>cl<strong>in</strong>ical</strong><br />
<strong>waste</strong> varies from 0.3 to 0.8 kg per occupied bed per<br />
day.<br />
Cl<strong>in</strong>ical Waste Problems <strong>and</strong> Obstacles<br />
In Brazil, law of environment No. 7 (1982), law of the<br />
atmospheric <strong>and</strong> air protection (1992), <strong>and</strong> law of<br />
transport of hazardous materials (2005) <strong>de</strong>als with<br />
<strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> management, but problems such as<br />
<strong>waste</strong> transported to on-site storage conta<strong>in</strong>ers via<br />
uncovered trolleys, conta<strong>in</strong>ers placed near the ma<strong>in</strong><br />
street with<strong>in</strong> the hospitals build<strong>in</strong>gs or located outsi<strong>de</strong><br />
at the street curb, <strong>waste</strong> simply dumped <strong>in</strong> the corner<br />
of hospital room until it could be transported off-site,<br />
use of open trucks <strong>and</strong> lack of tra<strong>in</strong><strong>in</strong>g (Sawalem,<br />
2008).<br />
Same goes to Korea. Korea National Assembly<br />
customized the Waste Management Act <strong>in</strong> 1999 to<br />
enhanced control from the generation of medical <strong>waste</strong><br />
to its f<strong>in</strong>al <strong>de</strong>st<strong>in</strong>ation. The Korea M<strong>in</strong>istry of<br />
Environment (MOE) had given the responsibility for<br />
implement<strong>in</strong>g the act. The Korea MOE spread several<br />
regulations for <strong>de</strong>f<strong>in</strong>ition, segregation, packag<strong>in</strong>g,<br />
track<strong>in</strong>g, <strong>and</strong> disposal of medical <strong>waste</strong> but yet, the<br />
mismanagement still arise that medical <strong>waste</strong> was<br />
often mixed <strong>and</strong> <strong>waste</strong> m<strong>in</strong>imization <strong>and</strong> recycl<strong>in</strong>g are<br />
still not well-promoted (Jang et al., 2006).<br />
In 2002, Croatia endorsed the Directive on the<br />
management of <strong>waste</strong> produced dur<strong>in</strong>g healthcare<br />
(Republic of Croatia, 2000). The Directive portrays an<br />
overall system of <strong>waste</strong> management; sort<strong>in</strong>g at the<br />
po<strong>in</strong>t of generation, collection, transportation, storage<br />
<strong>and</strong> treatment but only a small number of medical<br />
<strong>in</strong>stitutions report their <strong>waste</strong> to the Registry due to the<br />
weakness of Registry function.<br />
These <strong>de</strong>prived management si<strong>de</strong>s are due to the<br />
lack of sensitivity from the management of the<br />
facilities, mean<strong>in</strong>g there seems to be some lack of<br />
awareness concern<strong>in</strong>g health risks towards the<br />
community members as well as environmental issues,<br />
<strong>and</strong> due to economic problems <strong>in</strong> the country that<br />
prevent the government from a<strong>de</strong>quately support<strong>in</strong>g a<br />
healthcare policy (Silva et al., 2005; Abdulla et al.,<br />
2008).<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.47-54, 2010
Razali <strong>and</strong> Ishak<br />
49<br />
Concern<strong>in</strong>g this, this paper aimed to:<br />
(a) To exam<strong>in</strong>e the <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> management <strong>in</strong><br />
Selangor (one of the state <strong>in</strong> Malaysia) government’s<br />
hospitals.<br />
(b) To analyze problems faced <strong>in</strong> the management<br />
means.<br />
METHODOLOGY<br />
The methodological framework can be <strong>de</strong>scribed as<br />
follows (Fig. 1):<br />
Study site<br />
The study took place <strong>in</strong> Selangor, Malaysia. Located on<br />
the west coast of pen<strong>in</strong>sular Malaysia <strong>and</strong> cover<strong>in</strong>g<br />
8000 square kilometers, Selangor is boun<strong>de</strong>d on the<br />
north by Perak, on the east by Pahang <strong>and</strong> Negeri<br />
Sembilan, <strong>and</strong> on the west by the Straits of Malacca.<br />
Selangor has been called the gateway of Malaysia. It<br />
is also the <strong>in</strong>dustrial hub of Malaysia; the country’s<br />
largest <strong>in</strong>dustrial site is located <strong>in</strong> Shah Alam, the states<br />
capital, just 25 kilometers from Kuala Lumpur. It is the<br />
most populous state <strong>in</strong> the country with a total<br />
population of 2.7 million <strong>in</strong>habitants. Despite be<strong>in</strong>g<br />
<strong>in</strong>dustry-based, the state is blessed with natural forests,<br />
waterfalls, hills, <strong>and</strong> lakes to complement its many manma<strong>de</strong><br />
attractions.<br />
There are 10 government hospitals <strong>in</strong> Selangor,<br />
which are located at Kajang, Serdang, Kuala Kubu<br />
Bharu, Ampang, Selayang, Sungai Buloh, Bant<strong>in</strong>g,<br />
Sabak Bernam, Klang <strong>and</strong> Tanjung Karang.<br />
Data Collection<br />
The first step <strong>in</strong> collect<strong>in</strong>g the data was to <strong>de</strong>term<strong>in</strong>e the<br />
panelists. The panelists were the <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong><br />
supervisor <strong>in</strong> each hospital. Panelists were then<br />
<strong>in</strong>terviewed. Interviews were helpful <strong>in</strong> obta<strong>in</strong><strong>in</strong>g<br />
<strong>in</strong>formation about common practices <strong>in</strong> the management<br />
of the <strong>waste</strong> (Bdour, et al., 2007). Questions were asked<br />
basically about the management of <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> <strong>in</strong><br />
terms of collection, transportation, treatment, disposal<br />
<strong>and</strong> tra<strong>in</strong><strong>in</strong>g. The <strong>in</strong>formants were also asked about the<br />
<strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> management dur<strong>in</strong>g H1N1 acute <strong>and</strong> the<br />
challenges <strong>in</strong>volved as well as the problems faced. The<br />
number of <strong>waste</strong>s generated was also asked.<br />
A form for the panelists to provi<strong>de</strong> the <strong>in</strong>formation<br />
regard<strong>in</strong>g <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> generation <strong>and</strong> tra<strong>in</strong><strong>in</strong>g was<br />
also distributed.<br />
Data Analysis<br />
The recor<strong>de</strong>d <strong>in</strong>terviews were then transferred <strong>in</strong>to<br />
transcripts. Then, the transcripts were sent back to the<br />
<strong>in</strong>terviewees for the process of validation. In this<br />
process, the panelists may make any correction if they<br />
wish to.<br />
RESULTS AND DISCUSSION<br />
Background <strong>in</strong>formation <strong>and</strong> <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong><br />
generation<br />
The quantity of <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> <strong>de</strong>pends upon several<br />
factors such as the size of hospital, the segregation<br />
program of <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> <strong>and</strong> the medical activities.<br />
Table 1 <strong>and</strong> Fig. 2 presents <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> generated<br />
from all general hospitals <strong>in</strong> Selangor throughout the<br />
year 2009. There are 10 government hospitals <strong>in</strong><br />
Selangor <strong>and</strong> the number of beds ranges from 93 to 864<br />
with a mean of 335.9. However, <strong>in</strong> this study, we will<br />
look upon to just n<strong>in</strong>e hospitals because one of the<br />
hospitals refused to give cooperation. Consi<strong>de</strong>r<strong>in</strong>g the<br />
occupancy bed is 100%, the average generation rates of<br />
total <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> <strong>in</strong> Selangor’s government hospitals<br />
were estimated to be 1.355 kg/bed/day.<br />
Fig. 1 Research framework.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.47-54, 2010
50<br />
Razali <strong>and</strong> Ishak<br />
Table 1 Generated <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> for Selangor’s government hospital for 2009 (<strong>in</strong> kg)<br />
A B C D E F G H I<br />
Jan 13 022.6 10 935.3 2 184.2 24 749.37 16 978.1 23 304.2 2 672.0 2 253.3 30 424.0<br />
Feb 12 866.5 10 932.7 1 904.0 23 506.5 17 202.1 22 451.3 2 362.5 1 985.8 29 549.9<br />
March 14 683.3 11 871.6 2 001.7 26 101.3 18 539.6 21 095.4 2 813.4 2 210.8 33 020.4<br />
Apr 1 4707 11 898.1 2 126.8 25 939.3 20 259.4 23 611.2 2 950.3 2 290.7 34 368.2<br />
May 14 878.3 11 553.5 1 688.8 25 233.0 20 482.9 25 452.5 2 965.2 2 535.1 34 787.3<br />
June 14 298.3 10 366.3 2 077.1 24 002.3 19 600.1 36 896.3 2 739.1 2 415.4 32 917.0<br />
July 17 397.3 11 490.7 2 286.9 25 363.7 20 958.1 27 572.3 3 154.9 2 495.4 34 164.5<br />
Aug 1 6906.0 12 848.1 2 465.0 27 630.9 20 361.5 26 459.0 3 073.1 2 597.9 36 435.4<br />
Sept 14 587.5 11 498.8 1 676.7 23 141.3 18 306.3 23 121.2 2 840.5 2 519.0 31 556.9<br />
Oct 17 701.8 13 953.3 2 383.9 26 959.6 21 693.7 25 528.7 2 999.1 2 685.4 36 815.7<br />
Nov 16 498.9 12 753.2 1 981.0 25 414.1 19 948.7 23 678.4 2 853.2 2 595.9 33 005.9<br />
Dec 17 381.4 13 698.6 2 010.9 26 757.5 20 901.1 24 702.4 3 013.1 2 666.1 33 709.1<br />
Table 2. Specialties <strong>in</strong> each hospital<br />
Services<br />
Anaesthesiology<br />
Dermatology<br />
Herpetology<br />
Nephrology<br />
Paediatric<br />
Psychiatry<br />
Haematology<br />
Obstetric <strong>and</strong> Genecology<br />
Orthopaedic <strong>and</strong> Traumatology<br />
Ophthalmology<br />
Teeth & Mouth Operation<br />
Radiology<br />
Haemodialysis<br />
Otorh<strong>in</strong>olaryngology (ENT)<br />
Urology<br />
Cardiothoracic<br />
Ear, Mouth & Throat<br />
Plastic Surgery<br />
Paediatric <strong>and</strong> Neonatology<br />
Rehabilitation<br />
Neurosurgery<br />
■ <strong>in</strong>dicates the service available at the hospital.<br />
Hospital<br />
A B C D E F G H I<br />
Fig. 2 Reported generation of <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> <strong>in</strong> Selangor’s<br />
government hospital.<br />
Fig. 3 Generated <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> presented <strong>in</strong> months.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.47-54, 2010
Razali <strong>and</strong> Ishak<br />
51<br />
181<br />
306<br />
150<br />
778<br />
382<br />
340<br />
114 93<br />
A B C D E F G H I<br />
Hospital's name<br />
Fig. 4 Number of beds <strong>in</strong> Selangor’s government hospital.<br />
Figure 4 illustrates the number of beds <strong>in</strong> Selangor’s<br />
government hospitals.<br />
Almost all hospitals offer same services; general<br />
medical, general surgery, emergency, pathology <strong>and</strong><br />
maternity. Hospital F offers a lot of specialties that are<br />
not available <strong>in</strong> other hospitals. Table 3 presents the<br />
specialties offered <strong>in</strong> all hospitals.<br />
From the table, the number of <strong>waste</strong> generated is<br />
<strong>de</strong>pend<strong>in</strong>g on the number of beds <strong>and</strong> <strong>de</strong>partments <strong>in</strong><br />
the hospitals. Hospital J, which has the highest number<br />
of beds (864 beds), produced the highest amount of<br />
<strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> (406 753.7 kg). Also, the location of the<br />
hospital <strong>in</strong>fluenced the number of <strong>waste</strong> produced. For<br />
example, Hospital D, H <strong>and</strong> I produced <strong>waste</strong>s less than<br />
50 000 kg because they are located at the rural area,<br />
compare to other hospitals that are located <strong>in</strong> urban area.<br />
There is no specific reason to why the number of<br />
<strong>waste</strong>s <strong>de</strong>creased <strong>in</strong> September but H1N1 outbreak<br />
could be the possible explanation of why the number of<br />
<strong>waste</strong>s <strong>in</strong>creased <strong>and</strong> became the month with the highest<br />
production of <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong>s throughout the whole year.<br />
Accord<strong>in</strong>g to the <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> Supervisors, there are<br />
a few units that had been i<strong>de</strong>ntified to generate the most<br />
<strong>waste</strong>s.<br />
Abu: Usually GOT (General Operation Theatre),<br />
maternity, <strong>and</strong> ICU (Intensive Care Unit). But the rest is<br />
normal; however, it is hard to say that. It <strong>de</strong>pends on the<br />
patients. Sometimes, there is no patient enter<strong>in</strong>g ICU,<br />
which means there is no <strong>waste</strong>.<br />
Ahmad: Ward 2 (<strong>de</strong>ngue cases), labor room, GOT,<br />
emergency.<br />
Based on the <strong>in</strong>terview, it can be conclu<strong>de</strong>d that high<br />
production of <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong>s comes from Operation<br />
Theatre, maternity, Intensive Care Unit, Emergency <strong>and</strong><br />
<strong>de</strong>ngue cases ward.<br />
Cl<strong>in</strong>ical <strong>waste</strong> collection, storage, transportation <strong>and</strong><br />
<strong>in</strong>c<strong>in</strong>eration<br />
The collection usually starts at 08:30 am until 11:30 am.<br />
Porter with well-equipped Personal Protection<br />
864<br />
Equipment (PPE) then pushes the yellow cart <strong>and</strong> start<br />
do<strong>in</strong>g collection from unit to unit. The PPE that are<br />
be<strong>in</strong>g used <strong>in</strong>clu<strong>de</strong> apron, glove <strong>and</strong> boot. The fullyfilled<br />
yellow bags with the <strong>waste</strong>s will then be<br />
transferred to the cart <strong>and</strong> the process will be repeated<br />
until all the <strong>waste</strong> has been collected. Then, they will<br />
send the <strong>waste</strong> to the cold storage. Here, they will wait<br />
for the liaison officer or the hospital verifier to witness<br />
the weigh<strong>in</strong>g process. The weigh<strong>in</strong>g process is usually<br />
conducted from 11:30 am to 12:30 pm. While do<strong>in</strong>g the<br />
collection, there are some porters that are be<strong>in</strong>g<br />
assigned to clean the b<strong>in</strong>. They will change the used b<strong>in</strong><br />
with a new b<strong>in</strong>. To clean the b<strong>in</strong>, 100 ml of a chemical<br />
named SteriQuat were mixed with 250 L of water.<br />
Then, the b<strong>in</strong> will be dried. The b<strong>in</strong> must be washed<br />
daily.<br />
While for the disposable sharps <strong>and</strong> needles, a<br />
specific b<strong>in</strong> that is called sharp b<strong>in</strong> was used. After it is<br />
¾ filled, porter will close the cap <strong>and</strong> it will be collected<br />
<strong>and</strong> transferred to the cart. Unlike other yellow b<strong>in</strong>s, this<br />
sharp b<strong>in</strong> will not go<strong>in</strong>g to be washed. It will be<br />
<strong>in</strong>c<strong>in</strong>erated <strong>and</strong> the <strong>waste</strong>s <strong>in</strong>si<strong>de</strong> it are not go<strong>in</strong>g to be<br />
transferred to other place.<br />
For a small hospital which produces less <strong>cl<strong>in</strong>ical</strong><br />
<strong>waste</strong>, the <strong>waste</strong> will be store <strong>in</strong> a refrigerator with the<br />
temperature range of -1 to -5°C while for the hospital<br />
which produce large amount of <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong>, there is<br />
no refrigerator provi<strong>de</strong>d. This is because for the hospital<br />
which produce small amount of <strong>waste</strong>, the lorry will<br />
collect the <strong>waste</strong> three times per week while for the<br />
hospital which produce large amount of <strong>waste</strong>, the lorry<br />
will come daily to collect the <strong>waste</strong>. The lorry will send<br />
the <strong>waste</strong> to the <strong>in</strong>c<strong>in</strong>erator plant <strong>in</strong> Teluk Panglima<br />
Garang. The time for the lorry to come <strong>and</strong> collect the<br />
<strong>waste</strong> is from 12:00 noon to 05:30 pm.<br />
Dur<strong>in</strong>g the H1N1 outbreak, the management of<br />
<strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> <strong>in</strong> the H1N1 ward is slightly different.<br />
The hospital briefed to the porter who was assigned to<br />
collect the <strong>waste</strong>. Only one person can be assigned. At<br />
one time, he then will be given a H1N1 immunization.<br />
The <strong>waste</strong>s that have a H1N1 contact will be placed <strong>in</strong> a<br />
yellow bag <strong>and</strong> be<strong>in</strong>g stored <strong>in</strong> a labeled yellow cart.<br />
When the lorry comes, this yellow cart has the priority<br />
to be collected <strong>and</strong> <strong>in</strong>c<strong>in</strong>erated when they reach the<br />
<strong>in</strong>c<strong>in</strong>eration plant at Teluk Panglima Garang.<br />
A flowchart of the <strong>in</strong>c<strong>in</strong>eration system is be<strong>in</strong>g<br />
<strong>de</strong>scribed <strong>in</strong> Fig. 4. The system consists of a <strong>waste</strong><br />
fee<strong>de</strong>r, a primary combustion chamber (rotary type), a<br />
secondary combustion chamber, a <strong>waste</strong> heat boiler, dry<br />
air pollution control, <strong>and</strong> a 24-h gas-emission<br />
monitor<strong>in</strong>g <strong>de</strong>vice. The melt<strong>in</strong>g of the ash occurs <strong>in</strong> the<br />
secondary combustion chamber at 1 200°C (Azni et al.,<br />
2005).<br />
With the capacity of 500 kg/h, this <strong>in</strong>c<strong>in</strong>erator<br />
operates 24 hours. Table 3 summarizes the specification<br />
of the <strong>in</strong>c<strong>in</strong>erator.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.47-54, 2010
Razali <strong>and</strong> Ishak<br />
52<br />
Fig. 5 Flow diagram of <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> <strong>in</strong>c<strong>in</strong>erator.<br />
Table 3 Specifications of the <strong>in</strong>c<strong>in</strong>erator<br />
Operation time<br />
24h<br />
Type of fuel<br />
Diesel<br />
Type of feed<br />
Hospital <strong>waste</strong><br />
Treatment capacity<br />
500 kg/h<br />
Amount of flue gas<br />
20 000 m³/h<br />
Amount of slag produced 10%<br />
Temperature <strong>in</strong> Primary Chamber<br />
950° C<br />
Temperature <strong>in</strong> Secondary Chamber<br />
1 200° C<br />
Source: Azni et al. (2005).<br />
The <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> management <strong>in</strong> Selangor’s<br />
government hospital can be simplified as follows:<br />
Tra<strong>in</strong><strong>in</strong>g <strong>and</strong> Awareness<br />
The concession company is the one who is responsible<br />
to conduct the tra<strong>in</strong><strong>in</strong>gs. There are two groups of people<br />
who are <strong>in</strong>volved <strong>in</strong> the tra<strong>in</strong><strong>in</strong>gs. They are the porter<br />
<strong>and</strong> the user (hospital staffs). The porter will go for<br />
tra<strong>in</strong><strong>in</strong>gs at least three times a year. The tra<strong>in</strong><strong>in</strong>g<br />
<strong>in</strong>volves <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> collection procedure, procedure<br />
on wear<strong>in</strong>g Personal Protection Equipment (PPE),<br />
<strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> service management, procedure on<br />
wash<strong>in</strong>g the bag hol<strong>de</strong>r, procedure on fill<strong>in</strong>g the record<br />
sheet, procedure on fill<strong>in</strong>g bag hol<strong>de</strong>r wash<strong>in</strong>g form,<br />
customer relationship program, <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong><br />
segregation, procedure on h<strong>and</strong>le the acci<strong>de</strong>nt <strong>and</strong><br />
health. The tra<strong>in</strong><strong>in</strong>g will be given throughout the year or<br />
when acci<strong>de</strong>nt occurs.<br />
For the user, the tra<strong>in</strong><strong>in</strong>g will be given three times a<br />
year or upon request. Usually, the hospital will ask the<br />
concession company to conduct tra<strong>in</strong><strong>in</strong>g for user when<br />
there is a new staff. The tra<strong>in</strong><strong>in</strong>gs <strong>in</strong>volve i<strong>de</strong>ntify<strong>in</strong>g<br />
the types of <strong>waste</strong>, <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> segregation, <strong>cl<strong>in</strong>ical</strong><br />
<strong>waste</strong> <strong>h<strong>and</strong>l<strong>in</strong>g</strong>, safety <strong>and</strong> health.<br />
Unfortunately, there is no specific tra<strong>in</strong><strong>in</strong>g for the<br />
public. The only way to educate the public, especially<br />
on the segregation of <strong>waste</strong> is through posters. The<br />
poster will be p<strong>in</strong>ned-up on top of the b<strong>in</strong> to give<br />
awareness to the public regard<strong>in</strong>g the segregation.<br />
Problems <strong>and</strong> Obstacles<br />
Fig. 6 Cl<strong>in</strong>ical Waste Pathway <strong>in</strong> Selangor’s Government Hospital.<br />
Although there is a management system <strong>in</strong> <strong>cl<strong>in</strong>ical</strong><br />
<strong>waste</strong> <strong>h<strong>and</strong>l<strong>in</strong>g</strong> <strong>in</strong> Malaysia, there are several problems<br />
<strong>and</strong> <strong>obstacles</strong> that had been found throughout the study.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.47-54, 2010
Razali <strong>and</strong> Ishak<br />
53<br />
Ahmad: Usually, doctors always throw needles<br />
everywhere, sometimes <strong>in</strong> the yellow bag. That will<br />
cause problems.<br />
Ali: The problems arise while we are do<strong>in</strong>g the<br />
collection, the people from the district cl<strong>in</strong>ics come, <strong>and</strong><br />
the problem is that they like to store the <strong>waste</strong> outsi<strong>de</strong><br />
the fence but they are not allowed to do so.<br />
Abu: If every user is be<strong>in</strong>g given a room to place the<br />
b<strong>in</strong>s <strong>and</strong> do the clean<strong>in</strong>g process, it would be easier,<br />
avoid<strong>in</strong>g transportation <strong>in</strong> the elevator <strong>and</strong> contact with<br />
public, which would cause trouble to publics.<br />
Abu: The price should be <strong>in</strong>creased because the<br />
groceries’ price is <strong>in</strong>creas<strong>in</strong>g, plastic <strong>and</strong> all the b<strong>in</strong>s.<br />
The price is not the same as 10 or 15 years ago. Look at<br />
the current economy; I th<strong>in</strong>k we should <strong>in</strong>crease the<br />
price.<br />
Abu: There is no dry<strong>in</strong>g mach<strong>in</strong>e.<br />
A<strong>in</strong>i: Porters have no <strong>in</strong>fectious control towards patients<br />
or hospital. They open the door <strong>and</strong> collect the <strong>waste</strong> by<br />
us<strong>in</strong>g the same h<strong>and</strong>.<br />
A<strong>in</strong>i: The problem we face now is that even we had<br />
given the tra<strong>in</strong><strong>in</strong>g, not all patients know about it. In the<br />
yellow b<strong>in</strong>, they are not allowed to throw the <strong>cl<strong>in</strong>ical</strong><br />
<strong>waste</strong>. We cannot open the bag back because it is<br />
already mixed. There is no awareness from the public.<br />
Adila: If we do the collection <strong>in</strong> the ward, we will carry<br />
the plastic bag to the outsi<strong>de</strong>. The large wheel b<strong>in</strong><br />
cannot be pull <strong>in</strong>si<strong>de</strong> the ward. That is the weakness<br />
because we have to carry the yellow bag to the outsi<strong>de</strong>.<br />
It would spill on the floor.<br />
Mala: for the <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong>, there are too much<br />
procedures <strong>and</strong> too much contact. There are too much<br />
th<strong>in</strong>gs that we need to follow.<br />
From the <strong>in</strong>terview, there are several problems <strong>in</strong><br />
manag<strong>in</strong>g <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong>. The user’s attitu<strong>de</strong> has been<br />
i<strong>de</strong>ntified as one of it. Even though tra<strong>in</strong><strong>in</strong>g had been<br />
given, their self awareness still consi<strong>de</strong>red as low. The<br />
second problem is regard<strong>in</strong>g the <strong>waste</strong> from district<br />
health centre. For district health centre, there is no<br />
specific concession company to manage <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong><br />
as government hospitals does. They need to send their<br />
<strong>waste</strong> to the nearest government hospital by their own.<br />
The problem arises when they tend to put the <strong>waste</strong><br />
outsi<strong>de</strong> the fence of the <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> storage. This is<br />
dangerous because it will expose the <strong>waste</strong> to the public<br />
as well as animals such as dogs <strong>and</strong> cats. The vaporized<br />
<strong>waste</strong> might spread the <strong>in</strong>fectious disease through the<br />
air. Space had been i<strong>de</strong>ntified as one of the problems <strong>in</strong><br />
manag<strong>in</strong>g <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong>.<br />
There is no enough space provi<strong>de</strong>d for the hospital<br />
support services <strong>in</strong> the hospital. For <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong>, they<br />
need a large space especially to wash <strong>and</strong> dry the b<strong>in</strong>.<br />
Because of this limitation, some b<strong>in</strong>s are not be<strong>in</strong>g<br />
washed properly or not be<strong>in</strong>g washed at all. Also, us<strong>in</strong>g<br />
a traditional approach to dry the b<strong>in</strong>s proved to be a<br />
major problem too. A dry<strong>in</strong>g mach<strong>in</strong>e for the b<strong>in</strong>s is<br />
nee<strong>de</strong>d especially for the hospitals which produces high<br />
amount of <strong>waste</strong>s.<br />
When the agreement was first signed, both parties<br />
agreed to update the service charge from time to time.<br />
Unfortunately, s<strong>in</strong>ce it was signed <strong>in</strong> 1992, there is no<br />
revision ma<strong>de</strong> regard<strong>in</strong>g the price. The current price is<br />
RM5.20 per kilograms. The price should be revised due<br />
to the <strong>in</strong>creas<strong>in</strong>g price of raw materials <strong>and</strong><br />
transportation costs.<br />
The problem also occurs from the <strong>h<strong>and</strong>l<strong>in</strong>g</strong> of the<br />
<strong>cl<strong>in</strong>ical</strong> <strong>waste</strong>. By us<strong>in</strong>g the same h<strong>and</strong> to collect the<br />
<strong>waste</strong> <strong>and</strong> to open the door, the probability for the<br />
<strong>in</strong>fectious disease to spread becomes higher. When the<br />
porter carries the bags to the <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> cart outsi<strong>de</strong><br />
the ward, there is a probability of the <strong>in</strong>fection<br />
spread<strong>in</strong>g through the spillage on the floor occurs.<br />
Public awareness has been i<strong>de</strong>ntified as one of the<br />
problem. Because of the low public awareness on the<br />
management or <strong>h<strong>and</strong>l<strong>in</strong>g</strong> of <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong>, they tend to<br />
mix up the <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> <strong>and</strong> the general <strong>waste</strong>. This<br />
will <strong>in</strong>crease the disposal cost because the general <strong>waste</strong><br />
will be treated as <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> once it is contacted with<br />
the <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong>.<br />
There are several agencies that responsible to the<br />
<strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> management. Due to this, a lot of<br />
requirements <strong>and</strong> procedures need to be followed. This<br />
seems as the problems to the concession company <strong>in</strong><br />
<strong>de</strong>al<strong>in</strong>g with <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> management.<br />
CONCLUSION<br />
This study helped <strong>in</strong> establish<strong>in</strong>g database, <strong>in</strong>formation<br />
<strong>and</strong> statistics on the <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> sources, generation,<br />
collection, transportation, treatment <strong>and</strong> disposal. It also<br />
highlighted the problems <strong>and</strong> <strong>obstacles</strong> faced dur<strong>in</strong>g its<br />
<strong>h<strong>and</strong>l<strong>in</strong>g</strong>. Also, it has provi<strong>de</strong>d suggestions for policy<br />
makers <strong>and</strong> further <strong>in</strong>formation to facilitate policy<br />
<strong>de</strong>velopment <strong>and</strong> improve <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> management.<br />
The management of <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> has been of major<br />
concern due to its potential high risks to human health<br />
<strong>and</strong> the environment. The current practices for the<br />
<strong>h<strong>and</strong>l<strong>in</strong>g</strong>, transportation, storage <strong>and</strong> disposal of <strong>waste</strong>s<br />
generated at the hospitals needs to be changed <strong>and</strong><br />
improved. Generally, the problems <strong>and</strong> <strong>obstacles</strong> arise<br />
from the <strong>h<strong>and</strong>l<strong>in</strong>g</strong> of <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> are non-segregated<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.47-54, 2010
Razali <strong>and</strong> Ishak<br />
54<br />
<strong>waste</strong>, awareness <strong>and</strong> attitu<strong>de</strong>s among hospital’s staffs<br />
<strong>and</strong> patients, collection of <strong>waste</strong> from district healthcare<br />
centre, facilities <strong>and</strong> spaces provi<strong>de</strong>d, service charge<br />
<strong>and</strong> documentations.<br />
Hospital’s <strong>de</strong>sign is very important. It is the first step<br />
to allocate spaces for the <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> management;<br />
either from collection, storage <strong>and</strong> disposal. Abu had<br />
stated before, “If every user is be<strong>in</strong>g given a room to<br />
place the b<strong>in</strong>s <strong>and</strong> do the clean<strong>in</strong>g process, it would be<br />
easier. No elevator’s problem, contact with public <strong>and</strong><br />
caus<strong>in</strong>g trouble to publics”. It seems that tra<strong>in</strong><strong>in</strong>g <strong>and</strong><br />
supervision is very important to educate not only the<br />
porter, hospital’s staff but also all the patients to make<br />
sure there’s no misuse of b<strong>in</strong>s that had be<strong>in</strong>g provi<strong>de</strong>d.<br />
Other than that, improvements <strong>and</strong> research <strong>and</strong><br />
<strong>de</strong>velopment activity (R&D) should be carried out<br />
cont<strong>in</strong>uously to combat the problem that might not<br />
occur before but tend to be hazard such as what has<br />
be<strong>in</strong>g highlighted by Ani <strong>and</strong> Adila about <strong>in</strong>fectious<br />
control. The policy also should be revised to make it<br />
more un<strong>de</strong>rst<strong>and</strong>able <strong>and</strong> will not cause the difficulty as<br />
what Mala had told, “For the <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong>, there are<br />
too much procedures <strong>and</strong> too much contact. There are<br />
too much th<strong>in</strong>gs that we need to follow.”<br />
Improvements of <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> management require<br />
all parties to <strong>in</strong>volve. It also requires strategic <strong>and</strong><br />
systematic plann<strong>in</strong>g that aim the goal to controll<strong>in</strong>g<br />
costs, educate the users <strong>and</strong> publics, un<strong>de</strong>rst<strong>and</strong>able<br />
policy <strong>and</strong> to manage the <strong>cl<strong>in</strong>ical</strong> <strong>waste</strong> <strong>in</strong> proper<br />
manner as it can reduce the hazards <strong>and</strong> risks to the<br />
ecosystem <strong>and</strong> the community.<br />
REFERENCES<br />
Abdulla, F., Qdais, H.A. & Rabi, A. (2008) Site <strong>in</strong>vestigation on<br />
medical <strong>waste</strong> management practices <strong>in</strong> Northern Jordan. Waste<br />
Manag. 28(2), 450–458. doi: 10.1016/j.wasman.2007.02.035.<br />
Azni, I., Katayon, S., Ratnasamy, M. & Johari, M.M.N.M (2005)<br />
Stabilization <strong>and</strong> utilization of hospital <strong>waste</strong> as road <strong>and</strong> asphalt<br />
aggregate. J. Mater. Cycles Waste Manag. 7(1), 33–37. doi:<br />
10.1007/s10163-004-0123-0.<br />
Baveja, G., Muralidhar, S. & Aggarwal, P. (2000) Medical <strong>waste</strong><br />
management – an overview. Hospital Today 5(9), 485–486.<br />
Birp<strong>in</strong>ar, M. E., Bilgili, M. S. & Erdorgan, T. (2008). Medical <strong>waste</strong><br />
management <strong>in</strong> Turkey: a case study of Istanbul. Waste Manag.<br />
29(1), 445–448. doi: 10.1016/j.wasman.2008.03.015.<br />
Chaerul, M., Tanaka, M. & Shekdar, A.V. (2008) A Syatem dynamic<br />
approach for hospital <strong>waste</strong> management. Waste Manag. 28(2),<br />
442–449. doi: 10.1016/j.wasman.2007.01.007.<br />
Jang, Y.C., Lee, C., Yoon, O.S. & Kim, H. (2006) Medical <strong>waste</strong><br />
management <strong>in</strong> Korea. J. Environ. Manag. 80(2), 107–115. doi:<br />
10.1016/j.jenvman.2005.08.018.<br />
Sawalem, M., Selic, E. & Herbell, J.-D. (2008). Hospital <strong>waste</strong><br />
management <strong>in</strong> Libya: a case study. Waste Manag. 29(4), 1370–<br />
1375. doi: 10.1016/j.wasman.2008.08.028.<br />
Sh<strong>in</strong>ee, E., Gombajav, E., Nishimura, A., Hamajima, N. & Ito, K.<br />
(2008) Healthcare <strong>waste</strong> management <strong>in</strong> the capital city of<br />
Mangolia. Waste Manag. 28(2), 435–444. doi:<br />
10.1016/j.wasman.2006.12.022.<br />
Silva, C.E.D., Hoppe, A.E., Ravanello, M.M. & Mello, N. (2005)<br />
Medical Wastes Management <strong>in</strong> the South of Brazil. Waste<br />
Manag. 25(6), 600–605. doi: 10.1016/j.wasman.2004.03.002.<br />
Tsanoka, M., Anagnostopoulou, E. & Gidarakos, E. (2007) Hospital<br />
<strong>waste</strong> management <strong>and</strong> toxicity evaluation: a case study. Waste<br />
Manag. 27(7), 912–920. doi: 10.1016/j.wasman.2006.04.019.<br />
Yong, Z., Gang, X., Guanx<strong>in</strong>g, W., Tao, Z. & Dawei, J (2009).<br />
Medical <strong>waste</strong> management <strong>in</strong> Ch<strong>in</strong>a (2009). Waste Manag.<br />
29(4), 1376–1382. doi: 10.1016/j.wasman.2008.10.023.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.47-54, 2010
Massara <strong>and</strong> Udaeta<br />
55<br />
J U E E<br />
Journal of Urban <strong>and</strong> Environmental<br />
Eng<strong>in</strong>eer<strong>in</strong>g, v.4 , n.2, p.55-62<br />
ISSN 1982-3932<br />
doi: 10.4090/juee.2010.v4n2.055062<br />
Journal of Urban <strong>and</strong><br />
Environmental Eng<strong>in</strong>eer<strong>in</strong>g<br />
www.journal-uee.org<br />
MULTI-CRITERIA EVALUATION OF THE EXPANSION OF<br />
NATURAL GAS DISTRIBUTION NETWORK BY THE URBAN<br />
DYNAMICS<br />
Vanessa M. Massara 1∗ , Miguel E. M. Udaeta 1<br />
1 Department of Petroleum <strong>and</strong> Natural Gas, Energy <strong>and</strong> Electrotechnics Institute, University of São Paulo, Brazil<br />
Received 28 June 2009; received <strong>in</strong> revised form 30 September 2009; accepted 02 January 2010<br />
Abstract:<br />
Keywords:<br />
The objective of this work is to analyze the expansion of the <strong>in</strong>frastructure of natural<br />
gas distribution, i<strong>de</strong>ntify<strong>in</strong>g priorities from large metropolis us<strong>in</strong>g the energy plann<strong>in</strong>g<br />
based on urban <strong>de</strong>sign tools like urban dynamics <strong>and</strong> techniques like AHP (analytic<br />
hierarchy process). The methodology proposed uses matrices consi<strong>de</strong>r<strong>in</strong>g the relations<br />
between the concept of urban dynamics, quality of life <strong>and</strong> the possibilities of natural<br />
gas displac<strong>in</strong>g other energy forms. The matrices are ma<strong>de</strong> up of <strong>in</strong>formation about<br />
social <strong>and</strong> urban <strong>de</strong>velopment, costs of establish<strong>in</strong>g the <strong>in</strong>frastructure <strong>and</strong> projections of<br />
the consumption potential <strong>in</strong> various sectors. Relat<strong>in</strong>g the consumption to urban<br />
<strong>de</strong>velopment parameters <strong>and</strong> the real estate future of the areas <strong>in</strong> study, the<br />
methodology allows <strong>in</strong>dicat<strong>in</strong>g for each district, the viability of implement<strong>in</strong>g a gas<br />
network. As conclusion, the mo<strong>de</strong>l presents the <strong>in</strong>tegration between the cities profile<br />
<strong>and</strong> the natural gas use, by means of a growth natural gas on districts of São Paulo City<br />
as a specific case study.<br />
Energy, natural gas, <strong>in</strong>frastructure, urban <strong>de</strong>velopment, analytic hierarchy process<br />
© 2010 Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE). All rights reserved.<br />
∗ Correspon<strong>de</strong>nce to: Vanessa Meloni Massara, Tel.: +55 11 9860-8116.<br />
E-mail: vanessa.massara@gmail.com<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.55-62, 2010
Massara <strong>and</strong> Udaeta<br />
56<br />
INTRODUCTION<br />
Natural gas (NG) is currently the third largest source of<br />
primary energy <strong>in</strong> the world, surpassed only by oil <strong>and</strong><br />
coal (EIA, 2008).<br />
In the ANP (Brazilian National Agency of Oil,<br />
Natural Gas <strong>and</strong> Biofuels) surveys for the next years,<br />
new natural gas fields will be <strong>in</strong>com<strong>in</strong>g <strong>in</strong> Brazil. For<br />
<strong>in</strong>stance, the Manati field will duplicate <strong>de</strong> actual gas<br />
production <strong>in</strong> state of Bahia solv<strong>in</strong>g <strong>in</strong> the short-term<br />
the problem of natural gas unsatisfied <strong>de</strong>m<strong>and</strong>. In the<br />
Espírito Santo state, the Peroá-Cangoa field emerges as<br />
an opportunity of complement<strong>in</strong>g the gas for the<br />
Brazilian northeastern <strong>and</strong> southeastern regions <strong>and</strong> to<br />
reliability the Campos Bas<strong>in</strong> production <strong>in</strong> the state of<br />
Rio <strong>de</strong> Janeiro. But the Mexilhão field <strong>in</strong> the São Paulo<br />
state, Santos Bas<strong>in</strong>, as a prelu<strong>de</strong> of the context of the<br />
pre-salt reservoirs, will be anticipat<strong>in</strong>g the gas<br />
production as <strong>de</strong>f<strong>in</strong>ition of the fe<strong>de</strong>ral Brazilian<br />
adm<strong>in</strong>istration (BNDES, 2006). Also <strong>in</strong> January of<br />
2008, <strong>in</strong> the pre-salt hydrocarbons offshore Brazilian<br />
reservoirs of Santos Bas<strong>in</strong> (state of São Paulo) another<br />
well was discovered. The reserve that was entitled<br />
Jupiter is a large natural <strong>de</strong>posit for natural gas <strong>and</strong><br />
light-oil, <strong>in</strong> <strong>de</strong>ep waters estimated <strong>in</strong> very large gas<br />
mega-field that, <strong>de</strong>spite <strong>de</strong>m<strong>and</strong><strong>in</strong>g large <strong>in</strong>vestments,<br />
will make Brazil double its reserves (GASBRASIL,<br />
2008). With this favorable forecast of natural gas supply<br />
<strong>and</strong> consi<strong>de</strong>r<strong>in</strong>g its various uses, this work focuses on<br />
market gas <strong>de</strong>m<strong>and</strong> <strong>in</strong> sectors like resi<strong>de</strong>ntial,<br />
commercial, services <strong>and</strong> <strong>in</strong>dustrial.<br />
In this work, an analytical methodology that<br />
<strong>in</strong>tegrates the un<strong>de</strong>rst<strong>and</strong><strong>in</strong>g of the urban dynamics to<br />
the strategies of expansion <strong>in</strong> the natural gas distribution<br />
network is consi<strong>de</strong>red, characteriz<strong>in</strong>g gas consumption<br />
possibilities <strong>and</strong> attractiveness for the districts,<br />
compos<strong>in</strong>g a city.<br />
The methodology is <strong>de</strong>veloped by gather<strong>in</strong>g<br />
<strong>in</strong>formation such as family <strong>in</strong>come, <strong>de</strong>mographic<br />
<strong>de</strong>nsity <strong>and</strong> construction area, percentage of l<strong>and</strong> use,<br />
number of households as well as commercial, service<br />
<strong>and</strong> <strong>in</strong>dustrial establishments, number of real estate as<br />
well as <strong>in</strong>dicative <strong>in</strong>formation released by the Urban<br />
Plan of the city regard<strong>in</strong>g the <strong>in</strong>crements <strong>in</strong> the<br />
peripheral districts. By relat<strong>in</strong>g the gas consumption<br />
estimated by each type of l<strong>and</strong> occupation <strong>and</strong> the cost<br />
for exp<strong>and</strong><strong>in</strong>g the gas distribution network, the mo<strong>de</strong>l<br />
will <strong>in</strong>dicate, for each neighborhood, the viability of<br />
implement<strong>in</strong>g a gas network as well as the places with<br />
potential for grow<strong>in</strong>g <strong>de</strong>nsity <strong>in</strong> the exist<strong>in</strong>g gas<br />
distribution system. In this paper, examples of essential<br />
<strong>in</strong>formation that compose the methodology are<br />
presented for six districts of São Paulo city (Brazil):<br />
Ipiranga,Tatuapé, Penha, Vila Matil<strong>de</strong>, Socorro <strong>and</strong> Vila<br />
Formosa, which have different socioeconomic <strong>and</strong><br />
geographical profiles. F<strong>in</strong>ally, the mo<strong>de</strong>l tested for São<br />
Paulo will be generalized <strong>in</strong> a computer mo<strong>de</strong>l, which<br />
allows its use <strong>in</strong> other Brazilian cities, po<strong>in</strong>t<strong>in</strong>g out the<br />
possibilities of natural gas as a f<strong>in</strong>al option of energy <strong>in</strong><br />
the urban uses, besi<strong>de</strong>s present<strong>in</strong>g gui<strong>de</strong>l<strong>in</strong>es for the<br />
Urban Plans <strong>and</strong> the susta<strong>in</strong>able gas <strong>in</strong>frastructure<br />
<strong>in</strong>corporation <strong>in</strong> the cities (Udaeta et al., 2004).<br />
METHODOLOGY<br />
The methodology based on urban <strong>in</strong>dicators (Massara,<br />
2007) has the ma<strong>in</strong> objective of <strong>de</strong>velop<strong>in</strong>g procedures<br />
that allow to analyze <strong>and</strong> to gui<strong>de</strong> the expansion <strong>and</strong> the<br />
grow<strong>in</strong>g <strong>de</strong>nsity of the natural gas network <strong>in</strong>si<strong>de</strong> a<br />
municipal district through the study of the <strong>de</strong>velopment<br />
among several parameters <strong>and</strong> also to analyze the urban<br />
dynamics that <strong>de</strong>term<strong>in</strong>es the expansion of the<br />
metropolitan natural gas <strong>in</strong>frastructure (Williams, 1962;<br />
Forrester, 1969). The systemic mo<strong>de</strong>l was elaborated<br />
accord<strong>in</strong>g to the follow<strong>in</strong>g stages:<br />
• I<strong>de</strong>ntification, characterization <strong>and</strong> organization of<br />
the ma<strong>in</strong> <strong>in</strong>terventionary factors;<br />
• Def<strong>in</strong>ition of the study cells (streets, districts,<br />
suburbs) accord<strong>in</strong>g to the availability of<br />
<strong>in</strong>formation about the parameters;<br />
• Hierarchization of quantitative <strong>and</strong> qualitative<br />
parameters through the attribution scale of<br />
priorities that unifies the parameters units, so that<br />
they can mathematically be treated;<br />
• Evaluation of the system with the purpose of<br />
direct<strong>in</strong>g the choice based <strong>in</strong> theoretical gui<strong>de</strong>l<strong>in</strong>es<br />
<strong>and</strong> also <strong>in</strong> the application of the Analytic<br />
Hierarchy Process (AHP) support<strong>in</strong>g the computer<br />
mo<strong>de</strong>l (Saaty & Vargas, 1982);<br />
• Validation of the <strong>de</strong>cision-mak<strong>in</strong>g process for the<br />
plann<strong>in</strong>g of the expansion of natural gas<br />
distribution <strong>in</strong>frastructure through case study <strong>in</strong><br />
the city of São Paulo <strong>and</strong> comparison with the<br />
results obta<strong>in</strong>ed through the mo<strong>de</strong>l <strong>and</strong> the<br />
mapp<strong>in</strong>g of the existent implanted net.<br />
• Validation through comparison with the Analytic<br />
Hierarchy Process (Saaty, 2006).<br />
In this prelim<strong>in</strong>ary analysis, all the parameters have<br />
been used with <strong>in</strong>fluence distribution <strong>in</strong>dices <strong>in</strong> <strong>in</strong>tervals<br />
from 1 to 5. The attribution of this scale is related to the<br />
use of the card<strong>in</strong>al <strong>and</strong> semantic scales (Allen, 1987). At<br />
the end of the study, it will be possible to <strong>de</strong>term<strong>in</strong>e<br />
whether this l<strong>in</strong>ear scale is valid or not, via test <strong>and</strong><br />
comparison with analysis of multiple-criteria mo<strong>de</strong>ls<br />
that also use the priority scale as a tool <strong>in</strong> the <strong>de</strong>cisionmak<strong>in</strong>g<br />
process (Saaty, 1980). Similarly, the algorithm<br />
used <strong>in</strong> this paper for the <strong>de</strong>term<strong>in</strong>ation of the<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.55-62, 2010
Massara <strong>and</strong> Udaeta<br />
57<br />
attractiveness <strong>in</strong><strong>de</strong>x must be revised <strong>and</strong> compared with<br />
the results obta<strong>in</strong>ed accord<strong>in</strong>g to the AHP methods.<br />
The chosen hierarchization is the same used by the<br />
AHP (Saaty, 1980; 2006), which allows the verification<br />
between algorithms of this method <strong>and</strong> that one of the<br />
proposal mo<strong>de</strong>l<strong>in</strong>g (<strong>in</strong> this work), accord<strong>in</strong>g to the<br />
association presented <strong>in</strong> Table 1.<br />
The proposed mo<strong>de</strong>l <strong>and</strong> the <strong>in</strong>formation systems<br />
The sets of data are composed of four systems referr<strong>in</strong>g<br />
to the study area, accord<strong>in</strong>g to the Brazilian official<br />
<strong>de</strong>nom<strong>in</strong>ation “SEMPLA” (São Paulo, 2002). These<br />
systems establish an attractiveness <strong>in</strong><strong>de</strong>x for natural gas<br />
<strong>in</strong>frastructure expansion with regards to the urban<br />
<strong>de</strong>velopment of the cities.<br />
First System: Life Quality Indicators (LQ)<br />
This system is represented by three factors, all of which<br />
represent numerical values that are distributed <strong>in</strong> the<br />
five groups that have been <strong>in</strong>itially <strong>de</strong>scribed.<br />
Social Exclusion In<strong>de</strong>x (SEI): it has the objective of<br />
i<strong>de</strong>ntify<strong>in</strong>g the social <strong>de</strong>velopment <strong>de</strong>gree of the<br />
districts (Sposati, 2000) consi<strong>de</strong>r<strong>in</strong>g the existence of<br />
social equipments (green areas, bus l<strong>in</strong>es, number of<br />
schools <strong>and</strong> hospitals). The f<strong>in</strong>al <strong>in</strong><strong>de</strong>x attributed to each<br />
one of the districts lies with<strong>in</strong> the <strong>in</strong>terval that ranges<br />
between -1.00 (reflect<strong>in</strong>g the worst exclusion situation,<br />
i.e., the first group) <strong>and</strong> +1.00 (reflect<strong>in</strong>g the best<br />
<strong>in</strong>clusion situation, i.e., the fifth group).<br />
Human Development In<strong>de</strong>x (HDI): This is an adaptation<br />
of the <strong>in</strong><strong>de</strong>x created by the United Nations Organization<br />
(UNO), with the objective of compar<strong>in</strong>g the <strong>de</strong>gree of<br />
human <strong>de</strong>velopment accord<strong>in</strong>g to the districts. It<br />
comprises factors such as education <strong>and</strong> basic<br />
conditions of health <strong>in</strong> each district (IBGE, 2008).<br />
These <strong>in</strong>dicators are transformed <strong>in</strong> <strong>in</strong>tervals that range<br />
between 0 (first group, the worst <strong>de</strong>velopment<br />
condition) <strong>and</strong> 1 (fifth group, referr<strong>in</strong>g to the best<br />
conditions).<br />
Priority Infrastructure (WS, SS, SI): water supply, sewer<br />
system <strong>and</strong> streets illum<strong>in</strong>ation are consi<strong>de</strong>red “priority<br />
<strong>in</strong>frastructures” (São Paulo, 2002). This parameter<br />
corresponds to the i<strong>de</strong>a that a given district will not be<br />
attractive to natural gas network if it still does not<br />
possess such <strong>in</strong>frastructure. The <strong>in</strong><strong>de</strong>x consi<strong>de</strong>rs the<br />
mean percentage of the actual conditions of the three<br />
nets, expressed <strong>in</strong> five groups <strong>in</strong> <strong>in</strong>tervals of 20%.<br />
Table 1. Adaptation of AHP scale to natural gas study<br />
Group Semantic scale for natural gas<br />
AHP<br />
scale<br />
1 Low attractivity to network <strong>in</strong>stallation 1<br />
2<br />
Low to medium attractivity to the<br />
<strong>in</strong>stallation of the network<br />
3<br />
3<br />
Medium attractivity to the <strong>in</strong>stallation<br />
of the network<br />
5<br />
4<br />
Medium to high attractivity to the<br />
<strong>in</strong>stallation of the network<br />
7<br />
5<br />
High attractivity to the <strong>in</strong>stallation of<br />
the network<br />
9<br />
Source: Massara (2007) based on Saaty (2006).<br />
Second System: Urban Plan Indicators (UP)<br />
These <strong>in</strong>dicators comb<strong>in</strong>e qualitative parameters (l<strong>and</strong><br />
use, urban <strong>de</strong>velopment <strong>and</strong> zon<strong>in</strong>g) <strong>and</strong> quantitative<br />
(urbanization rate <strong>and</strong> real estate). For the analysis of<br />
non-numerical values, the systems are based on the<br />
mapp<strong>in</strong>g <strong>and</strong> classification of the Urban Plan of São<br />
Paulo city <strong>and</strong> its adaptation for any city, <strong>in</strong> or<strong>de</strong>r to<br />
verify the types of l<strong>and</strong> occupation (resi<strong>de</strong>ntial,<br />
commercial, <strong>in</strong>dustrial <strong>and</strong> services) <strong>and</strong> their expansion<br />
perspectives. This is carried out with the purpose of<br />
effect<strong>in</strong>g the elaboration of a neighborhood profile with<br />
larger ten<strong>de</strong>ncy to <strong>in</strong>dustrial <strong>de</strong>velopments (higher<br />
attractiveness for natural gas), <strong>and</strong> follows the natural<br />
gas consumption projection listed below.<br />
L<strong>and</strong> Use (Lures, LUcom, LUserv, LU<strong>in</strong>d): consi<strong>de</strong>r<strong>in</strong>g<br />
the largest percentage of streets with certa<strong>in</strong> use type,<br />
this concept is based <strong>in</strong> the occupation characteristic of<br />
the city districts represented <strong>in</strong> the urban plan map (São<br />
Paulo, 2006). Although the municipality consi<strong>de</strong>rs<br />
several occupation categories <strong>in</strong> view of the natural gas<br />
attractiveness, groups have been classified accord<strong>in</strong>g to<br />
five ma<strong>in</strong> uses:<br />
1st. Group: horizontal resi<strong>de</strong>ntial occupation;<br />
2nd. Group: mixed use (commercial <strong>and</strong> resi<strong>de</strong>ntial<br />
horizontal);<br />
3rd. Group: vertical resi<strong>de</strong>ntial occupation;<br />
4th. Group: mixed use (commercial, services <strong>and</strong><br />
resi<strong>de</strong>ntial vertical);<br />
5th. Group: mixed use (resi<strong>de</strong>ntial <strong>and</strong> <strong>in</strong>dustrial).<br />
Urban Development (UD): the basis of this concept is<br />
the Urban Plan of São Paulo city (São Paulo, 2002;<br />
2006), correspond<strong>in</strong>g to five “macroareas” that<br />
comprehend the present specifications <strong>and</strong> the future<br />
urban <strong>de</strong>velopments, as represented <strong>in</strong> the follow<strong>in</strong>g<br />
five groups:<br />
1st. Group: environmental protection – limits of public<br />
areas <strong>and</strong> preservation areas;<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.55-62, 2010
58<br />
Massara <strong>and</strong> Udaeta<br />
2nd. Group: urbanization <strong>and</strong> urban qualification – areas<br />
predom<strong>in</strong>antly occupied by low <strong>in</strong>come<br />
families with high concentration of<br />
irregular constructions;<br />
3rd. Group: requalification – areas with good<br />
<strong>in</strong>frastructure although present<strong>in</strong>g many<br />
empty properties;<br />
4th. Group: urbanization <strong>in</strong> consolidation – areas <strong>in</strong><br />
condition to attract real estate <strong>in</strong>vestments<br />
<strong>in</strong> resi<strong>de</strong>nces, services <strong>and</strong> commercial<br />
establishments;<br />
5th. Group: consolidated urbanization – areas formed by<br />
consolidated neighborhoods <strong>in</strong>habited by<br />
population of medium <strong>and</strong> high <strong>in</strong>come<br />
<strong>and</strong> good urbanization conditions.<br />
L<strong>and</strong> Use Rule (Zon<strong>in</strong>g - Z): This refers to the rule<br />
imposed by the Municipality (São Paulo, 2002; 2006)<br />
that limits l<strong>and</strong> use <strong>and</strong> the <strong>de</strong>st<strong>in</strong>ation of the several<br />
sections of the city for <strong>de</strong>term<strong>in</strong>ed uses (tra<strong>de</strong>, services,<br />
hous<strong>in</strong>g, <strong>in</strong>dustries). The natural gas attractiveness<br />
predom<strong>in</strong>ance is, <strong>in</strong> this case, represented by the<br />
follow<strong>in</strong>g five groups:<br />
1st. Group: resi<strong>de</strong>ntial zone of low <strong>de</strong>nsity / zone of<br />
environmental protection;<br />
2nd. Group: resi<strong>de</strong>ntial zone of medium <strong>de</strong>nsity /<br />
mixed zone of low <strong>de</strong>nsity;<br />
3rd. Group: resi<strong>de</strong>ntial zone of high <strong>de</strong>nsity / mixed<br />
zone of medium <strong>de</strong>nsity / special uses;<br />
4th. Group: mixed zone of high <strong>de</strong>nsity;<br />
5th. Group: great <strong>in</strong>dustrial zone occupation.<br />
The <strong>de</strong>f<strong>in</strong>ition of “mixed zone” corresponds to the<br />
comb<strong>in</strong>ation of the resi<strong>de</strong>ntial use, services <strong>and</strong><br />
commercial uses.<br />
Urbanization Rate (UR): corresponds to the percentage<br />
of the total district area occupied by urban use (blen<strong>de</strong>d<br />
resi<strong>de</strong>nces with commerce, services <strong>and</strong> <strong>in</strong>dustries) as<br />
compared to the total population (São Paulo, 2006;<br />
IBGE, 2008), which is represented <strong>in</strong> <strong>in</strong>tervals of 20%.<br />
Real Estate for resi<strong>de</strong>ntial <strong>and</strong> service market (REres,<br />
REserv): this concept is related to the Construction<br />
Co<strong>de</strong> about build<strong>in</strong>g facilities for natural gas <strong>in</strong> new<br />
constructions (International Co<strong>de</strong> Council, 2000), which<br />
must <strong>in</strong>crease the consumption of natural gas. In or<strong>de</strong>r<br />
to attribute the attractiveness <strong>in</strong><strong>de</strong>x, five groups have<br />
been selected which were based on the largest <strong>and</strong> the<br />
smallest number of district releases as reported <strong>in</strong> a<br />
survey by EMBRAESP (2008).<br />
Third System: Potential of Natural Gas<br />
Consumption Indicators (PNGC)<br />
In this data system, the characteristics of the<br />
neighborhoods <strong>in</strong> terms of street extensions are stored.<br />
All parameters are numerical values, thus simplify<strong>in</strong>g<br />
the creation of the five groups us<strong>in</strong>g the numerical<br />
simple division of values that are obta<strong>in</strong>ed by the<br />
Brazilian <strong>in</strong>stitutions IBGE (2008) <strong>and</strong> São Paulo<br />
(2002; 2006), as follows.<br />
Demographic Density (DD): corresponds to the ratio<br />
between the number of resi<strong>de</strong>nt people <strong>and</strong> the total area<br />
of the district (SEADE, 2008), consi<strong>de</strong>r<strong>in</strong>g that “larger<br />
people concentrations generate larger energy <strong>de</strong>m<strong>and</strong>”.<br />
Family Income (FI): corresponds to the m<strong>in</strong>imum wages<br />
<strong>in</strong> the homes of the district (IBGE, 2008; SEADE,<br />
2008), consi<strong>de</strong>r<strong>in</strong>g the relationship between “<strong>in</strong>come<br />
<strong>and</strong> energy consumption possibilities”.<br />
Stratification accord<strong>in</strong>g to the type of l<strong>and</strong> use,<br />
households <strong>and</strong> economic activities (Sres, Scom, Sserv,<br />
S<strong>in</strong>d): the basis of this method consi<strong>de</strong>rs that both the<br />
<strong>de</strong>m<strong>and</strong> for natural gas <strong>and</strong> the attractiveness of the<br />
district receiv<strong>in</strong>g the canalized gas system <strong>in</strong>crease with<br />
the number of domiciles or establishments with<br />
economic activities <strong>in</strong> a specific district (Groenendaal,<br />
1998). The division <strong>in</strong> establishments is ma<strong>de</strong> through<br />
the Economic Activities Cadaster (IBGE, 2003) <strong>and</strong> for<br />
the households through the “Census” (IBGE, 2008;<br />
SEADE, 2008). Those data enter <strong>in</strong> the file as “number<br />
of units” <strong>and</strong> the system automatically transforms them<br />
<strong>in</strong> <strong>in</strong>fluence <strong>in</strong>dices from 1 to 9, accord<strong>in</strong>g to the lowest<br />
<strong>and</strong> the highest numbers obta<strong>in</strong>ed for each district.<br />
Besi<strong>de</strong>s the <strong>in</strong>ternal scale, the program allows the<br />
<strong>in</strong>sertion by the user of one external <strong>in</strong>fluence that is<br />
multiplied by the attributed value. This comes already<br />
as a function of the number of units that dist<strong>in</strong>guish the<br />
activities that are more appropriate to the use of natural<br />
gas or, for other reasons, are more attractive to the<br />
natural gas concessionaire.<br />
In the follow<strong>in</strong>g example, the external <strong>in</strong>fluence has<br />
not been attributed. All parameters have only the<br />
<strong>in</strong>ternal scale from 1 to 9.<br />
Fourth System: Civil Construction Indicators (CC)<br />
The function of this system is to represent values<br />
associated with the cost of the un<strong>de</strong>rground<br />
<strong>in</strong>frastructure implementation. As for other numerical<br />
parameters, the five groups have been elaborated by<br />
simple division of the acquired values expressed by the<br />
four factors below.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.55-62, 2010
Massara <strong>and</strong> Udaeta<br />
59<br />
Natural Gas Infrastructure Extension (E): corresponds<br />
to the distance between areas that can be served <strong>and</strong><br />
others that are effectively atten<strong>de</strong>d by natural gas. The<br />
rule for the <strong>in</strong>fluence attribution is “the smaller the<br />
distance, the larger the <strong>in</strong>fluence attributed for natural<br />
gas attractiveness”.<br />
Natural Gas Distribution Ramification (D): corresponds<br />
to the total sum of the <strong>in</strong>ternal streets that may be<br />
atten<strong>de</strong>d by natural gas supply systems. The parameters<br />
are also consi<strong>de</strong>red <strong>in</strong> the rule: “the smaller the distance,<br />
the larger the <strong>in</strong>fluence”. Thus, districts partially located<br />
<strong>in</strong> served areas are attributed “five” as the <strong>in</strong>fluence<br />
<strong>in</strong><strong>de</strong>x, express<strong>in</strong>g the best attractiveness conditions.<br />
Built Density (BDres, BDcom, BDserv, BD<strong>in</strong>d): this<br />
factor is obta<strong>in</strong>ed by the ratio between the built area<br />
accord<strong>in</strong>g to the type of l<strong>and</strong> use (resi<strong>de</strong>ntial,<br />
commercial, services <strong>and</strong> <strong>in</strong>dustrial) <strong>and</strong> the total area<br />
of the district (São Paulo, 2006). In districts with high<br />
<strong>in</strong>dustrial <strong>de</strong>nsity, lower <strong>in</strong>vestments with ramifications<br />
are consi<strong>de</strong>red (“capillarity” of the distribution<br />
network).<br />
Avenues <strong>and</strong> Streets of Great Importance Traffic (T):<br />
even with the constructive process evolution, the<br />
parameter <strong>in</strong>dicates the importance of the district as a<br />
connection among neighborhoods <strong>and</strong> other municipal<br />
districts <strong>and</strong> the special attention that must be given to<br />
the <strong>in</strong>terdiction plan. The number of avenues (or streets)<br />
is obta<strong>in</strong>ed by simple consultation of any streets gui<strong>de</strong><br />
<strong>and</strong> by count<strong>in</strong>g avenues <strong>and</strong> streets of larger extension<br />
with<strong>in</strong> any given district.<br />
Accord<strong>in</strong>g to the <strong>de</strong>scription <strong>in</strong> the methodology, all<br />
the values gathered are unified to the same unit<br />
(percentage) <strong>and</strong> divi<strong>de</strong>d <strong>in</strong> five zones of attractiveness<br />
that are converted <strong>in</strong> the hierarchisation scale of 1 to 9.<br />
After the weights are attributed, they are submitted to<br />
the calculation of “Attractiveness <strong>in</strong><strong>de</strong>x”, whose<br />
algorithm is based on the simple sum of the weights for<br />
each area of study as summarized <strong>in</strong> Table 2.<br />
Table 2. Summary of the algorithm for the Attractiveness In<strong>de</strong>x.<br />
Attractivity In<strong>de</strong>x by Information Systems<br />
LQ= (ISE+HDI+WS+SS+SI)/n<br />
UP=<br />
LUres+LUcom+LUserv+LU<strong>in</strong>d+UD+Z+UR+REres+REserv)/n<br />
PNGC= (DD+FI+Sres+Scom+Sserv+S<strong>in</strong>d)/n<br />
CC= (E+D+BDres+BDcom+BDserv+BD<strong>in</strong>d+T)/n<br />
General Attractivity In<strong>de</strong>x<br />
IG= (LQ+UP+PNGC+CC)/n<br />
Source: Massara (2007). Note: n represents the number of<br />
parameters effectively used. The acronyms are <strong>de</strong>scribed <strong>in</strong><br />
methodology.<br />
RESULTS AND DISCUSSION<br />
As suggested <strong>in</strong> the recommendations about natural gas<br />
policy (Meier, 1997; IESP, 2004), the choice of São<br />
Paulo (capital of São Paulo State) as an all-gas city, is to<br />
<strong>de</strong>monstrate the complete utilization of natural gas <strong>in</strong> all<br />
its applications. In the context of this work, the choice<br />
has background <strong>in</strong>:<br />
• The possibility of <strong>in</strong>creas<strong>in</strong>g the <strong>de</strong>nsity of use of<br />
the current grids: the <strong>in</strong>crement of the factor of<br />
utilization of the <strong>in</strong>frastructure implanted <strong>in</strong> the<br />
exp<strong>and</strong>ed center, besi<strong>de</strong>s the extend<strong>in</strong>g the consumption<br />
market can leverage the f<strong>in</strong>anc<strong>in</strong>g of the network<br />
expansion.<br />
• The process of civil construction has evolved with<br />
the use of a no-dig method of cutt<strong>in</strong>g trenches, the<br />
trenchless technology (Istt, 2008; Najafi, 2005). It<br />
reduces the trouble caused by the <strong>in</strong>terdiction of traffic<br />
ways to rebuild the pavement, thus mak<strong>in</strong>g its execution<br />
cheaper, ma<strong>in</strong>ly <strong>in</strong> already consolidated urban areas.<br />
• As mentioned by Mout<strong>in</strong>ho dos Santos et al. (2002,<br />
p. 162): “In the São Paulo municipality, along the roads<br />
at the marg<strong>in</strong>s of Tietê <strong>and</strong> P<strong>in</strong>heiros rivers, the greatest<br />
commercial areas <strong>in</strong> the country are concentrated here,<br />
with various shopp<strong>in</strong>g centers <strong>and</strong> large office<br />
build<strong>in</strong>gs. All of them are located less than 2 km from<br />
the high pressure Comgás (gas utility) pipe-r<strong>in</strong>g, but<br />
rarely it is used”.<br />
This situation will be the same <strong>in</strong> the next years,<br />
accord<strong>in</strong>g to the map (Comgás, 2007). The map of the<br />
gas utility shows the natural gas grid distribution <strong>in</strong> the<br />
city of São Paulo, <strong>in</strong> other words, which district is<br />
already served by the network, permitt<strong>in</strong>g the selection<br />
of areas not served for test<strong>in</strong>g the mo<strong>de</strong>l proposed <strong>in</strong><br />
this work.<br />
Figure 1 shows the city of São Paulo with<strong>in</strong> the<br />
metropolitan region of São Paulo.<br />
Fig. 1. City of São Paulo (ash area) <strong>in</strong>serted <strong>in</strong> the Metropolitan<br />
Region. Source: SEADE (2008).<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.55-62, 2010
60<br />
Massara <strong>and</strong> Udaeta<br />
S<strong>in</strong>ce the case study has been used <strong>in</strong> the city of São<br />
Paulo, it has been divi<strong>de</strong>d <strong>in</strong> 96 districts <strong>in</strong> conformity<br />
with the <strong>in</strong>formation reported by the local Urban Plan<br />
Department (São Paulo, 2002; Massara, 2002).<br />
Figure 2 <strong>in</strong>troduces the position<strong>in</strong>g of the districts<br />
selected <strong>in</strong> this paper, emphasiz<strong>in</strong>g the different<br />
geographical <strong>and</strong> social characteristics of each one.<br />
The selected districts represent areas of the city with<br />
great urban transformations. Some of them face a<br />
process of <strong>in</strong>tensification of the resi<strong>de</strong>ntial real estate<br />
market (Ipiranga, Tatuapé <strong>and</strong> Penha), attract<strong>in</strong>g<br />
<strong>in</strong>vestments <strong>in</strong> several economic activities <strong>in</strong> or<strong>de</strong>r to<br />
respond to that new occupation style. Others have low<br />
family <strong>in</strong>come but possess <strong>in</strong>dustrial areas that must<br />
consume natural gas (Vila Formosa <strong>and</strong> Socorro).<br />
Others present great urban complexity, function<strong>in</strong>g<br />
as a “bedroom cities”, with l<strong>and</strong> use predom<strong>in</strong>antly<br />
composed by habitational groups (Vila Matil<strong>de</strong>).<br />
In or<strong>de</strong>r to <strong>de</strong>monstrate the use of the mo<strong>de</strong>l<br />
<strong>de</strong>veloped <strong>in</strong> this work, for <strong>in</strong>stance, one application is<br />
presented across the Table 3.<br />
Therefore, <strong>in</strong> this example, the follow<strong>in</strong>g<br />
simplifications were consi<strong>de</strong>red:<br />
• Calculation of the attractiveness <strong>in</strong><strong>de</strong>x, us<strong>in</strong>g the<br />
“General” group<strong>in</strong>g (without the division <strong>in</strong><br />
<strong>in</strong>formation systems) <strong>and</strong> its respective rank<strong>in</strong>g;<br />
• Study area: districts;<br />
• Extra weight to emphasize the segments for the<br />
natural gas use = 1.0 for all the activities.<br />
• Algorithm application accord<strong>in</strong>g to the Table 2.<br />
LEGEND:<br />
1 - District of Ipiranga<br />
2 - District of Tatuapé<br />
3 - District of Penha<br />
4 - District of Vila Matil<strong>de</strong><br />
5 - District of Socorro<br />
6 - District of Vila Formosa<br />
Fig 2. The 6 selected districts <strong>in</strong> São Paulo municipality (capital<br />
city). (Note: unscaled figure).<br />
Source: SEADE (2008).<br />
Table 3. Attribution of weights accord<strong>in</strong>g to the four <strong>in</strong>formation<br />
systems<br />
Selected Districts of São Paulo City<br />
Parameters Vila<br />
Ipiranga Tatuapé Penha<br />
Matil<strong>de</strong> Socorro Vila<br />
Formosa<br />
ISE 7 9 7 7 9 9<br />
HDI 3 5 5 5 7 7<br />
WS 9 9 7 9 9 9<br />
SS 9 9 9 9 1 9<br />
SI 9 9 9 5 9 7<br />
LU res 3 5 5 7 3 9<br />
LU com 5 7 7 5 3 3<br />
LU serv 3 9 9 5 1 1<br />
LU <strong>in</strong>d 5 1 1 1 7 1<br />
Z 9 7 7 5 5 7<br />
UD 9 5 5 5 9 9<br />
UR 9 9 9 9 9 9<br />
RE res 5 9 5 1 1 3<br />
RE serv 3 9 9 1 5 1<br />
DD 3 3 5 5 1 5<br />
FI 3 5 1 1 3 3<br />
S res 5 3 5 3 1 3<br />
S com 9 7 9 3 9 5<br />
S serv 5 7 9 5 5 5<br />
S <strong>in</strong>d 9 5 5 1 9 3<br />
D 9 9 3 5 9 3<br />
E 5 7 1 5 9 5<br />
T 1 1 1 3 1 3<br />
BD res 5 9 7 7 3 9<br />
BC com 9 3 7 5 3 9<br />
BC serv 9 3 7 5 3 9<br />
BC <strong>in</strong>d 1 5 3 1 9 1<br />
Source : Massara (2007) based on Sea<strong>de</strong> (2008), São Paulo (2006).<br />
Table 4. Results for the case-study – The general attractiveness<br />
<strong>in</strong><strong>de</strong>x <strong>and</strong> its respective rank<strong>in</strong>g<br />
District<br />
AHP<br />
General<br />
General AHP<br />
<strong>in</strong><strong>de</strong>x<br />
<strong>in</strong><strong>de</strong>x<br />
Rank<strong>in</strong>g rank<strong>in</strong>g<br />
(%)<br />
Ipiranga 6,1 18,9% 2º 2º<br />
Tatuapé 6,3 22,2% 1º 1º<br />
Penha 5,9 16,8% 3º 3º<br />
Vila Matil<strong>de</strong> 4,7 10% 6º 6º<br />
Socorro 5,4 15,6% 5º 5º<br />
Vila Formosa 5,6 16,5% 4º 4º<br />
Source: Massara (2007) based on Saaty (2006).<br />
Table 4 resumes the attractiveness rank<strong>in</strong>g for five<br />
municipal districts for the natural gas expansion<br />
projection. From the results listed there<strong>in</strong>, one may<br />
conclu<strong>de</strong> that the districts with better positions<br />
contemplate mixed l<strong>and</strong> use (Tatuapé is the best<br />
example of which) <strong>and</strong> equality of consumption<br />
projection <strong>in</strong> all uses, associated to a larger<br />
<strong>de</strong>mographic <strong>de</strong>nsity <strong>and</strong> larger family <strong>in</strong>come.<br />
The first units <strong>in</strong> the rank<strong>in</strong>g of the mo<strong>de</strong>l are<br />
districts with the largest <strong>de</strong>velopment <strong>in</strong> the real-estate<br />
market, largest leeway of zon<strong>in</strong>g. These districts are<br />
nearest to the areas already served <strong>and</strong>, of course, have<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.55-62, 2010
Massara <strong>and</strong> Udaeta<br />
61<br />
the largest forecast of natural gas consumption, ma<strong>in</strong>ly,<br />
<strong>in</strong> the vertical resi<strong>de</strong>ntial use (with highlight for the<br />
follow<strong>in</strong>g districts: Ipiranga, Tatuapé <strong>and</strong> Penha).<br />
The rest of the districts concentrate the least <strong>in</strong>come,<br />
horizontal resi<strong>de</strong>ntial use <strong>and</strong> a small commercial,<br />
services <strong>and</strong> <strong>in</strong>dustrial concentration (Vila Formosa <strong>and</strong><br />
Socorro). Apart from the characteristics of the<br />
<strong>in</strong>termediate group, the last classification (Vila Matil<strong>de</strong>)<br />
is <strong>in</strong> an area with environmental limitations to<br />
expansion <strong>and</strong>, therefore, of small urban <strong>de</strong>velopment.<br />
The comparison between the rank<strong>in</strong>g us<strong>in</strong>g AHP<br />
(Saaty, 2006) <strong>and</strong> the mo<strong>de</strong>l rank<strong>in</strong>g (Massara, 2007)<br />
has satisfactory results (presented <strong>in</strong> Table 4).<br />
The mo<strong>de</strong>l based on urban <strong>in</strong>dicators has presented<br />
coherent results when tested <strong>in</strong> the city of São Paulo. It<br />
has been <strong>de</strong>monstrated that the mo<strong>de</strong>l proved to be a<br />
good calculation tool with a reasonable precision<br />
<strong>de</strong>gree, be<strong>in</strong>g easily un<strong>de</strong>rstood <strong>and</strong> function<strong>in</strong>g as an<br />
auxiliary <strong>in</strong> the <strong>de</strong>cision-mak<strong>in</strong>g process for the natural<br />
gas expansion <strong>in</strong>frastructure <strong>in</strong> the Brazilian cities.<br />
However, <strong>in</strong> or<strong>de</strong>r to have greater potential of<br />
expansion of the natural gas grid service, it is necessary<br />
that there exists strengthen<strong>in</strong>g of the relation between<br />
the utilities, network eng<strong>in</strong>eers of natural gas pipel<strong>in</strong>e<br />
facilities <strong>and</strong> all the civil construction related to the<br />
urban plans. In this case, it is good look<strong>in</strong>g for the<br />
<strong>de</strong>velopment of equipment for the diverse purposes that<br />
can utilize natural gas, <strong>in</strong>clud<strong>in</strong>g the formal assist<strong>in</strong>g<br />
from the public entities <strong>in</strong> the implementation of<br />
<strong>in</strong>frastructure <strong>and</strong> <strong>in</strong> advertis<strong>in</strong>g of natural gas use as an<br />
agent for urban <strong>de</strong>velopment as well.<br />
Thus, the creation of a resi<strong>de</strong>ntial <strong>and</strong> commercial<br />
market is based on the current use of PLG (petroleum<br />
liquefied gas) <strong>and</strong> on the gradual <strong>in</strong>troduction of natural<br />
gas <strong>in</strong> daily activities, thus <strong>in</strong>duc<strong>in</strong>g customers to the<br />
use of natural gas <strong>and</strong> enabl<strong>in</strong>g the <strong>de</strong>velopment of<br />
natural gas canalization <strong>and</strong> distribution.<br />
Acknowledgment This work received f<strong>in</strong>ancial<br />
support from the Brazillian agencies Agência Nacional<br />
do Petróleo – ANP (PRH-04), F<strong>in</strong>anciadora <strong>de</strong> Estudos<br />
e Projetos – FINEP <strong>and</strong> M<strong>in</strong>istério da Ciência e<br />
Tecnologia – MCT <strong>and</strong> FAPESP (Fundação <strong>de</strong> Amparo<br />
à Pesquisa do Estado <strong>de</strong> São Paulo) through the research<br />
project Process 03/06441-7 (“New Instruments for<br />
Regional Energy Plann<strong>in</strong>g seek<strong>in</strong>g Susta<strong>in</strong>able<br />
Development”) that permitted the utilization of the<br />
Decision Lens Program.<br />
REFERENCES<br />
Allen, E. (1987) Techniques of atitu<strong>de</strong> scale construction, Appleton<br />
Century Crofts, New York.<br />
BNDES – Banco Nacional <strong>de</strong> Desenvolvimento Econômico e Social.<br />
(2006) Setor <strong>de</strong> Petróleo e Gás: Perfil dos Investimentos, (Gas<br />
<strong>and</strong> Petroleum Sector: Investiment Profile), BNDES, Brasília. (<strong>in</strong><br />
portuguese).<br />
COMGAS – Companhia <strong>de</strong> Gas <strong>de</strong> São Paulo. (2007) Mapa da área<br />
Coberta pela Re<strong>de</strong>, (Distribution Network Map), COMGÁS, São<br />
Paulo. (<strong>in</strong> portuguese).<br />
EMBRAESP - Empresa Brasileira <strong>de</strong> Estudos <strong>de</strong> Patrimônio. (2008)<br />
Relatório anual sobre número <strong>de</strong> lançamentos imobiliários por<br />
região, região (Annual report on the number of real estate<br />
<strong>in</strong>vestments by region), EMBRAESP, São Paulo. (<strong>in</strong> portuguese).<br />
EIA – Energy Information Adm<strong>in</strong>istration. (2008) Annual Energy<br />
Outlook 2008 with Projections to 2030. Available <strong>in</strong>:<br />
http://www.eia.doe.gov/.<br />
Forrester, J.W. (1969) Urban Dynamics,M.I.T. Press, Cambridge.<br />
Gas Brasil. (2008) Newsletter Semanal [on l<strong>in</strong>e]. Available <strong>in</strong>:<br />
http://www.gasbrasil.com.br. (<strong>in</strong> portuguese).<br />
Groenendaal, W.V. (1998) The economic appraisal of natural gas<br />
projects, Oxford University Press for the Oxford Institute for<br />
Energy Studies, New York.<br />
IBGE – Instituto Brasileiro <strong>de</strong> Geografia e Estatística. (2003)<br />
Cadastro Nacional das Ativida<strong>de</strong>s Econômicas (CNAE),<br />
(National cadaster of economic activities) IBGE, Rio <strong>de</strong> Janeiro.<br />
(<strong>in</strong> portuguese).<br />
IBGE – Instituto Brasileiro <strong>de</strong> Geografia e Estatística. (2008)<br />
Cida<strong>de</strong>s, (Cities), IBGE, Rio <strong>de</strong> Janeiro. (<strong>in</strong> portuguese).<br />
IESP – Instituto <strong>de</strong> Engenharia <strong>de</strong> São Paulo. (2004) Recomendações<br />
sobre a política para o gás natural, (Recommendations about<br />
natural gas policy), IESP, São Paulo. (<strong>in</strong> portuguese).<br />
International Co<strong>de</strong> Council. (2000) International build<strong>in</strong>g co<strong>de</strong>.<br />
Falls Church, New York.<br />
ISTT - International Society for Trenchless Technology (2008)<br />
Trenchless techniques.Available <strong>in</strong> http://www.istt.com/.<br />
Massara, V.M. (2002) O perfil da <strong>in</strong>fra-estrutura no Município <strong>de</strong><br />
São Paulo e sua relação com as transformações <strong>de</strong> uso do solo: o<br />
centro exp<strong>and</strong>ido e a região <strong>de</strong> São Miguel Paulista, (The<br />
<strong>in</strong>frastructure profile <strong>in</strong> the city of São Paulo with regards to the<br />
transformations on the l<strong>and</strong> uses: the exp<strong>and</strong>ed center <strong>and</strong> the<br />
region of São Miguel Paulista), Dissertação (Mestrado), Escola<br />
Politécnica da Universida<strong>de</strong> <strong>de</strong> São Paulo. (<strong>in</strong> portuguese).<br />
Massara, V.M. (2007) A D<strong>in</strong>âmica Urbana na Otimização da Infra-<br />
Estrutura para o Gás Natural, (Urban Dynamics <strong>in</strong> the natural<br />
gas <strong>in</strong>frastructure), PhD Thesis, Instituto <strong>de</strong> Eletrotécnica e<br />
Energia da Universida<strong>de</strong> <strong>de</strong> São Paulo.<br />
Méier, E.P. (1997) Memória da <strong>in</strong>fra-estrutura urbana na área<br />
central <strong>de</strong> São Paulo, (Urban Infrastructure Memory <strong>in</strong> tnhe<br />
central área of São Paulo City), Associação Viva o <strong>Centro</strong>, São<br />
Paulo. (<strong>in</strong> portuguese).<br />
Mout<strong>in</strong>ho dos Santos E., Fagá, M.T.W., Villanueva, L.D. &<br />
Zamalloa, G. (2002) Gás natural: estratégias para uma nova<br />
energia no Brasil, (Natural Gas: strategy for a new energy <strong>in</strong><br />
Brasil), Editora Anamblume, FAPESP, PETROBRAS, São Paulo.<br />
(<strong>in</strong> portuguese).<br />
Najafi, M. (2005) Trenchless technology: pipel<strong>in</strong>e <strong>and</strong> utility <strong>de</strong>sign,<br />
construction, <strong>and</strong> renewal, McGraw-Hill, New York.<br />
Saaty, T. (2006) Decision Lens. Programa <strong>de</strong> Computador, London.<br />
Saaty T. (1980) The Analytic Hierarchy Process: Plann<strong>in</strong>g, Priority<br />
Sett<strong>in</strong>g, Resource Allocation, McGraw-Hill, London.<br />
Saaty, T.; Vargas, L.G. (1982). The logic of priorities, applications<br />
<strong>in</strong> bus<strong>in</strong>ess, energy, healthy, transportation. Nijhoff, Boston.<br />
São Paulo – Secretaria Municipal <strong>de</strong> Planejamento Urbano. (2002)<br />
Plano Diretor estratégico do Município <strong>de</strong> São Paulo 2002-2012,<br />
(Strategic directory plann<strong>in</strong>g <strong>in</strong> the citybof São Paulo), SEMPLA,<br />
São Paulo. (<strong>in</strong> portuguese).<br />
São Paulo – Secretaria Municipal <strong>de</strong> Planejamento Urbano. (2006)<br />
D<strong>in</strong>âmica Urbana, SEMPLA, São Paulo. (<strong>in</strong> portuguese).<br />
SEADE – Fundação Sistema Estadual <strong>de</strong> Análise <strong>de</strong> Dados. (2008)<br />
Pesquisa Municipal Unificada, (Unified municipal research).<br />
SEADE, São Paulo. (<strong>in</strong> portuguese).<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.55-62, 2010
62<br />
Massara <strong>and</strong> Udaeta<br />
Sposati, A. (2000) Mapa da Exclusão Social da cida<strong>de</strong> <strong>de</strong> São<br />
Paulo, (Social Exclusion Map), São Paulo, CDRom. (<strong>in</strong><br />
portuguese).<br />
Udaeta, M.E.M;., Grimoni, J.A.B. & Galvão, L.C.R. (2004)<br />
Iniciação a Conceitos <strong>de</strong> Sistemas Energéticos para o<br />
Desenvolvimento Limpo, (Initiation to Energy Systems Concepts<br />
for the Clean Development), EDUSP, São Paulo. (<strong>in</strong> portuguese).<br />
Williams, N. (1966) The structure of urban zon<strong>in</strong>g <strong>and</strong> its dynamics<br />
<strong>in</strong> urban plann<strong>in</strong>g <strong>and</strong> <strong>de</strong>velopment, Buttenheim Pub. Corp., New<br />
York.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.55-62, 2010
Gonçalves Neto <strong>and</strong> Lollo<br />
63<br />
J U E E<br />
Journal of Urban <strong>and</strong> Environmental<br />
Eng<strong>in</strong>eer<strong>in</strong>g, v.4 , n.2, p.63-73<br />
ISSN 1982-3932<br />
doi: 10.4090/juee.2010.v4n2.063073<br />
Journal of Urban <strong>and</strong><br />
Environmental Eng<strong>in</strong>eer<strong>in</strong>g<br />
www.journal-uee.org<br />
ANALYSIS OF NEIGHBORHOOD IMPACTS ARISING FROM<br />
IMPLEMENTATION OF SUPERMARKETS IN CITY OF SÃO<br />
CARLOS<br />
Pedro Silveira Gonçalves Neto* <strong>and</strong> José Augusto <strong>de</strong> Lollo<br />
Urban Eng<strong>in</strong>eer<strong>in</strong>g Post-graduation Program, São Carlos University, Brazil<br />
Received 18 March 2010; received <strong>in</strong> revised form 27 October 2010; accepted 16 November 2010<br />
Abstract:<br />
Keywords:<br />
The study <strong>in</strong>clu<strong>de</strong>d supermarkets of different sizes (small, medium <strong>and</strong> large - <strong>de</strong>f<strong>in</strong>ed<br />
based on the area occupied by the project <strong>and</strong> volume of activity) located <strong>in</strong> São Carlos<br />
(São Paulo state, Brazil) to evaluate the <strong>in</strong>fluence of the size of the project impacts<br />
neighborhood generated by these supermarkets. It was consi<strong>de</strong>red the <strong>in</strong>fluence of<br />
factors like the location of enterprises, size of the build<strong>in</strong>g, <strong>and</strong> areas of <strong>in</strong>fluence<br />
contribute to the <strong>in</strong>creased population <strong>de</strong>nsity <strong>and</strong> change of use of build<strong>in</strong>gs s<strong>in</strong>ce it<br />
was post-<strong>de</strong>ployment analysis. The relationship between the variables of the spatial<br />
impacts was ma<strong>de</strong> possible by the use of geographic <strong>in</strong>formation system. It was noted<br />
that the legislation does not have suitable conditions to gui<strong>de</strong> the studies of urban<br />
impacts due to the complex <strong>in</strong>tegration between the urban <strong>and</strong> impact<strong>in</strong>g components.<br />
Urban plann<strong>in</strong>g, Law 10257/2001, neighborhood impacts, supermarket cluster<br />
© 2010 Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE). All rights reserved.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.63-73, 2010
Gonçalves Neto <strong>and</strong> Lollo<br />
64<br />
INTRODUCTION<br />
São Carlos is a Brazilian municipality located <strong>in</strong> the<br />
state of São Paulo, near its geographic center <strong>and</strong> <strong>in</strong> a<br />
distance of 231 km from the state capital. With a<br />
population of around 220 463 <strong>in</strong>habitants (IBGE, 2009),<br />
distributed a total area of 1 141 km², is the 14 th largest<br />
city <strong>in</strong> the state <strong>in</strong> number of resi<strong>de</strong>nts. The city is an<br />
important regional center, with the economy based on<br />
<strong>in</strong>dustrial <strong>and</strong> agricultural activities (<strong>in</strong> this sector, there<br />
is the production of sugarcane, oranges, milk <strong>and</strong><br />
chicken).<br />
Served by a several bus <strong>and</strong> rail systems, São Carlos<br />
has units of production of some mult<strong>in</strong>ational<br />
companies, among them Volkswagen, Faber-Castell,<br />
Electrolux, Husqvarna <strong>and</strong> Tecumseh <strong>and</strong> some units of<br />
domestic companies, among which Toalhas São Carlos,<br />
Tapetes São Carlos, Papel São Carlos, Brazil Prom<strong>in</strong>as,<br />
Card<strong>in</strong>ali, Opto Electronics <strong>and</strong> Lat<strong>in</strong>a.<br />
In the light of local needs, <strong>and</strong> <strong>in</strong> certa<strong>in</strong> aspects,<br />
regional, the city has a network of tra<strong>de</strong> <strong>and</strong> services<br />
distributed <strong>in</strong> high street shops, gas stations<br />
convenience, several supermarket cha<strong>in</strong>s <strong>and</strong> a shopp<strong>in</strong>g<br />
center of Iguatemi group. In the field of research, they<br />
present the Fe<strong>de</strong>ral University of São Carlos (UFSCar)<br />
<strong>and</strong> the University of São Paulo (USP), also are present<br />
<strong>in</strong> the city two centers of technical <strong>de</strong>velopment at<br />
Embrapa.<br />
Despite these qualifications, the occupation of urban<br />
space took place <strong>in</strong> a disor<strong>de</strong>rly way, due to high<br />
population growth <strong>and</strong> the absence of laws to propose a<br />
direction to the use <strong>and</strong> occupation. The result can be<br />
seen <strong>in</strong> the segregation experienced by the population<br />
<strong>and</strong> the <strong>in</strong>effectiveness of urban <strong>in</strong>frastructure for<br />
subsequent <strong>de</strong>ployment of large build<strong>in</strong>gs without the<br />
proper review with<strong>in</strong> the current legal parameters.<br />
Accord<strong>in</strong>g Lollo (2006), s<strong>in</strong>ce the adoption of the<br />
Neighborhood Impact Study as a tool for urban<br />
management by the City Statute, the technical means<br />
has an <strong>in</strong>strument whose purpose is to analyze the<br />
behavior of high <strong>de</strong>nsity, however, the fact there is an<br />
adm<strong>in</strong>istrative mechanism of this magnitu<strong>de</strong>, does not<br />
represent a significant advance <strong>in</strong> public adm<strong>in</strong>istration<br />
s<strong>in</strong>ce there is no st<strong>and</strong>ardized evaluation system for<br />
such analysis <strong>and</strong> the means are not known for the<br />
correct application of such <strong>in</strong>struments.<br />
Mol<strong>in</strong>a (2008) states that the City Statute is subject<br />
to many discussions of how to apply the study of<br />
neighborhood impacts <strong>and</strong> what requirements that<br />
should be consi<strong>de</strong>red for the analysis of impacts, s<strong>in</strong>ce<br />
this document is of recent creation. It is emphasized that<br />
the proposition <strong>and</strong> use of <strong>in</strong>struments of control <strong>and</strong><br />
supervision of urban <strong>de</strong>nsification activity <strong>in</strong> Brazil is<br />
typical of a certa<strong>in</strong> size of cities (especially capitals <strong>and</strong><br />
major cities with<strong>in</strong> the states), while <strong>in</strong> small <strong>and</strong><br />
medium cities, such <strong>in</strong>struments do not exist or are not<br />
applied.<br />
The Neighborhood Impact Study has been practiced<br />
by most government (Mol<strong>in</strong>a, 2008) to <strong>in</strong>dicate the<br />
<strong>de</strong>velopments of greatest potential impact on the<br />
dimensions of these, but is overlooked aspects which<br />
relate to the use <strong>and</strong> occupation of l<strong>and</strong>, traffic<br />
generation <strong>and</strong> urban plann<strong>in</strong>g.<br />
In view of this, this study addressed the urban<br />
impacts aris<strong>in</strong>g from Neighborhood Impact Study post<strong>de</strong>ployment<br />
supermarket which fit this group<strong>in</strong>g,<br />
(Decree n. 19.915/98 apud L<strong>in</strong>dau et al., 2009), <strong>in</strong> the<br />
municipality of São Carlos <strong>and</strong> propose strategies to<br />
support impact assessments of the neighborhood.<br />
For this, we used the methodology <strong>de</strong>veloped by<br />
Lollo (2006) which produced an impact study of<br />
neighborhood <strong>in</strong> the same region whose focus was on<br />
bus<strong>in</strong>ess <strong>in</strong>formation technology. It was based on the<br />
parameters required by Fe<strong>de</strong>ral Law 10.257/2001,<br />
report<strong>in</strong>g <strong>and</strong> questionnaires to st<strong>and</strong>ardize the<br />
quantification of variables that had more than one<br />
<strong>in</strong>terpretation.<br />
THE NEIGHBORHOOD IMPACT STUDY (NIS)<br />
The Neighborhood Impact Study un<strong>de</strong>r the City Statute<br />
guarantees the negotiation between the private <strong>in</strong>terests<br />
of entrepreneurs <strong>and</strong> the right to urban environmental<br />
quality of city resi<strong>de</strong>nts, especially those who live <strong>and</strong><br />
travel <strong>in</strong> the vic<strong>in</strong>ity of the project un<strong>de</strong>r study.<br />
The NIS, when established <strong>in</strong> the Organic Law of<br />
Municipalities, provi<strong>de</strong>s for the participation of society<br />
<strong>in</strong> the analysis of each new <strong>de</strong>velopment to be built,<br />
thereby exercis<strong>in</strong>g <strong>de</strong>mocracy <strong>and</strong> <strong>in</strong>clusion of local<br />
communities that participate <strong>in</strong> the <strong>de</strong>cisions of what is<br />
best for their city. Some authors such as Leão-Aguiar et<br />
al. (2005, p. 2) argue that “the charge required for the<br />
preparation of the NIS, fills an important gap <strong>in</strong> the<br />
legislation, with regard to build<strong>in</strong>gs or activities that<br />
damage the local environment”.<br />
Urban growth has <strong>in</strong>creased <strong>de</strong>m<strong>and</strong> <strong>and</strong> need for<br />
new jobs, hous<strong>in</strong>g options public transportation <strong>and</strong><br />
pollution control, as well as the provision of basic<br />
services like water, sanitation, education <strong>and</strong> health.<br />
Regard<strong>in</strong>g the social content, urban growth contributed<br />
with the <strong>in</strong>tensification of segregation s<strong>in</strong>ce the<br />
<strong>in</strong>habitants, accord<strong>in</strong>g to Gregori (2004, p. 6) “are<br />
forced to live <strong>in</strong> poor neighborhoods of large cities,<br />
which have given the poor quality of life <strong>in</strong> cities, <strong>and</strong><br />
directly contributed to environmental <strong>de</strong>gradation <strong>and</strong><br />
<strong>in</strong>creas<strong>in</strong>g poverty <strong>in</strong> urban society”.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.63-73, 2010
Gonçalves Neto <strong>and</strong> Lollo<br />
65<br />
It is noteworthy that the Statute, the collective<br />
<strong>in</strong>terest should prevail over the use of private property<br />
(Brazil, 2001). With regard to the social function of the<br />
city <strong>and</strong> urban property, people’s participation <strong>in</strong> the<br />
public <strong>in</strong>terest, must ensure <strong>de</strong>mocratic management<br />
<strong>and</strong> equitable distribution of benefits <strong>and</strong> bur<strong>de</strong>ns<br />
result<strong>in</strong>g from the urbanization process.<br />
S<strong>in</strong>ce the population should have access to<br />
<strong>in</strong>tervention, it must be aware about the recovery of<br />
public <strong>in</strong>vestment that provi<strong>de</strong>d the valuation of urban<br />
property <strong>and</strong> appropriateness of economic policy<br />
<strong>in</strong>struments, tax <strong>and</strong> f<strong>in</strong>ancial public spend<strong>in</strong>g on<br />
<strong>de</strong>velopment urban to reach the goals.<br />
The aim of the Neighborhood Impact Study is to<br />
<strong>de</strong>mocratize the system for mak<strong>in</strong>g <strong>de</strong>cisions on major<br />
projects to be un<strong>de</strong>rtaken <strong>in</strong> the city, listen<strong>in</strong>g to<br />
neighborhoods <strong>and</strong> communities that are exposed to the<br />
impacts of large ventures.<br />
The Art. 36 of the City Statute (Brazil, 2001) states<br />
that “the bylaws must <strong>in</strong>clu<strong>de</strong> criteria to <strong>de</strong>f<strong>in</strong>e which<br />
constructions <strong>de</strong>pend of a previous neighborhood<br />
impact study as a condition for approval. These criteria<br />
may vary <strong>de</strong>pend<strong>in</strong>g on the characteristics <strong>and</strong> urban<br />
<strong>in</strong>frastructure of the municipality, <strong>and</strong> may be based on,<br />
for example, the impact of traffic generated, stra<strong>in</strong> on<br />
<strong>in</strong>frastructure, population <strong>de</strong>nsity, shadow<strong>in</strong>g of<br />
neighbor<strong>in</strong>g properties, noise, etc.” (p. 1).<br />
Un<strong>de</strong>r the same Act, art. 37, “the NIS will run to<br />
consi<strong>de</strong>r the positive <strong>and</strong> negative effects of the activity<br />
on the quality of life of people liv<strong>in</strong>g <strong>in</strong> the area <strong>and</strong> its<br />
vic<strong>in</strong>ity, <strong>in</strong>clud<strong>in</strong>g the analysis, at least the follow<strong>in</strong>g<br />
questions: (i) <strong>de</strong>nsity of population, (ii) urban <strong>and</strong><br />
community facilities, (iii) use <strong>and</strong> l<strong>and</strong> cover, (iv)<br />
valuation of real property, (v) generation of traffic <strong>and</strong><br />
<strong>de</strong>m<strong>and</strong> for public transportation, (vi) ventilation <strong>and</strong><br />
light<strong>in</strong>g; ( vii) urban l<strong>and</strong>scape <strong>and</strong> natural <strong>and</strong> cultural<br />
heritage” (p. 1).<br />
The NIR (Neighborhood Impact Report) should be<br />
elaborated from a series of i<strong>de</strong>ntifiers that <strong>in</strong>dicate the<br />
level of impact on the <strong>de</strong>ployment of any new venture<br />
on the urban l<strong>and</strong>scape, human activities that can be<br />
<strong>in</strong>stalled, urban <strong>in</strong>frastructure <strong>and</strong> on natural resources<br />
of the neighborhood (Moreira, 1999).<br />
Moreira (1999) also says that the factors can be<br />
quantified <strong>and</strong> <strong>de</strong>monstrated through mathematical<br />
mo<strong>de</strong>ls, computational, <strong>and</strong> diagnostic reports (or<br />
prelim<strong>in</strong>ary studies).<br />
Accord<strong>in</strong>g to Moreira (1999), the roadmap for the<br />
<strong>de</strong>velopment of a NIR must bear the thought of the<br />
follow<strong>in</strong>g questions: “(i) the impact of new<br />
<strong>de</strong>velopment on the urban l<strong>and</strong>scape, (ii) located on<br />
human activities, (iii) on the move people <strong>and</strong> goods,<br />
(iv) on urban <strong>in</strong>frastructure, <strong>and</strong> (v) on the natural<br />
resources of the neighborhood” (p. 110).<br />
Moreira (1999) also <strong>de</strong>scribes what should be the<br />
end products of a NIR that the script needs to <strong>in</strong>dicate:<br />
“(i) the <strong>de</strong>monstration of the compatibility of the road<br />
system <strong>and</strong> transport, (ii) the <strong>de</strong>monstration of the<br />
compatibility of the dra<strong>in</strong>age system with <strong>in</strong>creased<br />
volume <strong>and</strong> velocity of storm water generated by the<br />
seal<strong>in</strong>g area of <strong>in</strong>tervention, (iii) the <strong>de</strong>monstration of<br />
the feasibility of water supply, sewage collection,<br />
electricity supply, (iv) an <strong>in</strong>dication of urban<br />
transformations <strong>in</strong>duced by the project, <strong>and</strong> (v) the<br />
<strong>in</strong>sertion of the work <strong>in</strong> the l<strong>and</strong>scape” (p. 111).<br />
The observation of the po<strong>in</strong>ts raised by Moreira<br />
(1999) compared to the analytical criteria of the City<br />
Statute (2001) suggests that the discussion on the NIR<br />
tends to spread, because the legislation is not complete<br />
on all possible parameters of analysis urban only serves<br />
“as a gui<strong>de</strong>l<strong>in</strong>e for the preparation of bylaws that<br />
address the evaluation of neighborhood impacts” (Lollo<br />
& Röhm, 2005, p. 39).<br />
About these items of the Statute, Lollo & Röhm<br />
(2005) produced a material that po<strong>in</strong>ts to the aspects<br />
that are not consi<strong>de</strong>rate <strong>in</strong> some neighborhood impact<br />
studies. It was found that the ma<strong>in</strong> problems occur or<br />
disability legislation or disability of the methodology<br />
applied. Follow<strong>in</strong>g the reason<strong>in</strong>g of Lollo & Rohm<br />
(2005, p. 41), “The consequences of the <strong>de</strong>velopment of<br />
Neighborhood Impact Studies that do not properly<br />
<strong>de</strong>scribe or evaluate the conditions of the project,<br />
neighborhood, or components subject to impact, create<br />
bad consequences <strong>in</strong> four spheres, namely: the<br />
environment, for local resi<strong>de</strong>nts, for the general<br />
population <strong>and</strong> to the public government”.<br />
SUGGESTION TO URBAN IMPACT<br />
EVALUATION<br />
The NIR does not yet have a st<strong>and</strong>ardized methodology;<br />
therefore, the discussion over the horizon of reach of<br />
neighborhood impacts is still open. However, over<br />
recent years <strong>and</strong> given the speed at which technology<br />
evolves, one sees a ten<strong>de</strong>ncy to simplify the procedures<br />
for the method of assess<strong>in</strong>g impacts on urban areas, “<strong>in</strong><br />
or<strong>de</strong>r to remove subjectivity from this process <strong>and</strong> thus<br />
shape, streaml<strong>in</strong>e <strong>de</strong>cision-mak<strong>in</strong>g processes” (Lollo &<br />
Röhm, 2005, p. 7).<br />
Lollo <strong>and</strong> Röhm (2005) also argue that this trend is<br />
not unique to the evaluation of neighborhood impacts,<br />
<strong>and</strong> falls with<strong>in</strong> a global context of establish<strong>in</strong>g criteria<br />
for evaluation <strong>and</strong> rank<strong>in</strong>g of environmental impacts as<br />
an alternative to st<strong>and</strong>ardization of the language issue.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.63-73, 2010
66<br />
Gonçalves Neto <strong>and</strong> Lollo<br />
In this sense, the use of Geographic Information<br />
Systems tends to be an excellent alternative to<br />
technological <strong>de</strong>velopment of such activities, because<br />
their use permits, <strong>in</strong> addition to encod<strong>in</strong>g <strong>in</strong>formation<br />
that facilitates their treatment, large sav<strong>in</strong>gs <strong>in</strong> storage<br />
process, updat<strong>in</strong>g <strong>and</strong> analysis data.<br />
The implementation of an urban mo<strong>de</strong>l <strong>in</strong> a<br />
geographic <strong>in</strong>formation system is one of the best ways<br />
to convey to users of the facilities of the urban structure<br />
as the changes could <strong>in</strong>terfere both positively <strong>and</strong><br />
negatively to the <strong>in</strong>habitants of any area <strong>de</strong>pend<strong>in</strong>g on<br />
the facilities that current technology provi<strong>de</strong>s.<br />
An “Urban GIS” can be used as a tool for automation<br />
of the municipal adm<strong>in</strong>istration, therefore, contributes to<br />
urban plann<strong>in</strong>g with <strong>de</strong>scriptive analysis of urban space<br />
which serves to support local government which<br />
<strong>de</strong>m<strong>and</strong>s efficient data transmission due to the grow<strong>in</strong>g<br />
process<strong>in</strong>g public services (Falcoski, 1997).<br />
Falcoski (1997, p. 12) <strong>de</strong>scribes that “<strong>in</strong> general, <strong>in</strong><br />
studies of urban impacts, us<strong>in</strong>g the computational mo<strong>de</strong>l<br />
proposed levels of study <strong>in</strong> the urban scale has<br />
consi<strong>de</strong>rable improvements <strong>in</strong> both macro-analysis well<br />
as <strong>in</strong> micro-analysis, therefore, it is possible to simulate<br />
potential of high <strong>de</strong>nsity, <strong>in</strong>dices to build, compare the<br />
volume <strong>and</strong> verify the proper application of coefficients<br />
<strong>de</strong>term<strong>in</strong>ed by law <strong>and</strong> can create comparisons of urban<br />
mo<strong>de</strong>ls <strong>de</strong>sirable, <strong>and</strong> likely present. With regards to<br />
urban <strong>de</strong>sign, the emphasis is on mechanisms of<br />
regulatory control <strong>and</strong> management (monitor<strong>in</strong>g) of<br />
dynamic spatial <strong>in</strong>formation”.<br />
DATA COLLECTION<br />
The procedure for obta<strong>in</strong><strong>in</strong>g the data followed the<br />
proposal prepared by Lollo (2006), which suggests the<br />
<strong>de</strong>limitation of areas of <strong>in</strong>fluence through the process of<br />
comparison matrix between the variables selected for<br />
the study. Accord<strong>in</strong>g to Leopold et al. (1971) apud<br />
Lollo (2006, p. 173), the matrix system st<strong>and</strong>s out<br />
among the available techniques of impact evaluation<br />
“due to the agility, flexibility <strong>and</strong> simplicity that allows<br />
the survey <strong>and</strong> impact assessment. It’s also wi<strong>de</strong>ly used<br />
when try<strong>in</strong>g to make i<strong>de</strong>ntifications <strong>and</strong> assessments of<br />
environmental impacts if it is agile”. The basic proposal<br />
of an “impact matrix” (Lollo, 2006) suggested <strong>in</strong> this<br />
study is to cross the proposed actions with<br />
environmental factors, <strong>and</strong> assign these crosses<br />
(characteriz<strong>in</strong>g impacts urban) values that represent the<br />
relative importance.<br />
Aim<strong>in</strong>g to represent the <strong>in</strong>fluences of the<br />
<strong>de</strong>velopment phase of the project impacts <strong>in</strong> the urban<br />
environment <strong>and</strong> consi<strong>de</strong>r<strong>in</strong>g the peculiarities of each of<br />
these steps, <strong>in</strong> this work, it was chosen to follow the<br />
mo<strong>de</strong>l proposed by Lollo (2006) who dist<strong>in</strong>guished<br />
stage of <strong>de</strong>velopment where the impact occurs <strong>in</strong> or<strong>de</strong>r<br />
to properly weigh the importance of each impact.<br />
The <strong>de</strong>limitation of the area of <strong>in</strong>fluence of the<br />
companies is extremely important for correct evaluation<br />
of the effects stemm<strong>in</strong>g from neighborhood impacts <strong>and</strong><br />
implications of conflict <strong>in</strong> urban l<strong>and</strong> use, <strong>and</strong> <strong>de</strong>pends,<br />
accord<strong>in</strong>g Lollo (2006, p. 20): “(1) conditions of the<br />
space, (2) the occupation <strong>in</strong> question, (3) impact<br />
analysis”.<br />
Lollo & Gonçalves Neto (2006) also highlight the<br />
importance of a thorough assessment of <strong>in</strong>terventions<br />
<strong>and</strong> the means consi<strong>de</strong>red to establish specific criteria<br />
for <strong>de</strong>f<strong>in</strong><strong>in</strong>g the area of <strong>in</strong>fluence of each factor.<br />
Accord<strong>in</strong>g Lollo (2006, p. 20.), “the <strong>de</strong>term<strong>in</strong><strong>in</strong>g factors<br />
for <strong>de</strong>f<strong>in</strong><strong>in</strong>g the area of <strong>in</strong>fluence are those related to<br />
traffic <strong>and</strong> park<strong>in</strong>g issues because they are the most farreach<strong>in</strong>g”.<br />
For the quantification of the impacts were possible to<br />
be exam<strong>in</strong>ed, was <strong>de</strong>signed a questionnaire based on<br />
studies of Lollo (2006), able to quantify the variables<br />
regar<strong>de</strong>d as quality by implement<strong>in</strong>g a new<br />
<strong>de</strong>velopment <strong>in</strong> the urban area.<br />
The step of the valuation of the impacts i<strong>de</strong>ntified<br />
<strong>and</strong> listed <strong>in</strong> the questionnaire established the relative<br />
values for each result of each impact consi<strong>de</strong>r<strong>in</strong>g their<br />
or<strong>de</strong>r (direct or <strong>in</strong>direct), magnitu<strong>de</strong> (high, medium <strong>and</strong><br />
low) <strong>and</strong> duration (temporary or permanent) shown <strong>in</strong><br />
Table 1. For each of the conditions <strong>de</strong>scribed were<br />
established values (weights) that consi<strong>de</strong>r the relative<br />
importance of impact, allow<strong>in</strong>g the assessment of<br />
impacts (Lollo, 2006).<br />
From this, it was possible to establish a scale for<br />
quantification of the data that were collected <strong>in</strong> the field<br />
by means of questionnaires.<br />
The field survey was the option to ref<strong>in</strong>e the data<br />
volume, because the experiments reported <strong>in</strong> the work<br />
of Lollo (2006) <strong>de</strong>monstrated that <strong>in</strong><strong>de</strong>pen<strong>de</strong>nt of<br />
cooperation by the companies (<strong>in</strong> this case,<br />
supermarkets), some k<strong>in</strong>d of <strong>in</strong>formation was not<br />
reported, which required data collection by visit<strong>in</strong>g the<br />
field.<br />
Table 1. Values adopted for quantify<strong>in</strong>g the impacts i<strong>de</strong>ntified<br />
Classification<br />
Classes<br />
Values<br />
Or<strong>de</strong>r<br />
Direct 3<br />
Indirect 1<br />
High 3<br />
Magnitu<strong>de</strong><br />
Medium 2<br />
Low 1<br />
Duration<br />
Permanent<br />
3<br />
Temporary<br />
1<br />
Source: Lollo (2006).<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.63-73, 2010
Gonçalves Neto <strong>and</strong> Lollo<br />
67<br />
Table 2. Schematic representation of the structure of the matrix of impacts used<br />
Planned<br />
Phase of Company<br />
Impact Available Components Measures Proposed<br />
Intervention<br />
Info Info I+ H+ M- D+ Info<br />
Info Info T+ D- L+ P+ Info<br />
Plann<strong>in</strong>g<br />
Info Info M+ P- T- I- Info<br />
Info Info H- T+ I+ M- Info<br />
Info Info P+ I+ L+ D+ Info<br />
Info Info H+ M T+ P- Info<br />
Info Info D- T+ H+ I- Info<br />
Construction/Adaptation<br />
Info Info L- D- P+ H- Info<br />
Info Info I+ P+ D+ L- Info<br />
Info Info P- H- I- T+ Info<br />
Info Info P- L+ M+ D- Info<br />
Info Info M- I+ T- H- Info<br />
Operation<br />
Info Info D+ H+ P- M- Info<br />
Info Info H- L- T- I- Info<br />
Info Info T- I+ H+ L+ Info<br />
Source: Lollo (2006).<br />
Info Info M+ P+ L- D- Info<br />
Thus, based on the proposed system were consi<strong>de</strong>red<br />
three steps: (1) Plann<strong>in</strong>g, (2) Construction / Adaptation /<br />
Occupancy (as appropriate), <strong>and</strong> (3) Operation, each<br />
represent<strong>in</strong>g different possibility of generat<strong>in</strong>g potential<br />
impacts as shown below.<br />
Still, accord<strong>in</strong>g with Lollo (2006, p. 9): “The<br />
structure <strong>in</strong> module of the matrix gives two clear<br />
advantages. The first is that it allows the impacts are<br />
assessed at each stage, allow<strong>in</strong>g the discussion of the<br />
importance that each phase represents the environmental<br />
quality of the neighborhood, <strong>and</strong> provid<strong>in</strong>g alternatives<br />
to dist<strong>in</strong>guish between bus<strong>in</strong>esses that represent<br />
construction, those result<strong>in</strong>g from adaptations or<br />
extensions to build<strong>in</strong>gs already exist. This situation also<br />
allows treat equally the two situations <strong>de</strong>scribed<br />
(construction × adaptation/upgra<strong>de</strong>), the impacts<br />
generated by the second group are <strong>de</strong>spised. In case of<br />
upgra<strong>de</strong>/expansion, such a proposal still allows only<br />
<strong>in</strong>terventions are valued consistent with the process”.<br />
Aim<strong>in</strong>g to represent the <strong>in</strong>fluences of the<br />
<strong>de</strong>velopment phase of the impacts i<strong>de</strong>ntified <strong>in</strong> the<br />
project, <strong>and</strong> consi<strong>de</strong>r<strong>in</strong>g the peculiarities of each of<br />
these steps, we chose to dist<strong>in</strong>guish the stage of<br />
<strong>de</strong>velopment where the impact occurs <strong>in</strong> or<strong>de</strong>r to<br />
properly weigh the importance of each impact.<br />
The structure <strong>in</strong>clu<strong>de</strong>s the categories of <strong>in</strong>formation<br />
(aspects of the occupation, its consequences,<br />
environmental components assessed, mitigation<br />
measures <strong>and</strong> compensation arrangements) <strong>in</strong> columns,<br />
while the actions themselves are <strong>de</strong>scribed <strong>and</strong> their<br />
impacts are assessed <strong>in</strong> rows of the matrix.<br />
DATA PROCESSING<br />
This phase <strong>in</strong>clu<strong>de</strong>d the selection <strong>and</strong> ref<strong>in</strong>ement of<br />
<strong>in</strong>formation obta<strong>in</strong>ed <strong>and</strong> <strong>de</strong>f<strong>in</strong>ed <strong>in</strong> the stage of data<br />
collection, <strong>in</strong> addition to proper storage of <strong>in</strong>formation<br />
collected <strong>in</strong> a database compatible with the Geographic<br />
Information System from the referenc<strong>in</strong>g of the<br />
enterprises to enable spatial analysis.<br />
Such results of evaluation of responses obta<strong>in</strong>ed by<br />
the use of the questionnaires were fed <strong>in</strong>to a database <strong>in</strong><br />
Access <strong>and</strong> subsequently <strong>in</strong>serted <strong>in</strong>to the environment<br />
of ESRI ArcGIS 9 which <strong>in</strong>terface controls <strong>and</strong><br />
<strong>h<strong>and</strong>l<strong>in</strong>g</strong> are similar to others st<strong>and</strong>ards software of<br />
eng<strong>in</strong>eer<strong>in</strong>g, facilitat<strong>in</strong>g data manipulation <strong>and</strong> vectors.<br />
The structur<strong>in</strong>g of the data followed the proposal<br />
already <strong>in</strong>clu<strong>de</strong>d <strong>in</strong> the work Mart<strong>in</strong>etti (2006), Sacute<br />
(2006) <strong>and</strong> Gonçalves Neto (2006) which related the<br />
type, the relationship of the work, time <strong>and</strong> <strong>in</strong>tensity of<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.63-73, 2010
68<br />
Gonçalves Neto <strong>and</strong> Lollo<br />
activity, but was kept the same geographic <strong>in</strong>formation<br />
system.<br />
The selection of companies whose impacts would be<br />
consi<strong>de</strong>red <strong>in</strong> this proposal was based on the criteria<br />
proposed by Lollo (2006, p. 8) <strong>de</strong>scribed as follows:<br />
“(1) can be i<strong>de</strong>ntified neighborhood impacts occurred<br />
by the time of <strong>de</strong>ployment of bus<strong>in</strong>ess or dur<strong>in</strong>g the<br />
execution of their activities; (2) companies should have<br />
characteristics that allow their classification as<br />
supermarkets, <strong>and</strong> (3) firms should be <strong>in</strong>stalled <strong>in</strong> urban<br />
areas”. Figure 1 shows the companies selected on this<br />
criterion.<br />
The survey of the surround<strong>in</strong>gs of each company on<br />
the map of the city was through sketches prepared<br />
dur<strong>in</strong>g field visits. The visits to the site of each<br />
company had as its <strong>in</strong>itial objective, to answer some<br />
questions that employees of these establishments did not<br />
know or did not respond, but was also serv<strong>in</strong>g to<br />
provi<strong>de</strong> a more reliable location of each company, with<br />
the realization of these sketches.<br />
The simplicity of spatial representation of<br />
neighborhood impacts (<strong>de</strong>f<strong>in</strong>ed area of <strong>in</strong>fluence<br />
surround<strong>in</strong>g each company consi<strong>de</strong>red) significantly<br />
facilitated the structur<strong>in</strong>g of the data mo<strong>de</strong>l <strong>in</strong> GIS.<br />
Thus, tak<strong>in</strong>g up the company as spatial reference data,<br />
the data structure <strong>in</strong> ArcGIS was composed of three<br />
fundamental layers: “Streets”, “Po<strong>in</strong>ts” <strong>and</strong> “Buffers”.<br />
The spatial representation of the companies was<br />
done by us<strong>in</strong>g po<strong>in</strong>ts, based <strong>in</strong> sketches ma<strong>de</strong> <strong>in</strong> field.<br />
The category “Companies” (with mo<strong>de</strong>l data object<br />
type) absorbed the <strong>in</strong>formation about the spatial location<br />
of each company (represented by po<strong>in</strong>ts referenced to<br />
the network of streets), <strong>and</strong> a set of data (as data table)<br />
with all <strong>in</strong>formation obta<strong>in</strong>ed <strong>in</strong> the surveys, which<br />
allowed the evaluation of neighborhood impacts.<br />
The category that has been the basis used for<br />
mapp<strong>in</strong>g the location of firms <strong>and</strong> evaluation of their<br />
areas of <strong>in</strong>fluence was named “Network Logradouros”<br />
(with cadastral data mo<strong>de</strong>l) <strong>and</strong> corresponds to the<br />
spatial representation of the axes of public areas (roads<br />
centers or centerl<strong>in</strong>es) with <strong>de</strong>tails of their name <strong>and</strong><br />
direction of traffic flow.<br />
The database relat<strong>in</strong>g to this category was already<br />
available <strong>in</strong> the laboratory of the Center for GIS (NGeo<br />
- UFSCar), which is used for other aca<strong>de</strong>mic work <strong>and</strong><br />
its upgra<strong>de</strong> has performed steadily.<br />
The category <strong>de</strong>signated as buffers (thematic data<br />
mo<strong>de</strong>l) has the spatial representation of areas of<br />
<strong>in</strong>fluence of the companies studied. A circular buffer<br />
was the choice to st<strong>and</strong>ardize the treatment given to the<br />
spatial area of <strong>in</strong>fluence of all bus<strong>in</strong>esses <strong>and</strong> consi<strong>de</strong>r<br />
that <strong>in</strong> this case, the <strong>in</strong>tensity of impacts can be<br />
consi<strong>de</strong>red uniform <strong>in</strong> all directions.<br />
Fig. 1 Companies selected.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.63-73, 2010
Gonçalves Neto <strong>and</strong> Lollo<br />
69<br />
The location of each company was represented on<br />
the map us<strong>in</strong>g coord<strong>in</strong>ates of the UTM system which<br />
<strong>in</strong>dicates the central po<strong>in</strong>t of the frontage portion<br />
which belongs to the company.<br />
After the conclusion of this Step, the sum of the<br />
weights of the issues data were transported to an<br />
Access base <strong>and</strong> after to ArcGIS. Thus, the goal was<br />
reached by the proposed matrix method: analyze not<br />
just isolated to quantify the impact of a s<strong>in</strong>gle<br />
company, but also enable its comparison with another<br />
company <strong>and</strong> also compare the results with the set of<br />
all companies with respect to the variable of urban<br />
<strong>in</strong>frastructure.<br />
Regard<strong>in</strong>g the classification of neighborhood<br />
impacts, Lollo (2006) proposed the separation of<br />
variables directly related to the impact on urban<br />
<strong>in</strong>frastructure to be evaluated <strong>in</strong> the array: high<br />
<strong>de</strong>nsity population, realty valorization, ventilation<br />
<strong>and</strong> light<strong>in</strong>g, cultural heritage <strong>and</strong> urban<br />
transformations, traffic generation, <strong>de</strong>m<strong>and</strong> for public<br />
transport <strong>and</strong> urban <strong>and</strong> community facilities.<br />
The next phase (impact evaluation) correspon<strong>de</strong>d<br />
to the treatment <strong>in</strong> GIS, of the <strong>in</strong>formation that<br />
com<strong>in</strong>g from the previous steps, so as to allow three<br />
types of analysis: (i) from impact isolated<br />
(<strong>in</strong>terpretation of the causes <strong>and</strong> consequences of<br />
each impact consi<strong>de</strong>red, for each of the areas<br />
assessed); (ii) <strong>in</strong> groups of impacts, accord<strong>in</strong>g to the<br />
classification already given (check<strong>in</strong>g for all areas<br />
assessed, the existence of a group that is more<br />
significant impacts), <strong>and</strong> (iii) the comb<strong>in</strong>ation of<br />
several related impacts (assess the full extent of<br />
impacts generated by each area, to classify the areas<br />
accord<strong>in</strong>g to the total impact generated by each one<br />
<strong>and</strong> see if there are portions of the urban area more<br />
weaker on this type of proposal occupancy).<br />
ANALYSIS AND RESULTS<br />
The work done by Lollo & Gonçalves Neto (2006),<br />
showed that urban impacts may have a direct<br />
relationship to the size of the build<strong>in</strong>g, but <strong>in</strong> the case<br />
of supermarkets <strong>and</strong> even more specifically on the<br />
issue of urban impacts that reality exists, but not<br />
shown be the largest contributor to the <strong>in</strong>tensity of<br />
impacts, partly due to the impossibility of assess<strong>in</strong>g<br />
post-<strong>de</strong>ployment of temporary impacts <strong>de</strong>pend<strong>in</strong>g on<br />
the size of the build<strong>in</strong>g.<br />
The data obta<strong>in</strong>ed showed the prevalence of some<br />
types of impact, especially those related to real estate<br />
appreciation, the high <strong>de</strong>nsity <strong>and</strong> traffic flow.<br />
Table 3 shows the results obta<strong>in</strong>ed through the<br />
use of questionnaires. The companies were i<strong>de</strong>ntified<br />
by symbols <strong>and</strong> each column represents a type of<br />
impact associated with putt<strong>in</strong>g the build<strong>in</strong>g <strong>in</strong> the<br />
urban system.<br />
This table was <strong>in</strong>serted <strong>in</strong>to the database after the<br />
process<strong>in</strong>g of the results obta<strong>in</strong>ed <strong>in</strong> the field <strong>in</strong><br />
numbers <strong>and</strong> <strong>in</strong> the GIS environment, allow<strong>in</strong>g the<br />
Table 3. Urban impacts i<strong>de</strong>ntified <strong>in</strong> the area<br />
ID<br />
Question<br />
Super 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31<br />
M-01 × × × × × × × × × × × × ×<br />
G-01 × × × × × × × × × × × × × × × × ×<br />
G-02 × × × × × × × × × × × × × × ×<br />
G-03 × × × × × × × × × × × × × ×<br />
G-04 × × × × × × × × × × × × × × × ×<br />
M-02 × × × × × × × × × × × × ×<br />
M-03 × × × × × × × × × × × × × ×<br />
P-01 × × × × × × × × × × × × × × × ×<br />
G-05 × × × × × × × × × ×<br />
M-04 × × × × × × × × × × × × × × × ×<br />
G-06 × × × × × × × × × × × × ×<br />
M-05 × × × × × × × × × × × × ×<br />
M-06 × × × × × × × × × × ×<br />
G-07 × × × × × × × × × × × × × × × ×<br />
G-08 × × × × × × × × × × × × × × × × ×<br />
M-07 × × × × × × × × × × × ×<br />
ID SUPER (Supermarket ID) – P: Small Companies; M: Medium Companies; G: Large Companies; Questions: 1 to 10 – Real Estate<br />
Valuations; 11 to 14 – Urban L<strong>and</strong>scapes; 15 to 19 – Public Heritages; 20 to 24 – Traffic Volumes; 25 to 31 – Urban Facilities.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.63-73, 2010
Gonçalves Neto <strong>and</strong> Lollo<br />
70<br />
graphical representation <strong>and</strong> spatial impacts. It is<br />
emphasized here a curious fact with respect to the<br />
results by the field survey. Observ<strong>in</strong>g Table 3, it is<br />
clear that the largest contribution to the amount of<br />
work has been grouped <strong>in</strong>to two ma<strong>in</strong> cores of the<br />
components of analysis: the real estate valuation <strong>and</strong><br />
traffic volumes. The first component of this<br />
observation is directly related to the shift of people to<br />
nearby supermarkets.<br />
In some areas <strong>in</strong> the municipality, the supermarket<br />
was the factor responsible for the attraction of<br />
people, given the existence of some as the Dotto<br />
supermarket that exists <strong>in</strong> the city more than 40<br />
years, this season which saw the promotion of<br />
technological growth of the country, promoted by the<br />
then presi<strong>de</strong>nt, Juscel<strong>in</strong>o Kubitschek, <strong>and</strong> his slogan<br />
of “grow 50 years <strong>in</strong> 5”.<br />
With the com<strong>in</strong>g of mult<strong>in</strong>ationals <strong>in</strong>to the<br />
country, the Brazilian citizen could live further away<br />
from the workplace, therefore the second component<br />
of the observation results, the acquisition of a<br />
material as a vehicle has become easier, moreover,<br />
the urban <strong>de</strong>velopment grown creat<strong>in</strong>g more<br />
pathways to l<strong>in</strong>k the neighborhoods to the<br />
commercial centers.<br />
Therefore, there was a swell<strong>in</strong>g population which<br />
had the opportunity to resi<strong>de</strong> <strong>in</strong> areas far beyond the<br />
commercial centers evi<strong>de</strong>nced by Figs 2 <strong>and</strong> 3.<br />
Fig. 2 Representation <strong>in</strong> "pie charts" of the relationship between the groups “real estate valuation” <strong>and</strong> the sum total of variables.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.63-73, 2010
Gonçalves Neto <strong>and</strong> Lollo<br />
71<br />
Fig. 3 Representation <strong>in</strong> “pie charts” of the relationship between the groups “traffic volume” <strong>and</strong> the sum total of variables.<br />
About the <strong>in</strong>fluence areas of selected supermarkets,<br />
it was observed that <strong>in</strong> several cases, there is the<br />
<strong>in</strong>terpenetration or overlapp<strong>in</strong>g of the l<strong>in</strong>es (Fig. 4).<br />
Accord<strong>in</strong>g Chasco Yrigoyen & Uceta (1998) apud<br />
Sharma et al. (2006), “<strong>in</strong>fluence area or commercial<br />
area of a municipality or a commercial equipment is the<br />
area that has a strong tra<strong>de</strong> <strong>de</strong>pen<strong>de</strong>nce on the place of<br />
study” (p. 112).<br />
The ma<strong>in</strong> occurrences of this group<strong>in</strong>g of <strong>in</strong>fluence<br />
areas appeared frequently <strong>in</strong> the central part of the<br />
municipality (which has different uses of urban l<strong>and</strong>)<br />
<strong>and</strong> near to the Cida<strong>de</strong> Jardim neighborhood which has<br />
a resi<strong>de</strong>ntial character.<br />
This overlap of buffers <strong>de</strong>monstrates the <strong>in</strong>tention of<br />
different companies reach<strong>in</strong>g geographically the same<br />
customer. This can be seen graphically by the existence<br />
of <strong>in</strong>tersections when there is conflict <strong>in</strong> different<br />
spheres of <strong>in</strong>fluence.<br />
In these regions confirmed that the urban impacts<br />
cannot be treated separately because, <strong>de</strong>pend<strong>in</strong>g on the<br />
proximity of the objects of study, the variables of each<br />
company began to <strong>in</strong>terfere with the behavior of others.<br />
The Master Plan of the city of São Carlos does not<br />
refer to possible <strong>in</strong>terferences grouped <strong>in</strong> the period<br />
before the implementation of the venture, <strong>de</strong>al<strong>in</strong>g with<br />
the variables for the study of urban isolation without<br />
admitt<strong>in</strong>g the existence of <strong>in</strong>tegrat<strong>in</strong>g factors for the use<br />
of urban facilities.<br />
This is a problem that occurs <strong>in</strong> the most of the laws<br />
about the subject, because the municipal government<br />
has not been able to produce some material about NIS.<br />
Moreover the NIS does not have a <strong>de</strong>f<strong>in</strong>ite pattern of<br />
fe<strong>de</strong>ral legislation.<br />
We must consi<strong>de</strong>r the efforts ma<strong>de</strong> by municipal<br />
authorities to tread a path of urban sprawl without a<br />
back<strong>in</strong>g of the Fe<strong>de</strong>ration, but the results showed a real<br />
need to reach a st<strong>and</strong>ard that consi<strong>de</strong>rs the variables of<br />
<strong>in</strong>tegration of urban occupation, which must be ad<strong>de</strong>d<br />
the mitigation measures <strong>in</strong> the period prior to<br />
<strong>in</strong>stallation of new <strong>de</strong>velopments.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.63-73, 2010
Gonçalves Neto <strong>and</strong> Lollo<br />
72<br />
Fig. 4 Influence areas of supermarkets.<br />
Fig. 5 Representation <strong>in</strong> proportional symbols of the relationship between the variables “supplier vehicles” <strong>and</strong> the sum of variables.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.63-73, 2010
Gonçalves Neto <strong>and</strong> Lollo<br />
73<br />
CONCLUSIONS<br />
The use of st<strong>and</strong>ardized questionnaire on field<br />
surveys, attached to the <strong>in</strong>formation provi<strong>de</strong>d by the<br />
enterprises <strong>and</strong> ad<strong>de</strong>d the field visits, allowed the<br />
i<strong>de</strong>ntification of impacts on the study.<br />
The ventures consolidated prior to the draft<strong>in</strong>g of<br />
the City Statute were those who obta<strong>in</strong>ed the highest<br />
measurements of impacts because the surround<strong>in</strong>g<br />
urban growth <strong>in</strong> the region, unlike the supermarkets<br />
built or acquired after the Fe<strong>de</strong>ral Law, which have<br />
found a <strong>de</strong>veloped urban structure <strong>and</strong> legal<br />
obligations to be met.<br />
The urban sprawl around the supermarkets was<br />
aggravated by the excessive number of cars<br />
purchased by local resi<strong>de</strong>nts <strong>in</strong> recent years <strong>and</strong><br />
consequently <strong>in</strong>crease of the flow of vehicles on<br />
roads which were not scaled enough to support such<br />
a volume of traffic.<br />
This new condition prevents the <strong>de</strong>livery of<br />
products to the supermarket is done effectively,<br />
which contributes directly to the f<strong>in</strong>al assessment of<br />
results, therefore, the variables related to the flow of<br />
vehicles were the largest contributors <strong>in</strong> the f<strong>in</strong>al<br />
results.<br />
Figure 5 shows the quantification of the variable<br />
related to the logistical vehicles com<strong>in</strong>g from<br />
suppliers.<br />
The Master Plan of the city of São Carlos has a<br />
pretty advanced legislation regard<strong>in</strong>g the NIS, but<br />
still treats the components of analysis alone neither a<br />
st<strong>and</strong>ard method for the quantification of the impacts<br />
(of any type), because is backed by the Statute of the<br />
City, which also lacks a st<strong>and</strong>ard for <strong>de</strong>term<strong>in</strong><strong>in</strong>g the<br />
variables nee<strong>de</strong>d for the preparation of any NIR <strong>in</strong><br />
the municipality.<br />
F<strong>in</strong>ally, the NIS is still evolv<strong>in</strong>g <strong>and</strong> needs to be<br />
<strong>de</strong>epened <strong>in</strong> or<strong>de</strong>r to provi<strong>de</strong> conditions to <strong>de</strong>velop<br />
methods that consi<strong>de</strong>r the <strong>in</strong>ter<strong>de</strong>pen<strong>de</strong>nce of<br />
variables <strong>and</strong> its <strong>in</strong>tegration with the urban<br />
environment.<br />
The geographic <strong>in</strong>formation system <strong>in</strong> this case,<br />
was the best option to generate relationships between<br />
variables <strong>and</strong> their spatial representation effectively.<br />
Clearer legislation, sett<strong>in</strong>g st<strong>and</strong>ards, coupled with<br />
a GIS which relates to the <strong>in</strong>tegration of urban<br />
elements contribute to the improvement <strong>in</strong> <strong>de</strong>cisionmak<strong>in</strong>g<br />
by the government.<br />
REFERENCES<br />
Brasil. Law 10.257, July 10 th (2001) Regulates arts. 182 <strong>and</strong> 183<br />
of the Constitution set out the general gui<strong>de</strong>l<strong>in</strong>es of urban<br />
policy <strong>and</strong> other matters. Official Gazette, 10 July 2001.<br />
Falcoski, L.A.N. (1997) Dimensões Morfológicas <strong>de</strong><br />
Desempenho: Instrumentos Urbanísticos <strong>de</strong> Planejamento e<br />
Desenho Urbano. Doctoral Dissertation submitted to the<br />
Graduate Program <strong>in</strong> Environmental Structures College of<br />
Architecture <strong>and</strong> Urbanism, University of São Paulo, USP,<br />
São Paulo, Brazil, 1997.<br />
Gonçalves Neto, P.S. & Röhm, S.A. (2006) Avaliação <strong>de</strong><br />
Impactos na Infra-estrutura Urbana Decorrentes da<br />
Implantação <strong>de</strong> Empresas do Pólo <strong>de</strong> Alta Tecnologia <strong>de</strong> São<br />
Carlos.<br />
Gonçalves Neto, P.S. (2006) Estudo <strong>de</strong> impacto <strong>de</strong> viz<strong>in</strong>hança<br />
por meio do planejamento por <strong>de</strong>sempenho na Região Central<br />
<strong>de</strong> São Carlos. Trabalho <strong>de</strong> Graduação Integrado, UFSCar,<br />
São Carlos.<br />
Gregori, I.C. (2004) Mesa Temática: Cida<strong>de</strong>s e Gestão Urbana,<br />
O Estatuto da Cida<strong>de</strong>: perspectivas e <strong>de</strong>safios dos novos<br />
<strong>in</strong>strumentos <strong>de</strong> reforma urbana. Universida<strong>de</strong> <strong>de</strong> Santa Cruz<br />
do Sul – RS.<br />
IBGE – Brazilian Institue of Geography <strong>and</strong> Statistics (2009)<br />
Cida<strong>de</strong>s@: <strong>in</strong>formations from data base of IBGE,<br />
Rio <strong>de</strong> Janeiro, 2009, Available <strong>in</strong> http://www.ibge.gov.br/<br />
cida<strong>de</strong>sat/<strong>de</strong>fault2.php Accessed on 14 August 2009.<br />
Leão-Aguiar, L., Campos, L.R., Moraes, L.R.S. & Borja, P.C.<br />
(2005) Estudo <strong>de</strong> Impacto <strong>de</strong> Viz<strong>in</strong>hança: Instrumento <strong>de</strong><br />
Gestão Pública para a Cida<strong>de</strong> <strong>de</strong> Salvador.<br />
L<strong>in</strong>dau, L.A., Diógenes, M.C. & P<strong>in</strong>to, A.B. (2009)<br />
Quantificação dos impactos <strong>de</strong> pólos geradores <strong>de</strong> tráfego.<br />
Available <strong>in</strong>: http://www.producao.ufrgs.br/arquivos/<br />
discipl<strong>in</strong>as/412_impactos_polo_gerador_versao_l<strong>in</strong>dau_rev.do<br />
c Accessaed <strong>in</strong> 10 September 2009.<br />
Lollo, J.A. (2006) Utilização <strong>de</strong> Sistema <strong>de</strong> Informações<br />
Geográficas em Estudo <strong>de</strong> Impacto <strong>de</strong> Viz<strong>in</strong>hança: o caso do<br />
Pólo Tecnológico <strong>de</strong> São Carlos.<br />
Lollo, J.A. & Röhm, S.A. (2005) Proposta <strong>de</strong> Matriz para<br />
Levantamento e Avaliação <strong>de</strong> Impactos <strong>de</strong> Viz<strong>in</strong>hança.<br />
Mart<strong>in</strong>etti, T.H. & Röhm, S.A. (2006) Avaliação <strong>de</strong> Impactos <strong>de</strong><br />
Viz<strong>in</strong>hança no Meio Físico us<strong>and</strong>o Sistema <strong>de</strong> Informações<br />
Geográficas – o caso <strong>de</strong> Pólo <strong>de</strong> Alta Tecnologia <strong>de</strong> São<br />
Carlos (SP).<br />
Mol<strong>in</strong>a, V. (2008) Efeito do Porte do Município em Impactos <strong>de</strong><br />
Viz<strong>in</strong>hança Gerados por Supermercados. Research Plan<br />
forwar<strong>de</strong>d to PPGEU/UFSCar, Universida<strong>de</strong> Fe<strong>de</strong>ral <strong>de</strong> São<br />
Carlos, São Carlos, 2008, 3–8.<br />
Moreira, A.C.M.L. (1999) Parâmetros para elaboração do<br />
relatório <strong>de</strong> impacto <strong>de</strong> viz<strong>in</strong>hança. Magaz<strong>in</strong>e of the Graduate<br />
Program <strong>in</strong> Architecture <strong>and</strong> Urbanism of FAU/USP, 7, p.<br />
107–118.<br />
Sacute, B.C.A. & Röhm, S.A. (2006) Avaliação <strong>de</strong> Impactos<br />
Urbanísticos Decorentes da Implantação <strong>de</strong> Empresas do Polo<br />
<strong>de</strong> Alta Tecnologia <strong>de</strong> São Carlos.<br />
Silva, L.R., Kneib, E.C. & Silva, P.C.M. (2006) Proposta<br />
metodológica para <strong>de</strong>f<strong>in</strong>ição da área <strong>de</strong> <strong>in</strong>fluência <strong>de</strong> pólos<br />
geradores <strong>de</strong> viagens consi<strong>de</strong>r<strong>and</strong>o características próprias e<br />
aspectos d<strong>in</strong>âmicos <strong>de</strong> seu entorno. Article Published <strong>in</strong> the J.<br />
Civil Engng, 27, 111–126.<br />
Unifra (2009) Manual <strong>de</strong> Normas para Apresentação <strong>de</strong><br />
Trabalhos Científicos. Available <strong>in</strong> http://www.unifra.br/<br />
cursos/economia/downloads/Normas_UNIFRA_versao1.pdf<br />
Accessed <strong>in</strong> 23 December 2009.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.63-73, 2010
L<strong>in</strong>s, Cruz, Vieira, C. Neto <strong>and</strong> Mir<strong>and</strong>a<br />
74<br />
J U E E<br />
Journal of Urban <strong>and</strong> Environmental<br />
Eng<strong>in</strong>eer<strong>in</strong>g, v.4, n.2, p.74-80<br />
ISSN 1982-3932<br />
doi: 10.4090/juee.2010.v4n2.074080<br />
Journal of Urban <strong>and</strong><br />
Environmental Eng<strong>in</strong>eer<strong>in</strong>g<br />
www.journal-uee.org<br />
DETERMINING INDICATORS OF URBAN HOUSEHOLD<br />
WATER CONSUMPTION THROUGH MULTIVARIATE<br />
STATISTICAL TECHNIQUE<br />
Gledsneli M.L. L<strong>in</strong>s 1 *, Walter S. Cruz 2 , Zedna M.C.L.Vieira 3 , Francisco A.C. Neto 4 <strong>and</strong><br />
Érico A.A. Mir<strong>and</strong>a 5<br />
1 Natural Resources Graduation Program, Fe<strong>de</strong>ral University of Camp<strong>in</strong>a Gr<strong>and</strong>e, Brazil<br />
2 Department of Civil Eng<strong>in</strong>eer<strong>in</strong>g, Fe<strong>de</strong>ral University of Camp<strong>in</strong>a Gr<strong>and</strong>e, Brazil<br />
3<br />
CAPES/National Postdoctoral Program Scholar, Fe<strong>de</strong>ral University of Camp<strong>in</strong>a Gr<strong>and</strong>e, Brazil<br />
4 Civil Eng<strong>in</strong>eer<strong>in</strong>g Un<strong>de</strong>rgraduate Stu<strong>de</strong>nt, Fe<strong>de</strong>ral University of Camp<strong>in</strong>a Gr<strong>and</strong>e, Brazil<br />
5 Department of Economy, Fe<strong>de</strong>ral University of Camp<strong>in</strong>a Gr<strong>and</strong>e, Brazil<br />
Received 09 September 2010; received <strong>in</strong> revised form 30 November 2010; accepted 15 December 2010<br />
Abstract:<br />
Keywords:<br />
Water has a <strong>de</strong>cisive <strong>in</strong>fluence on populations’ life quality – specifically <strong>in</strong> areas like<br />
urban supply, dra<strong>in</strong>age, <strong>and</strong> effluents treatment – due to its sound impact over public<br />
health. Water rational use constitutes the greatest challenge faced by water <strong>de</strong>m<strong>and</strong><br />
management, ma<strong>in</strong>ly with regard to urban household water consumption. This makes it<br />
important to <strong>de</strong>velop researches to assist water managers <strong>and</strong> public policy-makers <strong>in</strong><br />
plann<strong>in</strong>g <strong>and</strong> formulat<strong>in</strong>g water <strong>de</strong>m<strong>and</strong> measures which may allow urban water<br />
rational use to be met. This work utilized the multivariate techniques Factor Analysis<br />
<strong>and</strong> Multiple L<strong>in</strong>ear Regression Analysis – <strong>in</strong> or<strong>de</strong>r to <strong>de</strong>term<strong>in</strong>e the participation level<br />
of socioeconomic <strong>and</strong> climatic variables <strong>in</strong> monthly urban household consumption<br />
changes – apply<strong>in</strong>g them to two districts of Camp<strong>in</strong>a Gr<strong>and</strong>e city (State of Paraíba,<br />
Brazil). The districts were chosen based on socioeconomic criterion (<strong>in</strong>come level) so<br />
as to evaluate their water consumer’s behavior. A 9-year monthly data series (from year<br />
2000 up to 2008) was utilized, compris<strong>in</strong>g family <strong>in</strong>come, water tariff, <strong>and</strong> quantity of<br />
household connections (economies) – as socioeconomic variables – <strong>and</strong> average<br />
temperature <strong>and</strong> precipitation, as climatic variables. For both the selected districts of<br />
Camp<strong>in</strong>a Gr<strong>and</strong>e city, the obta<strong>in</strong>ed results po<strong>in</strong>t out the variables “water tariff” <strong>and</strong><br />
“family <strong>in</strong>come” as <strong>in</strong>dicators of these district’s household consumption.<br />
Household water consumption, factor analysis, l<strong>in</strong>ear regression, urban water<br />
<strong>de</strong>m<strong>and</strong>, water efficient use<br />
© 2010 Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE). All rights reserved.<br />
* Correspon<strong>de</strong>nce to: Gledsneli M. <strong>de</strong> L. L<strong>in</strong>s, Tel.: +55 83 3310 1157; Fax: + 55 83 3310 1388<br />
E-mail: gl<strong>in</strong>s16@yahoo.com.br<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p. 47-80, 2010
L<strong>in</strong>s, Cruz, Vieira, C. Neto <strong>and</strong> Mir<strong>and</strong>a<br />
75<br />
INTRODUCTION<br />
The access to potable water is one of the most serious<br />
problems that urban centres are fac<strong>in</strong>g today all over<br />
the world. Albeit Brazil is privileged <strong>in</strong> terms of<br />
freshwater resources – it <strong>de</strong>ta<strong>in</strong>s about 14% of the<br />
world’s water availability – this richness is very<br />
unequally distributed among its regions: Amazonia<br />
(North region) concentrates 70%; Central-West<br />
region, 15%; South <strong>and</strong> Southeast regions, 6% each<br />
one; <strong>and</strong> Northeast region, just 3% of these resources.<br />
In relation to the ma<strong>in</strong> consumptive water uses,<br />
Brazilian water resources are shared by irrigation<br />
(63%), human supply (18%), <strong>and</strong> <strong>in</strong>dustry (5%)<br />
(ANA, 2007).<br />
Whether these resources are utilized rationally, i.e.,<br />
with economic efficiency, social equity <strong>and</strong><br />
environmental susta<strong>in</strong>ability, <strong>in</strong> the long run they<br />
could become a competitive advantage which would<br />
put Brazil among those countries high on the human<br />
<strong>de</strong>velopment <strong>in</strong><strong>de</strong>x (Lanna, 2008). Nevertheless,<br />
Brazil’s population <strong>and</strong> economic growth occurred<br />
dur<strong>in</strong>g the 20 th century led to a predatory exploration<br />
of its natural resources <strong>in</strong> general, <strong>and</strong>, particularly, of<br />
its water resources (ANA, 2007).<br />
Thus, along with the problems result<strong>in</strong>g from water<br />
scarcity <strong>in</strong> the Northeast region <strong>and</strong> water pollution <strong>in</strong><br />
the Southeast region, Brazilian’s large urban centres –<br />
like São Paulo e Recife – as well as medium size ones<br />
– like Camp<strong>in</strong>a Gr<strong>and</strong>e, State of Paraíba – have faced<br />
water supply crises which subjected population to<br />
tight ration<strong>in</strong>g measures (Tomaz, 2001). This situation<br />
makes evi<strong>de</strong>nt the need for <strong>in</strong>tegrated water resources<br />
management, <strong>and</strong>, more emphatically, the need for<br />
urban water <strong>de</strong>m<strong>and</strong> management measures.<br />
Water <strong>de</strong>m<strong>and</strong> management (WDM) <strong>in</strong>volves<br />
actions <strong>and</strong> sound methods to push the community <strong>in</strong><br />
the direction of an appropriate use of water thus<br />
reduc<strong>in</strong>g water consumption by the f<strong>in</strong>al user <strong>and</strong>, at<br />
the same time, <strong>in</strong>crement<strong>in</strong>g new consumption habits<br />
that would br<strong>in</strong>g no impairment to comfort <strong>and</strong><br />
hygienic necessities as provi<strong>de</strong>d by exist<strong>in</strong>g systems<br />
(USEPA, 1998). Then, water <strong>de</strong>m<strong>and</strong> management<br />
goes beyond consumption management: rather than a<br />
question of organiz<strong>in</strong>g consumption data <strong>and</strong> of<br />
build<strong>in</strong>g graphics, WDM <strong>in</strong>sists on study<strong>in</strong>g these<br />
data <strong>and</strong> on guarantee<strong>in</strong>g the system feedback (L<strong>in</strong>s &<br />
Ribeiro, 2007).<br />
The need for urban water <strong>de</strong>m<strong>and</strong> management<br />
measures, which <strong>in</strong>clu<strong>de</strong>s public policies for<br />
stimulat<strong>in</strong>g household water consumption efficiency,<br />
exceed<strong>in</strong>gly justifies the <strong>de</strong>velopment of researches<br />
that can <strong>in</strong>dicate paths to address this objective.<br />
With<strong>in</strong> this context, this work aims to analyze<br />
possible relationships between socioeconomic<br />
variables, climatic factors, <strong>and</strong> household water<br />
consumption for two districts of Camp<strong>in</strong>a Gr<strong>and</strong>e,<br />
Paraíba, Brazil, based on the multivariate statistical<br />
techniques Factor Analysis <strong>and</strong> Multiple L<strong>in</strong>ear<br />
Regression Analysis <strong>and</strong> consi<strong>de</strong>r<strong>in</strong>g water<br />
potentiality <strong>and</strong> availability, so as to assist water<br />
managers <strong>and</strong> public policy-makers <strong>in</strong> plann<strong>in</strong>g <strong>and</strong><br />
regulat<strong>in</strong>g water consumption.<br />
STUDY CASE<br />
Study area characterization<br />
Camp<strong>in</strong>a Gr<strong>and</strong>e is the second largest city <strong>in</strong> the State<br />
of Paraíba, with a population of over 380 000<br />
<strong>in</strong>habitants (IBGE, 2010), which is divi<strong>de</strong>d along 49<br />
districts. It is an important educational, <strong>in</strong>dustrial <strong>and</strong><br />
technological centre situated <strong>in</strong> a semiarid region <strong>in</strong><br />
northeastern Brazil. Located <strong>in</strong> the highest portion of<br />
the Borborema Plateau’s oriental cliffs, with altitu<strong>de</strong>s<br />
rang<strong>in</strong>g from 500 to 600 m (average altitu<strong>de</strong> of<br />
551 m), the city presents a semiarid equatorial<br />
climate, average temperatures around 25°C, <strong>and</strong> an<br />
average annual ra<strong>in</strong>fall of 730 mm (PMCG, 2002).<br />
Situated with<strong>in</strong> the limits of the Paraíba River bas<strong>in</strong>,<br />
Camp<strong>in</strong>a Gr<strong>and</strong>e, however, lies closely to the bas<strong>in</strong><br />
boundaries <strong>and</strong>, consequently, is not served by the<br />
ma<strong>in</strong> river. Figure 1 shows the city’s localization.<br />
Fig.1 Maps: Brazil <strong>and</strong> Paraíba (Camp<strong>in</strong>a Gr<strong>and</strong>e city).<br />
Water supply has become a historical problem. In<br />
the period from 1997 to 1999, the severe drought that<br />
fell upon the northeastern region of Brazil aggravated<br />
the critical water storage of the Epitácio Pessoa<br />
Reservoir that supplies the city <strong>and</strong> brought about a<br />
very serious crisis <strong>in</strong> water supply, affect<strong>in</strong>g Camp<strong>in</strong>a<br />
Gr<strong>and</strong>e <strong>and</strong> other five cities <strong>in</strong> Borborema’s region<br />
<strong>and</strong> subject<strong>in</strong>g 500 000 people to tight ration<strong>in</strong>g<br />
measures (54% less water <strong>in</strong> its f<strong>in</strong>al phase) between<br />
November 1998 <strong>and</strong> April 2000 (Rêgo et al., 2000;<br />
Galvão et al., 2001).<br />
In 2001 <strong>and</strong> 2002, dur<strong>in</strong>g the dry season, new<br />
ration<strong>in</strong>g measures had to be adopted motivated by the<br />
very low storage levels <strong>in</strong> the reservoir; <strong>in</strong> 2003, due<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p. 47-80, 2010
L<strong>in</strong>s, Cruz, Vieira, C. Neto <strong>and</strong> Mir<strong>and</strong>a<br />
76<br />
to the same reason, a system of consumption quotas<br />
was adopted, for a few months, for all consumers;<br />
revert<strong>in</strong>g this situation, a consi<strong>de</strong>rable <strong>in</strong>crease <strong>in</strong>flow<br />
<strong>in</strong>to the reservoir, registered <strong>in</strong> January 2004, allowed<br />
the accumulation of the reservoir’s maxim storage<br />
volume (411 686 287 m 3 ) <strong>and</strong> avoi<strong>de</strong>d a new – <strong>and</strong><br />
already planned – water ration<strong>in</strong>g (Vieira et al., 2005).<br />
An analysis of the many factors that contributed to<br />
these crises – the city’s geographic location, the lack<br />
of proper water resources management, the adoption<br />
of ration<strong>in</strong>g policies only at times of impend<strong>in</strong>g crisis,<br />
among others – <strong>in</strong>dicates the necessity for effective<br />
water resources management, particularly consi<strong>de</strong>r<strong>in</strong>g<br />
the <strong>de</strong>m<strong>and</strong> si<strong>de</strong> management <strong>and</strong> its potential for<br />
<strong>in</strong>duc<strong>in</strong>g rational consumption, <strong>in</strong> or<strong>de</strong>r to avoid the<br />
occurrence of new crises <strong>in</strong> the city’s water supply<br />
(Vieira, 2002).<br />
Nevertheless, it is important to emphasize that, <strong>in</strong><br />
spite of these crises’ seriousness <strong>and</strong> except for the<br />
ration<strong>in</strong>g measures, no other water <strong>de</strong>m<strong>and</strong><br />
management <strong>in</strong>itiatives have been taken <strong>in</strong> Camp<strong>in</strong>a<br />
Gr<strong>and</strong>e. Besi<strong>de</strong>s some lack of political will, the need<br />
for knowledge about household consumption<br />
characteristics <strong>and</strong> relationships – which could gui<strong>de</strong><br />
public policies formulation – appears as one of the<br />
factors that have ma<strong>de</strong> it difficult to implement urban<br />
water <strong>de</strong>m<strong>and</strong> management <strong>in</strong> Camp<strong>in</strong>a Gr<strong>and</strong>e.<br />
As an example of the very few exist<strong>in</strong>g works that<br />
address this city’s necessity, L<strong>in</strong>s et al. (2008),<br />
utilized Factor Analysis <strong>and</strong> Simple L<strong>in</strong>ear Regression<br />
Analysis (based on data relative to the year 2000) to<br />
i<strong>de</strong>ntify the possible <strong>in</strong>terrelations between the<br />
variable “population per age group” <strong>and</strong> “household<br />
water consumption” <strong>in</strong> Camp<strong>in</strong>a Gr<strong>and</strong>e city; the<br />
obta<strong>in</strong>ed results <strong>in</strong>dicate that, for the predom<strong>in</strong>ant age<br />
group (0 to 49 years old), the city’s average household<br />
water consumption (103 L/<strong>in</strong>habitant·day) is <strong>in</strong>ferior<br />
to the present Brazilian average consumption.<br />
Based on socioeconomic criterion (family <strong>in</strong>come<br />
level), two districts of Camp<strong>in</strong>a Gr<strong>and</strong>e city were<br />
chosen as study area:<br />
• D<strong>in</strong>américa district, compris<strong>in</strong>g an area of<br />
1.33 km 2 , a population of 3626 <strong>in</strong>habitants<br />
(90.7% of which are <strong>in</strong>clu<strong>de</strong>d with<strong>in</strong> the age<br />
group “0 to 49 years old”) <strong>and</strong> an average family<br />
<strong>in</strong>come of US$1,612, which characterizes it as a<br />
medium <strong>in</strong>come district; <strong>and</strong><br />
• Mirante district, compris<strong>in</strong>g an area of 0.52 km 2 ,<br />
a population of 1056 <strong>in</strong>habitants (88% of which<br />
with<strong>in</strong> the age group “0 to 49 years old”) <strong>and</strong> an<br />
average family <strong>in</strong>come of US$6,741, which<br />
characterizes it as a high <strong>in</strong>come district (IBGE,<br />
2008–2009 & SEPLAN/CG, 2000). Figure 2<br />
shows these districts’ location <strong>in</strong> Camp<strong>in</strong>a<br />
Gr<strong>and</strong>e city’s map.<br />
Fig.2 Map of Camp<strong>in</strong>a Gr<strong>and</strong>e city, <strong>in</strong>dicat<strong>in</strong>g D<strong>in</strong>américa <strong>and</strong> Mirante districts location.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p. 47-80, 2010
L<strong>in</strong>s, Cruz, Vieira, C. Neto <strong>and</strong> Mir<strong>and</strong>a<br />
77<br />
METHODOLOGY<br />
Database<br />
The follow<strong>in</strong>g types of monthly data – relat<strong>in</strong>g to the<br />
selected districts – were used <strong>in</strong> this work, all of them<br />
referr<strong>in</strong>g to the period from 2000 to 2008 (9-year data<br />
series): socioeconomic (average family <strong>in</strong>come, water<br />
tariffs applied by the local water supply company, <strong>and</strong><br />
economies, i.e., the number of household connections to<br />
the public water supply system), climatic (average<br />
temperature <strong>and</strong> precipitation), <strong>and</strong> household water<br />
consumption. These data were ma<strong>de</strong> available by<br />
SEPLAN/CG (Camp<strong>in</strong>a Gr<strong>and</strong>e’s Municipal Secretariat<br />
of Plann<strong>in</strong>g), EMBRAPA (Brazilian Enterprise for<br />
Agricultural Research), <strong>and</strong> CAGEPA (Water <strong>and</strong><br />
Sewage Company of Paraíba, the local water supply<br />
company).<br />
Statistical Techniques<br />
This work utilized two multivariate statistical<br />
techniques – Factor Analysis <strong>and</strong> Multiple L<strong>in</strong>ear<br />
Regression Analysis – which were ma<strong>de</strong> viable by the<br />
Statistical Package for the Social Sciences – SPSS.<br />
At first, a Factor Analysis (FA) was carried out<br />
consi<strong>de</strong>r<strong>in</strong>g as variables, for both D<strong>in</strong>américa <strong>and</strong><br />
Mirante districts: monthly average temperature, monthly<br />
average precipitation, economies, monthly water tariffs,<br />
<strong>and</strong> monthly family <strong>in</strong>come.<br />
Factor Analysis is a multivariate statistical analysis<br />
technique which changes the orig<strong>in</strong>al variables set <strong>in</strong>to a<br />
smaller one composed by factors; the or<strong>de</strong>r <strong>in</strong> which<br />
these factors are obta<strong>in</strong>ed corresponds to their<br />
importance with respect to the quantity of variance of<br />
the orig<strong>in</strong>al variables set that each of them can expla<strong>in</strong>.<br />
This means that the first factor expla<strong>in</strong>s the greatest<br />
quantity of variance of the data set; the second factor<br />
expla<strong>in</strong>s the greatest possible quantity of rema<strong>in</strong><strong>de</strong>r<br />
variance; <strong>and</strong> so on. As this work <strong>de</strong>als with a 5-<br />
variable set, the maximum number of factors to be<br />
extracted is five, i.e., equal to the number of orig<strong>in</strong>al<br />
variables. Here, the factor<strong>in</strong>g (extraction of factors)<br />
method utilized was Pr<strong>in</strong>cipal Component Analysis (PCA).<br />
On the other h<strong>and</strong>, as the quantity of variance<br />
expla<strong>in</strong>ed by each factor is represented by the quotient<br />
between the factor’s eigenvalue <strong>and</strong> the number of<br />
variables (or the total number of factors), it is<br />
consi<strong>de</strong>red, <strong>in</strong> general, that only those factors with<br />
eigenvalues greater than or equal to 1 will take part <strong>in</strong><br />
the f<strong>in</strong>al solution. In other words, factors’ eigenvalues<br />
which are less than 1 are elim<strong>in</strong>ated from the f<strong>in</strong>al<br />
solution (Cruz, 1983).<br />
Generally, the factors thus extracted are not easily<br />
<strong>in</strong>terpreted. In or<strong>de</strong>r to facilitate their <strong>in</strong>terpretation, it is<br />
realized a rotation of the coord<strong>in</strong>ate axes which<br />
graphically represent the factors. Accord<strong>in</strong>g to the<br />
rotation method adopted – orthogonal or oblique – the<br />
variables’ coord<strong>in</strong>ates, also called “factors load<strong>in</strong>g,”<br />
represent the projection of the variables’ variance <strong>in</strong> this<br />
new coord<strong>in</strong>ate system. One of the most utilized<br />
methods is the so-called Varimax rotation which aims to<br />
maximize the variance projection of each variable on the<br />
factors. This was the rotation method used <strong>in</strong> this work.<br />
The other multivariate analysis technique utilized<br />
was Multiple L<strong>in</strong>ear Regression Analysis. This<br />
technique shows the l<strong>in</strong>ear comb<strong>in</strong>ation between a<br />
variable, called <strong>de</strong>pen<strong>de</strong>nt variable, <strong>and</strong> a set of<br />
variables called <strong>in</strong><strong>de</strong>pen<strong>de</strong>nt or explicative variables<br />
(Matos, 2000).<br />
This work adopted the significance level of 5% <strong>and</strong>,<br />
for both districts, consi<strong>de</strong>red the “household<br />
consumption” (calculated, for each district, as the<br />
quotient between the total district water consumption<br />
<strong>and</strong> the number of economies) as <strong>de</strong>pen<strong>de</strong>nt variable;<br />
the variables <strong>in</strong>itially suggested as <strong>in</strong><strong>de</strong>pen<strong>de</strong>nt<br />
variables were those which have presented the greatest<br />
load<strong>in</strong>gs on Factor 1, the most important factor. Based<br />
on this set of variables thus <strong>de</strong>f<strong>in</strong>ed, a Multiple L<strong>in</strong>ear<br />
Regression Analysis was performed, <strong>and</strong>, through the<br />
“stepwise method” (i.e., a method executed step-bystep,<br />
which <strong>in</strong>clu<strong>de</strong>s each <strong>in</strong><strong>de</strong>pen<strong>de</strong>nt variable<br />
accord<strong>in</strong>g to its contribution for <strong>in</strong>creas<strong>in</strong>g the<br />
<strong>de</strong>term<strong>in</strong>ation coefficient of the l<strong>in</strong>ear regression<br />
equation), the f<strong>in</strong>al l<strong>in</strong>ear regression equation was<br />
obta<strong>in</strong>ed.<br />
RESULTS FOR DINAMÉRICA DISTRICT<br />
Factor Analysis Application<br />
Table 1 shows the first obta<strong>in</strong>ed results for D<strong>in</strong>américa<br />
district, <strong>in</strong>dicat<strong>in</strong>g the <strong>in</strong>itial eigenvalues <strong>and</strong> the<br />
percentages of variance for each one of the five factors.<br />
Based on the criterion of eigenvalues greater than 1,<br />
two factors were extracted (Factor 1 <strong>and</strong> Factor 2)<br />
which expla<strong>in</strong> nearly 76% of the orig<strong>in</strong>al data set<br />
variance.<br />
Follow<strong>in</strong>g the methodological steps <strong>de</strong>scribed above,<br />
the Varimax method application resulted as shown <strong>in</strong><br />
Table 2. In this case, the socioeconomics variables<br />
(Tariff, Economies <strong>and</strong> Income) represent a high<br />
load<strong>in</strong>g on Factor 1, while the climatic variables<br />
(average temperature T, <strong>and</strong> average precipitation P)<br />
represent a high load<strong>in</strong>g on Factor 2.<br />
Table 1. Total variance expla<strong>in</strong>ed<br />
Factor Eigenvalue Variance (%)<br />
Cumulative<br />
(%)<br />
1 2.473 49.469 49.469<br />
2 1.338 26.768 76.238<br />
3 0.655 13.109 89.347<br />
4 0.481 9.629 98.976<br />
5 5.12.10 -2 1.024 100.000<br />
Extraction Method: Pr<strong>in</strong>cipal Component Analysis.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p. 47-80, 2010
L<strong>in</strong>s, Cruz, Vieira, C. Neto <strong>and</strong> Mir<strong>and</strong>a<br />
78<br />
Table 2. Rotated matrix<br />
Variables Factor 1 Factor 2<br />
T 6.882 × 10 -2 0.831<br />
P -2.458 × 10 -2 - 0.825<br />
Tariff 0.975 9.846 × 10 -2<br />
Economies 0.860 6.362 × 10 -2<br />
Income 0.855 -3.907 × 10 -2<br />
Rotation Method: Varimax rotation.<br />
Multiple L<strong>in</strong>ear Regression Analysis Application<br />
The Multiple L<strong>in</strong>ear Regression Analysis between<br />
“household consumption” <strong>and</strong> the explicative variables<br />
extracted from the Factor Analysis (Factor 1) presented<br />
the f<strong>in</strong>al regression equation as expressed by Eq. (1):<br />
Y = 1.933 + 0.131 X 2 – 8.246×10 -4 X 3 (1)<br />
where Y is the household consumption (m 3 /month); X 2 is<br />
the water supply company’s tariff; <strong>and</strong> X 3 is the monthly<br />
average family <strong>in</strong>come.<br />
This f<strong>in</strong>al regression equation’s <strong>de</strong>term<strong>in</strong>ation<br />
coefficient R 2 is equal to 0.896, which means that 89.6%<br />
of the <strong>de</strong>pen<strong>de</strong>nt variable Y’s variance are expla<strong>in</strong>ed by<br />
the <strong>in</strong><strong>de</strong>pen<strong>de</strong>nt variables X 2 (Tariff) <strong>and</strong> X 3 (Income).<br />
Table 3 shows the Analysis of Variance (ANOVA)<br />
for Eq.(1).<br />
Based on Table 3 analysis, one can verify that<br />
statistics F presents a value equal to 447. Admitt<strong>in</strong>g a<br />
given level of significance, this F’s value must be<br />
compared to a value obta<strong>in</strong>ed from a “Table of F’s<br />
values,” which consi<strong>de</strong>rs the regression <strong>and</strong> residual<br />
number of <strong>de</strong>grees of freedom (df). Only if the F’s value<br />
presented <strong>in</strong> Table 3 is greater than any F’s values<br />
tabulated, for the consi<strong>de</strong>red significance level <strong>and</strong><br />
<strong>de</strong>grees of freedom, it will be possible to say that Eq.(1)<br />
is statistically significant at that adopted significance level.<br />
It is important to emphasize that SPSS’ ANOVA<br />
straightly <strong>in</strong>dicates the value of the smallest significance<br />
level (the sixth column of Table 3) from which Eq. (1)<br />
is statistically significant. In other words, Table 3 (at its<br />
fifth column) <strong>in</strong>dicates the F’s value which is greater<br />
than any tabulated F for significance levels greater than<br />
that one presented at its sixth column.<br />
S<strong>in</strong>ce the adopted significance level (5%, i.e., 0.05)<br />
is greater than the smallest significance level (0.00)<br />
showed by Table 3, it is possible to affirm that F = 447<br />
Table 3 – ANOVA<br />
Mo<strong>de</strong>l<br />
Sum of Mean<br />
df<br />
squares square<br />
Regression 21,81 2 10,9<br />
Residual 2,53 104 ,024<br />
Total 24,34 106<br />
df = <strong>de</strong>grees of freedom; F = Fisher’s statistics;<br />
Sig = Significance level.<br />
F<br />
Sig<br />
447 .00<br />
Table 4. D<strong>in</strong>américa district: coefficients of l<strong>in</strong>ear regression<br />
equation mo<strong>de</strong>l<br />
Mo<strong>de</strong>l B t Sig<br />
Constant<br />
Tariff<br />
Income<br />
1.933<br />
.131<br />
-8.246×10 -4 39.746<br />
27.699<br />
-15.579<br />
.000<br />
.000<br />
.000<br />
B = Unst<strong>and</strong>ardized coefficient; t = t-Stu<strong>de</strong>nt;<br />
Sig = Significance level.<br />
is greater than the tabulated value of F for the 5% –<br />
significance level, <strong>and</strong>, consequently, the obta<strong>in</strong>ed f<strong>in</strong>al<br />
regression equation is statistically significant at this<br />
adopted significance level.<br />
Table 4 <strong>in</strong>dicates the constant part, the partial<br />
regression coefficients, as well as other <strong>in</strong>formation<br />
which are pert<strong>in</strong>ent to the statistical quality analysis of<br />
the <strong>in</strong><strong>de</strong>pen<strong>de</strong>nt variables <strong>in</strong> Eq. (1).<br />
As Table 4 shows, the values assumed by statistics t<br />
are greater than those tabulated for any significance<br />
level which is greater than 0.000 (result presented by<br />
SPSS at the fourth column of Table 4). Thus, the partial<br />
regression coefficients are statistically significant at the<br />
adopted significance level (5%).<br />
Therefore, for D<strong>in</strong>américa district, Eq. (1) can be<br />
consi<strong>de</strong>red as a l<strong>in</strong>ear regression equation mo<strong>de</strong>l which<br />
associates household consumption to the <strong>in</strong><strong>de</strong>pen<strong>de</strong>nt<br />
variables Tariff <strong>and</strong> Income.<br />
RESULTS FOR MIRANTE DISTRICT<br />
Factor Analysis Application<br />
Follow<strong>in</strong>g the methodological steps, Table 5 shows the<br />
first results for Mirante district, i.e., the <strong>in</strong>itial<br />
eigenvalues <strong>and</strong> the percentages of variance for each<br />
one of the five factors. Aga<strong>in</strong>, the criterion of<br />
eigenvalues greater than 1 implied <strong>in</strong> the extraction of<br />
two factors, which expla<strong>in</strong> about 79% of the orig<strong>in</strong>al<br />
data set variance.<br />
The rotated matrix (Varimax Method), shown by<br />
Table 6, presents a result which is similar to that<br />
obta<strong>in</strong>ed for D<strong>in</strong>américa district, i.e., the<br />
socioeconomic variables (Tariff, Economies <strong>and</strong><br />
Income) presented a high load<strong>in</strong>g on Factor 1, while<br />
the same occurred with the climatic variable (T <strong>and</strong><br />
P) <strong>in</strong> relation to Factor 2.<br />
Table 5. Total variance expla<strong>in</strong>ed<br />
Factor Eigenvalue Variance (%)<br />
Cumulative<br />
(%)<br />
1 2.614 52.285 52.285<br />
2 1.336 26.717 79.002<br />
3 0.640 12.800 91.802<br />
4 0.325 6.491 98.293<br />
5 8.53×10 -2 1.707 100.000<br />
Extraction Method: Pr<strong>in</strong>cipal Component Analysis.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p. 47-80, 2010
L<strong>in</strong>s, Cruz, Vieira, C. Neto <strong>and</strong> Mir<strong>and</strong>a<br />
79<br />
Table 6. Rotated matrix.<br />
Variables Factor 1 Factor 2<br />
T 5.491×10 -2 0.836<br />
P - 5.394×10 -2 - 0.817<br />
Tariff 0.941 0.175<br />
Economies 0.944 4.598×10 -2<br />
Income 0.876 -1.992×10 -2<br />
Rotation Method: Varimax rotation.<br />
Multiple L<strong>in</strong>ear Regression Analysis Application<br />
The l<strong>in</strong>ear regression analysis was performed by<br />
consi<strong>de</strong>r<strong>in</strong>g the district’s household consumption <strong>and</strong> its<br />
possible relations to the explicative variables extracted<br />
from the Factor Analysis (Factor 1). The follow<strong>in</strong>g f<strong>in</strong>al<br />
regression equation was obta<strong>in</strong>ed:<br />
Y = 5.970 + 0.104X 2 – 6.13×10 -4 X 3 (2)<br />
where Y is the household consumption (m 3 /month); X 2 is<br />
the water supply company’s tariff; <strong>and</strong> X 3 is the monthly<br />
average family <strong>in</strong>come.<br />
Equation (2) has a <strong>de</strong>term<strong>in</strong>ation coefficient R 2<br />
equal to 0.912, which means that 91.2% of the<br />
<strong>de</strong>pen<strong>de</strong>nt variable Y’s variance are expla<strong>in</strong>ed by the<br />
<strong>in</strong><strong>de</strong>pen<strong>de</strong>nt variables X 2 (Tariff) <strong>and</strong> X 3 (Income).<br />
Table 7 presents the results of the Analysis of<br />
Variance (ANOVA).<br />
The value of statistics F is equal to 537, which is<br />
greater than any F’s values tabulated for the 5% –<br />
significance level (<strong>in</strong> accordance with the explanation<br />
already given for Table 3 analysis). Consequently, this<br />
means that Eq. (2) is statistically significant at this<br />
adopted significance level.<br />
Table 8 shows the constant part, the partial<br />
regression coefficients, <strong>and</strong> other <strong>in</strong>formation related to<br />
the statistical quality analysis of the <strong>in</strong><strong>de</strong>pen<strong>de</strong>nt<br />
variables <strong>in</strong> Eq. (2).<br />
Table 7. ANOVA.<br />
Mo<strong>de</strong>l<br />
Regression<br />
Residual<br />
Total<br />
Sum of<br />
squares<br />
117.35<br />
11.36<br />
128.72<br />
df<br />
2<br />
104<br />
106<br />
Mean<br />
square<br />
58.68<br />
.109<br />
df = <strong>de</strong>grees of freedom; F = Fisher’s statistics;<br />
Sig = Significance level.<br />
F<br />
Sig.<br />
537 .000<br />
Table 8. Mirante district: coefficients of l<strong>in</strong>ear regression equation<br />
mo<strong>de</strong>l.<br />
Mo<strong>de</strong>l B t Sig<br />
Constant<br />
Tariff<br />
Income<br />
5.970<br />
.104<br />
-6.134×10 -4 55.059<br />
32.443<br />
-26.716<br />
.000<br />
.000<br />
.000<br />
B = Unst<strong>and</strong>ardized coefficient; t = t-Stu<strong>de</strong>nt;<br />
Sig = Significance level.<br />
As statistics t values are greater than those tabulated<br />
for any significance level greater than 0.000, the partial<br />
regression coefficients are statistically significant.<br />
Therefore, for Mirante district, Eq. (2) can be<br />
consi<strong>de</strong>red as a l<strong>in</strong>ear regression equation mo<strong>de</strong>l which<br />
associates household consumption to the variables<br />
Tariff <strong>and</strong> Income.<br />
CONCLUSION<br />
Household water consumption is a complex function of<br />
several factors, amongst which are socioeconomic <strong>and</strong><br />
climate ones. For the study case consi<strong>de</strong>red <strong>in</strong> this work,<br />
<strong>and</strong> specifically for the two selected districts, it was<br />
verified that the variables “average temperature” <strong>and</strong><br />
“average precipitation” did not take part <strong>in</strong> the f<strong>in</strong>al<br />
solution, due to the fact that none of them have<br />
presented a high load<strong>in</strong>g on Factor 1, the most important<br />
factor that has been <strong>in</strong>dicated by Factor Analysis<br />
application.<br />
Each obta<strong>in</strong>ed l<strong>in</strong>ear regression equation, Eq. (1) <strong>and</strong><br />
Eq. (2), respectively for D<strong>in</strong>américa <strong>and</strong> Mirante<br />
districts, can be consi<strong>de</strong>red as a l<strong>in</strong>ear regression mo<strong>de</strong>l<br />
which associates household consumption to water tariffs<br />
<strong>and</strong> family <strong>in</strong>come.<br />
Then, the obta<strong>in</strong>ed results po<strong>in</strong>t out the<br />
socioeconomics variables (water tariffs <strong>and</strong> family<br />
<strong>in</strong>come) as <strong>in</strong>dicators of urban household consumption<br />
for these Camp<strong>in</strong>a Gr<strong>and</strong>e city’s districts, so that they<br />
can be utilized to assist <strong>de</strong>cision-makers <strong>and</strong> to afford a<br />
more effective management of this consumption.<br />
The authors suggest that other <strong>in</strong>vestigations must be<br />
realized, <strong>in</strong>clud<strong>in</strong>g different districts <strong>and</strong>/or sectors of<br />
Camp<strong>in</strong>a Gr<strong>and</strong>e city, so as to po<strong>in</strong>t out l<strong>in</strong>ear<br />
regression mo<strong>de</strong>ls which can <strong>in</strong>corporate any variability<br />
with regard to this city’s household water consumption,<br />
as well as to support urban water <strong>de</strong>m<strong>and</strong> public<br />
policies formulation.<br />
Acknowledgements The third author gratefully<br />
acknowledges the scholarship provi<strong>de</strong>d by the Brazilian<br />
Agency CAPES, through the National Postdoctoral<br />
Program (PNPD).<br />
REFERENCES<br />
ANA – Brazilian National Water Agency (2007) Relatório nacional<br />
sobre o gerenciamento da água no Brasil- GEO Brasil Recursos<br />
Hídricos Available <strong>in</strong>: http://www.ana.gov.br. Accessed <strong>in</strong>: 15<br />
May 2010.<br />
Cruz, W.S. (1983) Métodos estatísticos multivariados aplicados à<br />
análise <strong>de</strong> transportes. MSc Thesis, Universida<strong>de</strong> Fe<strong>de</strong>ral da<br />
Paraíba, Campus II, Camp<strong>in</strong>a Gr<strong>and</strong>e, Paraíba, Brasil.<br />
Galvão, C.O., Rêgo, J.C., Ribeiro, M.M.R. & Albuquerque, J.P.T.<br />
(2001) Susta<strong>in</strong>ability characterization <strong>and</strong> mo<strong>de</strong>l<strong>in</strong>g of water<br />
supply management practices. IAHS Publ. 268, 81–88.<br />
IBGE – Instituto Brasileiro <strong>de</strong> Geografia e Estatística (2010). Dados<br />
da cida<strong>de</strong> <strong>de</strong> Camp<strong>in</strong>a Gr<strong>and</strong>e – PB. Available <strong>in</strong>:<br />
http://www.ibge.org.br/cida<strong>de</strong>s. Accessed <strong>in</strong>: 20 May 2010.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p. 47-80, 2010
L<strong>in</strong>s, Cruz, Vieira, C. Neto <strong>and</strong> Mir<strong>and</strong>a<br />
80<br />
IBGE – Instituto Brasileiro <strong>de</strong> Geografia e Estatística (2008-2009)<br />
Pesquisa <strong>de</strong> orçamentos domiciliares. Available <strong>in</strong>:<br />
http://www.ibge.org.br/estadosat/. Accessed <strong>in</strong>: 20 May 2010.<br />
Lanna, A.E.L. (2008) A economia dos recursos hídricos: os <strong>de</strong>safios<br />
da alocação eficiente <strong>de</strong> um recurso (cada vez mais) escasso.<br />
Estudos Avançados, 22(2), 113–130. doi: 10.1590/S0103-<br />
40142008000200008.<br />
L<strong>in</strong>s, G.M.L. & Ribeiro, M.M.R. (2007) Gestão da <strong>de</strong>m<strong>and</strong>a <strong>de</strong> água<br />
em centros urbanos do semiárido nor<strong>de</strong>st<strong>in</strong>o. Proc. XVII<br />
Sem<strong>in</strong>ário Brasileiro <strong>de</strong> Recursos Hídricos, ABRH, São Paulo,<br />
SP, Brasil.<br />
L<strong>in</strong>s, G.M.L., Cruz, W.S. & Mir<strong>and</strong>a, E.A.A (2008) O uso <strong>de</strong><br />
técnicas estatísticas multivariadas na <strong>de</strong>term<strong>in</strong>ação <strong>de</strong> <strong>in</strong>dicadores<br />
<strong>de</strong> consumo doméstico urbano <strong>de</strong> água. Proc. IX Simpósio <strong>de</strong><br />
Recursos Hídricos do Nor<strong>de</strong>ste, ABRH, Salvador, Brazil, 25–28<br />
Nov.<br />
Matos, O.C. (2000) Econometria Básica: teoria e aplicações. São<br />
Paulo: Atlas.<br />
PMCG – Prefeitura Municipal <strong>de</strong> Camp<strong>in</strong>a Gr<strong>and</strong>e (2002) Perfil do<br />
município <strong>de</strong> Camp<strong>in</strong>a Gr<strong>and</strong>e. Available <strong>in</strong>:<br />
http://www.pmcg.pb.gov.br. Accessed <strong>in</strong>: 27 May 2010.<br />
Rêgo, J.C., Ribeiro, M.M.R. & Albuquerque, J.P.T. (2000) Uma<br />
análise da crise <strong>de</strong> 1998–2000 no abastecimento d’água <strong>de</strong><br />
Camp<strong>in</strong>a Gr<strong>and</strong>e-PB. Proc. V Simpósio <strong>de</strong> Recursos Hídricos do<br />
Nor<strong>de</strong>ste, ABRH, Natal, RN, Brasil.<br />
SEPLAN/CG – Secretaria <strong>de</strong> Planejamento <strong>de</strong> Camp<strong>in</strong>a Gr<strong>and</strong>e<br />
(2000) Dados dos bairros – geral/ In: Perfil do município <strong>de</strong><br />
Camp<strong>in</strong>a Gr<strong>and</strong>e. Available <strong>in</strong>: http://www.pmcg.pb.gov.br.<br />
Accessed <strong>in</strong>: 27 May, 2010.<br />
Tomaz, P. (2001) Economia <strong>de</strong> água para empresas e residência:<br />
Um estudo atualizado sobre o uso racional da água. São Paulo:<br />
Editora Navegar, 112p.<br />
USEPA – U.S. Environmental Protection Agency (1998) Water<br />
Conservation Plan Gui<strong>de</strong>l<strong>in</strong>es. Available <strong>in</strong>:<br />
http://www.epa.gov/own. Accessed <strong>in</strong>: 12 December 2000.<br />
Vieira, Z.M.C.L. (2002) Análise <strong>de</strong> conflitos na seleção <strong>de</strong><br />
alternativas <strong>de</strong> gerenciamento da <strong>de</strong>m<strong>and</strong>a urbana <strong>de</strong> água. MSc<br />
Thesis, Fe<strong>de</strong>ral University of Camp<strong>in</strong>a Gr<strong>and</strong>e, 132p.<br />
Vieira, Z.M.C.L., Braga, C.F.C. & Ribeiro, M.M.R. (2005) Conflict<br />
analysis as a <strong>de</strong>cision support tool <strong>in</strong> urban water <strong>de</strong>m<strong>and</strong><br />
management. IAHS Publ. 293, 65–72.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p. 47-80, 2010
Martínez <strong>and</strong> Poleto<br />
81<br />
J U E E<br />
Journal of Urban <strong>and</strong> Environmental<br />
Eng<strong>in</strong>eer<strong>in</strong>g, v.4, n.2, p.81-104<br />
ISSN 1982-3932<br />
doi: 10.4090/juee.2010.v4n2.081104<br />
Journal of Urban <strong>and</strong><br />
Environmental Eng<strong>in</strong>eer<strong>in</strong>g<br />
www.journal-uee.org<br />
COMPARATIVE ANALYSIS OF EROSION MODELING<br />
TECHNIQUES IN A BASIN OF VENEZUELA<br />
Adriana M. Márquez ∗ <strong>and</strong> Edilberto Guevara-Pérez<br />
Centre of Environmental <strong>and</strong> Hydrologic Research, Carabobo University, Venezuela<br />
Received 20 May 2010; received <strong>in</strong> revised form 20 October 2010; accepted 16 December 2010<br />
Abstract:<br />
Keywords:<br />
This paper <strong>in</strong>vestigates a comparative analysis of erosion mo<strong>de</strong>l<strong>in</strong>g techniques<br />
based on <strong>de</strong>term<strong>in</strong>istic mo<strong>de</strong>ls, both empirically-based <strong>and</strong> process-based<br />
mo<strong>de</strong>ls. The empirical mo<strong>de</strong>l<strong>in</strong>g is based on statistical <strong>and</strong> artificial <strong>in</strong>telligence<br />
techniques. In the first, two types of statistical regression mo<strong>de</strong>l structures are<br />
<strong>in</strong>vestigated, a l<strong>in</strong>ear multiple regression mo<strong>de</strong>l structure <strong>and</strong> a nonl<strong>in</strong>ear<br />
multiple regression mo<strong>de</strong>l structure. In the second, tools such as artificial neural<br />
networks (ANN) <strong>and</strong> fuzzy <strong>in</strong>ference system (FIS) are used. The physical<br />
process-based mo<strong>de</strong>l<strong>in</strong>g <strong>in</strong>volves the calibration, validation <strong>and</strong> test<strong>in</strong>g of the<br />
mo<strong>de</strong>ls components: WEPP, EUROSEM <strong>and</strong> CIHAM-UC. The <strong>in</strong>put <strong>and</strong> output<br />
variables of mo<strong>de</strong>ls were collected dur<strong>in</strong>g ra<strong>in</strong>y <strong>and</strong> dry (irrigation) seasons <strong>in</strong><br />
Chirgua river bas<strong>in</strong>, Venezuela for two years (2008–2009). N<strong>in</strong>ety-seven ra<strong>in</strong>fall<br />
storms <strong>and</strong> 300 irrigation events were measured. Satisfactory fit was found <strong>in</strong> the<br />
techniques <strong>in</strong>vestigated, R 2 close to 0.7.<br />
Sediments; hydrological mo<strong>de</strong>ls; soil erosion mo<strong>de</strong>l; regression mo<strong>de</strong>l;<br />
hydrology; runoff; artificial neural networks; fuzzy <strong>in</strong>ference systems<br />
© 2010 Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE). All rights reserved.<br />
∗ Correspon<strong>de</strong>nce to: Adriana Márquez. Telefax: 0058-414-4165856.<br />
E-mail: ammarquez@uc.edu.ve<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.81-104, 2010
Márquez <strong>and</strong> Guevara-Pérez<br />
82<br />
INTRODUCTION<br />
The conventional techniques of erosion mo<strong>de</strong>l<strong>in</strong>g are<br />
empirically <strong>and</strong> process-based. The empirical mo<strong>de</strong>ls<br />
<strong>in</strong>clu<strong>de</strong> the Universal Soil Loss Equation (USLE)<br />
(Wischmeier & Smith, 1958), <strong>and</strong> its latter<br />
<strong>de</strong>velopments, Revised Soil Loss Equation (RUSLE)<br />
<strong>and</strong> Modified Universal Soil Loss Equation (MUSLE)<br />
(Renard et al., 1997a, 1997b). The process-based<br />
mo<strong>de</strong>ls are represented, e.g., by the WEPP (Water<br />
Erosion Prediction Project) (Flanagan et al., 2001),<br />
DWEPP (Dynamic Water Erosion Prediction Project)<br />
(Bulyg<strong>in</strong>a et al., 2007) <strong>and</strong> EUROSEM (European Soil<br />
Erosion Mo<strong>de</strong>l) (Morgan et al., 1998). There have been<br />
a large number of studies to calibrate <strong>and</strong> validate the<br />
runoff <strong>and</strong> erosion components of process-based mo<strong>de</strong>ls<br />
(e.g., Flanagan et al., 1995; Klik et al., 1995; Savabi et<br />
al., 1996; Zhang et al., 1996; Near<strong>in</strong>g et al., 1998;<br />
Zeleke, 1999; Ranieri et al., 1999; Santoro et al., 2002;<br />
Bulyg<strong>in</strong> et al., 2002; Laflen et al., 2004; Silva et al.,<br />
2007; Santos et al., 2003). Many erosion mo<strong>de</strong>ls require<br />
a large amount of data for calibration <strong>and</strong> validation as<br />
well as high performance, mean<strong>in</strong>g, the complex<br />
algorithms, execution times that eventually cause<br />
problems, <strong>and</strong> a programm<strong>in</strong>g language necessary to<br />
m<strong>in</strong>imize the time compute. Additionally, the ANN <strong>and</strong><br />
FIS have been proposed as efficient tools for mo<strong>de</strong>l<strong>in</strong>g<br />
<strong>and</strong> hydrology forecast<strong>in</strong>g (e.g., Bishop, 1994; Ja<strong>in</strong> &<br />
Ormsbee, 2002; Panigrahi & Mujundar, 2000). In this<br />
work, process-based mo<strong>de</strong>ls, statistical regression,<br />
ANN’s <strong>and</strong> FIS’s are used to mo<strong>de</strong>l the furrow erosion<br />
process.<br />
It is generally assumed that furrow erosion processes<br />
are similar to rill erosion processes that occur un<strong>de</strong>r<br />
ra<strong>in</strong>fall. There are similarities <strong>in</strong> the processes, but there<br />
are also differences <strong>in</strong> the conditions (Trout & Niebl<strong>in</strong>g,<br />
1993). Most ra<strong>in</strong>fall erosion occurs dur<strong>in</strong>g a few highly<br />
erosive events, while furrow erosion occurs at low-tomo<strong>de</strong>rate<br />
rates dur<strong>in</strong>g several irrigations with controlled<br />
water application. Most irrigation furrows are on slopes<br />
of less than 3%, while most rill erosion research is<br />
carried out on slopes greater than 3%. However, <strong>in</strong> spite<br />
of the controlled <strong>in</strong>flows <strong>and</strong> relatively low slopes, <strong>in</strong><br />
some areas with highly erodible soils, there is<br />
significant furrow irrigation caus<strong>in</strong>g erosion damage<br />
(Koluvec et al., 1993; Trout, 1999). The present study<br />
had two ma<strong>in</strong> objectives: (1) calibrate, validate <strong>and</strong> test<br />
mo<strong>de</strong>ls for the estimation of furrow erosion processes<br />
<strong>and</strong> (2) compare experimentally observed furrow<br />
erosion processes with simulated results by different<br />
methods.<br />
MODEL DESCRIPTION<br />
Water Erosion Prediction Project<br />
WEPP mo<strong>de</strong>l was <strong>de</strong>veloped by the United States<br />
Department of Agriculture – Agricultural Research<br />
Service (USDA-ARS). This mo<strong>de</strong>l is based on steadystate<br />
equation assum<strong>in</strong>g a uniform downslope overl<strong>and</strong><br />
flow; therefore, variables are expressed on a total width<br />
or total area basis (Foster & Meyer, 1975; Foster, 1982;<br />
Foster & Lane, 1987; Foster, 1990). WEPP<br />
dist<strong>in</strong>guishes erosion processes <strong>in</strong>volved <strong>in</strong> <strong>in</strong>terrill <strong>and</strong><br />
rill erosion. The basic relationship is given by:<br />
∂ ( CA)<br />
∂(<br />
CQ<br />
+<br />
) = S<br />
(1)<br />
∂t<br />
∂x<br />
where C is the concentration of sediment <strong>in</strong> the flow<br />
(kg/m 3 ), A the cross-sectional area of flow (m 2 ), Q the<br />
flow discharge (m 3 /s), t the time (s), x the distance<br />
downslope (m), the term ∂ ( CQ)<br />
/ ∂x<br />
the rate of change<br />
of sediment with the distance, ∂ ( CA)<br />
/ ∂t<br />
the storage<br />
rate of sediment with<strong>in</strong> the flow <strong>de</strong>pth, S the source/s<strong>in</strong>k<br />
term for sediment (kg/s/m 2 ) is given by:<br />
S = D I<br />
+ D R<br />
(2)<br />
By assum<strong>in</strong>g quasi-steady sediment movement, Eq.<br />
(1) is reduced to:<br />
∂ ( CQ)<br />
= D I<br />
+ D R<br />
∂x<br />
( wTc<br />
− CQ) , CQ ≥ wTc<br />
(3)<br />
⎛ CQ ⎞<br />
D<br />
⎜1<br />
⎟<br />
c − , CQ ≤ wTc<br />
= ⎝ wT<br />
D<br />
c ⎠<br />
R<br />
(4)<br />
0.5V<br />
Q<br />
s<br />
where D I is the <strong>in</strong>terrill sediment <strong>de</strong>livery to the rill<br />
(kg/s/m 2 ), D R the flow erosion rate (kg/s/m 2 ), T c the flow<br />
transport capacity <strong>in</strong> the rill (kg/s/m), w the rill width, V s<br />
the particle fall velocity (m/s), Q the flow discharge<br />
(m 3 /s), <strong>and</strong> D c the flow soil <strong>de</strong>tachment capacity<br />
(kg/s/m 2 ) is computed as:<br />
D<br />
c<br />
Kr<br />
( τ − τc<br />
),<br />
0<br />
= ,<br />
τ ≥ τ<br />
c<br />
τ ≤ τ<br />
c<br />
(5)<br />
τ = ρ W<br />
gSR<br />
(6)<br />
where K r is the rill erodibility (s/m), τ the flow shear<br />
stress act<strong>in</strong>g on the soil (Pa), τ c the critical shear stress<br />
(Pa), ρ w the water <strong>de</strong>nsity (kg/m 3 ), g the acceleration<br />
due to gravity (m/s 2 ), S slope (m/m), R hydraulic radius<br />
(m). Rills are assumed to be rectangular with widths that<br />
<strong>de</strong>pend on flow rate.<br />
The simplified form of the equation of Yal<strong>in</strong>, (1963)<br />
to estimate the sediment transport capacity has been<br />
adapted by Foster & Meyer, (1975) as:<br />
T<br />
b<br />
c<br />
= aτ<br />
(7)<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.81-104, 2010
Márquez <strong>and</strong> Guevara-Pérez<br />
83<br />
where T c is the sediment transport capacity, a <strong>and</strong> b are<br />
coefficients obta<strong>in</strong>ed by empirical fitt<strong>in</strong>g.<br />
Dynamic Water Erosion Prediction Project<br />
Bulyg<strong>in</strong>a et al. (2007) <strong>de</strong>veloped a dynamic version of<br />
the WEPP mo<strong>de</strong>l, which quantifies the dynamics of<br />
sediments with<strong>in</strong> a storm event. This mo<strong>de</strong>l is based on<br />
Eqs (1)–(6). Many equations of sediment transport were<br />
<strong>de</strong>veloped for stream flow, <strong>and</strong> later on applied to<br />
shallow overl<strong>and</strong> flow <strong>and</strong> channel flow. Accord<strong>in</strong>g to<br />
Alonso et al. (1981) <strong>and</strong> Foster & Meyer, (1972), the<br />
Yal<strong>in</strong> equation applies better for shallow flows<br />
associated with upl<strong>and</strong> erosion. In DWEPP, the Yal<strong>in</strong><br />
flow transport capacity equation is given by:<br />
⎡ 1 ⎤<br />
T c = 0.635Gs<br />
gd ρwτδ⎢1<br />
− ln( 1+ β) ⎥⎦ (8)<br />
⎣ β<br />
δ =<br />
β =<br />
Y<br />
Ycr<br />
0,<br />
−1,<br />
Y < Y<br />
cr<br />
Y ≥ Y<br />
−0.4<br />
0. 5<br />
2<br />
s cr<br />
cr<br />
(9)<br />
.45G Y δ<br />
(10)<br />
τ<br />
Y = (11)<br />
ρ ( G −1 gd<br />
w s )<br />
where G s is the particle specific gravity, d the particle<br />
diameter (m), δ <strong>and</strong> β are calculated by Eqs (8) <strong>and</strong> (9),<br />
respectively, Y is the dimensionless shear stress from<br />
the Shields diagram <strong>and</strong> Y cr the dimensionless critical<br />
shear stress from the Shields diagram.<br />
European Soil Erosion Mo<strong>de</strong>l<br />
The European Soil Erosion Mo<strong>de</strong>l (EUROSEM) was<br />
<strong>de</strong>veloped by Morgan et al. (1998). It is based on the<br />
Eqs (1) <strong>and</strong> (2), which simulate sediment transport,<br />
erosion <strong>and</strong> <strong>de</strong>position over the l<strong>and</strong> surface by rill <strong>and</strong><br />
<strong>in</strong>terill processes <strong>in</strong> s<strong>in</strong>gle storms for both, <strong>in</strong>dividual<br />
fields <strong>and</strong> small catchments. Mo<strong>de</strong>l output for one<br />
s<strong>in</strong>gle storm <strong>in</strong>clu<strong>de</strong>s total runoff, total soil loss,<br />
hydrographs <strong>and</strong> sediment graphs. The soil <strong>de</strong>tachment<br />
by runoff is mo<strong>de</strong>led <strong>in</strong> terms of a generalized erosion<strong>de</strong>position<br />
theory proposed by Smith et al. (1995).<br />
Accord<strong>in</strong>g to it, D R is the flow erosion rate (kg/s/m 2 ),<br />
(positive for erosion <strong>and</strong> negative for <strong>de</strong>position); the<br />
flow transport capacity concentration (T c ) represents the<br />
sediment concentration at which the flow erosion rate<br />
<strong>and</strong> correspond<strong>in</strong>g <strong>de</strong>position rate are <strong>in</strong> balance. A<br />
general equation for flow erosion rate is expressed <strong>in</strong><br />
terms of the settl<strong>in</strong>g velocity <strong>and</strong> flow transport capacity<br />
concentration, as follows:<br />
D<br />
R<br />
= V ( T − C)<br />
(12)<br />
s<br />
c<br />
The runoff capacity to transport <strong>de</strong>tached soil<br />
particles is expressed <strong>in</strong> terms of the concentration, T c<br />
(kg/m 3 ) is mo<strong>de</strong>led as a function of unit stream power<br />
us<strong>in</strong>g the relationship of Govers (1985):<br />
b<br />
c<br />
= a( ω− ωc<br />
(13)<br />
T )<br />
ω = VS<br />
(14)<br />
where ω c is the critical value of unit stream power, ω<br />
the unit stream power (m/s), V s the particle fall velocity<br />
(m/s), a <strong>and</strong> b are experimentally <strong>de</strong>rived coefficients<br />
<strong>de</strong>pend<strong>in</strong>g on particle size, S the slope (%) <strong>and</strong> V the<br />
mean flow velocity (m/s).<br />
Erosion mo<strong>de</strong>l of CIHAM-UC<br />
At the present, an erosion mo<strong>de</strong>l is be<strong>in</strong>g <strong>de</strong>veloped at<br />
the Centre of Environmental <strong>and</strong> Hydrologic Research,<br />
based on Eqs (1)–(3) <strong>and</strong> Eq. (15) of Simons et al.<br />
(1981), which is based on power relationships that<br />
estimate sediment transport based on the flow <strong>de</strong>pth h<br />
<strong>and</strong> velocity V.<br />
These power relationships were <strong>de</strong>veloped from a<br />
computer solution of the bedload transport equation of<br />
Meyer-Peter & Müller, (1948). Guevara & Márquez,<br />
(2009) propose it to estimate the sediment transport<br />
capacity <strong>in</strong> rill, <strong>and</strong> can be expressed as follows:<br />
b c<br />
T<br />
c<br />
= ah V<br />
(15)<br />
where T c is the flow transport capacity (kg/s/m), a, b<br />
<strong>and</strong> c are parameters estimated by empirical adjustment,<br />
V the flow velocity (m/s) <strong>and</strong> h the flow <strong>de</strong>pth (m).<br />
The flow erosion rate D R (kg/s/m 2 ) is <strong>de</strong>scribed <strong>in</strong><br />
terms of the generalized theory of erosion-<strong>de</strong>position<br />
modify<strong>in</strong>g proposal by Smith et al. (1995). This<br />
condition can be expressed as:<br />
DR = T w − f CV<br />
(16)<br />
c<br />
where V s is the particle fall velocity (m/s), w the furrow<br />
width (m), f s =A ec /A rill, similarity factor, A ec experimental<br />
cyl<strong>in</strong><strong>de</strong>r area from hydrometer test method (D422)<br />
(American Society for Test<strong>in</strong>g <strong>and</strong> Materials, 2007).<br />
L<strong>in</strong>ear Multiple Regression Mo<strong>de</strong>l<br />
The L<strong>in</strong>ear Multiple Regression Mo<strong>de</strong>l (LMRM) used is<br />
based on the structure given by:<br />
s<br />
D P P P P (17)<br />
R<br />
= β1 t<br />
+ β2<br />
t −1<br />
+ β3<br />
t −2<br />
+ ... + β12<br />
t−6<br />
where D R is the flow erosion rate (mg/s/m 2 ) at time t, β s<br />
represents the regression coefficients to be <strong>de</strong>term<strong>in</strong>ed;<br />
P s represents the ra<strong>in</strong>falls correspond to <strong>de</strong>f<strong>in</strong>ed time<br />
s<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.81-104, 2010
Márquez <strong>and</strong> Guevara-Pérez<br />
84<br />
<strong>in</strong>tervals with<strong>in</strong> a s<strong>in</strong>gle storm; <strong>and</strong> t <strong>in</strong><strong>de</strong>x represent<strong>in</strong>g<br />
time.<br />
Nonl<strong>in</strong>ear Multiple Regression Mo<strong>de</strong>l<br />
The Non-L<strong>in</strong>ear Multiple Regression Mo<strong>de</strong>l (NLMRM)<br />
used is based on the structure given by:<br />
n n<br />
n<br />
n<br />
D<br />
R<br />
= β1 P t + β2P<br />
t−1<br />
+ β3P<br />
t−2<br />
+ ... + β12P<br />
t−6<br />
(18)<br />
where n is the or<strong>de</strong>r of polynomial regression. The rest<br />
of the variables are expla<strong>in</strong>ed earlier.<br />
Artificial Neural Networks<br />
Neural networks are composed of simple elements<br />
operat<strong>in</strong>g <strong>in</strong> parallel, <strong>in</strong>spired by biological nervous<br />
systems. As <strong>in</strong> nature, the connections between<br />
elements largely <strong>de</strong>term<strong>in</strong>e the network function. A<br />
neural network can be tra<strong>in</strong>ed to perform a particular<br />
function by adjust<strong>in</strong>g the values of the connections<br />
(weights) between elements. Typically, neural networks<br />
are adjusted, or tra<strong>in</strong>ed, so that a particular <strong>in</strong>put leads<br />
to a specific target output. Figure 1 illustrates such a<br />
situation. The network is adjusted, based on a<br />
comparison of the output <strong>and</strong> the target, until the<br />
network output matches the target (Demuth et al.,<br />
2009).<br />
Target<br />
In this network, each element of the <strong>in</strong>put vector p is<br />
connected to each neuron <strong>in</strong>put through the weight<br />
matrix w. By consi<strong>de</strong>r<strong>in</strong>g a neuron with a s<strong>in</strong>gle R-<br />
element <strong>in</strong>put vector, p, the <strong>in</strong>dividual element <strong>in</strong>puts<br />
p 1 , p 2 ,…p R , are multiplied by weights w 1,1 , w 1,2 , ... w 1,R<br />
<strong>and</strong> the weighted values are fed to the summ<strong>in</strong>g<br />
junction, as shown <strong>in</strong> Fig. 2. This sum is Wp, the dot<br />
product of the (s<strong>in</strong>gle row) matrix W <strong>and</strong> the vector p.<br />
The neuron has a bias b, which is summed with the<br />
weighted <strong>in</strong>puts to form the net <strong>in</strong>put n. This sum, n, is<br />
the argument of the transfer function f <strong>de</strong>f<strong>in</strong>ed as<br />
follows:<br />
n<br />
= ,<br />
w1 ,1 p1<br />
+ w1,2<br />
p2<br />
+ ... + w1<br />
R pR<br />
+ b (19)<br />
This expression can be written as:<br />
n ( i)<br />
= Wp + b<br />
(20)<br />
where, the ith neuron has a summ<strong>in</strong>g that gathers its<br />
weighted <strong>in</strong>puts <strong>and</strong> bias to form its own scalar output n<br />
(i). The various n(i) taken together form an S-element<br />
net <strong>in</strong>put vector n.<br />
The sigmoid transfer function is commonly used <strong>in</strong><br />
backpropagation networks, <strong>in</strong> part because it is<br />
differentiable. As shown <strong>in</strong> Fig. 3, the sigmoid transfer<br />
function takes the <strong>in</strong>put, which can have any value<br />
between plus <strong>and</strong> m<strong>in</strong>us <strong>in</strong>f<strong>in</strong>ity, <strong>and</strong> squashes the<br />
output <strong>in</strong>to the range 0 to 1.<br />
Input<br />
Adjust<br />
Weights<br />
Neural<br />
Network<br />
Output<br />
Compare<br />
a<br />
+1<br />
Fig. 1 Scheme Input/Output to ANN.<br />
0<br />
n<br />
Neuron Mo<strong>de</strong>l<br />
One or more of the neurons can be comb<strong>in</strong>ed <strong>in</strong> a layer,<br />
<strong>and</strong> a particular network could conta<strong>in</strong> one or more such<br />
layers. A one-layer network with R <strong>in</strong>put elements <strong>and</strong> S<br />
neurons is shown <strong>in</strong> Fig. 2 as follows:<br />
p 1<br />
p 2<br />
p R<br />
Inputs<br />
w 1,1<br />
w 1,R<br />
Layer of Neurons<br />
Σ<br />
1<br />
Σ<br />
1<br />
Σ<br />
1<br />
b 1<br />
b 2<br />
b s<br />
Fig. 2 S<strong>in</strong>gle layer network architecture of neurons.<br />
n 1<br />
n 2<br />
n s<br />
f<br />
f<br />
f<br />
a = f (Wp + b)<br />
a 1<br />
a 2<br />
a s<br />
-1<br />
a = logsig(n)<br />
Fig. 3 Log-Sigmoid transfer function.<br />
a<br />
+1<br />
-1<br />
a = purel<strong>in</strong>(n)<br />
Fig. 4 L<strong>in</strong>eal transfer function.<br />
Occasionally, the L<strong>in</strong>ear Transfer Function Purel<strong>in</strong> is<br />
used <strong>in</strong> backpropagation networks. The l<strong>in</strong>ear transfer<br />
function is illustrated <strong>in</strong> Fig. 4.<br />
The symbol <strong>in</strong> the square to the right of each transfer<br />
function graph shown represents the associated transfer<br />
function. These icons replace the general f <strong>in</strong> boxes of<br />
network diagrams to show the particular transfer<br />
function be<strong>in</strong>g used.<br />
n<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.81-104, 2010
Márquez <strong>and</strong> Guevara-Pérez<br />
85<br />
Feedforward Network<br />
Feedforward networks often have one or more hid<strong>de</strong>n<br />
layers of sigmoid neurons followed by an output layer<br />
of l<strong>in</strong>ear neurons. Multiple layers of neurons with<br />
nonl<strong>in</strong>ear transfer functions allow the network to learn<br />
nonl<strong>in</strong>ear <strong>and</strong> l<strong>in</strong>ear relationships between <strong>in</strong>put <strong>and</strong><br />
output vectors. The l<strong>in</strong>ear output layer lets the network<br />
produce values outsi<strong>de</strong> the range –1 to +1. To <strong>de</strong>scribe<br />
networks hav<strong>in</strong>g multiple layers, the notation must be<br />
exten<strong>de</strong>d. Specifically, it needs to make a dist<strong>in</strong>ction<br />
between weight matrices that are connected to <strong>in</strong>puts<br />
<strong>and</strong> weight matrices that are connected between layers.<br />
A network can have several layers <strong>and</strong> it uses layer<br />
weight (LW) matrices as well as <strong>in</strong>put weight (IW)<br />
matrices. The network shown <strong>in</strong> Fig. 5 has R <strong>in</strong>puts, S 1<br />
neurons <strong>in</strong> the first layer, <strong>and</strong> S 2 neurons <strong>in</strong> the second<br />
layer. A constant <strong>in</strong>put 1 is fed to the bias for each<br />
neuron. Note that the outputs of each <strong>in</strong>termediate layer<br />
are the <strong>in</strong>puts to the follow<strong>in</strong>g layer. Thus layer 2 can be<br />
analyzed as a one-layer network with S 1 <strong>in</strong>puts, S 2<br />
neurons, <strong>and</strong> an S 2 ×S 1 weight matrix W 2 . The <strong>in</strong>put to<br />
layer 2 is a 1 ; the output is a 2 . Now that all the vectors<br />
<strong>and</strong> matrices of layer 2 have been i<strong>de</strong>ntified, it can be<br />
treated as a s<strong>in</strong>gle-layer network on its own. This<br />
approach can be taken with any layer of the network.<br />
Fuzzy Logic<br />
Input<br />
Inputs Terms<br />
Fig. 6 Diagram of a fuzzy <strong>in</strong>ference system.<br />
Fuzzy logic has two different mean<strong>in</strong>gs. In a narrow<br />
sense, fuzzy logic is a logical system, which is an<br />
extension of multivalued logic. However, <strong>in</strong> a wi<strong>de</strong>r<br />
sense fuzzy logic (FL) is almost synonymous with the<br />
theory of fuzzy sets, a theory which relates to classes of<br />
objects with unsharp boundaries, <strong>in</strong> which membership<br />
is a matter of <strong>de</strong>gree. The po<strong>in</strong>t of fuzzy logic is to map<br />
an <strong>in</strong>put space to an output space, <strong>and</strong> the primary<br />
mechanism for do<strong>in</strong>g this is a list of if-then statements<br />
called rules. All rules are evaluated <strong>in</strong> parallel, <strong>and</strong> the<br />
or<strong>de</strong>r of the rules is unimportant. The general<br />
<strong>de</strong>scription of a fuzzy system is shown <strong>in</strong> Fig. 6<br />
(MATLAB, 2010).<br />
Fuzzy Inference System<br />
Rules<br />
Output<br />
Outputs Terms<br />
Input<br />
p<br />
Rx1<br />
1<br />
R<br />
I<br />
W<br />
S 1 x<br />
R<br />
b 1<br />
S 1 x<br />
Hid<strong>de</strong>n Layer<br />
+<br />
n<br />
1<br />
S 1 x1<br />
S 1<br />
a 1<br />
S 1 x1<br />
Output<br />
Layer<br />
S 2 x1<br />
a 2 =purel<strong>in</strong> (LW 2, 1<br />
a 1 +b 2 )<br />
Fig. 5 Archictecture of a feedforward network.<br />
1<br />
L<br />
W<br />
S 2 x<br />
S 1<br />
b<br />
+<br />
n<br />
2<br />
S 2 x<br />
1<br />
a 2 =y<br />
S 2 x1<br />
S 2<br />
Fuzzy <strong>in</strong>ference is the process of formulat<strong>in</strong>g the<br />
mapp<strong>in</strong>g from a given <strong>in</strong>put to an output us<strong>in</strong>g fuzzy<br />
logic. The process of fuzzy <strong>in</strong>ference <strong>in</strong>volves three<br />
components: (1) membership functions, (2) logical<br />
operations <strong>and</strong> (3) If-Then rules. (1) A membership<br />
function (MF) is a curve that <strong>de</strong>f<strong>in</strong>es how each po<strong>in</strong>t <strong>in</strong><br />
the <strong>in</strong>put space is mapped to a membership value (or<br />
<strong>de</strong>gree of membership) between 0 <strong>and</strong> 1. Membership<br />
functions commonly used are as follows: the piece-wise<br />
l<strong>in</strong>ear functions, the Gaussian distribution function, the<br />
sigmoid curve, quadratic <strong>and</strong> cubic polynomial curves,<br />
(2) the logical operations, one particular<br />
correspon<strong>de</strong>nce between two-valued <strong>and</strong> multivalued<br />
logical operations can be <strong>de</strong>f<strong>in</strong>ed by the fuzzy<br />
<strong>in</strong>tersection or conjunction (AND), fuzzy union or<br />
disjunction (OR), <strong>and</strong> fuzzy complement (NOT). (3) If-<br />
Then rules are used to formulate the conditional<br />
statements that comprise fuzzy logic.<br />
Table 1. Description of study area<br />
Subbas<strong>in</strong><br />
Geographic Location Area L<strong>and</strong> Use<br />
Zone<br />
Agriculture Avian<br />
North Coord<strong>in</strong>ate West Coord<strong>in</strong>ate (ha)<br />
Domestic % Other<br />
%<br />
%<br />
High<br />
I 10º 13′ 55″ 10º 15′ 00″ 68º 12′ 10″ 68º 11′ 05″ 244.38 95.30 2.70 2 0<br />
II 10º 13′ 00″ 10º 14′ 00″ 68º 11′ 10″ 68º 12′ 00″ 209.21 87.99 4.46 6.13 1.42<br />
Medium III 10º 10′ 10″ 10º 11′ 50″ 68º 11′ 10″ 68º 10′ 20″ 320.64 93.04 0.74 6.22 0<br />
IV 10º 11′ 50″ 10º 12′ 20″ 68º 11′ 30″ 68º 10′ 30″ 162.58 77.78 5.56 16.66 0<br />
Low V 10º 12′ 25″ 10º 13′ 10″ 68º 11′ 10″ 68º 11′ 50″ 131.74 83.33 5.56 16.67 0<br />
VI 10º 12′ 18″ 10º 15′ 45″ 68º 11′ 67″ 68º 15′ 07″ 250.64 94.04 0.84 7 0<br />
Percentage Average 88.58 3.31 9.11 0.24<br />
Total 1319.19 1169 44 120 3.16<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.81-104, 2010
Márquez <strong>and</strong> Guevara-Pérez<br />
86<br />
Two types of fuzzy <strong>in</strong>ference systems can be<br />
implemented: Mamdani-type <strong>and</strong> Sugeno-type.<br />
These two types of <strong>in</strong>ference systems vary somewhat<br />
<strong>in</strong> the way outputs are <strong>de</strong>term<strong>in</strong>ed. In this<br />
<strong>in</strong>vestigation Sugeno-type system is used, which can<br />
be used to mo<strong>de</strong>l any <strong>in</strong>ference system <strong>in</strong> which the<br />
output membership functions are either l<strong>in</strong>ear or<br />
constant.<br />
Fuzzy Mo<strong>de</strong>l<strong>in</strong>g Scenario<br />
To apply fuzzy <strong>in</strong>ference to a system for which a<br />
collection of <strong>in</strong>put/output data mo<strong>de</strong>ls, a mo<strong>de</strong>l<br />
structure based on characteristics of variables <strong>in</strong> a<br />
system does not necessarily have been<br />
pre<strong>de</strong>term<strong>in</strong>ed. In some mo<strong>de</strong>l<strong>in</strong>g situations, the<br />
membership functions cannot be discerned from<br />
look<strong>in</strong>g at data. Rather than choos<strong>in</strong>g the parameters<br />
associated with a given membership function<br />
arbitrarily, these parameters could be chosen so as to<br />
tailor the membership functions to the <strong>in</strong>put/output<br />
data <strong>in</strong> or<strong>de</strong>r to account for these types of variations<br />
<strong>in</strong> the data values. In such cases, neuro-adaptive<br />
learn<strong>in</strong>g techniques <strong>in</strong>corporated <strong>in</strong> the ANFIS<br />
comm<strong>and</strong> of MATLAB can be used. The acronym<br />
ANFIS <strong>de</strong>rives its name from adaptive neuro-fuzzy<br />
<strong>in</strong>ference system. Us<strong>in</strong>g a given <strong>in</strong>put/output data<br />
set, the toolbox function ANFIS constructs a fuzzy<br />
<strong>in</strong>ference system (FIS) whose membership function<br />
parameters are tuned (adjusted) us<strong>in</strong>g either a<br />
backpropagation algorithm alone or <strong>in</strong> comb<strong>in</strong>ation<br />
with a least squares type of method. This adjustment<br />
allows to fuzzy systems to learn from the data they<br />
are mo<strong>de</strong>l<strong>in</strong>g.<br />
FIS structure <strong>and</strong> parameter adjustment A<br />
network-type structure similar to that of a neural<br />
network, which maps <strong>in</strong>puts through <strong>in</strong>put<br />
membership functions <strong>and</strong> associated parameters,<br />
<strong>and</strong> then through output membership functions <strong>and</strong><br />
associated parameters to outputs, can be used to<br />
<strong>in</strong>terpret the <strong>in</strong>put/output map. The parameters<br />
associated with the membership functions changes<br />
through the learn<strong>in</strong>g process. The computation of<br />
these parameters (or their adjustment) is facilitated<br />
by a gradient vector. This gradient vector provi<strong>de</strong>s a<br />
measure of how well the fuzzy <strong>in</strong>ference system is<br />
mo<strong>de</strong>l<strong>in</strong>g the <strong>in</strong>put/output data for a given set of<br />
parameters. When the gradient vector is obta<strong>in</strong>ed,<br />
any of several optimization rout<strong>in</strong>es can be applied<br />
<strong>in</strong> or<strong>de</strong>r to adjust the parameters to reduce some<br />
error measure. This error measure is usually <strong>de</strong>f<strong>in</strong>ed<br />
by the sum of the squared difference between actual<br />
<strong>and</strong> <strong>de</strong>sired outputs. ANFIS uses either back<br />
propagation or a comb<strong>in</strong>ation of least squares<br />
estimation <strong>and</strong> backpropagation for membership<br />
function parameter estimation.<br />
MATERIALS AND METHODS<br />
Measurements were ma<strong>de</strong> <strong>in</strong> Chirgua River Bas<strong>in</strong>,<br />
located <strong>in</strong> the north central region of Venezuela dur<strong>in</strong>g<br />
2008–2009. Each subbas<strong>in</strong> was divi<strong>de</strong>d <strong>in</strong>to zones,<br />
which were classified as follows: high: I <strong>and</strong> II,<br />
medium: III, low: IV, V <strong>and</strong> VI. Each zone is ma<strong>in</strong>ly<br />
used for agricultural purposes. Information on the<br />
location, area <strong>and</strong> l<strong>and</strong> use is shown <strong>in</strong> Table 1. Five<br />
fields were selected with the follow<strong>in</strong>g slopes <strong>in</strong> tillage<br />
orientation: 0.008 ± 0.0055 m/m; 0.01 ± 0.00197 m/m;<br />
0.015 ± 0.0006 m/m; 0.025 ± 0.0033 m/m <strong>and</strong> 0.13 ±<br />
0.0156 m/m, whose soils range from silty s<strong>and</strong> to silty<br />
clay. Two types of crops usually are grown <strong>in</strong> rotation,<br />
accord<strong>in</strong>g to the season: dry (potato: Solanum<br />
Tuberosum) <strong>and</strong> ra<strong>in</strong>y (corn: Zea Mays); whose<br />
<strong>de</strong>velopment cycles have the follow<strong>in</strong>g time periods: 12<br />
<strong>and</strong> 18 weeks, respectively.<br />
The irrigation events for the whole cycle lasted as<br />
follows: two hours for the first ten weeks <strong>and</strong> one hour<br />
for the last two weeks. A spr<strong>in</strong>kler irrigation system was<br />
used to provi<strong>de</strong> water to crops. The <strong>in</strong>flow rate that was<br />
applied to furrows varied between 0.4 <strong>and</strong> 0.6 L/s. A<br />
plot of 64 wheel-compacted furrows on each field was<br />
split <strong>in</strong>to eigth, 8-furrow blocks. The tests were carried<br />
out <strong>in</strong> three furrows with a width <strong>and</strong> length as varied as<br />
follows: 0.3 <strong>and</strong> 0.35 m; 100 <strong>and</strong> 200 m, respectively.<br />
Furrows were divi<strong>de</strong>d <strong>in</strong>to four equal-length sections<br />
(¼, ½, ¾, <strong>and</strong> at the end of the furrow), <strong>and</strong><br />
measurement stations were established at the<br />
downstream end of each section. Flows were measured<br />
us<strong>in</strong>g a steel plate with a 60º V-notch weir, by apply<strong>in</strong>g<br />
the volumetric method. The number of measurements<br />
carried out dur<strong>in</strong>g the dry <strong>and</strong> ra<strong>in</strong>y seasons were 300<br />
<strong>and</strong> 162, respectively, which <strong>in</strong>clu<strong>de</strong>d: a) Soil physical<br />
properties: undisturbed samples were captured to<br />
<strong>de</strong>term<strong>in</strong>e through laboratory analysis saturated<br />
hydraulic conductivity, specific gravity, shear stress,<br />
moisture content, Atterberg limits <strong>and</strong> particle size<br />
(hydrometer <strong>and</strong> sieve analysis). Infiltration rates were<br />
measured with a s<strong>in</strong>gle-r<strong>in</strong>g <strong>in</strong>filtrometer (diameter 30<br />
cm), which was driven up to 5 cm <strong>de</strong>ep <strong>in</strong>to the soil.<br />
(Table 2 <strong>and</strong> Fig. 7), b) Flow physical properties: flow<br />
rates <strong>and</strong> sediment concentrations were measured at<br />
time <strong>in</strong>tervals rang<strong>in</strong>g from 20 to 120 m<strong>in</strong> at a time,<br />
after the arrival of water to each measur<strong>in</strong>g station.<br />
Sediment concentration was <strong>de</strong>term<strong>in</strong>ed by apply<strong>in</strong>g the<br />
2 540 B method (American Public Health Association,<br />
1995). 120 irrigation events were measured <strong>in</strong> five<br />
different agricultural fields (5 fields × 24 events / field =<br />
120 irrigation events). Dur<strong>in</strong>g the ra<strong>in</strong>y season the<br />
follow<strong>in</strong>g events were measured: 73 for 2 hours, 24 for<br />
an hour <strong>and</strong> 20 for half an hour, with a precipitation<br />
gauge attached to an electronic storage unit, which has<br />
follow<strong>in</strong>g geographic location: °W 68º 10’ 50’’<strong>and</strong><br />
°N 10º 11’ 45’’. Total data were divi<strong>de</strong>d <strong>in</strong>to three<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.81-104, 2010
Márquez <strong>and</strong> Guevara-Pérez<br />
87<br />
groups: 60% for calibration, 20% for validation <strong>and</strong><br />
20% for test<strong>in</strong>g.<br />
Mo<strong>de</strong>ls data<br />
Table 2 shows soil physical properties. Soil is<br />
composed of soil particles, where approximately less<br />
than 3% represents < 2.00 mm, 46.1% represents <<br />
0.074 mm (f<strong>in</strong>e fraction) <strong>and</strong> 50.9% represents >0.074<br />
mm (f<strong>in</strong>est fraction: silt <strong>and</strong> clay). Saturated hydraulic<br />
conductivity is close to 1 mm/h <strong>in</strong> most areas, which is<br />
classified as low to lower (Terzaghi & Peck, 1967). The<br />
friction angle varies from 29.8 to 32.7º. The cohesion<br />
varies from 0.1 to 0.43 kgf/cm 2 . The <strong>in</strong>itial <strong>and</strong> f<strong>in</strong>al<br />
moisture contents of the soil vary from 7.1 to 21.8%;<br />
from 26.1 to 54%, respectively. The porosity varies<br />
from 0.4 to 0.5 cm 3 /cm 3 . The liquid limit, plastic limit<br />
<strong>and</strong> plasticity <strong>in</strong><strong>de</strong>x range respectively as follows: 26.2<br />
<strong>and</strong> 33.2%; 23 <strong>and</strong> 28%; 4.3 <strong>and</strong> 5.2%. Specific gravity<br />
of solids varies from 2.5 to 2.6. Soil type varies from<br />
organic silt (OL) to plasticity low silty s<strong>and</strong> (SM) (e.g.,<br />
Lambe & Whitman, 1972; Juarez & Rico, 1991) (Table<br />
2).<br />
Infiltration rates for silty soil are highly variable.<br />
Figure 7(a) shows the cumulative <strong>in</strong>filtration versus<br />
time for 582 <strong>in</strong>filtrometer tests over one period, which<br />
last 2.28 h each. Infiltration <strong>de</strong>pths at the end of this<br />
time have a mean of 22.39 mm <strong>and</strong> a st<strong>and</strong>ard <strong>de</strong>viation<br />
of 21.02 mm. Infiltration <strong>de</strong>pths greater than 22.39 mm<br />
are measured dur<strong>in</strong>g the ra<strong>in</strong>y season, where ra<strong>in</strong>fall<br />
events occur less frequently than those of irrigation,<br />
therefore the storage capacity of the soil is higher than<br />
those available dur<strong>in</strong>g the irrigation period. Figure 7(b)<br />
shows the histogram for the cumulative <strong>in</strong>filtration at<br />
t = 2.28 h, together with the correspond<strong>in</strong>g probability<br />
<strong>de</strong>nsity functions. Exponential <strong>and</strong> Weibull functions<br />
result <strong>in</strong> the best fit.<br />
Table 2. Soil physical properties<br />
Soil physical<br />
Zone<br />
Properties I II III IV V VI<br />
µ σ CV µ σ CV µ σ CV µ σ CV µ σ CV µ σ CV<br />
Texture<br />
PReta<strong>in</strong>ed Percentage on Sieve<br />
>9.51mm<br />
(Sieve 3/8″) 3.9 4.4 112.9 0.0 0.0 - 1.2 1.1 93.2 1.2 1.5 126.5 0.0 0.0 - 0.3 0.5 146.9<br />
9.51 – 4.76 mm<br />
(Sieve # 4) 2.6 1.9 70.1 4.0 4.5 110.6 1.3 0.8 63.3 1.4 1.3 94.1 0.9 0.3 38.2 2.2 2.1 96.8<br />
4.76 – 2.00 mm<br />
(Sieve # 10) 4.0 2.5 61.8 5.5 5.3 95.3 2.7 1.1 39.0 2.7 1.5 55.9 3.7 2.8 75.9 0,7 1.3 188.2<br />
2.00 – 0.84 mm<br />
(Sieve # 20) 6.7 3.6 54.0 10.9 7.8 71.7 5.7 2.1 36.5 5.6 2.0 34.9 8.3 4.1 49.2 7.2 3.5 48.2<br />
0.84 – 0.42 mm<br />
(Sieve # 40) 10.9 5.2 47.3 14.5 6.4 44.2 8.7 2.8 32.1 8,3 1.8 22.1 13.8 3.9 28.7 10.4 3.7 36.1<br />
0.42 – 0,149 mm<br />
(Sieve # 100) 22.6 5.2 23.0 31.2 9.8 31.6 15.7 2.9 18.4 18.8 3.1 16.4 19.0 3.3 17.5 32.5 8.0 24.7<br />
0.149 – 0.074 mm<br />
(Sieve #200) 14.4 6.3 43.8 28.3 12,6 44.6 13.0 3.4 26.1 15.1 2.5 16.6 15.7 5.1 32.2 28.5 11.5 40.2<br />
< 0.074 mm<br />
Plate 34.9 9.3 26.6 5.9 3.5 60.3 51.7 9.5 18.4 48.2 8.6 17.9 38.7 7.8 20.2 18.3 12.0 65.8<br />
Total Percentage Average<br />
S<strong>and</strong> (%) 46.4 4.2 9.1 30.7 5.6 18.4 46.1 1.6 3.5 46.2 2.1 4.6 46.1 1.6 3.5 46.1 1.6 3.5<br />
Silt (%) 49.5 2.7 5.5 36.7 11.7 31.9 46.9 2.3 4.9 46.9 2.3 4.9 49.5 2.7 5.5 46.9 2.3 4.9<br />
Clay (%) 4.1 2.0 48.4 19.9 8.2 41.3 7.0 2.4 33.9 6.9 2.5 36.2 6.6 1.8 27.6 7.0 2.4 33.9<br />
Hydraulic Properties<br />
K s ¹ (mm/h) 0.4 0.1 33.3 24.0 22.7 94.3 0.9 0.6 60.6 1.4 0.6 41.8 0.9 0.6 60.6 0.9 0.6 60.6<br />
Structural Properties<br />
Friction Angle (º) 29.9 3.9 12.9 29.8 1.1 3.8 29.8 1.1 3.8 32.7 1.1 3.4 31.8 1.9 5.9 32.7 1.1 3.4<br />
Cohesion<br />
(kgf/cm 2 ) 0.3 0.0 15.4 0.1 0.0 60.0 0.1 0.0 60.0 0.3 0.0 6.7 0.43 0.03 7.0 0.3 0,0 6.7<br />
Phase Relations<br />
Initial moisture<br />
content (%) 21.8 11.5 52.6 7.1 3.5 49.2 18.9 11.8 62.5 15.0 7.1 47.2 20.7 10.6 51.3 18.9 11.8 62.5<br />
F<strong>in</strong>al moisture<br />
content (%) 26.1 11.5 44.0 51.0 11.7 22.9 40.5 7.7 19.0 54.0 12.6 23.4 40.5 7.7 19.0 40.5 7.7 19.0<br />
Porosity<br />
(cm 3 /cm 3 ) 0.4 0.1 13.5 0.4 0.1 11.6 0.4 0.1 14.6 0.4 0.1 18.9 0.5 0.1 16.7 0.4 0.1 14.6<br />
Liquid limit (%) 26.2 2.5 9.4 30.7 4.1 13.3 32.4 4.4 13.4 29.6 2.0 6.8 33.2 4.2 12.8 32.4 4.4 13.4<br />
Plastic limit (%) 23.1 5.0 21.6 26.2 3.7 14.0 27.1 2.9 10.6 25.2 3.4 13.6 28.3 3.9 13.9 27.1 2.9 10.6<br />
Plasticity <strong>in</strong><strong>de</strong>x<br />
(%) 5.1 2.2 43.8 4.5 2.5 54.3 5.2 4.0 76.0 4.3 2.6 60.0 4.9 1.2 25.0 5.2 4.0 76.0<br />
Specific gravity 2.6 0.1 3.9 2.5 0.1 4.8 2.5 0.1 4.3 2.6 0.2 7.7 2.7 0.1 3.7 2.5 0.1 4.3<br />
¹ Satured hydraulic conductivity, µ: mean, σ: st<strong>and</strong>ard <strong>de</strong>viation, CV: coefficient of variation <strong>in</strong> percentage (%)<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.81-104, 2010
Márquez <strong>and</strong> Guevara-Pérez<br />
88<br />
Fig. 7 Cumulative <strong>in</strong>filtration <strong>de</strong>pth: (a) temporal variation, <strong>and</strong> (b) frequency histogram <strong>and</strong> predicted Weibull probability<br />
<strong>de</strong>nsity function.<br />
Discharge, Q (L/s)<br />
Time (h)<br />
Fig. 8 Measured hydrographs for different slope furrows dur<strong>in</strong>g each irrigation.<br />
Dur<strong>in</strong>g the dry season, runoff is stabilized after the<br />
first half hour when the irrigation is applied. The time<br />
for flow stabilization is approximately one hour; which<br />
occurs faster <strong>in</strong> the furrows where the slope was equal<br />
to 0.8% (Fig. 8). Flow rate at steady state varies from<br />
0.3 to 0.5 L/s, <strong>de</strong>creas<strong>in</strong>g from 0.05 to 0.1 L/s after the<br />
seventh irrigation. The effect of the coverage radio of<br />
the spr<strong>in</strong>kler on flow vary<strong>in</strong>g between the different<br />
furrows is low with<strong>in</strong> an irrigation event.<br />
A comparison of the sediment load dur<strong>in</strong>g irrigation<br />
<strong>and</strong> ra<strong>in</strong> events of one <strong>and</strong> two hours is shown <strong>in</strong> Fig. 9.<br />
The maximum sediment load observed dur<strong>in</strong>g the<br />
irrigation <strong>and</strong> ra<strong>in</strong>fall events of one hour varies<br />
approximately from 100 to 400 mg/s, from 600 to<br />
100 mg/s, respectively (Fig. 9(a)). The maximum<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.81-104, 2010
Márquez <strong>and</strong> Guevara-Pérez<br />
89<br />
sediment load observed for the two-hour event varies<br />
from 500 to 2000 mg/s for irrigation <strong>and</strong> from 2500 to<br />
7000 mg/s for ra<strong>in</strong> (Fig. 9(b)). Therefore, there is<br />
significant variation <strong>in</strong> sediment load between the<br />
events of ra<strong>in</strong> <strong>and</strong> irrigation, where the first is 5 to 10<br />
times more erosive than the latter.<br />
A comparison of ra<strong>in</strong>fall events <strong>in</strong> both one <strong>and</strong> two<br />
hours is shown <strong>in</strong> Fig. 10, which shows that the ra<strong>in</strong>fall<br />
for these events ranges as follows: from 20 to 30 mm,<br />
<strong>and</strong> from 25 to 40 mm, respectively (Figs 10(a) <strong>and</strong><br />
(b)).<br />
Fig. 9 Measured sediment load dur<strong>in</strong>g ra<strong>in</strong>fall <strong>and</strong> irrigation events.<br />
Cumulative Precipitation, P (mm)<br />
Time (h)<br />
Fig. 10 Precipitation <strong>de</strong>pth.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.81-104, 2010
Márquez <strong>and</strong> Guevara-Pérez<br />
90<br />
RESULTS<br />
Mo<strong>de</strong>l<strong>in</strong>g of furrow erosion based on physical<br />
processes<br />
Process of particle <strong>de</strong>tachment<br />
Table 3 shows the parameters of Eq. (5), used to<br />
estimate the capacity of soil particle <strong>de</strong>tachment <strong>in</strong><br />
furrows. In general, the follow<strong>in</strong>g statements can be<br />
ma<strong>de</strong> about the parameters K c <strong>and</strong> τc: (a) K c varies from<br />
1.31E–06 to 1.525E–06 s m –1 ; (b) τc varies from 0.704<br />
to 1.489 Pa; (c) estimate st<strong>and</strong>ard error of the<br />
parameters is low <strong>in</strong> all cases.<br />
Table 4 shows fit statistics of Eq. (5) to the<br />
observations obta<strong>in</strong>ed from field test<strong>in</strong>g <strong>in</strong> the furrows<br />
of different slopes. The results are summarized as<br />
follows: R 2 is equal to 0.682, R 2 adj. is equal to 0.681, its<br />
reduction is not significant <strong>in</strong> relation to R 2 . Mallows C p<br />
<strong>de</strong>creases slightly compared to the number of<br />
<strong>in</strong><strong>de</strong>pen<strong>de</strong>nt variables <strong>in</strong> Eq. (5). Durb<strong>in</strong>-Watson<br />
statistic is equal to 0.66. By compar<strong>in</strong>g the errors dur<strong>in</strong>g<br />
the calibration <strong>and</strong> validation stages, with emphasis on<br />
the Average Absolute Relative Error (AARE), <strong>and</strong> the<br />
Average Relative Error (ARE) the follow<strong>in</strong>g results are<br />
found: AARE varies from 47.87 to 57.9, ARE varies<br />
from –20.52 to –33.74. In general, the errors do not vary<br />
significantly between the stages of calibration <strong>and</strong><br />
validation. The rest of errors can be seen <strong>in</strong> Table 4.<br />
Process of sediment transport<br />
Table 5 shows the parameters of Eqs 7, 13 <strong>and</strong> 15 for<br />
estimat<strong>in</strong>g the sediment transport capacity <strong>in</strong> furrows. In<br />
general, the follow<strong>in</strong>g statements can be ma<strong>de</strong> about the<br />
parameters K t , b <strong>and</strong> c: Coefficient of transport (K t ):<br />
(a) the units differ between Eqs 7, 13 <strong>and</strong> 15, (b) the<br />
ranges found for Eqs 7, 13 <strong>and</strong> 15 vary as follows:<br />
9.929E–08 <strong>and</strong> 1.526E–07; 7.634 <strong>and</strong> 34.583; 0.421 <strong>and</strong><br />
0.892, respectively, (c) estimate st<strong>and</strong>ard error of<br />
parameters is mo<strong>de</strong>rately low <strong>in</strong> all cases. Parameter b:<br />
(a) the ranges found for Eqs 7, 13 <strong>and</strong> 15 vary as<br />
follows: 1.519 <strong>and</strong> 1.705; 0.398 <strong>and</strong> 1.299; 2.087 <strong>and</strong><br />
2.228; respectively, (b) estimate st<strong>and</strong>ard error of<br />
parameters is low <strong>in</strong> all cases. Parameter c: (a) the<br />
range found for Eq. (15) varies from 1.097 to 1.193; (b)<br />
estimate st<strong>and</strong>ard error of parameters is low <strong>in</strong> all cases.<br />
Table 6 shows fit statistics of Eqs 7, 13 <strong>and</strong> 15 to<br />
the observations obta<strong>in</strong>ed from field test<strong>in</strong>g <strong>in</strong> the<br />
furrows. R 2 : varies from 0.72 to 0.88. R 2 adj.: its<br />
reduction is not significant <strong>in</strong> relation to R 2 . Mallows<br />
C p : <strong>de</strong>creases slightly compared to the number of<br />
<strong>in</strong><strong>de</strong>pen<strong>de</strong>nt variables <strong>in</strong> each equation. Durb<strong>in</strong>-Watson<br />
statistic: values found to Eqs 7, 13 <strong>and</strong> 15 are equal to<br />
0.719, 1.09 <strong>and</strong> 0.74, respectively. By compar<strong>in</strong>g the<br />
errors dur<strong>in</strong>g the calibration <strong>and</strong> validation stages with<br />
emphasis on the Average Absolute Relative Error<br />
(AARE) <strong>and</strong> the Average Relative Error (ARE) the<br />
follow<strong>in</strong>g results <strong>in</strong> Eqs 7, 13 <strong>and</strong> 15 are found: AARE<br />
varies from 56.319 to 63.081; from 12.125 to 13.482;<br />
from 29.585 to 29.894. ARE varies from –13.365 to<br />
–1.831; from –1.831 to 1.974; from –16.597 to –18.547.<br />
In general, the errors do not vary significantly between<br />
calibration <strong>and</strong> validation stages. The rest of errors can<br />
be seen <strong>in</strong> Table 6.<br />
Comb<strong>in</strong>ation of furrow erosion processes<br />
The representation of furrow erosion processes for<br />
follow<strong>in</strong>g slopes: 0.8%, 1%, 1.5% <strong>and</strong> 2.5% is shown <strong>in</strong><br />
Fig. 11, us<strong>in</strong>g the Eqs 7, 13 <strong>and</strong> 15 for estimat<strong>in</strong>g the<br />
transport capacity of sediments (Figs 11a, 11c, 11e <strong>and</strong><br />
11h, <strong>and</strong> Eqs 4 <strong>and</strong> 12 to estimate the net <strong>de</strong>tachment<br />
capacity (Figs 11b, 11d, 11f <strong>and</strong> 11g). The cycle un<strong>de</strong>r<br />
irrigation has been divi<strong>de</strong>d <strong>in</strong>to three stages; each stage<br />
represents 1/3 of the total duration for 12 weeks,<br />
obta<strong>in</strong><strong>in</strong>g data average values for each phase.<br />
Table 3. Regression coefficients of particle <strong>de</strong>tachment capacity mo<strong>de</strong>l<br />
Eq. Variable Parameter Unit Average St<strong>and</strong>ard Error M<strong>in</strong>imum Limit Maximum Limit<br />
5 D c K r s/m 0.000 001 418 5.43 497E–8 0.00 000 131 163 0.000 0015 255<br />
(kg/s/m 2 ) τ c Pa 1.09 707 0.199 697 0.704 172 1.48 997<br />
Table 4. Performance criteria of particle <strong>de</strong>tachment capacity mo<strong>de</strong>l<br />
Eq. p n R 2 (R 2 ) adj. C p SEE d MSSE AAE AARE AE ARE n AARE ARE<br />
(a) Dur<strong>in</strong>g Calibration<br />
(b) Dur<strong>in</strong>g Validation<br />
5 1 320 0.682 0.681 0 2.45E–6 0.66 6.04E–12 2.08E–06 47.87 –7.72E–13 –20.52 84 57.9 –33.74<br />
Eq.: equation; p: number of <strong>in</strong><strong>de</strong>pen<strong>de</strong>nt variables <strong>in</strong> the mo<strong>de</strong>l; n: number of observed data; R 2 : <strong>de</strong>term<strong>in</strong>ation coefficient; (R 2 ) adj. : adjusted<br />
<strong>de</strong>term<strong>in</strong>ation coefficient; C p : coefficient of Mallows; SEE: St<strong>and</strong>ard error of estimate; d: statistic of Durb<strong>in</strong>-Watson; MSSE: mean of sum of<br />
squared error; AAE: average absolute error; AARE: average absolute relative error (%); AE: average error; ARE: average relative error (%).<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.81-104, 2010
Márquez <strong>and</strong> Guevara-Pérez<br />
91<br />
Fig. 11 Comparisons of simulated furrow erosion processes with observed data.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.81-104, 2010
Márquez <strong>and</strong> Guevara-Pérez<br />
92<br />
Table 5. Regression coefficients of sediment transport capacity mo<strong>de</strong>l<br />
Eq. Variable Parameter Unit Average St<strong>and</strong>ard Error M<strong>in</strong>imum Limit Maximum Limit<br />
7 T c K t kg 1–b m –(1–b) s –(1–2b) 1.2598E–7 1.35 793E–8 9.92 906E–8 1.52 669E–7<br />
kg/m/ s b 1.61 238 0.0 473 371 1.51 934 1.70 542<br />
13 T c K t kg m –(b+1) s –(1–b) 21.1093 6.843 7.634 34.583<br />
b 0.8489 0.2287 0.3985 1.2993<br />
kg/m/s ω c m s –1 –0.05 414 0.0221 –0.0978 –0.0104<br />
15 T c K t kg m (–1–b–c) s –(1+b) 0.657 032 0.120 287 0.421 274 0.89 279<br />
b 2.15 807 0.0 359 035 2.0877 2.22 844<br />
kg/m/ s c 1.14 551 0.0 244 874 1.09 751 1.1935<br />
Table 6. Performance criteria of sediment transport capacity mo<strong>de</strong>l<br />
Eq. p n R 2 R 2 adj C p SEE d MSSE AAE AARE AE ARE n AARE ARE<br />
(a) Dur<strong>in</strong>g Calibration<br />
(b) Dur<strong>in</strong>g Validation<br />
7 1 437 0.73 0.732 0 1.102E–06 0.719 1.22E–12 9.16E–07 63.08 4.24E–08 –31.831 100 56.31 –13.36<br />
13 1 320 0.72 0.720 0 0.34 1.09 0.115 0.287 12.12 0.0028 –1.831 96 13.48 1.974<br />
15 2 700 0.88 0.882 1 1.644E–06 0.741 2.70E–12 1.08E–06 29.58 –5.75E–08 –16.597 165 29.89 –18.54<br />
Comparison between the transport capacities of<br />
sediments observed <strong>and</strong> estimated by Eqs 7, 13 <strong>and</strong> 15<br />
shows <strong>in</strong> Figs 11a, 11c, 11e <strong>and</strong> 11h. The solid l<strong>in</strong>e<br />
represents the observed sediment load <strong>in</strong> furrows, while<br />
the bars represent estimates by Eqs 7, 13 <strong>and</strong> 15, where<br />
the follow<strong>in</strong>g observations can be ma<strong>de</strong>: (a) Eq. (13)<br />
overestimated the values of the observed transport<br />
capacity, but the rest of equations un<strong>de</strong>restimated them<br />
<strong>and</strong> (b) the observed sediment load <strong>in</strong> stages from 1 to 3<br />
varies as follows: from 1000 to 2500 mg/s, from 2000 to<br />
500 mg/s; less than 500 mg/s.<br />
Values estimated by Eq. (7) <strong>in</strong> stages from 1 to 3<br />
vary as follows: from 500 to 1000 mg/s, from 0 to 500<br />
mg/s, less than 500 mg/s. Values estimated by Eq. (13)<br />
<strong>in</strong> stages from 1 to 3 vary as follows: from 1000 to 2500<br />
mg/s, from 2000 to 500 mg/s, less than 500 mg/s.<br />
Estimates from Eq. (15) <strong>in</strong> steps 1 through 3 vary as<br />
follows: from 1500 to 3500 mg/s, from 500 to 2000<br />
mg/s, from 500 to 1000 mg/s, (c) the difference between<br />
observed <strong>and</strong> estimated values is significant <strong>in</strong> the first<br />
stage of the cycle <strong>and</strong> <strong>de</strong>creases towards the second <strong>and</strong><br />
third stages.<br />
Accord<strong>in</strong>g to the theory of mo<strong>de</strong>ls based on physical<br />
processes, when the sediment transport capacity is<br />
excee<strong>de</strong>d by the flow sediment load occurs the<br />
<strong>de</strong>position process (Foster <strong>and</strong> Meyer, 1975, Morgan et<br />
al., 1998).<br />
Deposition process estimated by Eqs 4 <strong>and</strong> 12 is<br />
shown <strong>in</strong> Figs 11(b), 11(d), 11(f) <strong>and</strong> 11(g), where the<br />
follow<strong>in</strong>g results <strong>de</strong>term<strong>in</strong>e: (a) most times Eq. (12)<br />
estimates higher values than Eq. (4), (b) the estimated<br />
values by Eq. (4) <strong>in</strong> the stages from 1 to 3 vary as<br />
follows: from 2 to 6 kg/m 2 , from 0 to 4 kg/m 2 , from 0 to<br />
2 kg/m 2 , respectively. Estimates from Eq. (12) <strong>in</strong> stages<br />
from 1 to 3 vary as follows: from 6 to 12 kg/m 2 , from 2<br />
to 6 kg/m 2 , from 2 to 4 kg/m 2 , (c) the difference<br />
between observed <strong>and</strong> estimated values is significant<br />
dur<strong>in</strong>g the first stage of the cycle <strong>and</strong> <strong>de</strong>creases towards<br />
the second <strong>and</strong> third stages.<br />
Mo<strong>de</strong>l<strong>in</strong>g furrow erosion based on regressions<br />
Table 7 shows the parameter values <strong>in</strong> the furrow<br />
erosion mo<strong>de</strong>ls based on l<strong>in</strong>ear regressions, Eqs 17(a)<br />
<strong>and</strong> (b), as well as nonl<strong>in</strong>ear regressions, Eqs 18(a)–(e).<br />
In general, the follow<strong>in</strong>g statements can be ma<strong>de</strong> <strong>in</strong><br />
relation to the parameters β 1 to β 6 : <strong>in</strong> Eq. (17a) the<br />
parameters vary from –4.267 to 4.382, <strong>in</strong> Eq. (17b) they<br />
vary from –7.5 to 9.14, <strong>in</strong> Eq. (18a) they vary from –<br />
0.031 <strong>and</strong> 0.2697, <strong>in</strong> Eq. (18b) they vary from –0.014 to<br />
0.093, <strong>in</strong> Eq. (18c) they vary from –0.629 to 1.938, <strong>in</strong><br />
Eq. (18d) they vary from –0.046 to 1.012.<br />
Table 8 shows the fitted statistics of erosion mo<strong>de</strong>ls<br />
based on regression from the observations obta<strong>in</strong>ed<br />
when the field was exam<strong>in</strong>ed <strong>in</strong> furrows, <strong>in</strong> the<br />
follow<strong>in</strong>g slopes: 0.8%, 1%, 1.5%, <strong>and</strong> 2.5%. R 2 : varies<br />
from 0.90 to 0.53, <strong>and</strong> it <strong>de</strong>creases when the polynomial<br />
power is <strong>in</strong>creased. R 2 adj its reduction is not significant<br />
<strong>in</strong> relation to R 2 . Mallows C p : <strong>de</strong>creases slightly<br />
compared to the number of <strong>in</strong><strong>de</strong>pen<strong>de</strong>nt variables <strong>in</strong><br />
each equation. Durb<strong>in</strong>-Watson Statistics: varies<br />
between 0.7 <strong>and</strong> 1.68. By compar<strong>in</strong>g the errors dur<strong>in</strong>g<br />
the calibration <strong>and</strong> validation stages, with emphasis on<br />
the average absolute relative error (AARE), <strong>and</strong> the<br />
average relative error (ARE) the follow<strong>in</strong>g results are<br />
found: AARE varies from 6.21 to 12.57, ARE varies<br />
from –0.71 to 6.78. In general, (a) errors ten<strong>de</strong>d to be<br />
lower <strong>in</strong> the calibration stage with respect to validation<br />
<strong>and</strong> (b) errors <strong>in</strong>crease when the polynomial power is<br />
<strong>in</strong>creased. The rest of errors can be seen <strong>in</strong> Table 8.<br />
Figure 12 shows a comparison of the estimated<br />
erosion results us<strong>in</strong>g Eqs 17(a), 17(b), 18(a), 18(b),<br />
18(c) <strong>and</strong> 18(d) to the observations un<strong>de</strong>r irrigation <strong>in</strong><br />
the furrows with follow<strong>in</strong>g slopes: 0.8%, 1%, 1.5% <strong>and</strong><br />
2.5%. Additionally, it <strong>in</strong>clu<strong>de</strong>s a comparison with the<br />
physical process mo<strong>de</strong>l represented by Eq. (16), <strong>and</strong> the<br />
artificial <strong>in</strong>telligence technique based on neural<br />
networks, which will be wi<strong>de</strong>ly discussed <strong>in</strong> the next<br />
section.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.81-104, 2010
Márquez <strong>and</strong> Guevara-Pérez<br />
93<br />
Table 7. Regression coefficients of erosion mo<strong>de</strong>l based on LMR <strong>and</strong> NLMR<br />
Eq. Depen<strong>de</strong>nt<br />
variable<br />
In<strong>de</strong>pen<strong>de</strong>nt<br />
variable<br />
Parameter Unit Average St<strong>and</strong>ard<br />
Error<br />
M<strong>in</strong>imum<br />
Limit<br />
Maximum<br />
Limit<br />
17(a) D Pt β mg m –2 s –1 mm –1 1.398583862 1.442760413 –4.267676785 1.47050906<br />
-<br />
LMRM<br />
r<br />
mg m –2 s –1 Pt-1<br />
1<br />
β 2 mg m –2 s –1 mm –1 0.833406797 1.784933149 –2.716135469 4.382949064<br />
n=1 Pt-2 β 3 mg m –2 s –1 mm –1 0.731227333 1.392642416 –2.03820026 3.500654927<br />
D=1H Pt-3 β 4 mg m –2 s –1 mm –1 0.183129882 1.814769332 –3.425745022 3.792004785<br />
Pt-4 β 5 mg m –2 s –1 mm –1 0.737812187 1.250395401 –1.748741065 3.224365439<br />
Pt-5 β 6 mg m –2 s –1 mm –1 0.559520883 0.731816989 –0.895780305 2.014822071<br />
17(b) D r Pt β 1 mg m –2 s –1 mm –1 0.794050629 1.4177439 –3.585879266 1.997778008<br />
-<br />
LMRM mg m –2 s –1 Pt-1 β mg m –2 s –1 mm –1 2.195990758 2.69409942 –7.5012253 3.109243783<br />
-<br />
n=1 Pt-2<br />
2<br />
β 3 mg m –2 s –1 mm –1 0.780383734 3.769700299 –6.642929634 8.203697103<br />
D=2H Pt-3 β 4 mg m –2 s –1 mm –1 2.177019018 1.357605748 –0.496385258 4.850423294<br />
Pt-4 β 5 mg m –2 s –1 mm –1 4.541647609 2.338266456 –0.062878815 9.146174033<br />
Pt-5 β 6 mg m –2 s –1 mm –1 0.099997046 0.333760338 –0.557245543 0.757239634<br />
18(a) D r Pt β 1 mg m –2 s –1 mm –1 0.119031048 0.073909116 –0.031707895 0.269769991<br />
NLMR<br />
M mg m –2 s –1 Pt-1 β 2 mg m –2 s –1 mm –1 0.055276888 0.084390798 –0.116839631 0.227393406<br />
n=2 Pt-2 β 3 mg m –2 s –1 mm –1 0.066816629 0.061319061 –0.058244676 0.191877933<br />
D=1H Pt-3 β 4 mg m –2 s –1 mm –1 0.050906233 0.084917164 –0.122283818 0.224096283<br />
Pt-4 β 5 mg m –2 s –1 mm –1 0.075651656 0.055759098 –0.038070006 0.189373318<br />
Pt-5 β 6 mg m –2 s –1 mm –1 0.05458908 0.025644464 0.002286744 0.106891417<br />
18(b) D r Pt β 1 mg m –2 s –1 mm –1 0.065622334 0.013487094 0.038182579 0.093062088<br />
NLMR<br />
M mg m –2 s –1 Pt-1 β 2 mg m –2 s –1 mm –1 0.014849305 0.014240642 –0.014123558 0.043822168<br />
n=3 Pt-2 β 3 mg m –2 s –1 mm –1 0.002274855 0.007430699 –0.012843047 0.017392756<br />
D=1H Pt-3 β 4 mg m –2 s –1 mm –1 0.01600253 0.014611824 –0.013725509 0.045730568<br />
Pt-4 β 5 mg m –2 s –1 mm –1 0.008392722 0.009444179 –0.010821644 0.027607088<br />
Pt-5 β 6 mg m –2 s –1 mm –1 0.006471389 0.003122043 0.000119533 0.012823245<br />
18 (c) D r Pt β 1 mg m –2 s –1 mm –1 0.565467014 0.170384074 0.227999229 0.902934798<br />
NLMR<br />
M mg m –2 s –1 Pt-1 β 2 mg m –2 s –1 mm –1 0.060856563 0.260216264 –0.454535539 0.576248665<br />
n=2 Pt-2 β 3 mg m –2 s –1 mm –1 1.110633143 0.418014423 0.282701296 1.938564991<br />
D=2H Pt-3 β 4 mg m –2 s –1 mm –1 0.334619992 0.145250564 0.046932389 0.622307595<br />
Pt-4 β 5 mg m –2 s –1 mm –1 –1.53E–01 0.240668154 –0.629188284 0.324160779<br />
Pt-5 β 6 mg m –2 s –1 mm –1 0.042533197 0.020393795 0.002140637 0.082925756<br />
18(d) D r Pt β 1 mg m –2 s –1 mm –1 0.305101494 0.068845599 0.168692499 0.44151049<br />
NLMR<br />
M mg m –2 s –1 Pt-1 β 2 mg m –2 s –1 mm –1 0.1154 0.0815 –0.0460 0.2769<br />
n=3 Pt-2 β 3 mg m –2 s –1 mm –1 0.691301164 0.161871969 0,370571984 1.012030344<br />
D=2H Pt-3 β 4 mg m –2 s –1 mm –1 0.08363529 0.049106013 –0.013662171 0.18093275<br />
Pt-4 β 5 mg m –2 s –1 mm –1 –0.3610 0.0819 –0.5233 –0.1988<br />
Pt-5 β 6 mg m –2 s –1 mm –1 0.01057652 0.003464163 0.003712711 0.017440329<br />
Table 8. Performance criteria of erosion mo<strong>de</strong>l based on LMR <strong>and</strong> NLMR<br />
Eq. Or<strong>de</strong>r D p N R 2 (R 2 ) adj Cp SEE d MSSE AAE AARE AE ARE n AARE ARE<br />
(a) Dur<strong>in</strong>g Calibration<br />
(b) Dur<strong>in</strong>g Validation<br />
17(a) 1 1 6 90 0.900 0.894 5 2.85 0.146 8.179 2.121<br />
17(b) 1 2 6 264 0.895 0.893 5 4.22 0.160 17.874 3.179<br />
18(a) 2 1 6 37 0.717 0.672 5 0.78 0.181 0.609 0.593 7.184 0.0008 –0.71 8 9.09 6.78<br />
18(b) 3 1 6 39 0.705 0.660 5 0.84 0.708 0.909 0.708 8.486 0.0024 –0.78 6 6.21 –2.31<br />
18(c) 2 2 6 122 0.725 0.713 5 1.50 0.945 2.252 1.183 10.060 0.0685 –0.45 10 6.83 4.06<br />
18(d) 3 2 6 118 0.538 0.518 5 1.90 1.680 3.626 1.430 11.859 0.2065 0.90 14 12.57 3.31<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.81-104, 2010
Márquez <strong>and</strong> Guevara-Pérez<br />
94<br />
Erosion, Dr (kg / m 2 )<br />
Irrigation Stage<br />
Fig. 12 Comparison of furrow simulated erosion by regression mo<strong>de</strong>ls with observed data.<br />
The cycle un<strong>de</strong>r irrigation has been divi<strong>de</strong>d <strong>in</strong>to<br />
three stages; each stage represents 1/3 of the total<br />
duration for 12 weeks, obta<strong>in</strong><strong>in</strong>g an averaged data for<br />
each phase. In Fig. 12, solid l<strong>in</strong>es represent the mo<strong>de</strong>ls<br />
by which a satisfactory approximation to the furrow<br />
erosion observed data is obta<strong>in</strong>ed, <strong>and</strong> the bars represent<br />
the rest of mo<strong>de</strong>ls, where the estimated values differ<br />
significantly from the observed data. In general, Eq.<br />
(17b) achieves the best approximation with respect to<br />
the rest of the l<strong>in</strong>ear <strong>and</strong> nonl<strong>in</strong>ear mo<strong>de</strong>ls.<br />
Mo<strong>de</strong>l<strong>in</strong>g of furrow erosion based on artificial<br />
<strong>in</strong>telligence techniques<br />
Table 9 shows the fitted statistics to sigmoidal function<br />
us<strong>in</strong>g the Artificial Neural Network (ANN) technique to<br />
the observations of ra<strong>in</strong>fall-erosion process for the onehour<br />
<strong>and</strong> two-hour events, both with different amounts<br />
of data. Statistics <strong>in</strong> the stages of tra<strong>in</strong><strong>in</strong>g, validation<br />
<strong>and</strong> test<strong>in</strong>g vary for the networks mentioned before as<br />
follows: R²: For ANN1 <strong>in</strong> each stage is equal to 0.686;<br />
0.897 <strong>and</strong> 0.876; for ANN2 <strong>in</strong> each stage is equal to<br />
0.39; 0.615 <strong>and</strong> 0.49.<br />
For ANN3 <strong>in</strong> each stage is equal to 0.758; 0.831 <strong>and</strong><br />
0.786. By compar<strong>in</strong>g the errors dur<strong>in</strong>g the calibration<br />
<strong>and</strong> validation stages, with emphasis on the average<br />
absolute error (AAE), the follow<strong>in</strong>g results are found:<br />
AAE varies from 0.1 to 0.14. In general, (a) the error<br />
does not vary significantly between stages of<br />
calibration, validation <strong>and</strong> test<strong>in</strong>g, (b) the error tends to<br />
be lower as the variability <strong>in</strong> the data is lower, which<br />
promotes a better network learn<strong>in</strong>g with respect to the<br />
observations ma<strong>de</strong>.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.81-104, 2010
Márquez <strong>and</strong> Guevara-Pérez<br />
95<br />
(a)<br />
D=1 h<br />
N=144<br />
(b) D=1 h<br />
N=144<br />
(c) D=1 h<br />
N=144<br />
(d) D=2h<br />
N=438<br />
(e) D= 2h<br />
N=438<br />
(f) D=2h<br />
N=438<br />
Simulated Erosion (Dr) mg /m 2 /s<br />
(g) D=2 h<br />
N=219<br />
(h) D=2h<br />
N=219<br />
(i)D= 2h<br />
N=219<br />
Observed Erosion (DR) mg/m 2 /s<br />
Observed Data<br />
Tra<strong>in</strong><strong>in</strong>g Best L<strong>in</strong>ear Fit<br />
Validation Best L<strong>in</strong>ear Fit<br />
Test Best L<strong>in</strong>ear Fit<br />
Fig. 13 Comparison of simulated erosion by Artificial Neural Network with observed 1 h <strong>and</strong> 2h ra<strong>in</strong>fall events.<br />
Comparison of estimated values by the application of<br />
ANN’s, to the observed data of furrow erosion un<strong>de</strong>r<br />
ra<strong>in</strong>fall <strong>and</strong> irrigation events of one <strong>and</strong> two hours,<br />
dur<strong>in</strong>g the stages of tra<strong>in</strong><strong>in</strong>g or calibration, validation<br />
<strong>and</strong> test<strong>in</strong>g, is shown <strong>in</strong> Fig. 13. Vertical axis represents<br />
the simulated erosion (D R sim), <strong>and</strong> the horizontal axis<br />
represents the erosion observed (D R obs.). Each pair of<br />
coord<strong>in</strong>ates is represented by a dot.<br />
L<strong>in</strong>es <strong>in</strong> Figs 13(a)–(i) represent the ratio of 1:1,<br />
where the approximation of po<strong>in</strong>ts to the l<strong>in</strong>es is a<br />
measure of how well the correlation between the set of<br />
erosion estimated values us<strong>in</strong>g the ANN technique, <strong>and</strong><br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.81-104, 2010
Márquez <strong>and</strong> Guevara-Pérez<br />
96<br />
the obta<strong>in</strong>ed values by field tests <strong>in</strong> furrows <strong>in</strong> different<br />
slopes is. Figures 13a–13c <strong>and</strong> Figs 13g–13i show the<br />
approximate po<strong>in</strong>ts r<strong>and</strong>omly around the 1:1 l<strong>in</strong>e,<br />
<strong>in</strong>dicat<strong>in</strong>g that the network has successfully learned the<br />
pattern shown by observations dur<strong>in</strong>g ra<strong>in</strong>fall events.<br />
Figures 13d–13f show that there is a fairly satisfactory<br />
approximation of the po<strong>in</strong>ts to the 1:1 l<strong>in</strong>e, which<br />
expla<strong>in</strong>s that there is some variability <strong>in</strong> the<br />
observations that the network was not able to simulate.<br />
Fuzzy Logic<br />
Mo<strong>de</strong>l<strong>in</strong>g furrow erosion us<strong>in</strong>g a Fuzzy Inference<br />
System<br />
Table 10 shows the two membership functions<br />
coefficients for the two <strong>in</strong>put variables to each Fuzzy<br />
Inference System (FIS). Characterized FIS’s are three:<br />
FIS 1 corresponds to natural ra<strong>in</strong>fall events that last an<br />
hour, FIS 2 corresponds to the natural ra<strong>in</strong>fall events<br />
that last two hours, <strong>and</strong> FIS 3 corresponds to irrigation<br />
events that last two hours. Input 1 is represented by<br />
measurements of natural <strong>and</strong> simulated ra<strong>in</strong>fall<br />
(irrigation) <strong>in</strong> those events that last an hour every 10<br />
m<strong>in</strong>utes, <strong>and</strong> two hours every 20 m<strong>in</strong>utes. Input 2 is<br />
represented by the <strong>in</strong>filtration values range as shown <strong>in</strong><br />
Fig. 7.<br />
As an example, FIS 1 is <strong>de</strong>scribed, represented by<br />
<strong>in</strong>put membership functions of trapezoidal type. The<br />
trapezoidal functions for the ra<strong>in</strong>fall representation have<br />
the follow<strong>in</strong>g parameters: trapmf1: 1.827; 2.974; 4.557<br />
<strong>and</strong> 5.928. Trapmf2: 4.677; 5.705; 7.562 <strong>and</strong> 8.709. The<br />
trapezoidal functions for the <strong>in</strong>filtration representation<br />
have the follow<strong>in</strong>g parameters: trapmf1: –0.508; 0.0293,<br />
0.859; 1.381. Trapmf2: 3.637; 4.676; 7.46 <strong>and</strong> 9.147.<br />
The l<strong>in</strong>ear output surface has the follow<strong>in</strong>g parameters:<br />
2.081 for the ra<strong>in</strong>fall; –2.675 for <strong>in</strong>filtration <strong>and</strong> 0.0487<br />
for the constant term. The rest of FIS’s can be seen <strong>in</strong><br />
Table 10. Table 11 shows the fitted statistics of erosion<br />
mo<strong>de</strong>l<strong>in</strong>g un<strong>de</strong>r ra<strong>in</strong>fall <strong>and</strong> irrigation events through<br />
the various FIS’s from all the observations obta<strong>in</strong>ed <strong>in</strong><br />
field test<strong>in</strong>g <strong>in</strong> furrows of different slopes. By compar<strong>in</strong>g<br />
the Average Absolute Errors dur<strong>in</strong>g the calibration,<br />
validation <strong>and</strong> test<strong>in</strong>g stages, the follow<strong>in</strong>g results are found:<br />
FIS 1: 1.108; 1.5147 <strong>and</strong> 1.489; FIS 2: 2.403; 2.5977 <strong>and</strong><br />
2.676; FIS 3: 1.4; 2.396 <strong>and</strong> 1.565, respectively. In general,<br />
errors do not vary significantly between the stages of the<br />
calibration, validation <strong>and</strong> test<strong>in</strong>g.<br />
Table 9. Performance criteria of erosion mo<strong>de</strong>l based on ANN<br />
Mo<strong>de</strong>l N D AAE R 2<br />
(a) Dur<strong>in</strong>g calibration/tra<strong>in</strong><strong>in</strong>g<br />
ANN 1: 6-20-6 144 1 0.13962 0.686<br />
ANN 2: 6-100-6 438 2 0.1750 0.390<br />
ANN 3: 6-100-6 219 2 0.1283 0.758<br />
(b) Dur<strong>in</strong>g validation<br />
ANN 1: 6-20-6 144 1 0.146 0.897<br />
ANN 2: 6-100-6 438 2 0.166 0.615<br />
ANN 3: 6-100-6 219 2 0.103 0.831<br />
(c) Dur<strong>in</strong>g test<strong>in</strong>g<br />
ANN 1: 6-20-6 144 1 0.146 0.876<br />
ANN 2: 6-100-6 438 2 0.166 0.490<br />
ANN 3: 6-100-6 219 2 0.1039 0.786<br />
N: number of observed data; D: irrigation <strong>and</strong> ra<strong>in</strong>fall events duration; AAE: average absolute error; R 2 : <strong>de</strong>term<strong>in</strong>ation coefficient; ANN:<br />
artificial neural network<br />
Table 10. Coefficients of membership function from various FIS<br />
Mo<strong>de</strong>l Inputs Membership Function Parameter- 1 Parameter- 2 Parameter- 3 Parameter- 4<br />
Input 1 Trapmf-1 1.827 2.974 4.557 5.928<br />
FIS 1<br />
Input 1 Trapmf-2 4.677 5.705 7.562 8.709<br />
Input 2 Trapmf-1 –0.508 0.0293 0.859 1.381<br />
Input 2 Trapmf-2 3.637 4.676 7.460 9.147<br />
Output l<strong>in</strong>ear 2.801 –2.675 0.048<br />
Input 1 Trapmf-1 –0.9727 0.7139 3.089 5.373<br />
FIS 2<br />
Input 1 Trapmf-2 3.637 4.676 7.46 9.147<br />
Input 2 Trapmf-1 –0.5081 0.0293 0.859 1.381<br />
Input 2 Trapmf-2 3.637 4.676 7.46 9.147<br />
Output L<strong>in</strong>ear 5.13 0.1209 –0.179<br />
Input 1 Trapmf-1 –5.071 –1.601 4.226 7.563<br />
FIS 3<br />
Input 1 Trapmf-2 5.466 7.789 12.28 15.75<br />
Input 2 Trapmf-1 –0.501 0.021 0.859 1.381<br />
Input 2 Trapmf-2 0.979 1.398 2.179 2.717<br />
Output L<strong>in</strong>ear 2.298 2.073 4.518<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.81-104, 2010
Márquez <strong>and</strong> Guevara-Pérez<br />
97<br />
Table 11. Performance criteria of erosion mo<strong>de</strong>l based on FIS<br />
Mo<strong>de</strong>l N D AAE<br />
(a) Dur<strong>in</strong>g calibration/tra<strong>in</strong><strong>in</strong>g<br />
FIS 1 (trapmf) 88 1 1.108<br />
FIS 2 (trapmf) 262 2 2.007<br />
FIS 3 (trapmf) 180 2 1.400<br />
(b) Dur<strong>in</strong>g validation<br />
FIS 1 (trapmf) 28 1 1.5147<br />
FIS 2 (trapmf) 88 2 2.1407<br />
FIS 3 (trapmf) 60 2 2.396<br />
(c) Dur<strong>in</strong>g test<strong>in</strong>g<br />
FIS 1 (trapmf) 28 1 1.489<br />
FIS 2 (trapmf) 88 2 2.367<br />
FIS 3 (trapmf) 60 2 1.565<br />
N: number of observed data; D: irrigation <strong>and</strong> ra<strong>in</strong>fall events duration; AAE: average absolute error; R 2 : <strong>de</strong>term<strong>in</strong>ation coefficient; ANN:<br />
artificial neural network<br />
(a)<br />
Erosion DR (mg m ‐2 s ‐1 )<br />
(b)<br />
(c)<br />
FIS Data Po<strong>in</strong>ts * Tra<strong>in</strong><strong>in</strong>g Data Po<strong>in</strong>ts o Validation Data Po<strong>in</strong>ts + Test Data Po<strong>in</strong>ts<br />
Fig. 14 Comparison of simulated erosion by Fuzzy Inference System with observed 2h ra<strong>in</strong>fall events.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.81-104, 2010
Márquez <strong>and</strong> Guevara-Pérez<br />
98<br />
A comparative example of the estimated values by<br />
apply<strong>in</strong>g the FIS to the observed data of furrow erosion<br />
un<strong>de</strong>r ra<strong>in</strong>fall events that last two hours, dur<strong>in</strong>g the<br />
stages of tra<strong>in</strong><strong>in</strong>g, validation <strong>and</strong> test<strong>in</strong>g; us<strong>in</strong>g a sample<br />
of 35 data, is shown <strong>in</strong> Fig. 14. Total number of data <strong>in</strong><br />
each stage can be seen <strong>in</strong> Table 11. Vertical axis<br />
represents the simulated erosion (D R sim) <strong>and</strong> the<br />
observed (D R obs). The horizontal axis represents a<br />
number assigned to each value; each pair of coord<strong>in</strong>ates<br />
is represented by a dot. Figure 14(a) shows the results<br />
<strong>in</strong> the tra<strong>in</strong><strong>in</strong>g stage of the FIS, which shows that the<br />
data po<strong>in</strong>ts (circles) were satisfactorily approximate to<br />
the FIS (asterisks). Figures 14(b)–(c) show a good<br />
approximation of the po<strong>in</strong>ts found <strong>in</strong> the FIS.<br />
DISCUSSION OF RESULTS<br />
Comparison of furrow erosion processes simulated<br />
with experimental data<br />
Process of particles <strong>de</strong>tachment As shown <strong>in</strong> the<br />
results section, the range of values of concentrated flow<br />
erodibility K c <strong>and</strong> critical shear stress τ c , obta<strong>in</strong>ed from<br />
the fit of the Eq. (5) varied between 1.31E–06 <strong>and</strong><br />
1.525E–06 s/m, 0.704 <strong>and</strong> 1.489 Pa, respectively.<br />
Knapen et al. (2007) conducted a literature review to<br />
compile the values of K c <strong>and</strong> τ c reported for different<br />
soils <strong>and</strong> tillage conditions. Studies <strong>in</strong>clu<strong>de</strong>d: (a)<br />
concentrated flow experiments on field plots, <strong>and</strong> (b)<br />
flume experiments <strong>in</strong> the laboratory. Characteristics of<br />
the experimental methods used are summarized <strong>in</strong><br />
Tables 12 <strong>and</strong> 13. Values ranges have been <strong>de</strong>term<strong>in</strong>ed<br />
empirically for K c <strong>and</strong> τ c parameters from the excess<br />
shear stress l<strong>in</strong>ear equation (Eq. 5), <strong>in</strong>clud<strong>in</strong>g the size of<br />
the sample by us<strong>in</strong>g field <strong>and</strong> laboratory experiments,<br />
they varied as follows: 0.000 001 <strong>and</strong> 0.1 m/s (n = 151),<br />
0.01 <strong>and</strong> 20 Pa (n = 161), 0.00 001 <strong>and</strong> 0.7 s/m<br />
(n = 185), 0.03 <strong>and</strong> 60 Pa (n = 220), respectively. By<br />
compar<strong>in</strong>g these value ranges with those reported on<br />
Table 3 for Eq. (5), it was found: (a) K c values fell<br />
with<strong>in</strong> the range for the field experiments <strong>and</strong> were<br />
lower than those found from laboratory experiments, (b)<br />
τ c values were totally <strong>in</strong>clu<strong>de</strong>d with<strong>in</strong> the reported<br />
ranges for the field <strong>and</strong> laboratory experiments.<br />
From the above, comparisons show that the K c value<br />
range obta<strong>in</strong>ed <strong>in</strong> this study was significantly different<br />
from the one reported by Knapen et al. (2007), while a<br />
lower variation for τ c value range was found. There are<br />
several reasons for these differences, two of the ma<strong>in</strong><br />
are (1) differences <strong>in</strong> experimental conditions <strong>in</strong> which<br />
the data were collected, <strong>and</strong> (2) the variation of soil<br />
types <strong>and</strong> environmental conditions. By compar<strong>in</strong>g the<br />
experimental <strong>and</strong> environmental conditions of the<br />
research <strong>de</strong>scribed <strong>in</strong> the methods section with outl<strong>in</strong>ed<br />
<strong>in</strong> Tables 12 <strong>and</strong> 13, some similarities were found alike<br />
as those <strong>in</strong> field studies conducted by Bjorneberg et al.<br />
(1999). With regard to laboratory experiments, cases<br />
with which to make a comparison were not found.<br />
The value ranges found for K c <strong>and</strong> τ c by Bjorneberg<br />
et al. (1999) varied as follows: from 0.0003 to 0.006<br />
m/s <strong>and</strong> from 1.2 to 1.8 Pa, respectively. Despite the<br />
similarities <strong>in</strong> the experimental conditions shown <strong>in</strong><br />
Table 12, the value range of K c was significantly<br />
different. Causes of the differences appear to be due to<br />
environmental conditions, ma<strong>in</strong>ly <strong>in</strong> both climate <strong>and</strong><br />
agricultural practices, because the experiments of<br />
Bjorneberg et al. (1999) were ma<strong>de</strong> <strong>in</strong> beans <strong>and</strong> corn<br />
fields, where the wett<strong>in</strong>g <strong>and</strong> dry<strong>in</strong>g sequences,<br />
consolidation <strong>and</strong> residues could be contribut<strong>in</strong>g factors<br />
to variability of K c <strong>and</strong> τ c values.<br />
In general, the Eq. (5) statistics show a satisfactory<br />
fit between most of the equation <strong>and</strong> the observed data,<br />
for the follow<strong>in</strong>g reasons: (1) coefficient of<br />
<strong>de</strong>term<strong>in</strong>ation R 2 was estimated at 0.7 (Ramírez et al.,<br />
2004), (2) Durb<strong>in</strong>-Watson coefficient presented<br />
evi<strong>de</strong>nce that there is a slight r<strong>and</strong>omness with<strong>in</strong><br />
consecutive residues, (3) Average Absolute Relative<br />
Error <strong>and</strong> Average Relative Error were mo<strong>de</strong>rately low.<br />
Process of sediment transport The ranges of the<br />
parameters K t <strong>and</strong> b reported <strong>in</strong> the results section for<br />
mo<strong>de</strong>ls of sediment transport capacity represented by<br />
Eqs 7, 13 <strong>and</strong> 15 varied as follows: 9.929E–08 <strong>and</strong><br />
1.52 669E–07; 7.634 <strong>and</strong> 34.583; 0.4212 <strong>and</strong> 0.892,<br />
respectively. The value ranges reported for field tests by<br />
F<strong>in</strong>kner et al. (1989) <strong>and</strong> Trout (1999) for Eq. (7)<br />
(Table 14), differ significantly from those <strong>in</strong>dicated <strong>in</strong><br />
Table 5. As for Eq. (13), the ω c critical unit current<br />
power value was negative <strong>and</strong> slightly different from<br />
zero, which differs from that proposed by Yang, (1973),<br />
which is equal to 0.002 m/s.<br />
Despite the differences, it can be observed that <strong>in</strong><br />
both cases ω c ten<strong>de</strong>d to zero. Accord<strong>in</strong>g to theory, the<br />
ω c value should be positive. The results of negative ω c<br />
<strong>and</strong> its low st<strong>and</strong>ard error, along with the high<br />
correlation make it difficult to expla<strong>in</strong> the physical<br />
mean<strong>in</strong>g of the estimated parameter. However, the<br />
estimated negative value range of ω c was not<br />
significantly different from zero.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.81-104, 2010
Márquez <strong>and</strong> Guevara-Pérez<br />
99<br />
Table 12. Characteristics of field plot <strong>in</strong> which concentrated flow erosion was measured<br />
Country<br />
Soil Slope Q <strong>in</strong>flow I ra<strong>in</strong>fall τ Rill plot dim.<br />
(No.) (%) (l m<strong>in</strong> –1 ) (mm/h) (Pa) (m × m)<br />
Surface Condition<br />
Source<br />
Iran 1 n.a. 132–1693 n.a.<br />
2.2–<br />
I, vary<strong>in</strong>g vegetation<br />
15 × 0.3<br />
13.2<br />
cover<br />
A<strong>de</strong>lpour et al. (2004)<br />
USA 1<br />
0.5–<br />
Length.:110– I, residue removed<br />
30–40 n.a. n.a.<br />
1.33<br />
256<br />
<strong>and</strong> tilled<br />
Bjorneberg et al. (1999)<br />
Brazil 1 n.a. n.a. n.a. n.a. 9 × 0.5<br />
I, residue removed<br />
<strong>and</strong> tilled<br />
Braida & Cassol (1996)<br />
Brazil 1 6.7 0–50 74 n.a. n.a. n.a. Cantalice et al. (2005)<br />
USA 2 3–15 96–768 n.d.<br />
4.0–<br />
I, residue removed<br />
10 × 0.75<br />
37.3<br />
<strong>and</strong> tilled<br />
Franti et al. (1985; 1999)<br />
Brazil 1 n.a. 0 60 n.a. 9 × 0.5<br />
I, residue removed<br />
<strong>and</strong> tilled<br />
Giasson & Cassol, (1996)<br />
USA 30 4–13 7–35 62 n.a. 9 × 0.46<br />
I, residue removed<br />
<strong>and</strong> tilled<br />
Gilley et al. (1993)<br />
USA 4 0.5–3 n.a. n.a. 1–36 30.5×0.91 I, no cover Hanson (1989;1990a; 1990b)<br />
USA 1 1–3 4–17.10 n.a.<br />
12–<br />
55<br />
29×1.8 I, no cover Hanson & Cook (1999)<br />
USA 2 5–11 n.d. n.a. n.a. 10.7×3<br />
I, var<strong>in</strong>g canopy<br />
cover<br />
Husse<strong>in</strong> & Laflen (1982)<br />
USA 2 4 11–189 64<br />
0.7–<br />
I, vary<strong>in</strong>g tillage<br />
4×0.2<br />
14<br />
practices<br />
K<strong>in</strong>g et al. (1995)<br />
USA 2 3–15 n.a. n.a. n.a. n.a. n.a. Laflen (1987)<br />
USA 56 2–13 0.1–0.6 63 0–22 9–11×0.5–3<br />
I, residue removed Laflen et al. (1991), Elliot et<br />
<strong>and</strong> tilled<br />
al. (1989)<br />
USA 1 3–6 8–38 n.a. 2–10 5.5×2<br />
I, residue removed<br />
<strong>and</strong> tilled<br />
Mamo & Bubenzer (2001b)<br />
Canada 3 12–14 n.a. 25–30 n.a. 10×0.8<br />
I, vegetation free,<br />
seedbed conditions<br />
Merz & Bryan (1993)<br />
USA 1 4–6 8–38 51<br />
1.3–<br />
I, tilled <strong>and</strong> all<br />
68.6×6.1<br />
6.1<br />
residue buried<br />
Morrison et al. (1994)<br />
USA 2 3–5 8–53 64 0–6 6.1×0.76<br />
I, residue removed<br />
<strong>and</strong> tilled<br />
Norton & Brown, (1992)<br />
USA 1 27 16–23 n.a.<br />
24–<br />
I, vegetation clipped<br />
6×0.3<br />
192<br />
to various heights<br />
Prosser et al. (1995)<br />
Australia 1 1–12.7 1.7–8 n.a. n.a. 20×1<br />
I, vegetation clipped<br />
to various heights<br />
Prosser (1996)<br />
Brazil 1 10 12–120 65<br />
2.5–<br />
19<br />
6×0.2 n.a. Reichert et al. (2001)<br />
USA 1 2–31 8–60 60 2–7 4.6×0.3<br />
I, residue removed<br />
<strong>and</strong> plants clipped<br />
West et al. (1992)<br />
USA 1<br />
0.52–<br />
I, residue removed<br />
15–46 n.a. n.a. 204–256<br />
1.33<br />
<strong>and</strong> tilled<br />
Trout et al. (1999)<br />
n.a.: not available. Soils: number of soil types tested. Slope: slope of soil surface. Q <strong>in</strong>flow : simulated concentrated flow discharge. I ra<strong>in</strong>fall :<br />
simulated ra<strong>in</strong>fall <strong>in</strong>tensities. τ: Range of applied flow shear stresses. Rill plot dimensions: length×width. Surface condition: (S: smoothened,<br />
I:irregular)<br />
With respect to Eq. (15), comparisons can not be<br />
ma<strong>de</strong> with the values proposed by Simons et al. (1981)<br />
because the values reported <strong>in</strong> literature correspond to<br />
flow <strong>in</strong> rivers.<br />
In general, fitted statistics <strong>in</strong> Eqs (7), (13) <strong>and</strong> (15)<br />
<strong>in</strong>dicated a satisfactory approximation between observed<br />
<strong>and</strong> estimated values, for the follow<strong>in</strong>g reasons: (a)<br />
coefficient of <strong>de</strong>term<strong>in</strong>ation varied between 0.72 <strong>and</strong><br />
0.88, accord<strong>in</strong>g to Ramirez et al. (2004), (b) adjusted R 2<br />
coefficient <strong>de</strong>creased slightly <strong>in</strong> relation to R 2 , (c)<br />
Durb<strong>in</strong>-Watson coefficient presented evi<strong>de</strong>nce that there<br />
is a slight r<strong>and</strong>omness between consecutive residues, (3)<br />
Average Absolute Relative Error <strong>and</strong> Average Relative<br />
Error were mo<strong>de</strong>rately low.<br />
Comb<strong>in</strong>ation of furrow erosion processes It can be<br />
observed from Figure 11 that the sediment transport<br />
capacity values obta<strong>in</strong>ed by Eqs (7) <strong>and</strong> (13) were lower<br />
than the observed sediment load values <strong>in</strong> the stream for<br />
different slopes of furrows <strong>in</strong> the three stages with<strong>in</strong> the<br />
cycle un<strong>de</strong>r irrigation, which <strong>in</strong>dicates the occurrence of<br />
concentrated flow <strong>and</strong> <strong>de</strong>position process. As for Eq.<br />
(15), estimated values of the transport capacity were<br />
very close to the sediment load current, therefore,<br />
<strong>de</strong>position processes occur at low rates. Accord<strong>in</strong>g to<br />
theory, the highest proportion of <strong>de</strong>tached sediments is<br />
transported out of the furrow (e.g., Near<strong>in</strong>g et al., 1989,<br />
Morgan, 1998; Bulyg<strong>in</strong> et al., 2002).<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.81-104, 2010
Márquez <strong>and</strong> Guevara-Pérez<br />
100<br />
Table 13. Characteristics of laboratory flume experiments <strong>in</strong> which concentrated flow erosion was measured<br />
Country<br />
Soil Slope Q <strong>in</strong>flow I ra<strong>in</strong>fall<br />
Flume dim.<br />
(N°) (%) (l m<strong>in</strong> –1 τ (Pa)<br />
) (mm/h)<br />
(m × m × m)<br />
Source<br />
Remark<br />
Italy 6 10–40 2.4–18 / 1–12 1.5 × 0.2 Ciampall<strong>in</strong>i & Torri (1998) a<br />
Australia 1 2–7 2.5–40 / n.a. 1.8 × 0.6 × 0.2 Crouch & Novruzi (1989) b, e<br />
Mexico 1 0 n.a. / 1.1–38.2 4.9 × 0.3 × 2.5 Ghebreiyessus et al. (1994)<br />
Belgium 2 5–21 3.3–60 / 1–24 4 × 0.4 × 0.45 Giménez & Govers (2002)<br />
Belgium 2 9–21 6.7–60 / 1–16 2 × 0.10 × 0.09 Giménez & Govers (2002)<br />
Belgium 2 1.5–7 0.7–10.4 / 0,1–1,3 6 × 0.12 Govers (1985) b, e<br />
Belgium 1 20–30 18–114 / 4.4–22.4 2 × 0.1 × 0.09 Gyssels et al. (2006)<br />
Canada 5 n.a. n.a. / n.a. 9.1 × 0.15 Kamphuis & Hall (1983) b, e<br />
USA 5 0.2 n.a. / n.a. 18.3 × 0,77 × 0,46 Laflen & Beasley (1960) a, e<br />
USA 7 0.2 n.a. / n.a. 22 × 0.4 × 0.76 Lyle & Smerdon (1965) a, e<br />
USA 1 3–5 7.6–37.9 / n.a. 4 × 0.2 × 0.05 Mamo & Bubenzer (2001a)<br />
Canada 1 14 / 34 n.a. 10 × 0.8 × 0,2 Merz & Bryan (1993) b, c, e<br />
USA 8 n.a. 6.1–30.3 / 0–3.2 1.84 × 0.1 × 0,19 Moody et al. (2005) e<br />
Belgium 1 10–35 5.6–11.5 / 1.6–5.7 2 × 0.1 × 0.09 Nachtergaele & Poesen (2002)<br />
USA 1 n.a. n.a. / 0.1–1.3 18.3 × 0.3 × 0.38 Parthenia<strong>de</strong>s (1965) e<br />
S-Africa 12 2–20 0.08–0.32 / 0,2–4.9 0.5 × 0.05 × 0.13 Rapp (1998)<br />
Israel 3 2–20 0.08–0.32 / 0.2–4.9 0.5 × 0.05 × 0.13 Rapp (1998)<br />
Belgium 1 2.6–14 n.a. / n.a. 2 × 0.4 Rauws (1987)<br />
Belgium 1 3–33 n.a. 20 n.a. 2.5 × 0.6 Rauws & Govers (1988) b, c, e<br />
USA 1 2–20 0.04–0.2 / 0.4–4.8 0.5 × 0.05 × 0.12 Sha<strong>in</strong>berg et al. (1994) b, e<br />
Israel 3 5 0.04–0.32 / 0.8–1.8 0.5 × 0.05 × 0.12 Sha<strong>in</strong>berg et al. (1996)<br />
Australia 16 5–30 0.1–1.8 100 n.a. 3 × 0.8 × 0.15 Sheridan et al. (2000a,b) d<br />
Canada 1 9 n.a. 35 n.a. 15 m long. Slattery & Bryan (1992)<br />
USA 11 0–1 n.a. / n.a. 18.3 × 0.77 × 0.46 Smerdon & Beasley (1959)<br />
USA 5 0.1–0.2 n.a. 0–127 n.a. 22 × 0.8 Smerdon (1964)<br />
Italy 4 1–31 n.a. 15–110 0–3.3 2 × 0.5 × 0.1 Torri et al. (1987)<br />
USA 1 7 2–4 / 1.9–3.9 2.73 × 0.46 × 0.88 Van Klavern & McCool (1998<br />
Ch<strong>in</strong>a 1 10,5–21.2 2.5–6.5 / 1–10 5 × 0.3 Zhang et al. (2005) a<br />
USA 5 1.5–5 3.8–15.1 / 0.5–2.5 6.4 × 0.15 × 0.05 Zhu et al. (1995)<br />
USA 5 0–5.5 n.a. / 0.5–2.5 6.4 × 0.15 × 0.05 Zhu et al. (2001)<br />
n.a.: not available. /:not applicable. Soils: number of soil types tested. Slope: slope of soil surface. Q <strong>in</strong>flow : simulated concentrated flow<br />
discharge. I ra<strong>in</strong>fall : simulated ra<strong>in</strong>fall <strong>in</strong>tensities. τ: Range of applied flow shear stresses. Flume dimensions: length×width×<strong>de</strong>pth. Remarks:<br />
(a) values for K c <strong>and</strong>/or τ c <strong>de</strong>duced from graphs, (b) values calculated from critical shear velocity, (c) values for poorly cohesive or noncohesive<br />
soils, (d) tests on coal m<strong>in</strong>e soils <strong>and</strong> overbur<strong>de</strong>ns, (e) excess shear stress equation was not used to <strong>de</strong>duce K c <strong>and</strong>/or τ c .<br />
Table 14. Mo<strong>de</strong>ls parameters to estimate sediment transport capacity of flow <strong>in</strong> furrow <strong>and</strong> characteristics of field tests<br />
Tc= Kt(τ) b<br />
Source<br />
T c<br />
Slope Q I<br />
K<br />
(kg/m/s)<br />
t τ(Pa) b<br />
ra<strong>in</strong>fall Rill plot Surface<br />
(%) (L / m<strong>in</strong>) (mm/ h) dim. (m×m) condition<br />
F<strong>in</strong>kner 0.0006–0.2366 0,015–0,05 1–6 1.5 2–20 n.a. 80–83 10, 50, 100 n.a.<br />
et al.<br />
m long<br />
(1989)<br />
Trout<br />
(1999)<br />
0.002–0.035 0,017–<br />
0,065<br />
n.a. 2 1.3 6–18 n.a. 204 m long residue<br />
removed<br />
<strong>and</strong> tilled<br />
0.001–0.02 0,015–0,15 n.a. 4 0.5 6–45.6 n.a. 256 m long residue<br />
removed<br />
<strong>and</strong> tilled<br />
Source T c (ppm) C<br />
ω c<br />
(m/s)<br />
b<br />
Tc = c[(ω–ωc)/Vs] b<br />
Slope<br />
(%)<br />
Q <strong>in</strong>flow<br />
(L/m<strong>in</strong>)<br />
Ira<strong>in</strong>fall<br />
(mm/h)<br />
Rill plot<br />
dim.<br />
(m×m)<br />
Surface<br />
condition<br />
Loch 41.000–87.000 n.a. 0.002 n.a. 4 15–107 95 0.4×1.5–22.5 removed<br />
(1984)<br />
<strong>and</strong> tilled<br />
n.a.: not available. Remark: K t : (a) obta<strong>in</strong>ed by empirical fit, (b) obta<strong>in</strong>ed by equation of Yal<strong>in</strong> (1963).<br />
Remark<br />
b<br />
a<br />
a<br />
Remark<br />
b<br />
Trout (1996) found that at the end of the furrows of<br />
bean crops, erosion <strong>in</strong>creases from the beg<strong>in</strong>n<strong>in</strong>g to the<br />
end of the cycle from 1 to 5 Mg/ha, while the process<br />
was reversed <strong>in</strong> the corn crop due to the reduction of<br />
erosion from 2 to 0.1 Mg/ha at the end of the cycle. In<br />
this study, the variability <strong>in</strong> the rate of erosion along the<br />
furrows was not significant, based on the uniformity of<br />
both water supplies through the spr<strong>in</strong>klers as well as<br />
frequency of wett<strong>in</strong>g <strong>in</strong> each furrow. The erosion rate,<br />
from the beg<strong>in</strong>n<strong>in</strong>g to the end of the cycle, varied from<br />
0.6 to 0.1 Mg/ha (Fig. 12). This range is lower than that<br />
found by Trout (1996) for bean <strong>and</strong> corn fields.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.81-104, 2010
Márquez <strong>and</strong> Guevara-Pérez<br />
101<br />
Trout (1999) reported sediment transport rates <strong>in</strong><br />
furrows of slope at 1.3% <strong>and</strong> 5.2% for corn <strong>and</strong> bean,<br />
that varied from the beg<strong>in</strong>n<strong>in</strong>g to the end of the cycle as<br />
follows: from 20 000 to 4000 mg·s –1 m –1 , from 25 000 to<br />
5000 mg·s –1 m –1 , respectively. These ranges were higher<br />
than those found <strong>in</strong> this study, which varied from 1428<br />
to 10 000 mg·s –1 m –1 (mg of sediment per meter of<br />
wetted perimeter per second).<br />
Furrow erosion mo<strong>de</strong>ls based on regressions<br />
Estimated erosion values by the multiple l<strong>in</strong>ear mo<strong>de</strong>ls<br />
were closer to the field observations than those which<br />
were obta<strong>in</strong>ed by mo<strong>de</strong>ls based on nonl<strong>in</strong>ear<br />
polynomials. L<strong>in</strong>ear mo<strong>de</strong>l showed that the maximum<br />
loads of sediments dur<strong>in</strong>g events of natural <strong>and</strong> artificial<br />
ra<strong>in</strong> were successfully estimated. However, there was a<br />
lack of adjustment for values at the beg<strong>in</strong>n<strong>in</strong>g <strong>and</strong> at the<br />
end of such events, which co<strong>in</strong>ci<strong>de</strong>s with the results of<br />
Ja<strong>in</strong> et al. (2004), Kumar (2001) <strong>and</strong> Sr<strong>in</strong>ivasulu et al.<br />
(2006) (Fig. 12). F<strong>in</strong>d<strong>in</strong>gs <strong>in</strong> polynomial-based mo<strong>de</strong>ls<br />
<strong>in</strong>dicate that the estimation process of the observed<br />
variable tends to <strong>de</strong>crease as the power <strong>in</strong>creases.<br />
Mo<strong>de</strong>l<strong>in</strong>g of furrow erosion us<strong>in</strong>g artificial<br />
<strong>in</strong>telligence techniques<br />
Artificial Neural Networks Estimated erosion values<br />
by artificial neural networks became successfully closer<br />
to the observations dur<strong>in</strong>g the stages of calibration,<br />
validation, <strong>and</strong> test<strong>in</strong>g, <strong>in</strong> terms of the coefficient of<br />
<strong>de</strong>term<strong>in</strong>ation <strong>and</strong> the Average Absolute Error. The<br />
ANN estimate improved as the variability of data for<br />
tra<strong>in</strong><strong>in</strong>g <strong>and</strong> test<strong>in</strong>g of the network was lower, which<br />
helped the network to learn the pattern of data.<br />
Fuzzy Inference System Estimated erosion values by<br />
FIS’s became successfully closer to the observations<br />
dur<strong>in</strong>g the stages of calibration, validation <strong>and</strong> test<strong>in</strong>g,<br />
<strong>in</strong> terms of the Average Absolute Error. The pattern <strong>in</strong><br />
the <strong>in</strong>put data was better represented by us<strong>in</strong>g the<br />
trapezoidal function both for ra<strong>in</strong>fall, as well as for<br />
<strong>in</strong>filtration. The system output was represented by a<br />
multiple l<strong>in</strong>ear equation. Estimated erosion resulted<br />
closer to observations <strong>in</strong> the stages of tra<strong>in</strong><strong>in</strong>g,<br />
validation <strong>and</strong> test<strong>in</strong>g (Fig. 14). Estimation by ANN<br />
improved as the variability of data for tra<strong>in</strong><strong>in</strong>g <strong>and</strong><br />
test<strong>in</strong>g of the network was lower, which helped the<br />
network to learn the pattern of data. The application of<br />
FIS’s has been used successfully for mo<strong>de</strong>l<strong>in</strong>g of<br />
reservoir operation by Panigrahi (2000), which validates<br />
its use for simulat<strong>in</strong>g hydrological processes.<br />
Comparison of mo<strong>de</strong>ls to estimate rill erosion The<br />
differences between the dynamic versions of the WEPP<br />
<strong>and</strong> EUROSEM mo<strong>de</strong>ls were significant <strong>in</strong> the<br />
magnitu<strong>de</strong>s of the events; however, both of them<br />
estimated the erosion processes, namely, sediment<br />
transport <strong>and</strong> <strong>de</strong>position (Fig. 11). As for CIHAM-UC<br />
mo<strong>de</strong>l was different from WEPP <strong>and</strong> EUROSEM<br />
mo<strong>de</strong>ls because they produced smaller estimates of<br />
<strong>de</strong>position process.<br />
The comparison between the different mo<strong>de</strong>ls<br />
<strong>in</strong>dicates the follow<strong>in</strong>g or<strong>de</strong>r on the quality of the<br />
estimation <strong>in</strong> furrow erosion: Artificial <strong>in</strong>telligence<br />
techniques, mo<strong>de</strong>ls based on physical processes, l<strong>in</strong>ear<br />
regression, <strong>and</strong> f<strong>in</strong>ally, the non-l<strong>in</strong>ear regression (Fig.<br />
12).<br />
CONCLUSIONS<br />
It is possible to make estimates of the physical<br />
processes of erosion <strong>in</strong> furrows with a mo<strong>de</strong>rate to high<br />
accuracy, although it is believed that the estimation can<br />
be <strong>in</strong>creased <strong>and</strong> improved by mak<strong>in</strong>g adjustments of<br />
the mo<strong>de</strong>ls for each slope of furrow.<br />
Mo<strong>de</strong>ls based on l<strong>in</strong>ear regressions achieved better<br />
estimates at the observed values than the nonl<strong>in</strong>ear<br />
mo<strong>de</strong>ls based on polynomials. This should be<br />
emphasized <strong>in</strong> tra<strong>in</strong><strong>in</strong>g with new data to better estimate<br />
the values at the beg<strong>in</strong>n<strong>in</strong>g <strong>and</strong> end of natural <strong>and</strong><br />
artificial ra<strong>in</strong> events.<br />
Mo<strong>de</strong>l<strong>in</strong>g of furrow erosion us<strong>in</strong>g artificial<br />
<strong>in</strong>telligence techniques resulted <strong>in</strong> a satisfactory<br />
approximation of the estimated values to those<br />
observed, improv<strong>in</strong>g accuracy through the application<br />
of ANN’s with respect to the FIS’s. The estimation<br />
quality improves as variability <strong>in</strong> the mo<strong>de</strong>l data is low.<br />
The comparison between the different mo<strong>de</strong>ls<br />
<strong>in</strong>dicate the follow<strong>in</strong>g or<strong>de</strong>r on the quality of the<br />
estimation of furrow erosion: artificial <strong>in</strong>telligence<br />
techniques, mo<strong>de</strong>ls based on physical processes, l<strong>in</strong>ear<br />
regression <strong>and</strong> f<strong>in</strong>ally, the non-l<strong>in</strong>ear regression<br />
Acknowledgment The research was <strong>de</strong>veloped <strong>in</strong><br />
<strong>Centro</strong> <strong>de</strong> Investigaciones Hidrológicas y Ambientales<br />
<strong>de</strong> la Universidad <strong>de</strong> Carabobo (CIHAM-UC) with the<br />
f<strong>in</strong>ancial support of CDCH-UC <strong>and</strong> FONACIT.<br />
REFERENCES<br />
AMERICAN PUBLIC HEALTH ASSOCIATION (1995) St<strong>and</strong>ard<br />
Methods for The Exam<strong>in</strong>ation of Water <strong>and</strong> Wastewater.<br />
American Public Health Association, United States of America,<br />
Wash<strong>in</strong>gton, DC, 19th Edition. 2–53.<br />
AMERICAN SOCIETY FOR TESTING AND MATERIALS<br />
(2007) St<strong>and</strong>ard Test Methods for particle-size analysis of soils.<br />
American Society for Test<strong>in</strong>g <strong>and</strong> Materials International, United<br />
States of America, West Conshohocken. (Reaproved 2007), 1–8.<br />
A<strong>de</strong>lpour, A.A., Soufi, M. & Behnia, A.K. (2004) Channel erosion<br />
thresholds for different l<strong>and</strong> uses assessed by concentrated<br />
overl<strong>and</strong> flow on a silty loam. Proc. International Soil<br />
Conservation Organisation Conference. Brisbane, Australia.<br />
Alonso C.V., Neibl<strong>in</strong>g, W.H. & Foster, G.R. (1981) Estimat<strong>in</strong>g<br />
sediment transport capacity <strong>in</strong> watershed mo<strong>de</strong>l<strong>in</strong>g. Trans. ASAE<br />
24(5), 1211–1220.<br />
Bjorneberg, D.L., Trout, R.E., Sojka, R.E. & Aase, J.K. (1999)<br />
Evaluat<strong>in</strong>g WEPP predicted <strong>in</strong>filtration, runoff <strong>and</strong> soil erosion<br />
for furrow irrigation. Trans. ASAE 42(6), 1733–1741.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.81-104, 2010
Márquez <strong>and</strong> Guevara-Pérez<br />
102<br />
Braida, J.A. & Cassol, E.A. (1996) Rill <strong>and</strong> <strong>in</strong>terrill erodibility of a<br />
paleudult soil. Rev. Bras. Ciênc. Solo 20(1), 127–134.<br />
Bulyg<strong>in</strong>a, N.S., Near<strong>in</strong>g, M.A., Stone, J.J. & Nichols, M.H. (2007)<br />
DWEPP: a dynamic soil erosion mo<strong>de</strong>l based on WEPP source<br />
terms. Earth Surf. Processes L<strong>and</strong>f. 32(7), 998–1012. doi:<br />
10.1002/esp.1467.<br />
Bulyg<strong>in</strong>, S.Y., Near<strong>in</strong>g, M.A. & Achasov, A.B. (2002) Parameters of<br />
<strong>in</strong>terrill erodibility <strong>in</strong> the WEPP mo<strong>de</strong>l. Eurasian Soil Sci. 35(11),<br />
1237–1242.<br />
Bishop, C.M. (1994) Neural networks <strong>and</strong> their applications. Rev.<br />
Sci.Instrum. 65(6), 1803–1832.<br />
Cantalice, J.R.B., Cassol, E.A., Reichert, J.M. & Borges, A.L.D.<br />
(2005) Flow hydraulics <strong>and</strong> sediment transport <strong>in</strong> rills of a s<strong>and</strong>y<br />
clay loam soil. Rev. Bras. Ciênc. Solo 29(4), 597–607. (<strong>in</strong><br />
portuguese). doi: 10.1590/S0100-06832005000400012.<br />
Ciampall<strong>in</strong>i, R. & Torri, D. (1998) Detachment of soil particles by<br />
shallow flow: sampl<strong>in</strong>g methodology <strong>and</strong> observations. Catena<br />
32(1), 37–53. doi:10.1016/S0341-8162(97)00050-7.<br />
Crouch, R.J. & Novruzi, T. (1989) Threshold conditions for rill<br />
<strong>in</strong>itiation on a vertisol, Gunnedah, N.S.W., Australia. Catena<br />
16(1), 101–110. doi:10.1016/0341-8162(89)90007-6.<br />
Demuth, H., Beale, M., Hagan, M. (2009) Neural Network Toolbox.<br />
User Gui<strong>de</strong>. MATLAB.<br />
Elliot, W.J., Liebenow, A.M., Laflen, J.M. & Kohl, K.D. (1989) A<br />
compendium of soil erodibility data from WEPP cropl<strong>and</strong> soil<br />
field erodibility experiments 1987–1988. NSERL Rpt. No. 3.<br />
Ohio State University <strong>and</strong> Natural Soil Erosion Research<br />
Laboratory, Agricultural Research Service, U.S. Department of<br />
Agriculture, W. Lafayette, Indiana.<br />
Flanagan, D.C., Ascough, J.C., Near<strong>in</strong>g, M.A. & Laflen, J.M. (2001)<br />
The Water Erosion Prediction Project mo<strong>de</strong>l. In: Harmon, R.S. &<br />
Doe, W.W. (eds) L<strong>and</strong>scape Erosion <strong>and</strong> Evolution Mo<strong>de</strong>ll<strong>in</strong>g,<br />
Kluwer: New York, 145–199.<br />
Flanagan, D.C. & Near<strong>in</strong>g, M.A. eds. (1995) USDA–Water Erosion<br />
Prediction Project (WEPP), Hillslope Profile <strong>and</strong> Watershed<br />
Mo<strong>de</strong>l Documentation Technical Documentation, NSERL Report<br />
10. USDA-ARS National Soil Erosion Research Laboratory:<br />
West Lafayette, IN.<br />
F<strong>in</strong>kner, S.C., Near<strong>in</strong>g M.A., Foster G.R. & Gilley J.E. (1989) A<br />
simplified equation for mo<strong>de</strong>l<strong>in</strong>g sediment transport capacity.<br />
Trans. ASAE 32(5), 1545–1550.<br />
Foster, G.R. & Meyer, L.D. (1972) Transport of particles by shallow<br />
flow. Trans. ASAE 15(1), 99–102.<br />
Foster, G.R. & Meyer, L.D. (1975) Mathematical simulation of<br />
upl<strong>and</strong> erosion us<strong>in</strong>g fundamental erosion mechanics. Proc.<br />
Sediment Yield Workshop. U.S. Sedimentation Laboratory,<br />
Oxford, MI, 190–201.<br />
Foster, G.R. (1982) Mo<strong>de</strong>l<strong>in</strong>g the erosion process. In: Hahn, C.T.<br />
(Ed.). Hydrologic Mo<strong>de</strong>l<strong>in</strong>g of Small Watersheds, 295–380.<br />
Foster, G.R., Hugg<strong>in</strong>s, L.F. & Meyer, L.D. (1984) A laboratory<br />
study of rill hydraulics: shear stress relationships. Trans. ASAE<br />
27(3), 797–804.<br />
Foster, G.R. & Lane, L.J. (1987) User Requirements: USDA-WEPP<br />
(Draft 6.3) National Soil Erosion Research Laboratory, West<br />
Lafayette, Indiana.<br />
Foster, G.R. (1990) Major <strong>de</strong>velopments <strong>in</strong> prediction of soil erosion<br />
by water. In: Lal, R. & Pierce, F.J. (eds.) Soil Management for<br />
Susta<strong>in</strong>ability Soil <strong>and</strong> Water Conservation Society, Ames, Iowa.<br />
Franti, F.G., Laflen, J.M. & Watson, D.A. (1985) Soil erodibility<br />
<strong>and</strong> critical shear un<strong>de</strong>r concentrated flow. ASAE Paper No. 85-<br />
2033<br />
Franti, T.G., Laflen, J.M. & Watson, D.A. (1999) Predict<strong>in</strong>g soil<br />
<strong>de</strong>tachment from high discharge concentrated flow. Trans. ASAE<br />
42(2), 329–335.<br />
Ghebreiyessus, Y.T., Gantzer, C.J., Alberts, E.E. & Lentz, R. (1994)<br />
Soil erosion by concentrated flow: shear stress <strong>and</strong> bulk <strong>de</strong>nsity.<br />
Trans. ASAE 37(6), 1791–1797.<br />
Giménez, R. & Govers, G. (2002) Flow <strong>de</strong>tachment by concentrated<br />
flow on smooth <strong>and</strong> irregular beds. Soil Sci. Soc. Am. J. 66(5),<br />
1475–1483. doi:10.2136/sssaj2002.1475.<br />
Gyssels, G., Poesen, J., Van Dessel, W., Knapen, A. & Debaets, S.<br />
(2006) Effects of cereal roots on <strong>de</strong>tachment rates of s<strong>in</strong>gle <strong>and</strong><br />
doubledrilled topsoils dur<strong>in</strong>g concentrated flow. Eur. J. Soil Sci.<br />
57(3), 381–391. doi: 10.1111/j.1365-2389.2005.00749.x.<br />
Giasson, E. & Cassol, E.A. (1996) Rill erosion related to <strong>in</strong>flow<br />
rates <strong>and</strong> amounts of <strong>in</strong>corporated wheat straw <strong>in</strong> a s<strong>and</strong>y clay<br />
loam Paleudult soil. Rev. Bras. Ciênc. Solo 20(1), 117–125.<br />
Gilley, J.E., Elliot, W.J., Laflen, J.M. & Simanton, S.R. (1993)<br />
Critical shear stress <strong>and</strong> critical flow rates for <strong>in</strong>itiation of rill<strong>in</strong>g.<br />
J. Hydrol. 142(2), 251–271. doi:10.1016/0022-1694(93)90013-Y.<br />
Govers, G. (1985) Selectivity <strong>and</strong> transport capacity of th<strong>in</strong> flow <strong>in</strong><br />
relation to rill erosion. Catena 12(1), 35–49. doi:10.1016/S0341-<br />
8162(85)80003-5.<br />
Guevara, E. & Márquez, A. (2009) Mo<strong>de</strong>lación <strong>de</strong> la erosión y<br />
transporte <strong>de</strong> sedimentos en un campo agrícola bajo riego.<br />
Tecnología y Ciencias <strong>de</strong>l Agua (Submitted).<br />
Hanson, G.J. (1989) An <strong>in</strong>-situ erodibility test<strong>in</strong>g <strong>de</strong>vice. American<br />
Society of Agricultural Eng<strong>in</strong>eers - Canadian Society of<br />
Agricultural Eng<strong>in</strong>eers. Paper No. 89–2151.<br />
Hanson, G.J. (1990a) Surface erodibility of earthen channels at high<br />
stresses. Part II-Develop<strong>in</strong>g an <strong>in</strong> situ test<strong>in</strong>g <strong>de</strong>vice. Trans. ASAE<br />
33(1), 132–137.<br />
Hanson, G.J. (1990b) Surface erodibility of earthen channels at high<br />
stresses part I. Open channel test<strong>in</strong>g. Trans. ASAE 33(1), 127–<br />
131.<br />
Hanson, G.J., Cook, K.R. & Simon, A. (1999) Determ<strong>in</strong>ig erosion<br />
resistance of cohesive materials. American Society of Civil<br />
Eng<strong>in</strong>eers, Proc. International Water Resources Eng<strong>in</strong>eer<strong>in</strong>g<br />
Conference. Seattle, Wash<strong>in</strong>gton, USA.<br />
Husse<strong>in</strong>, M.H. & Laflen, J.M. (1982) Effects of crop canopy <strong>and</strong><br />
residue on rill <strong>and</strong> <strong>in</strong>terrill soil erosion. Trans. ASAE 25(5), 1310–<br />
1315.<br />
Ja<strong>in</strong>, A. & Ormsbee, L.E. (2002) Evaluation of short-term water<br />
<strong>de</strong>m<strong>and</strong> forecast mo<strong>de</strong>l<strong>in</strong>g techniques: conventional methods<br />
versus AI. J. American Water Works Assoc. 94(7), 64–72.<br />
Ja<strong>in</strong>, A., Sudheer, K.P. & Sr<strong>in</strong>ivasulu, S. (2004) I<strong>de</strong>ntification of<br />
physical processes <strong>in</strong>herent <strong>in</strong> artificial neural network ra<strong>in</strong>fall<br />
runoff mo<strong>de</strong>ls. Hydrol. Processes 118(3), 571–581. doi:<br />
10.1002/hyp.5502.<br />
Juarez, E. & Rico, A. (1991) Mecánica <strong>de</strong> Suelos. Ed. Limusa.<br />
Mexico.<br />
Kamphuis, J.W. & Hall, K.R. (1983) Cohesive material erosion by<br />
unidirectional current. J. Hydr. Engrg. 110(3), 370–370. doi:<br />
10.1061/(ASCE)0733-9429(1984)110:3(370).<br />
Knapen, A., Poesen, J., Govers, G., Gyssels, G. & Nachtergaele, J.<br />
(2007) Resistance of soils to concentrated<br />
flow erosion: A review. Earth-Science Rev. 80(2), 75–109.<br />
doi:10.1016/j.earscirev.2006.08.001.<br />
K<strong>in</strong>g, K.W., Flanagan, D.C., Norton, L.D. & Laflen, J.M. (1995)<br />
Rill erodibility parameters <strong>in</strong>fluenced by long-term management<br />
practices. Trans. ASAE 38(1), 159–164.<br />
Koluvek, P.K., Tanji, K.K. & Trout, T.J. (1993) Overview of soil<br />
erosion from irrigation. J. Irrig. Dra<strong>in</strong>age Engr. ASCE 119(6),<br />
929–946. doi: 10.1061/(ASCE)0733-9437(1993)119:6(929).<br />
Klik, A., Savabi, M.R., Norton, L.D. & Baumer, O. (1995)<br />
Application of WEPP hillslope mo<strong>de</strong>l on Austria. Proc. Annual<br />
Conference of the American Water Resources Association<br />
(AWRA). Houston, Texas, 313–322.<br />
Kumar, A. & M<strong>in</strong>ocha, K. (2001) Ra<strong>in</strong>fall runoff mo<strong>de</strong>l<strong>in</strong>g us<strong>in</strong>g<br />
artificial neural networks. J. Hydrol. Eng. 6(2), 176–177. doi:<br />
10.1061/(ASCE)1084-0699(2001)6:2(176).<br />
Laflen, J.M. & Beasley, R.P. (1960) Effects of compaction on<br />
critical tractive forces <strong>in</strong> cohesive soils. University of Missouri.<br />
Experiment Station Research Bullet<strong>in</strong>, 749.<br />
Laflen, J.M. (1987) Effect of tillage systems on concentrated flow<br />
erosion. In: Sentis, I.P. (Ed.), Soil Conservation <strong>and</strong> Productivity.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.81-104, 2010
Márquez <strong>and</strong> Guevara-Pérez<br />
103<br />
Proc. 40 International Conference on Soil Conservation, Maracay,<br />
República Bolivariana <strong>de</strong> Venezuela.<br />
Laflen, J.M., Elliot, W.J., Simanton, J.R., Holzhey, C.S. & Kohl,<br />
K.D. (1991) WEPP: Soil erodibility experiments for rangel<strong>and</strong><br />
<strong>and</strong> cropl<strong>and</strong> soils. J. Soil Water Conserv. 46(1), 39–44. doi:<br />
10.2489/jswc.46.1.39.<br />
Laflen, J.M., Flanagan, D.C. & Engel, B.A. (2004) Soil erosion <strong>and</strong><br />
sediment yield prediction accuracy us<strong>in</strong>g WEPP.<br />
J. Am. Water Res. Assoc. 40(2), 289–297. doi: 10.1111/j.1752-<br />
1688.2004.tb01029.x.<br />
Lyle, W.M. & Smerdon, E.T. (1965) Relation of compaction <strong>and</strong><br />
other soil properties to erosion resistance of soils. Trans. ASAE<br />
8(4), 419–422.<br />
Loch, R.J. (1984) Field ra<strong>in</strong>fall simulator studies on two caly spils of<br />
the Darl<strong>in</strong>g Downs, Queensl<strong>and</strong>, III, An evaluation of currents<br />
methods of <strong>de</strong>riv<strong>in</strong>g soil erodibilities (K factors), Aust. J. Soil<br />
Res. 22(4), 401–412.<br />
Mamo, M. & Bubenzer, G.D. (2001a) Detachment rate, soil<br />
erodibility <strong>and</strong> soil strength as <strong>in</strong>fluenced by liv<strong>in</strong>g plant roots<br />
part I: laboratory study. Trans. ASAE 44(5), 1167–1174.<br />
Mamo, M. & Bubenzer, G.D. (2001b) Detachment rate, soil<br />
erodibility <strong>and</strong> soil strength as <strong>in</strong>fluenced by liv<strong>in</strong>g plant roots<br />
part II: field study. Trans. ASAE 44(5), 1175–1181.<br />
MATLAB. (2010) Fuzzy Logic Toolbox. User’s Gui<strong>de</strong>. The<br />
Mathworks Inc.<br />
Merz, W. & Bryan, R.B. (1993) Critical threshold conditions for rill<br />
<strong>in</strong>itiation on s<strong>and</strong>y loam Brunisols: laboratory <strong>and</strong> field<br />
experiments <strong>in</strong> southern Ontario, Canada. Geo<strong>de</strong>rma 57(4), 357–<br />
385. doi: 10.1016/0016-7061(93)90050-U.<br />
Meyer-Peter, E. & Müller R. (1948) Formulas for bed-load transport.<br />
Proc. II Meet<strong>in</strong>g IARH, Stockholm, 39–64.<br />
Moody, J.A., Smith, J.D. & Ragan, B.W. (2005) Critical shear stress<br />
for erosion of cohesive soils subjected to temperatures<br />
typical of wildfire. J. Geophys. Res. 110, F01004. doi:<br />
10.1029/2004JF000141.<br />
Morrison, J.E., Richardson, C.W., Laflen, J.M. & Elliott, W.J.<br />
(1994) Rill erosion of a Vertisol with exten<strong>de</strong>d time s<strong>in</strong>ce tillage.<br />
Trans. ASAE 37(4), 1187–1196.<br />
Morgan, R.P.C., Qu<strong>in</strong>ton, J.N., Smith, R.E., Govers G., Poesen, J.W.<br />
A., Auerswald, K., Chisci, G., Torri, D. & Styczen, M.E. (1998)<br />
The European Soil Erosion Mo<strong>de</strong>l (EUROSEM): A dynamic<br />
approach for predict<strong>in</strong>g sediment transport from fields <strong>and</strong> small<br />
catchments. Earth Surf. Processes L<strong>and</strong>f. 23(6), 527–544. doi:<br />
10.1002/(SICI)1096-9837(199806)23:6.<br />
Nachtergaele, J. & Poesen, J. (2002) Spatial <strong>and</strong> temporal variations<br />
<strong>in</strong> resistance of loess-<strong>de</strong>rived soils to ephemeral gully erosion.<br />
European J. Soil Sci. 53(3), 449–463. doi: 10.1046/j.1365-<br />
2389.2002.00443.x.<br />
Near<strong>in</strong>g M.A., Foster G.R., Lane L.J. & F<strong>in</strong>kner S.C. (1989) A<br />
process-based soil erosion mo<strong>de</strong>l for USDA–Water Erosion<br />
Prediction Project technology. Trans. ASAE 32(5), 1587–1593.<br />
Near<strong>in</strong>g, M.A., Bulyg<strong>in</strong>, S.Y. & Kotova, M.M. (1998) Primary<br />
verification <strong>and</strong> adaptation of the WEPP mo<strong>de</strong>l for Ukra<strong>in</strong>ian<br />
conditions: problems, possible solutions, <strong>and</strong> perspectives.<br />
Pochvove<strong>de</strong>nie 31(2), 96–99.<br />
Norton, L.D. & Brown, L.C., (1992) Time-effect on water erosion<br />
for ridge tillage. Trans. ASAE 35(2), 473–478. 1992.<br />
Lambe, T.W. & Whitman, R.V. (1972) Mecanica <strong>de</strong> Suelos. Ed.<br />
Limusa. Mexico.<br />
Panigrahi, D.P. & Mujumdar, P.P. (2000) Reservoir operation<br />
mo<strong>de</strong>ll<strong>in</strong>g with Fuzzy Logic. Water Res. Manag. 14(2) 89–109.<br />
doi: 10.1023/A:1008170632582.<br />
Parthenia<strong>de</strong>s, E. (1965) Erosion <strong>and</strong> <strong>de</strong>position of cohesive soils.<br />
Journal of the Hydraulics Division. Proc. American Society of<br />
Civil Eng<strong>in</strong>eers, 105–139.<br />
Prosser, I.P., Dietrich, W.E. & Stevenson, J. (1995) Flow resistance<br />
<strong>and</strong> sediment transport by concentrated overl<strong>and</strong> flow <strong>in</strong> a<br />
grassl<strong>and</strong> valley. Geomorphol. 13(1), 73–86. doi:10.1016/0169-<br />
555x(95)00020-6.<br />
Prosser, I.P. (1996) Thresholds of channel <strong>in</strong>itiation <strong>in</strong> historical <strong>and</strong><br />
Holocene times. Adv. Hills. Processes 2(7), 687–708.<br />
Rapp, I. (1988) Effects of soil properties <strong>and</strong> experimental<br />
conditions on the rill erodibilities of selected soils. Ph. D. Thesis,<br />
Faculty of Biological <strong>and</strong> Agricultural Sciences, University of<br />
Pretoria, South Africa.<br />
Ramirez, H. & De La Vara, R. (2004) Análisis y diseño <strong>de</strong><br />
experimentos. Mc Graw Hill. México.<br />
Rauws, G. (1987) The <strong>in</strong>itiation of rills on plane beds of noncohesive<br />
sediments. In: Catena Supplement, 8, 107–118.<br />
Rauws, G. & Govers, G. (1988) Hydraulic <strong>and</strong> soil mechanic aspects<br />
of rill generation on agricultural soils. J. Soil Sci. 39(2), 111–124.<br />
Ranieri, S.B.L, Sparovek, G., Demaria, I.C. & Flanagan, D.C.<br />
(1999) Erosion rate estimation us<strong>in</strong>g USLE <strong>and</strong> WEPP on a<br />
Brazilian watershed. Proc. International Soil Conservation<br />
Organization Conference. West Lafayette, IN.<br />
Renard, K.G., Foster, G.R., Weesies, G.A, McCool D.K. & Yo<strong>de</strong>r,<br />
D.C. (1997a) Predict<strong>in</strong>g Ra<strong>in</strong>fall Erosion Losses: a Gui<strong>de</strong> to<br />
Conservation Plann<strong>in</strong>g with the Revised Universal Soil Loss<br />
Equation (RUSLE). USDA Agricultural H<strong>and</strong>book 703. US<br />
Government Pr<strong>in</strong>t<strong>in</strong>g Office: Wash<strong>in</strong>gton, DC.<br />
Renard, K.G., Foster, G.R., Weesies, G.A. & Porter J.P. (1997b)<br />
RUSLE – Revised Universal Soil Loss Equation. J. Soil Water<br />
Conserv. 49(3), 213–220.<br />
Reichert, J.M., Schäfer, M.J., Cassol, E.A. & Norton, L.D. (2001)<br />
Interrill <strong>and</strong> rill erosion on a tropical s<strong>and</strong>y loam soil affected by<br />
tillage <strong>and</strong> consolidation. In: Stott, D.E., Mohtar, R.H. &<br />
Ste<strong>in</strong>hardt, G.C. (eds.), Susta<strong>in</strong><strong>in</strong>g the Global Farm. Proc. 10th<br />
International Soil Conservation Organization Meet<strong>in</strong>g, May 24–<br />
29, 1999, Purdue University <strong>and</strong> the USDA-ARS Soil Erosion<br />
Research Laboratory, USA.<br />
Santoro, V.C., Amore, E., Modica, C. & Near<strong>in</strong>g, M.A. (2002)<br />
Application of two erosion mo<strong>de</strong>ls to a large Sicilian bas<strong>in</strong>. Proc.<br />
Int. Congress of European Soc. for Soil Conservation, Valencia.<br />
Santos, C.A.G., Sr<strong>in</strong>ivasan, V.S., Suzuki, K. & Watanabe, M. (2003)<br />
Application of an optimization technique to a physically based<br />
erosion mo<strong>de</strong>l. Hydrol. Processes 47, 989–1003, doi:<br />
10.1002/hyp.1176.<br />
Savabi, M.R., Klik, A., Grulich, K., Mitchell, J.K. & Near<strong>in</strong>g, M.A.<br />
(1996) Application of WEPP <strong>and</strong> GIS on small watersheds <strong>in</strong><br />
USA <strong>and</strong> Austria. Proc. HydroGIS 96: Application of Geographic<br />
Information Systems <strong>in</strong> Hydrology <strong>and</strong> Water Resources<br />
Management. IAHS Publication 235.<br />
Simons, D. B., Li, R.M. & Fullerton, L. (1981) Theoretically <strong>de</strong>rived<br />
sediment transport equations for Pima County, Arizona. Prepared<br />
for Pima County DOT <strong>and</strong> Flood Control District, Tucson,<br />
Arizona, Colorado.<br />
Sha<strong>in</strong>berg, I., Laflen, J.M., Bradford, J.M. & Norton, L.D. (1994)<br />
Hydraulic flow <strong>and</strong> water quality characteristics <strong>in</strong> rill<br />
erosion. Soil Sci. Soc. Am. J. 58(4), 1007–1012. doi:<br />
10.2136/sssaj1994.03615995005800040002x.<br />
Sha<strong>in</strong>berg, I., Goldste<strong>in</strong>, D. & Levy, G.J. (1996) Rill erosion<br />
<strong>de</strong>pen<strong>de</strong>nce on soil water content, ag<strong>in</strong>g, <strong>and</strong><br />
temperature. Soil Sci. Soc. Am. J. 60(3), 916–922.<br />
doi:10.2136/sssaj1996.03615995006000030034x.<br />
Sheridan, G.J., So, H.B., Loch, R.J., Pocknee, C. & Walker, C.M.<br />
(2000a) Use of laboratory-scale rill <strong>and</strong> <strong>in</strong>terrill erodibility<br />
measurements for the prediction of hillslope-scale erosion on<br />
rehabilitated coal m<strong>in</strong>e soils <strong>and</strong> overbur<strong>de</strong>ns. Aust. J. Soil Res.<br />
38(2), 285–297. doi:10.1071/SR99039.<br />
Sheridan, G.J., So, H.B., Loch, R.J. & Walker, C.M. (2000b)<br />
Estimation of erosion mo<strong>de</strong>l erodibility parameters from media<br />
properties. Australian J. Soil Res. 38(2), 256–284. doi:<br />
10.1071/SR99041.<br />
Silva, R.M., Santos, C.A.G. & Silva, L.P. (2007) Evaluation soil loss<br />
<strong>in</strong> Guaraira bas<strong>in</strong> by GIS <strong>and</strong> remote sens<strong>in</strong>g based mo<strong>de</strong>l. J.<br />
Urban <strong>and</strong> Environ. Eng. 1(2), 44–52, doi:<br />
10.4090/juee.2007.v1n2.044052.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.81-104, 2010
Márquez <strong>and</strong> Guevara-Pérez<br />
104<br />
Slattery, M.C. & Bryan, R.B. (1992) Hydraulic conditions for rill<br />
<strong>in</strong>cision un<strong>de</strong>r simulated ra<strong>in</strong>fall: a laboratory experiment. Earth<br />
Surf. Processes L<strong>and</strong>f. 17(1), 127–146. doi:<br />
10.1002/esp.3290170203.<br />
Smerdon, E.T. (1964) Effect of ra<strong>in</strong>fall on critical tractive forces <strong>in</strong><br />
channels with shallow flow. Trans. ASAE, Paper No. 63-700.<br />
Smerdon, E.T. & Beasley, R.P. (1959) The tractive force theory<br />
applied to stability of open channels <strong>in</strong> cohesive soils.<br />
Agricultural Experiment Station University of Missouri Research<br />
Bullet<strong>in</strong> 715.<br />
Smith, R. E., Goodrich, D. & Qu<strong>in</strong>ton, J. N. (1995) Dynamic<br />
distributed simulation of watershed erosion: the KINEROS2 <strong>and</strong><br />
EUROSEM mo<strong>de</strong>ls. J. Soil Water Conserv. 50(5), 517–520.<br />
Terzaghi, K. & Peck, R.B. (1967) Soil Mechanics <strong>in</strong> Eng<strong>in</strong>eer<strong>in</strong>g<br />
Practice, 2a. ed. John Wiley <strong>and</strong> Sons, New York.<br />
Torri, D. (1987) A theoretical study of soil <strong>de</strong>tachability. In: Catena<br />
Supplement, 10, 15–20.<br />
Trout, T.J. & Neibl<strong>in</strong>g, W.H. (1993) Erosion <strong>and</strong> sedimentation<br />
processes on irrigated fields. J. Irrig. Dra<strong>in</strong>age Engr. ASCE<br />
119(6), 947–963. doi: 10.1061/(ASCE)0733-<br />
9437(1993)119:6(947).<br />
Trout. T.J. (1996) Furrow erosion <strong>and</strong> sedimentation: on field<br />
distribution. Trans. ASAE 39(5), 1717–1723. 1996<br />
Trout, T.J. (1999) Sediment transport <strong>in</strong> irrigation furrows. In:<br />
Susta<strong>in</strong><strong>in</strong>g the Global Farm: Selected Papers from the 10th Intl.<br />
Soil Conservation Organization Meet<strong>in</strong>g, 710–716. In: Stott, D.E.,<br />
Mothar, R.H. & Ste<strong>in</strong>hardt, G.C. (eds.) West Lafayette, Ind.:<br />
Purdue University <strong>and</strong> USDA-ARS National Soil Erosion<br />
Research Laboratory.<br />
Van Klaveren, R.W. & MccooL, D.K. (1998) Erodibility <strong>and</strong> critical<br />
shear of a previously frozen soil. Trans. ASAE 41(5), 1315–1321.<br />
West, L.T., Miller, W.P., Bruce, R.R., Langdale, G.W., Laflen, J.M.<br />
& Thomas, A.W. (1992) Cropp<strong>in</strong>g system <strong>and</strong> consolidation<br />
effects on rill erosion <strong>in</strong> the Georgia Piedmont.<br />
Soil Sci. Soc. Am. J. 56(4), 1238–1243. doi:<br />
10.2136/sssaj1992.03615995005600040038x.<br />
Wischmeier W.H. & Smith D.D. (1958) Ra<strong>in</strong>fall energy <strong>and</strong> its<br />
relationship to soil loss. Trans. Am. Geophys. Union 39(2), 285–<br />
291.<br />
Yal<strong>in</strong> Y.S. (1963) An expression for bed-load transportation.<br />
Journal of Hydraulics Division ASCE 89(HY3), 221–250.<br />
Zeleke, G. (1999) Application <strong>and</strong> adaptation of WEPP to the<br />
traditional farm<strong>in</strong>g system of the Ethiopian highl<strong>and</strong>s. Proc.<br />
International Soil Conservation Organization Conference. West<br />
Lafayette, Indiana.<br />
Zhang, X.C., Near<strong>in</strong>g, M.A., Risse, L.M. & McGregor, K.C. (1996)<br />
Evaluation of runoff <strong>and</strong> soil loss predictions us<strong>in</strong>g natural runoff<br />
plot data. Trans. ASAE 39(3), 855–863.<br />
Zhang, X.C., Li, Z.B. & D<strong>in</strong>g, W.F. (2005) Validation of WEPP<br />
sediment feedback relationships us<strong>in</strong>g spatially distributed rill<br />
erosion data. Soil Sci. Soc. Am. J. 69(5), 1140–1147.<br />
doi:10.2136/sssaj2004.0309.<br />
Zhu, J.C., Gantzer, C.J., Peyton, R.L., Alberts, E.E. & An<strong>de</strong>rson,<br />
S.H. (1995) Simulated small-channel bed scour <strong>and</strong> head cut<br />
erosion rates compared. Soil Sci. Soc. Am. J. 59(1), 211–218.<br />
doi:10.2136/sssaj1995.03615995005900010032x.<br />
Zhu, J.C., Gantzer, C.J., An<strong>de</strong>rson, S.H., Peyton, R.L. & Alberts,<br />
E.E. (2001) Comparison of concentrated flow-<strong>de</strong>tachment<br />
equations for low shear stress. Soil Till. Res. 61(3), 203–212.<br />
doi:10.1016/S0167-1987(01)00207-0.<br />
Journal of Urban <strong>and</strong> Environmental Eng<strong>in</strong>eer<strong>in</strong>g (JUEE), v.4, n.2, p.81-104, 2010