Land Reclam. 37.indb - Annals of Warsaw University of Life ...

Contents 

ROMANOWICZ R.J. Data Based Mechanistic 

rainfall-fl ow models for climate change 

simulations 3 

RAVAZZANI G., MANCINI M., MERONI 

C. Design hydrological event and routing 

scheme for fl ood mapping in urban area 15 

BANASIK K., BYCZKOWSKI A. Estimation 

of T-year fl ood discharge for a small 

lowland river using statistical method 27 

WOODWARD D.E., SCHEER C.C., 

HAWKINS R.H. Curve number update used 

for Runoff Calculation 33 

Annals 

of Warsaw 

Agricultural 

University 

Land Reclamation No 37 

Warsaw 2006 

DYSARZ T., WICHER-DYSARZ J. Assessment 

of hydrologic regime changes induced 

by the Jeziorsko dam performance and morphodynamic 

processes in the Warta river 

43 

MAJEWSKI G., PRZEWOŹNICZUK W. 

Characteristics of the particulate matter 

PM10 concentration fi eld and an attempt to 

determine the sources of air pollution in the 

living district of Ursynów 55 

BARYŁA A. Runoff volume and slope gradient 

relationship – laboratory investigations 

69

PAŃKA D., ROLBIECKI R., RZE- 

KANOWSKI CZ. Infl uence of sprinkling 

irrigation and nitrogen fertilization on health 

status of potato grown on a sandy soil 75 

MUSIAŁ E., , BUBNOWSKA J., GĄSIOREK 

E., ŁABĘDZKI L. Heat balance and climatic 

water balance in vegetation period of spring 

wheat 83 

KASPERSKA-WOŁOWICZ W., ŁABĘDZKI 

L. Climatic and agricultural water balance 

for grasslands in Poland using the penman- 

-monteith method 93 

Series Editorial Board 

Elżbieta Biernacka, Warsaw Agricultural 

University, Chair 

Wojciech Bartnik, Cracow Agricultural 


Marian Granops, Warsaw Agricultural 

University, 

Waldemar Mioduszewski, Institute for 

Land Reclamation and Grassland Farming, 

Poland 

Marek Lechowicz, Warsaw Agricultural 

University, 

Józef Mosiej, Warsaw Agricultural 


Gunno Renman, Royal Institute of Technology, 

Stockholm, Sweden 

Michael Hirschi, University of Illinois at 

Urbana-Champaign, USA 

Lubos Jurik, Slovak Agriculture University, 

Nitra, Slovakia 

SERIES EDITORS 

Józef Mosiej – Chairman 

Gunno Renman 

Janusz Kubrak 

EDITORIAL STAFF 

Jadwiga Rydzewska 

Krystyna Piotrowska 

MUSIAŁ E., BUBNOWSKA J., GĄSIOREK 

E. Variation of climatic water balance and 

heat balance for various ecosystems in 

Wrocław in the years 1964–2000 101 

VABOLIENÉ G. Investigation for biological 

nitrogen removal from wastewater using simultaneous 

nitrifi cation/denitrifi cation technology 

111 

LITWIN U., JANUS J., ZYGMUNT M. 

Development of technologies used in agricultural 

engineering work on an example of 

selected stages of land consolidation process 

123 

WARSAW AGRICULTURAL UNIVERSITY PRESS 

e-mail: wydawnictwo@sggw.pl 

ISSN 0208-5771 

Edition 500 copies 

PRINT: Agencja Reklamowo-Wydawnicza A. Grzegorczyk

Annals of Warsaw Agricultural University – SGGW 

Land Reclamation No 37, 2006: 3–13 

(Ann. Warsaw Agricult. Univ. – SGGW, Land Reclam. 37, 2006) 

Data Based Mechanistic rainfall-fl ow models for climate change 

simulations 

RENATA J. ROMANOWICZ 

Lancaster University, United Kingdom. 

Abstract: Data Based Mechanistic rainfall- 

-fl ow models for climate change simulations. 

Analysis of climate change indicates an increased 

probability of high rainfalls and related fl ooding. 

Climate change may also lead to an increase in 

rainfall variability and the probability of extreme 

events. This forces researchers to focus on both 

high and low fl ow models. This paper presents the 

application of a Data Based Mechanistic (DBM) 

approach and Stochastic Transfer Function (STF) 

methods to rainfall-fl ow modelling, with special 

emphasis on low fl ows. We present two different 

models, one with nonlinearity applied to the input, 

and the second with nonlinearity on the output. In 

both cases, the application of stochastic methods of 

identifi cation and estimation of model parameters 

allows for the evaluation of predictive uncertainty 

of the estimated fl ow. The fi rst model introduces 

input nonlinearity in the form of effective rainfall. 

The second model applies the STF approach to 

log – transformed fl ow acting as a surrogate of 

water storage in the catchment. Each of the 

models has two modules: a groundwater storage 

module and a surface water module. Logarithmic 

transformation of the output introduces a bias 

towards low fl ows. In order to model also high 

fl ows, the fast fl ow component is transformed 

using the linear STF model. Both models are 

applied to a karstic catchment, the River Thet in 

the United Kingdom. 

Key words: DBM and STF models, low fl ow, 

karstic catchment. 

INTRODUCTION 

The increased variability of rainfall 

magnitude and the resulting increased 

variability of river fl ows observed in 

the past two decades may have been 

caused by climate changes related to 

NAO intensifi cation, NSO cycles and/ 

or global warming. This time period is 

not long enough to provide the statistical 

evidence supporting any particular thesis 

(Robson et al. 1998, 2002). However, 

globalisation of the world economy 

requires global thinking about the 

possible water crisis in order to avoid 

global disaster. Ongoing developments 

in modelling future climate and its 

infl uence on water resources requires 

incorporating models able to simulate 

rainfall-fl ow processes under changing 

climatic conditions. From the point of 

view of water resources management 

and planning the prediction of extreme 

(both high and low) fl ows is the most 

critical. 

Low-fl ow hydrology deals with river 

fl ows during the annual dry season. River 

fl ow is a result of complex processes of 

rainfall-water transport operating on 

a catchment scale. Usually we distinguish 

fast and slow components of river fl ow. 

The slow component, with a large time 

constant, is attributed to base fl ow, 

because the majority of the streamfl ow 

during the low fl ow period originates 

from groundwater storage. Groundwater 

storage may be related to the drainage 

of the saturated top soil zone or 

a groundwater aquifer. Another source of

4 R.J. Romanowicz 

low fl ow may be relatively slow moving 

groundwater drainage in fracture zones 

with signifi cant lateral components. Yet 

another source may be permanently 

wetted channel bank soils, the bottoms 

of alluvial valleys or wetlands. The 

existence of these different sources 

of low water depends on catchment 

geology. Smakhtin (2001) lists several 

different processes which he attributes 

to low fl ows, such as karst formations, 

lakes and glaciers. Additionally there 

are also anthropogenic impacts on low 

fl ows, among others, deforestation, 

afforestation, industrial and agricultural 

direct water abstraction, irrigation, return 

fl ows and dams and river regulation. 

The fi rst statistical approach towards 

the separation of quick and slow fl ow 

components was introduced by Young and 

Beven (1994), who developed a bilinear 

model for the rainfall-runoff process 

with input nonlinearity representing the 

effective rainfall. The main underlying 

concept was the development of a Data 

Based Mechanistic (DBM) approach to 

modelling, combining the data-based 

model with a mechanistic interpretation 

of its components (Young 1998). This 

approach applies Stochastic Transfer 

Function (STF) methods and uses fl ow 

as a surrogate of information on the soil 

water content in the catchment. Young 

(2003) extended this methodology to 

model rainfall-runoff processes “offline”, 

i.e. in a simulation mode applying 

a simple fi rst order linear model to 

describe the soil moisture storage in 

the catchment used for the nonlinear 

transformation of rainfall. In this paper 

we propose an extension of this approach, 

consisting of using a Transfer Function 

model representing the fl ow which is 

subsequently used as a surrogate of 

the catchment soil moisture storage. 

We then compare this model to a novel 

methodology for modelling low fl ows. It 

applies the STF approach to a logarithm 

of fl ow, representing water storage in 

the catchment (Romanowicz 2006). The 

logarithmic transformation enhances 

the model performance for low fl ow 

values. It introduces a bias to the model 

solution, as the mean of the logarithm 

is not equal to the logarithm of its mean 

value. We apply a separation of the fast 

and slow components of the estimated 

STF model and attribute the base-fl ow 

to the slow component only. The fast 

component, representing catchment 

quick response related to surface or 

near surface waters, may be modelled 

using the ordinary STF model without 

a logarithmic transformation of fl ow. The 

model structure has to vary depending on 

the fl ow values. We introduce potential 

evaporation as additional information on 

fl ow levels in the catchment to distinguish 

between high and low fl ow periods. It 

is clear that the base-fl ow module may 

be calibrated on information-rich data, 

i.e. in the catchment with a noticeable 

base fl ow component in the catchment 

outfl ow. The model is suitable for climate 

change simulations as it uses rainfall and 

evaporation as input variables, without 

the necessity of explicit modelling of 

soil moisture content in the catchment. 

Moreover, the application of stochastic 

methods of identifi cation and estimation 

of model parameters allows for the 

evaluation of predictive uncertainty 

of the estimated fl ow. It is interesting 

to note that water resource engineers 

use the logarithm of fl ow to test model 

performance for low fl ow, but this

Data Based Mechanistic rainfall-fl ow models for climate change simulations 5 

transformation has never been explicitly 

applied in a rainfall-runoff model with 

the aim of low fl ow modelling. 

Section 2 presents the structure of 

both DBM STF models. In section 3 

we present the application of a standard 

DBM rainfall-fl ow model developed for 

climate scenario simulations to the River 

Thet in south-east England. In section 4 

we describe the application of the logtransformed 

low fl ow model to the same 

catchment. Section 5 is a discussion of 

the results. 

DATA BASED MECHANISTIC 

(DBM) RAINFALL-FLOW 

MODEL 

In this section we present two DBM STF 

based models developed for the purpose 

of climate change simulations. The 

fi rst model follows the DBM approach 

to rainfall-fl ow modelling introduced 

by Young (1998, 2003) and Young 

and Beven (1994). This model applies 

the nonlinear transformation of the 

rainfall. The second model, described in 

Romanowicz (2006), applies a nonlinear 

transformation of the fl ow. Thus the fi rst 

model has an input nonlinearity, while the 

second uses the output nonlinearity. This 

nonlinear transformation is introduced 

in order to represent process nonlinear 

behaviour of the process thus assuming 

that its dynamics is linear. 

DBM model with nonlinear 

transformation of the rainfall 

The methodology follows the 

approach described in detail by Young 

(2003) with the difference of applying 

a Multiple Input Single Output (MISO) 

Stochastic Transfer Function (STF) 

model to obtain an estimate of the fl ow 

variable, which issubsequently used as a 

surrogate of soil moisture content in the 

catchment. It is assumed that the model 

structure has the fi rst order: 

b 

s 1 

t = r 

1 t−δ 

+ 

− 1 

1− 

az 1 

b 

+ 2 ev 

1 t−δ +ζ 

(1) 

− 2 t 

1− 

az 2 

where: 

s t – is the observed fl ow used as a 

surrogate for soil moisture at the end of 

sample time t; 

r t, ev t – denote, respectively, rainfall and 

evaporation at the same sample time t; 

δ 1, δ 2 – denote any pure, ‘advective’ time 

delay for each input; 

ζ t – represents the noise (which in some 

cases can be considered as zero mean, 

serially uncorrelated Gaussian white 

noise); 

a 1, a 2, b 1, b 2 – model parameters and 

the operator denotes a backward shift in 

time, i.e. z –1 u t = u t–1 

The parameters are estimated using 

the Simplifi ed Refi ned Instrumental 

Variable (SRIV) algorithm in the 

CAPTAIN Matlab toolbox (e.g. Young 

1989, 2000, Young et al. 1999). 

Later on, the rainfall-runoff process is 

described by a second order model, which 

is decomposed into the fast and slow 

components (Young and Beven, 1994). 

Looking for a physical interpretation, 

the slow model component represents 

ground water storage and the fast 

component is related to quick response of 

the catchment. In the case of daily time 

periods the third component, representing


direct runoff, may also appear. The model 

has the following form: 

Fast component: 

β1,1 

Z1, t = ueff 

−1 

t−δ 

1+α1z 

Slow component: 

β 

Z 2 

2, t = ueff 

−1 

t−δ 

1+α2z 

Direct component: 

Z3, t =β3uefft 

p 

uefft = rt ⋅st 

yt = Z1, t + Z2, t + Z3, 

t +ξt 

where: 

α 1, α 2, β 1, β 2 – parameters derived from 

the identifi ed second order STF model; 

ueff t – denotes the effective rainfall 

obtained from the rainfall multiplied by 

a nonlinear gain expressed by the power 

of the fl ow estimate obtained from (1). 

All the parameters of both equations 

and the power p are estimated 

simultaneously using optimisation 

methods from MATLAB® optimisation 

toolbox. Due to the application of 

stochastic methods of estimation, this 

methodology gives estimates of the 

variance of the predictions, as well as the 

variance and the correlation structure of 

the model parameters. 

DBM model with logarithmic 

transformation of fl ow 

In a second approach we apply STF 

method to the logarithm of fl ow. The 

physical explanation of this procedure 

is described in detail in Romanowicz 

(2006). In brief, we look at the rate of 

change of the fl ow rather than the water 

balance in the catchment. As mentioned 

in the introduction, the resulting fl ow 

estimates are based towards low fl ows. 

The model structure is similar to that 

shown in equation (2) with second order 

model dynamics, thus allowing for the 

decomposition of the model into fast and 

slow components: 


β1,1 

Z1, t = r 

1 t−δ 

+ 

− 1 

1+α1z 

β1,2 

+ ev 

−1 

t−δ2 

1+α1z 


β 

Z 2 

2, t = r 

1 t−δ 

+ 

− 1 

1+α2z 

β2,2 

+ ev 

−1 

t−δ2 

(3) 

1+α2z 

yt = exp(Z1,t + Z2,t + ηt) where: α1, α2, β1,1, β1,2, β2,1, β2,2 – 

parameters derived from the identifi ed 

second order STF model. 

Due to the introduced logarithmic 

transformation of fl ow, the variance of 

model predictions is heteroscedastic and 

has the form: 

μ f 

2 

= exp[(2 μ s +σs) 

/ 2] 

2 

σ f 

2 2 

= exp(2μ s + 2 σs) −exp(2 μ s +σs) 

(4)


where: μs – denotes the mean value of 

the state variable Zt = Z1,t + Z 

2 

2,t, 

σ s – denotes the estimate of its prediction 

variance. 

In order to model also high fl ows we 

should include a high fl ow module in our 

model. The high fl ow module may consist 

of a linear STF model based on either the 

errors between the low fl ow component 

and the observed fl ow or directly on 

the observations (without logarithmic 

transformation) for the periods with 

high fl ows. Yet another option is a 

nonlinear transformation of a fast fl ow 

component. This module would require 

the use of additional information on the 

conditions in the catchment provided by 

temperature or potential evaporation. 

The low fl ow module can have the form 

of an STF model for log transformed 

fl ows or we can use the slow component 

as a representation of a base fl ow in the 

catchment. In other words, the difference 

lies in the way we construct the base fl ow 

module, either as a full module obtained 

from log-transformed fl ows or its slow 

component only. The choice will depend 

on the catchment characteristics. For dry 

catchments, dominated by base fl ow, the 

fi rst approach may be suitable. However, 

for catchments with predominantly high 

fl ows, the second approach will be more 

suitable. 

APPLICATION TO RAINFALL- 

-FLOW PROCESS IN THE RIVER 

THET, UK: DBM MODEL WITH 

EFFECTIVE RAINFALL 

The Thet catchment, situated in 

the SE of England, is characterised by 

predominant base fl ows, due to its chalk- 

based geology. The catchment has an area 

about 307 km 2 with mainly arable land and 

very mixed geology. There are reservoirs 

in the catchment affecting runoff as well 

as industrial and agricultural abstractions 

and effl uent returns. Additionally, runoff 

is affected by groundwater abstraction 

and recharge. 

In this section we apply the nonlinear 

rainfall-fl ow model (1–2) for climate 

change simulations with the nonlinearity 

on the input to the Thet catchment. In 

order to represent the nonlinear relation 

between fl ow and rainfall we apply 

a State Dependent Parameter (SDP) 

approach (Young 2000) to the daily 

rainfall and fl ow data. Figure 1 shows the 

resulting nonlinear gain for the rainfall 

as a function of fl ow. This nonlinearity 

is subsequently approximated using the 

power law, as shown in eq. 2. However, 

instead of the observed fl ow we use the 

estimate of fl ow obtained from the fi rst 

order MISO model (1) identifi ed from 

the data. 

FIGURE 1. SDP identifi ed nonparametric gain for 

Thet 

The fi rst order model for soil moisture 

has the form:


0.0278 

st= rt+ 

−1 

1− 0.9844z 

0.0131 

− ev 

−1 

t−1+ξt (5) 

1− 0.9844z 

The second order rainfall-fl ow model 

with nonlinearly transformed rainfall has 

the form: 


0.0532 

Z1, t = ueff 

−1 

t−δ 

1+ 0.7953z 


0.0107 

Z2, t = ueff 

−1 

t−δ 

1+ 0.979z 

Direct component: 

Z3,t = 0.0017 ueft 0.21 

uefft = rt ⋅st 

yt = Z1, t + Z2, t + Z3, 

t +ξt 

(6) 

This model has an additional direct 

component representing catchment 

response shorter than a day. The years 

1970–1972 were used for the calibration 

and the model explained 88% of the 

observed fl ow variance. Validation was 

performed on the years 1990–1993 

and only 55% of the fl ow variance was 

explained (Figure 2). 

DBM MODEL WITH LOG- 

-TRANSFORMED FLOW 

In this section we present an application 

of the DBM model with log-transformed 

fl ow to the Thet catchment. For the 

FIGURE 2. DBM model with effective rainfall, 

validation, Thet; 55% of data variation explained 

purpose of comparison with the DBM 

model with effective rainfall presented 

in the previous section, we use the same 

as before datasets for calibration and 

validation stages. 

Low fl ow model 

The best identifi ed MISO STF model has 

a structure described by [2 2 2 1 2 0], 

which can be decomposed into the fast 

and slow components for both rainfall 

and evaporation. The model has the 

following form: 

0.011r 

log( Q 

1 

1) = t− 

+ 

−1 

1− 0.9931z 

0.027rt 1 0.0061ev 

+ − + t−2 

+ 

−1 −1 

1−0.7350z 1−0.9931z 0.0078ev 

− t−2 

+ξ 

−1 

t 

1− 0.7350z 

(7) 

where r t denotes rainfall and ev t denotes 

evaporation.


The variance of the transformed model 

estimates (one step-ahead prediction) is 

0.038. The results of the calibration give 

a 92% explanation of the data variation 

for water storage and 89% of fl ow after 

back transformation. The decrease of 

effi ciency is related to the bias towards 

lower fl ows, which is introduced by the 

logarithmic transformation. 

The decomposition of the model 

into slow and fast component is shown 

in Figure 3. The slow component has a 

time constant of 144 days and describes 

the base fl ow in the catchment, while 

the fast component has a time constant 

of about 3 days. Due to the logarithmic 

transformation of fl ow, after back 

transformation both components are 

multiplied by each other. 

2 

The effi ciency ( RT ) of the model 

during the validation period (all 27 years) 

is 77.6%. The results of the validation on 

the years 1990–1993 are shown in Figure 

4. These three years were characterised 

by a very low fl ow, and the model gives 

worst fi t to the observations out of the 

whole validation period. 

Full DBM model with 

log-transformed fl ow 

For high fl ows the DBM low fl ow 

model output has to be augmented by the 

surface fl ow from saturated areas in the 

catchment. We use evaporation levels 

to distinguish between the low and high 

fl ow periods. For evaporation higher 

than a certain level, the model output is 

equal to the exponent of groundwater 

storage in the catchment. When the 

evaporation is lower than this threshold 

value, the fl ow is increased by the value 

obtained from the STF model identifi ed 

from the errors between the base fl ow 

and observed fl ow. The threshold value 

of evaporation is optimised together 

with the parameters of the linear in fl ow 

FIGURE 3. Decomposition of the model into fast (lower panel) and slow component (upper panel); the 

observations are shown by a dotted line; calibration stage, years 1970–1973


FIGURE 4. The DBM log-fl ow model: validation 

stage, years 1990–1993; 57.6% of the data variation 

explained; observed daily fl ows are shown by 

dots, continuous line denotes model simulations, 

shaded area represent 95% confi dence bands 

model. This model gives nonzero output 

only for the evaporation smaller than a 

threshold value. The form of the STF 

model, identifi ed using Captain toolbox 

IV (instrumental variable) methods is 

[1 1 1 2 5], it is a fi rst order model as 

follows: 

⎧ ⎫ 

⎪ 0.08 0.1233 

⎪ 

Q − Q = ⎨ r − 

ev ev < 0.4⎬ 

⎪ ⎪ 

⎩ > 0.4⎭ 

tsim , tobs , −1 t−2 −1 

t−5 t 

1−0.6886z 1−0.6886z 0 

evt 

where Q t,sim denotes the fl ow 

estimated using (7) and Q t,obs denotes 

observations. 

The full DBM rainfall-fl ow model 

is the sum of (7) and (8). The threshold 

value for the evaporation was estimated 

as 0.4, using an optimisation routine 

with simultaneous estimation of the 

model parameters. This model gave 

a 93% explanation of the fl ow variation 

for the calibration period (years 1970– 

–1972). The validation was performed 

on the years (1990–1993) and is shown 

together with 95% confi dence bands in 

Figure 5. 

FIGURE 5. Full DBM model – validation on the 

years 1990–1993, 65.1% of data variation explained; 

observed daily fl ows are presented by 

dots, continuous line denotes model simulations, 

shaded area represent 95% confi dence bands 

(8) 

Figure 6 shows the log-log plot of 

the full DBM model for all the available 

years 1970–1997. As might be expected, 

it differs from the groundwater storage 

model only for high fl ows. Compared to 

that model, the gain is 2% over the whole 

validation period and it considerably 

improves the high fl ow predictions. 

Further work should be done on the 

use two different time constants for 

evaporation and rainfall and on the use

FIGURE 6. Log-log results for the 

full DBM model, River Thet, validation 

on all the data from the years 

1970–1997. 


of temperature as a threshold value. In 

the model presented here the choice 

of the threshold, although based on 

data, is site specifi c. It is related to the 

seasonal temperature variation through 

evaporation and thus explains high fl ows 

which usually occur in winter in the 

UK. It might not be general enough for 

other countries with different climatic 

conditions. 

CONCLUSIONS 

We have presented two DBM models 

suitable for off-line climate change 

simulations. The fi rst rainfall-fl ow 

model applies nonlinearity on the input, 

following the methodology of Young 

(2003). The second DBM model for low 

fl ows applies log-transformed fl ows in 

order to obtain better estimates of the 

base fl ow. The model uses evaporation as 

additional information on soil moisture 

conditions in the catchment. 

The full DBM model consists of basefl 

ow and high-fl ow components. Basefl 

ow is determined by the separation of 

a low fl ow module into slow and quick 

catchment responses. The logarithm of 

fl ow is treated here as a substitute for 

groundwater storage in the catchment. 

In hydrology the assumption of lognormal 

distribution for the fl ow is well 

justifi ed (Yevjevich 1972). In the result 

of logarithmic transformation of fl ow 

variable, the variance of the predictions 

of fl ow is heteroscedastic, depending 

on the predicted estimates of fl ow. The 

models have been tested on the River 

Thet catchment, UK. The low fl ow model 

performs better for this karstic catchment 

than the model with effective rainfall. In 

this application the simulated base fl ow 

matches well the observed low fl ow, thus 

showing that the approach developed 

is suitable for the climate change 

simulations and low fl ow predictions. 

Further research is underway on 

generalising the low fl ow model for 

catchments with larger differences 

between high and low fl ows. Experience 

gained so far points towards the 

transformation of the quick fl ow 

component of the low fl ow model


rather than to building a model based 

on the error between the observed fl ow 

and simulated base-fl ow. The simple 

structure of a surface store module may 

be extended by including a representation 

of off-line effective rainfall, as in the fi rst 

model, or can be made more general by 

introducing a complex expression for 

surface runoff. The important difference 

from all previous approaches lies in the 

way in which we condition our model 

using the available observations. Namely, 

we use the logarithmic transformation 

of fl ow to enhance information on low 

fl ows in order to better identify the base 

fl ow and we divide the dataset into 

periods characterised by high fl ows in 

order to identify the runoff component 

of the model. 

REFERENCES 

ROBSON A.J., JONES T.K., REED D.W., 

BAYLISS A.C. 1988: A study of national 

trend and variation in UK fl oods, 

“International Journal of Climatology”, 

165–182, 18. 

ROBSON A.J. 2002: Evidence for trends in 

UK fl ooding, “Philosphical Transactions 

of the Royal Society London A”, 1327– 

–1343, 360. 

ROMANOWICZ R.J., 2006: Data Based 

Mechanistic model for low fl ows: 

implications for the effects of climate 

change, submitted to Journal of 

Hydrology. 

SMAKHTIN V.U.: 2001 Low fl ow hydrology: 

A review, “J. Hydrol.” 147–186, 240. 

YEVJEVICH V.M. 1972 Stochastic 

Processes in Hydrology, Water Resources 

Publications, Fort Collins, Colorado. 

YOUNG P.C. 1989: Recursive estimation, 

forecasting and adaptive control, In: 

Leondes C. T. (Ed.), Control and Dynamic 

Systems: Advances in Algorithms and 

Computation Techniques in Dynamic 

Systems Control, 30, Academic Press, 

San Diego pp. 119–166. 

YOUNG P.C. 2000: Stochastic, dynamic 

modelling and signal processing: Time 

variable and state dependent parameter 

estimation. In: W. J. Fitzgerald, A. 

Walden, R. Smith, & P.C. Young (eds.), 

Nonstationary and Nonlinear Signal 

Processing, Cambridge University Press: 

Cambridge, 74 –114. 

YOUNG P.C. 1998: Data-based mechanistic 

modelling of environmental, ecological, 

economic and engineering systems, 

“Environmental Modelling and 

Software”, 105–122, 13. 

YOUNG P.C. 1999: Data-based mechanistic 

modelling, generalised sensitivity and 

dominant mode analysis, “Computer 

Phys. Communications”, 113–129, 117. 

YOUNG P.C. 2003: Top-down and data- 

-based mechanistic modelling of rainfallfl 

ow dynamics at the catchment scale, 

“Hydrological Processes”, 2195–2217, 17. 

YOUNG P.C., BEVEN K.J. 1994: Data- 

-based mechanistic modelling of rainfallfl 

ow nonlinearity, “Environmetrics”, 

335–363, 5. 

YOUNG, P.C., PEDREGAL D., TYCH W. 

1999: Dynamic harmonic regression, “J. 

Forecasting”, 369–394, 18. 

Streszczenie: Modelowanie niskich przepływów 

do analizy wpływu zmian klimatu na zasoby wodne. 

Racjonalna gospodarka zasobami wodnymi, 

jak również symulacja scenariuszy zmian klimatu 

nie są możliwe bez lepszego zrozumienia bilansu 

wody w zlewni, jak również lepszych metod do 

modelowania niżówek. Niniejszy artykuł przedstawia 

zastosowanie Mechanistycznego Bazującego 

na Danych (DBM) podejścia do modelowania 

procesu opad-odpływ. Przedstawione są dwa 

modele, pierwszy stosuje nieliniową transformację 

wejść (opadu), drugi, stosuje logarytmiczną 

transformację wyjść (przepływu). Obydwa modele 

wykorzystują liniową Stochastyczną Funkcję 

Przejścia (STF) do opisu dynamiki procesu. 

Wyprowadzone modele są zdekomponowane na 

dwa moduły: jeden reprezentujący szybką dyna-


mikę przepływu (wody powierzchniowe) i drugi, 

reprezentujący wolną dynamikę zwiazaną z wodami 

gruntowymi. Zastosowanie stochastycznych 

metod identyfi kacji i estymacji parametrów modeli 

pozwala na ocenę niepewności predykcji. 

W związku z logarytmiczną transformacją wyjścia, 

uzyskana wariancja predykcji modelu z 

nieliniowością na wyjściu jest zmienna w czasie. 

W artykule przedstawiamy zastosowanie obydwu 

modeli do modelowania zlewni o podłożu karstycznym 

w Wielkiej Brytanii (rzeka Thet). 

MS. received November 2006 

Author’s address 

Renata J. Romanowicz 

Environmental Centre 

Lancaster University 

Lancaster, LA1 4YQ 

United Kingdom




Design hydrological event and routing scheme for fl ood mapping 

in urban area 

GIOVANNI RAVAZZANI1 , MARCO MANCINI1 and CLAUDIO MERONI2 1Politecnico di Milano, 

2MMI s.r.l., Milan (Italy) 

Abstract: Design hydrological event and 

routing scheme for fl ood mapping in urban area. 

Defi nition of fl ood risk maps is a task to which 

modern surface hydrology addresses a substantial 

research effort. Their impact on the government of 

the fl ood prone areas have increased the need for 

better investigation of the inundation dynamics 

[Fema 2002]. This identifi es open research 

problems such as: the defi nition of the design 

hydrograph, the identifi cation of the surface 

boundary conditions for the fl ood routing over 

the inundation plan, the choice of the hydraulic 

model that is the most close to the physical 

behaviour of the fl ood routing in the specifi c 

environment, such as urban areas or river valley. 

Most of academic and commercial mathematical 

models resolving the De Saint Venant equations 

in mono or bidimensional approach, fail on 

complex topography. Steep slopes, geometric 

discontinuities, mixed fl ow regimes, initially dry 

areas are just the main problems an hydraulic 

model should solve. In this study, we address two 

points: the defi nition of the critical event for an 

inundation area and a fl ood routing modelling 

technique for a highly urbanized fl at area. For this 

latter we show that, in urban areas, a modelling 

scheme of a network of connected channels 

and storages, gives a better representation of 

surface boundary conditions such as aggregation 

of buildings and road network and suffi cient 

accuracy for fl ood risk mapping purpose respect 

to a real 2-D hydraulic routing model. 

Key words: design hydrograph, distributed model, 

fl ood hazard maps, quasi-2D model, urban fl ood. 

INTRODUCTION 

Flooding is one of the most common 

environmental hazard, due to the 

widespread geographical distribution of 

river valleys and the attraction of human 

settlements to these areas. 

Floods can be generally considered 

in two categories [Castelli. 1994]: fl ash 

fl oods, the product of heavy localized 

precipitation in a short time period over 

a given location, and general fl oods, 

caused by precipitation over a longer 

time period and over a given river basin. 

Although fl ash fl ooding occurs 

often along mountain streams, it is also 

common in urban areas where much of 

the ground is covered by impervious 

surfaces. Fixed drainage channels in 

urban areas may be unable to contain 

the runoff that is generated by relatively 

small, but intense, rainfall events. 

Many modelling approaches exist to 

simulate fl oods [Leopardi et al. 2002]. 

The choice of the simulation approach 

depends on the questions to be answered 

[Horrit and Bates 2001, Ferrante et 

al. 2000]. In this work we focus on 

determination of fl ood hazard map for

16 G. Ravazzani, M. Mancini, C. Meroni 

an urban area located in the Liguria 

province in Italy. Traditional approach 

is based on determination of fl ood extent 

for a given frequency event, using steady 

fl ow 1-dimensional model. A new rule 

has been recently introduced for highly 

urbanized area; it identifi es the hazard 

according to both fl ow depth and fl ow 

velocity, claiming the necessity to use 

more accurate unsteady fl ow numerical 

models. 

The aim of the work is to assess the 

accuracy of a quasi-2D model versus 

pure 2D models for fl ood prediction in 

urban area. 

THE CASE STUDY 

The study area is located in the western 

Italy, in the province of Liguria (Fig. 

1). The area approaches the sea and 

has an extension of about 1.8 km 2 in 

which 6 rivers are encountered: Gorleri, 

Varcavello, S. Pietro, Pineta, Rodine and 

Madonna. The drainage basin ranges 

from 0.32 km 2 of the river Rodine to 

18.05 km 2 of the river S. Pietro. The 

maximum elevation is reached in the 

Evigno Peak at 988.5 m. a.s.l. 

The morphology is characterized by 

steep slopes in the upper part of the basins, 

decreasing while approaching the sea. 

The restricted coastal plain has attracted 

tourism activity with the effect of an 

exponential growth of urban density. As 

a consequence the rivers often have been 

channelized into artifi cial drainage. In 

this situation, we can frequently observe, 

during storm event, water overtopping 

the levees and fl owing into main roads 

and districts, dragging people and cars. 

The target area is crossed by a railway 

that divides it in two parts: the upper 

with a slope of about 1.5% and the lower 

with an average slope of about 0.8%. 

THE DIRECTIVES FOR BASIN 

PLANNING 

For the purpose of land use planning 

and management, fl ood risk maps are 

required. The Italian legislation leave 

the task of the assessment of fl ood risk 

maps to the River Basin Authority with 

FIGURE 1. Aerial photo of the study area. The six rivers and the railway are visible

the Basin Plan study. The traditional 

approach simulates the inundation 

processes with a steady 1-dimensional 

model. The analysis is iterated for 50 and 

200 year return period peak discharge. 

The 50 year fl ood extent is classifi ed as 

class-A hazard and 200 year fl ood extent 

is classifi ed as class-B hazard (Fig. 

2a). Class-A areas are subject to more 

restrictive rules than class-B areas. 

A new scheme recently proposes 

[FEMA, 2002; Rosso 2003] to determine 

the hazard map on the basis of both 

hydraulic depth and fl ow velocity. 

According to this scheme, if an area is 

frequently fl ooded but with low depth 

and low velocity, the level of hazard 

is reduced (Fig. 2b). Moreover a new 

hazard class is introduced: according to 

the new scheme, an area can be classifi ed 

as class-A, class-B and class-B0 hazard. 

The class-B0 area is the less restrictive 

one. 

The necessity to employ fl ow 

velocity data, requires the use of more 

sophisticated hydraulic model, with 

a b 

Design hydrological event and routing scheme for fl ood mapping... 17 

the ability to simulate unsteady fl ow in 

complex urban terrain. 

THE DESIGN STORM 

HYDROGRAPHS FOR 

INUDATION MAPS 

There are many types of design 

hydrographs that have been developed 

over the years [Maidment 1993; Chow et 

al. 1988], and the debate is open on the 

frequencies of the fl ood map due to the 

difference between the frequencies of 

the peak discharge and the hydrograph 

volume [De Michele et. al. 2005; 

Salvadori and De Michele 2004]. 

Moreover, if a rainfall runoff model is 

used for hydrograph computation, the 

hypothesis of the equivalence among the 

return period of the peak discharge and 

the rainfall is false too. 

Inundation maps are generally 

computed using peak discharge for 

given return period as input variable and 

simple steady 1-dimensional analysis 

FIGURE 2. Evaluation of hazard maps according to the actual rule (a) based on the fl ooding extent for 

given return period events and according to the new rule (b) which considers both hydraulic depth and 

fl ow velocity


for the river channel, that means that the 

hydrograph volume and the routing on 

the inundated areas are not considered. 

As well known, especially for plain 

urban area, the storaging volume can 

signifi cantly affects the routing and the 

extension of the inundated areas. 

According to this we think that it 

is more correct, in the assessing of 

inundation map, to take into account the 

inundation volume and its statistics as the 

key variable for map characterization. 

For ungaged basin, when the 

hydrograph is determined from rainfallrunoff 

transformation, we propose 

a methodology for the inundation volume 

defi nition based on research of the 

critical rainfall event for an inundation 

area. This is defi ned as the one that 

gives the maximum value of inundation 

volume for a given return period of 

the rainfall in the hypothesis that the 

intensity duration frequency (IDF) curve 

represents the rainfall behaviour. The 

maximum inundation volume is defi ned 

as the maximum value of the integral 

of the difference between the incoming 

hydrograph and the bankfull discharge. 

For this purpose, rainfall runoff 

distributed model, FEST [Mancini 1990; 

Montaldo et al. 2002; Rulli and Rosso 

2002; Montaldo et al. 2003; Montaldo 

et al. 2004; Salandin et al. 2004], 

was employed. FEST is a distributed 

hydrologic model especially developed 

at the Politecnico di Milano focusing 

on fl ash fl ood event simulation. As 

a distributed model, FEST can manage 

heterogeneity in hillslope and drainage 

network morphology (slope, roughness, 

etc..) and land use [Rosso 1994]. 

From a family of IDF curves it is 

possible to obtain an hydrograph for 

any duration at a given frequency and so 

a series of hydrographs for every return 

period (Fig. 3). Given the bankfull 

discharge for the examined river 

branch, the hydrograph that presents the 

maximum inundation volume identifi es 

the critical rainfall event (Fig. 4) for the 

FIGURE 3. Procedure for the search of the critical event defi ned as that event which is characterized by 

the maximum potential inundation volume. The results refer to river Varcavello for the 50 years return 

period

inundation map. In the end, according 

to retention pool analysis [Artina et 

al.1997], the critical event for a fl ood 

mapping is different from the critical 

event for the maximum peak discharge 

(Fig. 3) and the correlated hydrograph 

peak presents different return period for 

a given rainfall frequency (Tab. 1) . 

DESCRIPTION OF THE 

HYDRAULIC MODEL 

Urban areas are usually characterized by 

streets and aggregation of buildings (in 

the following termed blocks), that can 

be schematized, from a fl ood routing 

point of view, as network of channels 

where the fl ow velocity is greater than 

zero, and storages where the velocity is 

about zero. This latter hypothesis derives 

from the consideration on the friction 

induced by the macro roughness of the 

manmade obstacles present in a block. 

So that the implemented network model 

presents three main unit: the main rivers, 

the channels along the main streets 


FIGURE 4. Series of potential inundation volume for different return period. The design hydrograph for 

a given return period is the one characterized by the maximum value of the potential fl ooding volume 

and the storages for the aggregation of 

buildings. 

De Saint Venant equations are then 

integrated along the river branches and 

street channels using the Preissman 

implicit numerical scheme [Wallingford 

Software 2005]. Energy and continuity 

equation is verifi ed at nodes given by the 

channel intersection. Water discharge 

can fl ow in every direction in the channel 

network according to the hydraulic 

gradient. Reservoir equation is used to 

model diffusion in the blocks. 

The connection between storages and 

channels are discretized in the specifi c 

nodes and the weir equation controls 

the water fl ux to and from the reservoir 

according to the difference of water 

level. 

The river branches feed the channel 

network where the cross sections are 

insuffi cient respect to the fl ood discharge. 

When the riverbanks are overtopped, 

water enters the street channels. 

The channel network model is 

a very good representation of urban 

fl ood routing when the hydraulic depth


TABLE 1. Comparison between maximum peak discharge and peak discharge of the hydrograph with 

the maximum inundation volume computed for a given rainfall frequency. In the last column the return 

period of the peak discharge of the hydrograph with the maximum inundation volume is reported 

Rainfall return 

period (years) 

Maximum peak 

discharge (m 3 /s) 

and energy is lower than the height of 

the surrounding buildings which act as 

impervious boundary elements. 

The described channel network 

model was implemented (Fig. 5) in the 

Infoworks-CS software [Wallingford 

Software 2005]. Rivers and main roads 

are represented by conduits. Manholes 

allow the exchange of water between 

river and roads and are used to represent 

Peak discharge of the 

hydrograph with the maximum 

inundation volume (m 3 /s) 

Return period of the peak 

discharge of the hydrograph 

with the maximum inundation 

volume (years) 

50 94.04 79.82 25.5 

100 111.75 86.37 38 

200 132.82 95.61 54 

500 161.24 99.7 64.5 

crossroads. Conduits are linked with 

storages through weirs. 

The main model parameters are the 

roughness coeffi cient which controls 

fl ow velocity in the channels and the 

weir discharge coeffi cient which controls 

the amount of water exchanged between 

channels and storages. The adopted set 

of values are reported in Table 2. 

FIGURE 5. Network quasi-2D hydraulic model representing urban area: detail of the river Varcavello

THESIS VERIFICATION USING 

2-DIMENSIONAL MODEL 

The basic assumption of the network 

model is that velocity of fl ood fl ow 

over the urban area is greater than zero 

in the main streets while, in the blocks, 

velocity can be neglected. To verify 

this assumption, we analysed fl ood 

dynamic in the blocks, by means of 

a high resolution 2-D model. A subset 

of urban land in proximity to river 

Varcavello, has been extracted from the 

main domain. It is composed by two 

main blocks: one upstream the railway 

and the other downstream. The upper 

one is characterized by an average slope 

of about 1.5%, the lower by a slope of 

about 0.8%. These subsets are considered 

to be representative of the whole study 

area. A full 2-D model was implemented 

using the SMS software [www.bossintl. 

com] (Fig. 6). A steady analysis has been 

performed. Hydraulic depths deriving 

from channel network simulation have 

been taken as boundary condition of the 

blocks. 


TABLE 2. Parameters describing components of the network quasi-2D model 

Parameter Value 

Natural river channel Strickler roughness 30 m1/3 s –1 

Concrete river channel Strickler roughness 50 m 1/3 s –1 

Not asphalted street Strickler roughness 30 m 1/3 s –1 

Asphalted street Strickler roughness 50 m 1/3 s –1 

Weir discharge coeffi cient for districts upstream the railway (a) 

0.1 

Weir discharge coeffi cient for districts downstream the railway (a) 

0.28 

Weir length for districts upstream the railway (a) 

50 m 

Weir length districts downstream the railway (a) 

20 m 

(a) 

Weir formula adopted for discharge computation in the model: Q = CdBDu g( Du −Dd) 

, 

where Q is the discharge [m 3 /s], Cd is the discharge coeffi cient, B is the width of the weir [m], Du is 

the upstream depth with respect to the crest [m], Dd is the downstream depth with respect to the crest 

[m] and g is the acceleration due to gravity [m/s 2 ]. 

Buildings have been modelled in 

two different ways: as impervious area 

(buildings are excluded from the model 

domain) and as a high roughness surface 

(Strickler coeffi cient equal to 0.01 m 1/3 s –1 ). 

The latter way means that water is free 

to move in the whole domain, but the 

fl ow through buildings is made diffi cult 

because of an high value of roughness 

parameter. Two cases for the Strickler 

coeffi cient for gardens surrounding 

buildings were considered: in the fi rst 

case its value was fi xed to 5 m 1/3 s –1 and 

in the second case to 10 m 1/3 s –1 . 

Velocity fi eld deriving from 

simulations has been mapped to a regular 

grid and the cumulative frequency of the 

velocity values has been evaluated (Fig. 

7). In the upstream block (Fig. 7a), we 

can note that most of the cells (67%) 

has a value of velocity less than 0.6 m/ 

s even for the simulation with Strickler 

coeffi cient for gardens equal to 10 m 1/3 s –1 

and impervious buildings. 

In the other simulations, the 

percentage of cells with velocity under 

0.6 m/s increases nearly to 100%. Both


a b 

FIGURE 6. Subset of the study area (a) near river Varcavello upstream the railway and (b) representation 

in the 2-D model 

a b 

FIGURE 7. Cumulative frequency of velocity magnitude simulated by means of full 2D model in (a) 

upstream district and (b) downstream district; ks_g and ks_b denote, respectively, the Strickler roughness 

coeffi cient for gardens and buildings 

considering buildings impervious or as 

high roughness surface, the peak fl ow 

velocity reaches, anyway, a maximum 

value of 1.3 m/s in just a few cells. 

The observations are valid for the 

downstream block (Fig. 7b) too. As a 

consequence of a milder slope, fl ow 

velocity is lower indeed. 

We can conclude that motion of 

water in the aggregation of buildings is 

characterized by low values of velocity. 

The assumption to simulate blocks as 

storages in the channel network model 

seems reasonable.

THE FLOOD HAZARD MAPS 

Flow velocity and hydraulic depth values, 

computed respectively in the channels 

and in the nodes of the network model, 

have been interpolated over the entire 

study area to obtain a continuous map. 

For this purpose the borders of districts 

have been considered as barriers. The 

overlay of the velocity with the depth 

map, produced the hazard map (Fig. 8), 

according to the new directive of the 

River Basin Authority (§ 3). 

A comparison is made with the 

previous study performed by Basin 

Authority [Regione Liguria] which 

made use of steady fl ow computation 

model for delimitation of fl ood extent, 

not regarding fl ow velocity. 

Total amount of fl ooded area has 

increased in this new study (Fig. 9) from 

0.97 to 1.27 km 2 . The class-A areas and 

class-B areas have decreased, respectively 

from 0.56 to 0.24 km 2 and from 0.42 to 

0.16 km 2 . The complementary area is 

included in the class-B0 area which was 

not defi ned in the previous analysis. 

The differences in the framework of 

the present study respect to the previous 


one, have counterpart in the resulting 

fl ooding map (Fig. 8). In the network 

quasi-2D model, overtopping water is 

routed along main roads as far as distant 

blocks. This is good explanation for those 

areas that, in the previous study, are not 

even water logged by 200 year fl ood 

event: the pure 1D steady model can’t 

predict fl ood processes in urban area. 

The new framework can also predict 

water logging due to manmade obstacles 

orthogonal to the fl ow direction. The 

railway divides the city on north-south 

direction and behaves as an impervious 

levee to water fl ow. This does not seem 

to be completely represented by the 1D 

model. 

CONCLUSIONS 

A procedure is presented for estimation 

of design hydrographs. It is based on 

the search for the critical event which 

maximizes the potential fl ooding 

volume. The iterative process makes 

use of the FEST model for the rainfallrunoff 

transformation. The distributed 

model permits to well represent 

basins characterized by heterogeneous 

FIGURE 8. (a) Flood hazard map resulting in this study, compared to (b) actual hazard maps published 

by Basin Authority


FIGURE 9. Extension of class-hazard area as resulting from this study, compared to previous study 

performed by Basin Authority 

morphology and land use, as the ones in 

this work. The critical event guarantees 

that the hydrograph with the maximum 

effect on territory is assumed, even 

if peak discharge is lower than the 

maximum peak discharge for the given 

rainfall frequency. 

Classical one dimensional models 

are poor tools for fl ood analysis in urban 

area. They can be used as long as the main 

stream is not overtopped, as they fail to 

simulate fl ow component other than along 

river direction. Two dimensional models, 

on the other hand, are time consuming 

and, for unsteady fl ow simulations on 

complex topography, they fail on steep 

slopes, geometric discontinuities, mixed 

fl ow and initially dry areas. 

This work proposes an hybrid 

approach. The urban and drainage 

system is modelled by means of a 

network in which both rivers and roads 

are modelled as channels linked by 

nodes. The basic assumption is that high 

density urban blocks can be modelled as 

storages in which fl ow velocity is null. 

The assumption is verifi ed by a 2D model 

results which confi rm that fl ow velocity 

in the blocks is negligible. 

The channel network model seems 

to well represents fl ood routing in urban 

area, it is not computationally expensive 

as 2D model and, above all, is much more 

stable on complex topography. On the 

other hand it requires a deep knowledge 

of the territory and a good skills of the 

modeller to set up the hydraulic sketch. 

REFERENCES 

ARTINA S., CALENDA G., CALOMINO 

F., LA LOGGIA, G., MODICA C., 

PAOLETTI A., PAPIRI S., RASULO G., 

VELTRI P. 1997: Sistemi di Fognatura. 

Manuale di Progettazione. Hoepli, chap. 

X,XVIII, Milan. (In Italian). 

Boss International. http://www.bossintl.com 

CASTELLI F. 1994: Spatial scales of frontal 

precipitation. In Advances in Distributed 

Hydrology, R. Rosso, A. Peano, I. Becchi 

and G. A. Bemporad Editors, Water 

Resources Publications, 87–114. 

CHOW V.T., MAIDMENT D.R., MAYS 

L.V. 1988: Applied Hydrology. McGraw- 

Hill.

De MICHELE C., SALVATORI G., 

CANOSSI M., SETACCIA A., ROSSO 

R. 2005: Bivariate statistical approach to 

check adequacy of dam spillway. Journal 

of Hydrologic Engineering, 10 (1): 50–57. 

FEMA, Federal Emergency Management 

Agency, 2002: Guidelines and 

specifi cations for fl ood hazard mapping 

partners. FEMA Publications. 

FERRANTE M., NAPOLETANO F., 

UMBERTINI L. 2000: Optimization of 

transportation networks during urban 

fl ooding. Journal of the American Water 

Resources Association, 36 (5): 1115– 

–1120. 

HORRIT M.S., BATES P.D. 2001: 

Predicting fl oodplain inundation: raster- 

-based modelling versus the fi nite element 

approach. Hydrological Processes, 15, 

825–842. 

LEOPARDI A., OLIVERI E., GRECO M. 

2002: Two-dimensional modelling of 

fl oods to map risk-prone areas. Journal 

of Water Resources Planning and 

Management, 128(3),168–178. 

MAIDMENT D.R. 1993: Handbook of 

hydrology. McGraw-Hill. 

MANCINI M. 1990: La modellazione 

distribuita della risposta idrologica: 

effetti della variabilitŕ spaziale e della 

scala di rappresentazione del fenomeno 

dell’assorbimento. PhD thesis, Politecnico 

di Milano. (In Italian). 

MONTALDO N., MANCINI M., ROSSO 

R. 2004: Flood hydrograph attenuation 

induced by a reservoir system: analysis 

with a distributed rainfall-runoff model. 

Hydrological Processes, 18(3), 545–563. 

MONTALDO N., RAVAZZANI G., 

MANCINI M. 2003: The role of the 

antecedent soil moisture condition on the 

distributed hydrologic modelling of the 

Toce alpine basin fl oods. International 

Conference on Alpine Meteorology and 

MAP-Meeting 2003, http://www.map2. 

ethz.ch/icam2003/ICAM-MAP2003.htm 

MONTALDO N., TONINELLI V., MANCINI 

M., ROSSO R. 2002: Coupling Limited 

Area Models with Distributed Hydrologic 


Models for Flood Forecasting: the Toce 

Basin Study Case. IAHS, 274, 229–236. 

Regione Liguria, Autoritŕ di Bacino di 

regionale. Piano di bacino stralcio per 

la difesa idraulica ed idrogeologica, 

(Ambito di Bacino no 7 – DIANESE). 

(In Italian). 

ROSSO R. 1994: An introduction to spatially 

distributed modelling of basin response. 

In Advances in Distributed Hydrology, 

R. Rosso, A. Peano, I. Becchi and G.A. 

Bemporad Editors, Water Resources 

Publications, 3–30. 

ROSSO R. 2003: Consulenza tecnico 

scientifi ca per la defi nizione degli ambiti 

normativi relativi alle fasce di inondabilitŕ 

in funzione di tiranti idrici e velocitŕ di 

scorrimento. Abstract on http://www. 

regione.liguria.it (In Italian). 

RULLI M.C., ROSSO R. 2002: An integrated 

simulation method for fl ash-fl ood risk 

assessment: 1. Frequency predictions 

in the Bisagno River by combining 

stochastic and deterministic methods. 

Hydrology and Earth system Sciences, 

6(2), 267–283. 

SALANDIN A., RABUFFETTI D., 

BARBERO S., CORDOLA M., 

VOLONTČ G., MANCINI M. 2004: 

Monitoraggio e simulazione numerica 

del fenomeno fi nalizzata alla previsione 

e gestione dell’emergenza. Neve e 

Valanghe, 51. (In Italian). 

SALVADORI G., De MICHELE C. 2004: 

Analytical calculation of storm volume 

statistics involving Pareto-like intensityduration 

marginals. Geophysical 

Research Letters, 31 (4). 

Wallingford Software. http://www. 

wallingfordsoftware.com/products/ 

infoworks_cs/. Site visited on 2005-07-08. 

Streszczenie: Konstruowanie hydrogramów 

i schematów obliczeniowych do odwzorowania 

obszarów zagrożonych powodziami na terenach 

zurbanizowanych. Wyznaczanie map ryzyka powodziowego 

jest jednym z ważniejszych zagadnień 

we współczesnej hydrologii. Ich wpływ na 

zarządzanie terenami zalewowymi zyskał duże


znaczenie, co spowodowało potrzebę badań nad 

dynamiką powodzi. Było to również przyczyną 

intensyfi kacji badań nad następującymi zagadnieniami: 

zdefi niowanie hydrogramu projektowego, 

identyfi kacja powierzchniowych warunków brzegowych 

do obliczeń zasięgu powodzi, wybór modeli 

jak najlepiej opisujących zasięg powodzi w 

specyfi cznym środowisku, jakim jest obszar zurbanizowany 

lub doliny rzeczne. Większość modeli 

matematycznych zarówno szkoleniowych, 

jak i komercyjnych, rozwiązuje równania De 

Sait Venant’a w jednym lub dwu wymiarach, nie 

sprawdza się jednak dla skomplikowanych topografi 

cznie powierzchni. Duże spadki, nieciągłości 

geometryczne, różne zmieniające się reżimy przepływu, 

początkowo suche obszary są głównymi 

problemami, które powinny być rozwiązywane 

przez modele hydrauliczne. W tych badaniach skupiono 

się na dwóch zagadnieniach: zdefi niowane 

zdarzenia krytycznego dla terenów zalewowych 

i technice modelowania terenów zalewowych 

dla silnie zurbanizowanej płaskiej powierzchni. 

W tym przypadku schemat połączonych kanałów 

i zbiorników retencyjnych dał lepsze odwzorowanie 

powierzchniowych warunków brzegowych, 

takich jak kompleksy budynków, sieć ulic i wystarczająca 

dokładność w określaniu obszarów 

zagrożonych powodziami w porównaniu do dwuwymiarowego 

modelu hydraulicznego. 

MS received November 2006 

Authors’ addresses: 

Giovanni Ravazzani 

Marco Mancini 

Politecnico di Milano 

Plazza L. da Vinci 32, 20133 Milano 

Italy 

Claudio Meroni 

MMI s.r.l, 

Via Aselli 24, 20133 Milano 

Italy 

Corresponding author: 

e-mail: giovanni.ravazzani@polimi.it 

Fax +39 02 2399 6207




Estimation of T-year fl ood discharge for a small lowland river using 

statistical method 

KAZIMIERZ BANASIK, ANDRZEJ BYCZKOWSKI 

Warsaw Agricultural University – SGGW, 

Department of Water Engineering and Environmental Restoration 

Abstract: Estimation of T-year fl ood discharge 

for a small lowland river using statistical method. 

The 34-year series of daily discharges from 

a small agricultural river basin of the Zagożdżonka 

river at the gauging station of Płachty Stare (A = 

= 82.4 km 2 ), located in the center of Poland, 

were used in the investigation. The Pearson 

type 3 (i.e. 3-parameter gamma) distribution 

and the log-normal distribution were considered 

to fi nd the best fi t with the empirical data. The 

IMGW (Institute of Meteorology and Water 

Management) computer program, which applies 

the method of maximum likelihood, was used to 

estimate the parameters of the above distributions. 

Also commonly applied in engineering practice, 

the Pearson type 3 distribution with the method of 

quantile for estimating the distribution parameters 

was additionally used. Log-normal distribution 

gives better fi t with the measured annual fl ood 

fl ows, applying Akaike criterion, and the best fi t 

applying subjective visual criterion. 

Key words: fl ood fl ow, small river basin, peak 

fl ow, frequency curve. 

INTRODUCTION 

Estimation of fl ood discharges are 

needed in designing of hydraulic or road 

structures, as well as in fl ood protection 

activity. It is often assumed that in case 

of existence of long term hydrometric 

records, the procedure for estimating 

T-year fl ood discharge could be carried 

out without larger diffi culties, which 

however does not always take place, as 

selection of various probability function 

may produce signifi cant differences in 

the results (Maidment 1993, Pilgrim and 

Doran 1993). In case of the studied river 

basin, which is research river basin of 

Department of Water Engineering and 

Environmental Restoration of Warsaw 

Agricultural University, the results are 

also important as reference values for 

application and verifi cation of indirect 

methods for the T-year fl ood discharge 

estimation, as well as for rainfall-runoff 

model verifi cation. In the paper there has 

been analysis carried out for examining 

the following statistical distributions: 

a) the Pearson type 3 – P3 (i.e. 3- 

-parameter gamma) distribution and, b) 

the log-normal distribution – LN. For 

the statistical calculation the IMGW 

(2005; Institute of Meteorology and 

Water Management) computer program, 

which applies the method of maximum 

likelihood – MML, was used to estimate 

the parameters of the above distributions. 

Also commonly applied in engineering 

practice in Poland, the Pearson type 3 

distribution with the method of quantile 

– MQ, for estimating the distribution 

parameters was additionally used. Data 

of 34-year daily discharges (1969–2002) 

of small river basin of Zagożdżonka 

were used for the analyze.

28 K. Banasik, A. Byczkowski 

DATA USED 

Data of discharge of the Zagożdżonka 

river in the gauge of Plachty Stare, 

where area of the basin is 82.4 km 2 , 

has been colleted by the Department of 

Water Engineering and Environmental 

Restoration (former Department of 

Hydraulic Structures) of the Warsaw 

Agricultural University – SGGW since 

1962. Until 1980 water stages had been 

measured by observer three times a day, 

when additionally water stage recorder 

was installed. Location of the river basin 

is shown in the Figure 1. The mean annual 

precipitation and runoff are estimated at 

610 mm and 109 mm respectively. The 

Zagożdżonka watershed is of lowland 

type. The absolute relief is 37 m. The 

mean slops of the main streams are from 

2.5 to 3.5‰. The land use is dominated 

by arable land (small grain and potatoes). 

Sandy soils are the dominant soil types in 

the watershed (Byczkowski et al. 2001). 

Hydrometric discharge measurements, 

carried out at the Płachty Stare 

FIGURE 1. Map of the upper part of the Zagożdżonka River basin 

gauge up to ten times a year, have 

been basis for rating curve estimation 

and continue its verifi cation. Annual 

maximum fl ood fl ows from the period 

1969–2002 were used for the statistical 

analysis. This period has been accepted, 

as the whole 40-year, i.e. 1963–2002 

(Banasik et al. 2003) period of records did 

not fulfi ll the postulate of homogeneity 

according to Kraskal-Wallis (IMGW, 

2005). 

RESULTS OF WQ p% 

(T-YEAR FLOOD DISCHARGE) 

ESTIMATION AND 

CONCLUDING REMARKS 

Using the IMGW computer program 

(Ozga-Zielińska et al. 1999; IMGW, 

2005), which apply the method of 

maximum likelihood (MML) for 

parameter estimation, two distribution 

function were selected from four for 

further analyze, based on the Akaike 

criterion. The selected distributions have 

been as follow:

• 

the Pearson type 3 (P3; i.e. 3parameter 

gamma) distribution, and 

• the log-normal (LN) distribution. 

Very close values of the Akaike 

criterion for the both distribution (i.e. 

167,04 and 166,03 for the Pearson 

type 3 and the log-normal distributions 

respectively) infl uence the decision 

to include both of the functions in the 

analyze. 

The statistical series of the 34-year 

period have been also approximated 

using the most frequently used in Poland 

probability distribution according to 

the Pearson type 3, with the method of 

quantiles (MQ; Kaczmarek, Trykozko 

1964). The calculations of the T-year 

discharges (or discharges with p% 

probability of exceedance) are based on 

quantiles of the following range: 10%, 

50%, 90% and 100%. 

The maximal discharges are estimated 

from the equation (Byczkowski 1999): 

WQp = WQ50%[1 + cvФ(p; s)] (1) 

where: 

WQp – maximal annual discharge with 

the probability of exceedance p%, 

Estimation of T-year fl ood discharge for a small lowland river... 29 

WQ 50% – maximal discharge with the 

probability of exceedance p = 50%, 

c v – coeffi cient of variation of the series 

of maximal discharges, 

Ф(p; s) – probability function for a given 

type of distribution. 

The results of the calculations for the 

three distributions are shown on the fi g. 

2 and given in the Table 1. 

Results presented on the Figure 

2 and in the Table 1 show relative 

signifi cant differences in the fl ood 

discharges of small probability of 

exceedance, estimated according the two 

different statistical distributions, i.e. the 

lognormal distribution and Pearson type 

3 distribution. As the difference of the 

Akaike criterion for the above statistical 

distribution was relatively small (i.e. 

166.03 for the lognormal and 167.04 

for the Pearson type 3 distribution), 

one may consider to select also other 

criterion at choosing the propel statistical 

distribution. Visual assessment of 

agreement of the theoretical distributions 

with the empirical data, shown on 

the Figure 2, indicates also that the 

TABLE 1. Flood fl ows according various distributions and methods of parameter estimation for the 

Zagożdżonka River at Płachty Stare gauge station 

Return period 

T (year) 

Probability p 

(%) 

(1-CDF) 

Flood fl ows (m 3 /s) according to distribution and method of 

parameter estimation 

Lognormal -LN Pearson type 3 - P3 Pearson type 3 – P3 

method of maximum likelihood – MML 

method of quantiles 

- MQ 

1000 0.1 71.4 33.3 34.3 

100 1.0 31.1 21.8 22.1 

20 5.0 14.9 13.9 13.8 

10 10 10.2 10.5 10.3 

2 50 2.79 3.20 2.80 

1.11 90 0.98 0.94 0.75 

1.01 99 0.59 0.63 0.59

30 K. Banasik, A. Byczkowski 

Q [m3/s] 

1 2 5 10 20 50 100 200 500 1000 

60 

T 

55 

50 

45 

40 

35 

30 

25 

20 

15 

10 

5 

100 99 90 80 70 60 50 40 30 20 10 5 2 1 0.5 0.2 p [%] 0.1 

lognormal distribution fi t better with the 

measurement results. 

The similar results of fl ood discharges 

of various probabilities, estimation 

according to Pearson type 3 distribution, 

with the use of the two various method 

of parameter estimation (i.e. method 

of maximum likelihood and method of 

quantiles) indicate that the method has 

had small infl uence on the results of 

computation. 

As the fi ve largest fl ood fl ows, shown 

on the Figure 2, indicated heterogeneity 

with the remaining data, an analyze in 

which a division of the fl oods caused 

by rainfall events and snowmelt events, 

as various genesis for the river basin 

responses, seems to be needed. 

Acknowledgment. The study described 

in this paper has been carried out within 

research project Nr 1010/P01/2006/30, 

founded by Ministry of Science and 

Higher Education. The fi nancial support 

provided by this organization is gratefully 

acknowledged. The assistance of Mr J. 

Gładecki in the process of preparing the 

material for the analyses is also gratefully 

acknowledged. 

REFERENCES 

LN-MML 

P3-MML 

P3-Mq 

FIGURE 2. Annual peak discharges frequency curve for the Zagożdżonka River at Płachty Stare gauge 

for the period of 1969–2002 

BANASIK K., BYCZKOWSKI A., 

GŁADECKI J. 2003: Prediction of 

T-year fl ood discharge for a small river 

basin using direct and indirect methods. 

Ann. Warsaw Agricult. University – 

SGGW, Land Reclam. No 23, p. 3–8.

BYCZKOWSKI A. 1999: Hydrologia t. 2. 

Wyd. SGGW, Warszawa. 

BYCZKOWSKI A., BANASIK K., HEJ- 

DUK L., MANDES B. 2001: Wieloletnie 

tendencje zmian procesu opadu 

i odpływu w małych zlewniach nizinnych 

- na przykładzie rzeki Zagożdżonki 

(Long-term tendency in changes of precipitation 

and runoff in a small lowland river 

basin of Zagożdżonka). Instytut Meteorologii 

i Gospodarki Wodnej. Atlasy 

i monografi e. Warszawa, p. 43–52. 

IMGW, 2005: Zasady obliczania największych 

przepływów rocznych o określonym 

prawdopodobieństwie przewyższenia 

– Długie ciągi pomiarowe przepływów 

(Guidelines for computation of 

annual fl ood discharges with small probability 

of exceedance – long data sets). 

Instytut Meteorologii i Gospodarki Wodnej, 

Warszawa. 

KACZMAREK Z., TRYKOZKO E. 1964: 

Application of the method of quantiles 

to estimation of the Pearson distribution. 

Acta Geoph. Polon. T. 12, z. 1. 

MAIDMENT D.R. (ed.) 1993: Handbook 

of hydrology. McGraw-Hill, Inc. New 

York. 

OZGA-ZIELIŃSKA M., BRZEZIŃSKI J., 

OZGA-ZIELIŃSKI B. 1999: Zasady obliczania 

największych przepływów rocznych 

o określonym prawdopodobieństwie 

przewyższenia – przy projektowaniu 

obiektów budownictwa hydrotechnicznego. 

(Guidelines for computation 

of annual fl ood discharges with small 

probability of exceedance – for designe 

of hydrotechnical structures). Materiały 

Badawcze, Seria: Hydrologia i Oceanologia. 

IMGW, Warszawa. 

PILGRIM D.H., DORAN D.G. 1993: 

Practical criteria for the choice of 

method for estimating extreme design 

fl oods. Extreme Hydrological Events: 

Precipitation, Floods and Droughts 

(Proceedings of the Yokohama 

Symposium, July 1993). IAHS Publ. no 

213, p. 227–235. 

Estimation of T-year fl ood discharge for a small lowland river... 31 

Streszczenie. Wyznaczenie przepływów maksymalnych 

prawdopodobnych w małej zlewni 

nizinnej przy zastosowaniu metody statystycznej. 

Codzienne przepływy rzeki Zagożdżonki w 

Płachtach Starych (A = 82,4 km 2 ), położonej na 

Równinie Radomskiej z okresu 34 lat były podstawą 

wyboru wartości maksymalnych rocznych, 

wykorzystanych do wyznaczenia przepływów 

maksymalnych o małym prawdopodobieństwie 

przekroczenia. Rozkłady prawdopodobieństwa 

– gamma i logarytmiczno-normalny, wybrano 

do opisu własności losowych ciągu maksymalnych 

przepływów rocznych. Przeprowadzone 

obliczenia za pomocą programu komputerowego 

IMGW, wykorzystujacego metodę największej 

wiarygodności do wyznaczania parametrów, wykazały 

lepszą zgodność rozkładu lograytmiczno- 

-normalnego z danymi empirycznymi niż rozkładu 

gamma (Pearsona typu 3), zarówno przy zastosowaniu 

kryterium Akaike, jak również według 

oceny wizualnej (subiektywnej). Przy rozkładzie 

tym (logarytmiczno-normalnym) uzyskano znacznie 

wyższe wartości przepływów maksymalnych 

o niskim prawdopodobieństwie przekroczenia 

(1% i 0,1%). Wyznaczając przepływy maksymalne 

prawdopodobne metodą dotychczas powszechnie 

stosowaną w praktyce inżynierskiej tj. 

przyjmując rozkład Pearsona typu 3, z parametrami 

wyznaczonymi metodą kwantyli, wykazano 

znikomy wpływ metody szacowania parametrów 

na postać rozkładu prawdopodobieństwa. 


Authors’ address: 

Kazimierz Banasik, Andrzej Byczkowski 

Wydział Inżynierii i Kształtowania Środowiska 

SGGW 

02-787 Warszawa, ul. Nowoursynowska 166 

Poland 

e-mail: kazimierz_banasik@sggw.pl 

andrzej_byczkowski@sggw.pl




Curve Number update used for runoff calculation 

DONALD E. WOODWARD*, CLAUDIA C. SCHEER* 

RICHARD H. HAWKINS** 

*Natural Resources Conservation Service, USA 

**University of Arizona, USA 

Abstract: Curve Number update used for runoff 

calculation. The Natural Resources Conservation 

Service (NRCS), formerly the Soil Conservation 

Service (SCS), developed the runoff curve 

number procedure for estimating direct runoff 

from rainfall on ungaged agricultural watersheds 

in the late 1950s. This procedure is being used 

world wide to estimate direct runoff from 

rainfall events and, more recently, applied to 

continuous simulation models. In 1990 a work 

group of hydraulic engineers from NRCS and 

the Agricultural Research Service (ARS) began 

an effort to update the procedure for estimating 

direct runoff. The results of this joint effort will 

be explained in this paper. 

Key words: NRCS runoff curve, event analysis, 

model fi tting. 

INTRODUCTION 

The NRCS runoff curve number 

procedure was developed in the late 

1950s to estimate the runoff volumes. 

It was developed as a simple procedure 

for estimating direct runoff for use in the 

design of conservation practices. Curve 

numbers represent the runoff potential 

of various soil-cover combinations. 

Originally, curve numbers were 

developed by utilizing the maximum 

annual daily runoff and associated rainfall 

data from ARS watersheds located across 

the United States. The original curve 

numbers encompassed a wide variety of 

hydrologic soil groups, agricultural land 

uses and hydrologic conditions. Curve 

number values are found in the NRCS 

(NEH-630 1985). 

Over the years, additional curve 

numbers have been developed for new 

land uses, tillage practices and cover 

conditions. Since adequate rainfall-runoff 

data was not always available for these 

new situations, alternative methods of 

developing curve numbers were utilized. 

Many of the newer curve numbers were 

developed by comparison of values 

for similar land uses, or by weighting 

current land use curve numbers, a method 

particularly favored in the development 

of urban curve numbers. 

In the 1980s, NRCS recognized that 

an in-depth review of the runoff curve 

number procedure was needed in order 

to adequately address concerns with 

curve number usage. These concerns 

included the development of curve 

numbers for additional land use and 

cover conditions from limited data, the 

need of curve numbers for new land uses 

and cultivation practices, regional and 

seasonal variation of curve numbers. 

At the 1989 ARS/NRCS Hydraulic 

Engineers Workshop, the principle 

subjects discussed were the use and abuse

34 D.E. Woodward et al. 

of curve numbers, the application of 

curve numbers in continuous hydrologic 

models, and the future of the runoff curve 

number procedure. Meeting participants 

recommended that an ARS/NRCS Work 

Group be established to further the study 

of curve numbers and investigate the 

potential need to develop curve numbers 

for major regional land areas. 

At the 1992 meeting, the Curve 

Number Work Group decided that the 

procedures used to determine the fi rst 

curve numbers needed to be revised and 

updated. A limited test indicated that 

plotting the ordered values rather than the 

paired or natural values provided realistic 

results. In the initial and subsequent 

analyses, annual maximum daily rainfall 

and runoff values were used. It was noted 

that while paired values matched nature, 

ordered values better matched design 

and evaluation watershed studies. 

The Work Group concluded that they 

did not have the resources to analyze all 

of the data available for small watersheds 

in an orderly and timely manner. Hawkins 

(1998) had the most complete, organized 

set of ARS data from small agricultural 

watersheds and spent considerable time 

and effort to remove errors from this 

data and was interested in dedicating the 

time needed to complete the required 

analysis. 

About this same time period, the 

American Society of Civil Engineers 

(ASCE) established a task group whose 

purpose it was to review the NRCS 

rainfall-runoff estimation procedure 

known as the curve number system. 

The Curve Number Work Group 

revised eight chapters of National 

Engineering Handbook (NEH-630) 

Revised version l of NEH-630 may be 

found at: http://www.wcc.nrcs.usda. 

gov/water/quality/common/neh630/ 

4content.html . 

During the revision of the NEH- 

630 chapters incorrect or misleading 

statements were corrected and, in some 

cases, certain points were reemphasized 

for clarity. The most important changes 

were: 

• 

• 

• 

• 

• 

• 

The reference to Antecedent Moisture 

Conditions (AMC) was removed 

and the terminology was changed to 

Antecedent Runoff Condition (ARC). 

It was recognized that ARC II 

represents the average watershed 

condition of the watershed when 

fl ooding occurs. This means the 

average watershed conditions could 

be different for different geographical 

areas when fl ooding occurs. In other 

words, a particular runoff curve 

number for ARC II in Arizona 

represents a different condition than 

the same curve number for ARC II in 

Oregon. 

Explicit expression of the Runoff Curve 

Number Equation as a transformation 

of the rainfall frequency distribution 

to runoff frequency distribution was 

reemphasized. 

ARC I and ARC III were expressed 

as measures of dispersion about the 

central tendency (ARC II). This is 

a corollary to treating curve number 

as a random variable. 

Mathematical proof and demonstration 

that S does not include Ia was 

documented. 

The desirability of locally determined 

curve numbers was reiterated. 

Although this was part of the original 

documentation in NEH-630, local 

calibration has been seldom pursued.

ORDERED PAIRS AND 

ASYMPTOTES 

One of the goals of the Work Group was to 

standardize the procedure for calculating 

curve numbers from rainfall-runoff data. 

The accepted handbook method was to 

plot the annual series of rainfall-runoff 

on a scatter diagram and select the curve 

number that best fi t the data. However, 

this method ignored the many storms 

that were not the largest annual events. 

The runoff curve number equation 

is often used to transform a rainfall 

frequency distribution into a runoff 

frequency distribution. For example, the 

100-year rainfall is used to determine 

the 100-year runoff, etc. This practice, 

called frequency matching, led to the 

idea of ordered pairs (Hjelmfelt 1980). 

Hawkins (1993) followed this idea by 

sorting the rainfall and runoff depths 

separately and re-aligning them on a rank 

order basis, creating new sets of rainfallrunoff 

pairs. These rainfall-runoff pairs 

can be thought of as having equal return 

periods with the individual runoffs not 

necessarily associated with the original 

causative rainfall. 

Using all the storms in the data set, 

Hawkins calculated a curve number for 

each of the ordered pairs and plotted them 

against the rainfall. The curve number 

was found to vary with storm depth, but 

for most cases approached a constant 

value at higher rainfalls. The resulting 

curves were fi tted with asymptotic 

equations to approach this constant 

curve number value. The limiting 

curve number, approached as rainfall 

approaches infi nity, is taken as the best 

fi t curve number for the watershed. 

Curve Number update user for runoff calculation 35 

This method of determining a 

watershed curve number has the 

advantage of being mathematical and 

therefore programmable. Results are 

mostly infl uenced by the largest event, 

which is in keeping with the usual intended 

applications of the curve number. . This 

method appears to give results consistent 

with the present procedures and curve 

number tables in NEH-630. The curve 

number Work Group has adopted this 

procedure for future determination of 

curve numbers from local data. 

MODES OF APPLICATION OF 

THE CURVE NUMBER METHOD 

The Work Group recognized three 

distinctly different modes of application 

for curve numbers: 

1. Determination of runoff volume for a 

given return period, given total event 

rainfall for that return period. 

2. Determination of direct runoff for 

individual events. This acknowledges 

the variation between events and is 

the basis for the initial development. 

3. 

Process models, an inferred 

application as an infi ltration model, 

or a soil moisture-curve number 

relationship, or as a source area 

distribution. 

The fi rst application is the most 

widely used in engineering and uses the 

curve number to transform the rainfall 

frequency distribution into a runoff 

frequency distribution. The runoff volume 

that is computed is often overlooked 

and the peak discharge, which is more 

frequently the desired value, is calculated 

using a unit hydrograph model.


The second application, runoff from 

individual rainfall events, is the basis 

for the original development, as the 

rainfall versus runoff plots led to the 

curve number concept. There is a wide 

variation of runoff from rainfalls of 

the same magnitude which forces us to 

acknowledge that curve number varies 

among storms for a wide range of reasons. 

The original handbook, developed in 

large part for conditions in the humid 

east, south, and mid-west, designated 

Antecedent Moisture Condition, or 

AMC, as the most signifi cant variable 

in explaining this. Average moisture 

condition was called AMC II and 

applied to the curve number when 

fl ooding occurs. Dry conditions (AMC 

I) applied to the low curve number, and 

wet conditions (AMC III) applied to the 

high curve number. This condition is 

most often determined by prior rainfall 

since soil moisture conditions are not 

frequently monitored. 

The Work Group studied the effect 

of soil moisture on curve numbers by 

looking at infi ltrometer studies (Van 

Mullem 1992). Four studies with 162 data 

pairs on 86 different soils from across 

the United States were used. Although 

average curve number increased from 

9% to 40% between the studies from the 

initial test (dry to average condition) to 

the wet condition test (average to wet 

condition), no signifi cant relationship 

was found between soil moisture and 

curve number. The study indicated that 

the difference in curve number that 

might be related to soil moisture is much 

less than the variation between ARC I 

and ARC III. Similarly, Hawkins and 

Cate (1998) showed that 5-day prior 

rainfall was the only consistent factor in 

explaining deviations from the central 

trend of runoff in 11 of 25 agricultural 

watersheds which were studied, and at 

levels far below handbook expectations. 

Because prior rainfall explains only 

part of the variation of curve number, 

the terminology has been changed to 

Antecedent Runoff Condition, or ARC. 

More importantly, the ARC I and ARC 

III conditions have been shown to be 

the bounds on the distribution of curve 

number. Figure 1 shows the ARC I, II, and 

III curve numbers plotted on a rainfallrunoff 

scatter diagram. Upon review of 

this by Hjelmfelt (1980) indicated that 

there appears to be a correlation between 

exceedance limits and ARC. Figure 2 

shows that ARC I represents the 10% 

exceedance limit and ARC III represents 

the 90% exceedance limit for a number 

of agricultural watersheds (Hjelmfelt et 

al. 1982). 

It might be inferred that the curve 

number model is an infi ltration model 

because in its application it is used to 

determine runoff incrementally over 

the duration of the storm for input into 

a unit hydrograph model. With this use 

it becomes a surrogate for an infi ltration 

model. This has created much confusion. 

The model doesn’t behave like most 

infi ltration models/equations because 

the curve number losses don’t always 

decline with time or with prior rainfall 

and may actually increase when rainfall 

intensity increases (Hawkins 1980). The 

curve number model behavior in this 

regard is the same as from a partial area 

saturation model. Additionally, with the 

curve number model, the ultimate steady 

state rate is zero. It does not approach 

a steady-state non-zero infi ltration rate 

with time, as do the Horton or Green-

Q(in) 

5 

4 

3 

2 

1 

Ampt Equations. There doesn’t seem to 

be any consensus as to whether this is the 

way watershed losses occur. 

It should be emphasized that the curve 

number model is not a point infi ltration 

model and the difference between 


Hastings, Nebraska WS44028 (1941–1954) 

0 

0 1 2 3 4 5 

FIGURE 1. Rainfall-Runoff scatter diagram showing ARC I, II, and III curve numbers 

P(in) 

CN(III) = 94 

CN(II) = 85 

CN(I) = 70 

FIGURE 2. The 10% and 90% exceedance values compared to the AMC I and AMC III curve numbers 

rainfall and runoff is better defi ned as 

watershed “losses.” Watershed indices, 

such as the curve number, are lumped 

expressions of net basin performance. 

In this regard, the runoff curve number 

model performs well as an integrator of


all the losses from all the processes over 

the watershed, which was the original 

intention. 

It is also sometimes inferred that the 

parameter S, defi ned as the potential 

maximum retention, is a physical 

property of the site like a soil moisture 

storage parameter, and the water in it can 

be accounted for. This is has not been 

shown with any certainty. The parameter 

S (or curve number) is a model variable 

and is only constant for a particular storm. 

Although it is related to soil and cover 

characteristics, it is not an identifi able 

physical property. 

VARIATION OF CURVE NUMBER 

WITH SEASON AND LAND USE 

Curve numbers were derived from 

rainfall-runoff data for 15 distinct land 

uses on 177 small watersheds in the 

United States (Rietz and Hawkins 2000). 

Curve numbers for each land use on 

each watershed were calculated using 

the asymptotic method and evaluated 

at the local, regional and national scale. 

Signifi cant differences at the 5% level 

were found between the curve numbers 

of almost all of the different land uses 

tested. Signifi cant differences in curve 

number were also found on grazed and 

ungrazed paired watersheds, and on 

watersheds that had undergone land use 

conversions. 

The general magnitudes and rank 

order of the average land use curve 

numbers were in general agreement with 

expected handbook values. Meadows 

almost always produced the lowest 

curve number at both the local and 

regional level. Forestland produced the 

lowest overall average curve number 

at the national level, but also displayed 

the largest variability. No signifi cant 

differences could be determined between 

curve numbers for pasture and rangeland 

at the regional scale or between row 

crops and small grain at any scale. Where 

comparable, pastures usually had higher 

curve numbers than meadows. None of 

these comparisons considered hydrologic 

soil group or any other soil parameter. 

Seasonal variation of curve number 

has also been noted. This is seen more 

readily in the more humid settings, and 

is rare in arid and semi-arid watersheds. 

Where evident it follows a pattern with 

higher curve numbers in the dormant 

season when the ground has less cover 

and is likely to be wetter; and lower 

curve numbers during the summer when 

the ground is dryer and vegetation is 

in a high growth stage. Also, seasonal 

variation in forest curve numbers may 

be associated with leafi ng stages (Price, 

1998). 

THE INITIAL ABSTRACTION 

RATIO (IA/S RATIO) 

The Initial Abstraction Ratio (Ia/S or λ) 

was assigned a value of 0.2 in the original 

development of runoff curve numbers. 

Data from NEH-630 (1985) has been 

used to support the value of 0.2. As part 

of the effort to develop the required 

documentation for the curve number 

study, it became apparent that additional 

review of λ would be appropriate. Data 

analysis from NEH-630 (1985) indicated 

the statistical average (mean) of the 

plotted values was about 0.05.

Two techniques, Event Analysis and 

Model Fitting, were used for determining 

Ia/S from fi eld data sets. 

Event Analysis. Event analysis 

requires concurrent synchronized breakpoint 

records of both rainfall and runoff 

depth. The event rainfall depth recorded 

when the direct runoff hydrograph 

begins is taken as Ia. Knowing the total 

event rainfall P and the direct runoff Q, 

equation 1a is solved for S, and the ratio 

simply taken Ia/S = λ. Each event gives 

a separate value for λ, and the median for 

a large number of events is taken as the 

representative watershed value. 

General Model Fitting: Here the 

value of λ is simply determined by 

iterative least squares procedure fi tting 

for both λ and S of the general equation. 

Q = (P – λS) 2 /(P + (1 – λ)S) 

for P ≥ λS 

Direct Runoff Q (inch) 

6 

5 

4 

3 

2 

1 

0 


ARS WS26030 Coshocton, Ohio 

Q = 0 for P ≤ λS 

The objective of the fi tting is to fi nd 

the values of λ and S such that 

Σ{Q – [(P – λS) 2 /(P + (1 – λ)S)]} 2 

for P > λS 

is a minimum. Here each P:Q data set 

gives only one value of λ. An illustration 

of such fi tting is given in Figure 3. 

In each of the above two methods, only 

larger storms were used. This was done to 

avoid the biasing effects of small storms 

towards high curve numbers. With Event 

Analysis, only events with Pe = P – λS 

≥ 1 inch were used. With Model Fitting, 

only events with P ≥ 1 inch were used. As 

shall be seen, resulting values of Ia were 

often quite small, so that this difference 

between the two techniques was slight. 

For statistical analysis, only watersheds 

Q = P 

Ordered 

Natural 

0 1 2 3 4 5 6 

Rainfall P (inch) 

FIGURE 3. Model Fitting by least squares for WS26030 located at Coshocton, OH with a drainage area 

of 303 acres. For the natural data (squares): S = 4.0974 inches, CN = 70.8, λ = 0.0179, R 2 = 50.50%, 

and SE = 0.32 inch. For the ordered data (triangles): S = 2.0943 inches, CN = 82.6, λ = 0.1364, R 2 = 

= 99.17%, and SE = 0.0372 inches


with more than 20 events with P ≥ 1 inch 

or Pe ≥ 1 inch were used. 

In addition, for the model fi tting 

determinations, both “natural” and 

“ordered” data sets were used. Natural 

data pairs the P and Q as they naturally 

occurred in time, and thus displays 

considerable variety in runoff with 

rainfall. Ordered data matches (usually) 

rank-ordered P and Q values, so that 

each has approximately the same 

return period. This is in keeping with 

a major application of the method, 

which in design work at least - matches 

the frequency of the rainfall with the 

frequency of the runoff. For example, the 

100-year rainfall is assumed to produce 

the 100-year runoff. 

Rainfall-runoff data from 307 

watersheds or plots were used, 

originating from USDA-Agricultural 

Research Service, US Forest Service, 

US Geological Survey, and New Mexico 

State University. The data sets covered 

23 states, mainly in the east, Midwest, 

and south of the United States. There 

was no data from the northwestern 1/3 of 

the country, from roughly California to 

Minnesota. A total of 28,301 events were 

available that met the rainfall depth (P 

and Pe) criteria. For event analysis, only 

ARS data was applicable, insofar as it 

alone contained the needed detailed in- 

storm break-point information. All others 

were only rainfall and runoff depths P 

and Q. This is summarized in Table 1. 

These 307 watersheds all had 20 or 

more events which met the storm size 

criteria. The ARS data is available from 

Web site ftp://hydrolab.arsusda.gov/ 

pub/arswater/. The “USLE” plot data 

had been used in the development of the 

Universal Soil Loss Equation, and was 

downloaded from the web site: http:// 

topsoil.nserl.purdue.edu/usle/. Forest 

Service data was in large part supplied in 

reduced form to Hawkins (1980, 1993) 

by Hewlett (1977, 1984), who used it 

in earlier papers (Hewlett, et al., 1977; 

Hewlett and Fortson, 1984). The Jornada 

plot data, from a site north of Las Cruces 

NM, was supplied by Ward and described 

in Hawkins and Ward (1998). The USGS 

data was supplied from local sources 

for a number of urban and urbanizing 

watersheds in the Tucson area. 

In general, the results showed that λ 

is not a constant from storm to storm, 

or watershed to watershed, and that the 

assumption of λ = 0.20 is unusually 

high. 

In 202 out of the 307 watersheds 

studied, Ia/S = 0.05 had a lower standard 

error in predicted Q that Ia/S = 0.2. ( Jiang 

2001) 

TABLE 1. Data sets and sources 

Data source # Watersheds (w) or plots (p) Method used 

ARS 

134 (w) 

Event Analysis, Model Fitting 

USLE (ARS) 

137 (p) 

Model Fitting 

USFS 

26 (w) 

Model Fitting 

Jornada (NMSU) 

6 (p) 

Model Fitting 

USGS 

4 (w) 

Model Fitting

CONCLUSIONS 

1. 

2. 

Additional work is needed to develop 

a revised set of curve numbers.. An 

up-to-date soils maps needs to be 

prepared for each of the experimental 

watersheds used in the study. NRCS 

is reviewing the impacts of changing 

the Ia/S ratio. 

The use of curve numbers as a 

moisture counting technique needs 

additional study. While it appears 

to work, additional documentation 

needs to be developed. 

REFERENCES 

Agricultural Research Service Water Data 

Center, 1995: ARS Water Data: ARS/ 

Access CD. USDA-ARS Hydrology 

Lab, Beltsville, Maryland. Agricultural 

Research Service Water Data Center, 

http://www.hydrolab.arsusda.gov/ 

arswater.html. 

HAWKINS R.H. 1980: Infi ltration and 

Curve Numbers: Some Pragmatic and 

Theoretic Relationships”, Proceedings of 

Symposium on Watershed Management 

1980, ASCE, 925–37. 

HAWKINS R.H. 1993: “Asymptotic 

determination of runoff curve numbers 

from data”. Journal of Irrigation and 

Drainage Engineering. Amer Soc Civ 

Eng. 119(2): 334–345. 

HAWKINS R.H., WARD T.J. 1998: “Site 

and cover effects on event runoff, 

Jornada Experimental Range, New 

Mexico”. Proceedings from American 

Water Resource Association Conference 

on Rangeland Management and Water 

Resources. Reno, NV, 361–370. 

HAWKINS R.H., CATE A. 1998: “Secondary 

Infl uences in Curve Number Rainfall 

Runoff”, Proceeding of the International 

Water Resources Engineering Conference 

ASCE. 


HEWLETT J.D., CUNNINGHAM. G.B., 

TROENDLE C.A. 1977: “Predicting 

Storm Flow and Peak Flow from small 

basins in humid areas by the R-index 

method”. Water Resources Bulletin. 

13(2): 231–253. 

HEWLETT J.D., FORTSON J.C. 1984: 

“Additional tests on the effect of rainfall 

intensity on storm fl ow and peak fl ow 

from wild-land basins”. Water Resources 

Research. 20(7): 985–989. 

HJELMFELT A.T. 1980: “Empirical 

investigation of curve number technique”. 

Journal of the Hydraulics Division. 

American Society Civil Eng 106 (HY9): 

1471–1476. 

HJELMFELT A.T., KRAMER L.A., 

BURWELL R.E. 1982: “Curve Numbers 

as Random Variables, Rainfall-Runoff 

Relationship´ Resources Publications, 

Littleton, CO., 365–370. 

HJELMFELT A.T., WOODWARD D.A., 

CONAWAY G., QUAN Q.D., Van 

MULLEM J., HAWKINS R.H. 2001: 

“Curve Numbers, Recent Developments”. 

XXIX IAHR Congress Proceedings, 

Beijing, China. 

JIANG R. 2001: Investigation of Runoff 

Curve Number Initial Abstraction Ratio. 

MS thesis, Watershed Management, 

University of Arizona, 120 pp. 

PRICE M. 1998: “Seasonal Variation in 

Runoff Curve Numbers”. MS Thesis, 

Watershed Management, University of 

Arizona. 189 pp. 

RIETZ P.D., HAWKINS R.H. 2000: “Effects 

of land use on runoff curve numbers”. 

Watershed Management 2000, American. 

Society. Civil Engineers. Proceedings 

Watershed Management Symposium, 

Fort Collins CO. (CD ROM) 

Soil Conservation Service 1985: National 

Engineering Handbook, Section 4, 

Hydrology (NEH-630). 

USDA/USLE data web site: http://topsoil. 

nserl.purdue.edu/usle/. 

USDA, Agricultural Research Service: ftp:// 

hydrolab.arsusda.gov/pub/arswater/.


Van MULLEM J.A. (1992). “ Soil Moisture 

and Runoff–Another Look”. ASCE Water 

Forum ‘92, Proceedings of the Irrigation 

and Drainage Session, Baltimore, MD. 

Streszczenie: Aktualizacja parametru CN metody 

obliczania opadu efektywnego. W drugiej 

połowie lat 50. Służba Ochrony Gleb (ang. Soil 

Conservenation Sernice ) w USA (obecnie Służba 

Ochrony Zasobów Naturalnymi ang. Natural 

Resources Conservation Service – NRCS) 

opracowała procedurę nazywaną metodą Curve 

Number do wyznaczania opadu efektywnego dla 

zlewni nieobserwowanych. Procedura ta jest sto- 

sowana na całym świecie do wyznaczania opadu 

efektywnego ze zdarzeń opadowych oraz częściej 

w modelach do symulacji ciągłych. W 1990 roku 

zespół z NRCS i Służby Badań Rolniczych (ang. 

Agricultural Research Service) rozpoczął działania 

nad aktualizacją procedury określania opadu 

efektywnego. W artykule przedstawiono wyniki 

przeprowadzonych prac. 

MS received July 2006 

Authors address: 

Donald E. Woodward 

Natural Resources Conservation Service, USA 

e-mail: dew7718@comcast.net




Assessment of hydrologic regime changes induced by the Jeziorsko 

dam performance and morphodynamic processes in the Warta river 

TOMASZ DYSARZ, JOANNA WICHER-DYSARZ 

Department of Hydraulic Engineering, Agricultural University of Poznań, Poland 

Abstract: Assessment of hydrologic regime 

changes induced by the Jeziorsko dam 

performance and morphodynamic processes 

in the Warta river. In the paper the overview of 

morphological processes occurring in the selected 

Warta river reach is given. The processes were 

induced by the performance of the Jeziorsko dam, 

which was built in 1986. The main purpose of 

the presented research is the assessment of these 

processes impact on the hydrologic conditions in 

the river system. First time the Jeziorsko reservoir 

it was fi lled up to the maximum water level in 

1991. In 1995 the hydropower plant was put into 

operation. The sediment particles transported 

with fl owing water were accumulated in the inlet 

part of the reservoir. In this area very convenient 

conditions for wild life, especially for water plants 

and birds, appeared as the result of water stages 

increase. However, the risk related to seasonal 

fl oods and dike break possibility have threaten 

the local societies until now. The historical data 

and hydrodynamic simulation model are used 

for the assessment of the hydrologic changes. 

The data consists of fi eld measurements of the 

river morphology and hydrologic data from the 

selected river gages. The hydrodynamic model 

was used to reconstruct the fl ow conditions in 

particular states of the river system evolution. The 

statistical analysis of the obtained results enabled 

the assessment of the hydrologic characteristics 

transformation in the selected Warta river reach. 

Key words: hydrodynamic, simulation model, 

river morphology, river hydrology. 

INTRODUCTION 

In the paper the overview of 

morphological processes occurring 

in the selected Warta river reach is 

given. The processes were directly or 

indirectly induced by the performance 

of the Jeziorsko dam, which was built 

in 1986. The main purpose of the 

presented research is the assessment of 

these processes impact on the hydrologic 

conditions in the river system. Since the 

dam was built the velocity characteristics 

in the analyzed Warta river reach were 

changed what induced the deposition 

of the transported sediment in the inlet 

part of the reservoir. Next the sediment 

accumulation in the river channel induced 

one more process affecting hydrologic 

regime. Due to the increase of fl oodplain 

inundation frequency, the luxuriant 

vegetation start to grow there. These 

two processes, sediment accumulation 

and vegetation growth, cause additional 

increase of fl oodplain inundation. The 

processes induced by the dam built and 

its performance seems to be self-driven 

cycle causing irreversible alterations of 

hydrologic regime. 

The problem under consideration 

draws attention of many researchers 

and engineers today. The relationships 

between hydrologic regime conditions 

and the riparian vegetation are crucially 

important for the robust river management. 

The examples of such studies were given 

by Keeland et al. (1997), Surtevant 

(1998), Stromberg (2001), Nilsson and

44 T. Dysarz, J. Wicher-Dysarz 

Svedmark (2002) and many others. 

Amongst many anthropogenic actions in 

the water environment the damming of 

rivers plays special role due to complexity 

of induced ecological responses. One 

of the most interesting aspects of these 

responses is the dam impact on the 

upstream hydraulic conditions and 

related changes of hydrologic regime. 

The example of studies on this aspect 

was given by Mumba and Thompson 

(2005). They analyzed the changes in 

the fl oodplains located along the Kafue 

river (Zambia) reach between two dams, 

Itezhi-tezhi (upstream) and Kafue Gorge 

(downstream). The Authors proved that 

dams totally changed the hydrological 

and ecological conditions in the region. 

Almost permanent inundation of the area 

located between the dams was the direct 

reason. Such conditions were imposed 

by parallel infl uences of releases from 

upper dam and high water levels in the 

lower reservoir. 

The methodology used for assessment 

of hydrologic regime changes ranges 

from simple approach based on equal 

discharge studies (Pinter and Heine 

2005) to complex analysis including 

studies of discharges and water table 

frequency (e.g. King et al. 1998, Maingi 

and Marsh 2002, Magilligan and Nislow 

2005, Wellmeyer et al. 2005). This 

is also quite common to simulate the 

behavior of the water system by means 

of mathematical model (e.g. Kite 2001, 

Maingi and Marsh 2002). In each case 

the methodology applied should match 

the specifi c features of the system, 

processes occurring there as well as data 

availability. 

The main purpose of the presented 

paper is to present the hydrological 

regime changes in the Warta river 

reach located upstream of the Jeziorsko 

reservoir. According to the performed 

observations the hydrologic conditions 

are caused by three factors. The primary 

reason is water management in the 

reservoir. Two next elements are (1) 

sediment accumulation in the river 

and (2) vegetation growth. The fi eld 

measurement data and hydrodynamic 

model are used for our studies. Steady 

and unsteady fl ow simulation were run 

to reconstruct the fl ow conditions in the 

system before and after the analyzed 

changes. The fi nal results are presented in 

two way. First method consists of water 

stages in the selected river cross-section 

related to some fl ows determined for the 

infl ow gauge station. The chosen fl ows 

are related to the specifi c probabilities of 

excedance, e.g. 1%, 10%, or maximum 

observed discharge in the system is used. 

The second method is the statistical 

analysis of unsteady fl ow simulations 

presented as relative and cumulative 

frequency functions related to the 

particular stages of river evolution. 

The paper includes into six main 

parts, acknowledgments, references and 

two appendixes. In the fi rst part brief 

introduction to the investigated problem 

is given. In the second part the study 

area and processes occurring there are 

described. The data collected are briefl y 

presented in third part. Then the basic 

elements of the applied methodology 

are presented in fourth sections. In fi fth 

part the obtained results are presented 

and discusses. Finally the conclusions 

are formulated in the sixth part. The 

attached appendixes includes the details 

of applied methodology.

CASE STUDY SYSTEM 

Jeziorsko reservoir was built in 1986 

in central part of Poland. It is located 

in the Warta river, between the Sieradz 

(upstream) and Uniejów (downstream) 

gauge stations. The reservoir inlet with 

Warta river reach are shown in the attached 

photo map (Fig. 1). The Jeziorsko 

reservoir is relatively new object. First 

time it was fi lled up to the admissible 

maximum water level (121.50 m a.s.l.) 

in 1991. The hydropower plant was put 

into operation in 1995. 

The Warta catchment area in the 

inlet of the reservoir is 8450 km 2 . The 

Assessment of hydrologic regime changes induced by the Jeziorsko... 45 

km 499 

km 500 

km 501 

km 502 

km 503 

km 504 

average discharge in the nearest Sieradz 

gauge station is about 45 m 3 /s, but the 

discharges variability ranges from 10 m 3 /s 

to 440 m 3 /s (observed in 1997). The wet 

season during the year is January – May. 

The lowest discharges are observed 

during the summer and autumn time, 

from September to December. 

The Jeziorsko reservoir is 

multipurpose reservoir performing in 

the annual working scheme. During the 

months from September to December the 

water stages are kept on the minimum 

level of 116 m a.s.l. In spring months 

the reservoir is fi lled with water. The 

maximum water level 121.5 m a.s.l. is 

km 505 

km 506 

FIGURE 1. Photo-map of Jeziorsko reservoir inlet part and Warta river


hold during April – June period. In the 

period from July to August the water 

stored at the reservoir is used and the 

reservoir water stages gradually come 

back to the minimum level. The main 

purpose of the reservoir performance is 

the fl ood protection during wet season. 

The reservoir is also the source of 

water for the big cities in Great-Poland 

province: Poznań, Koło, Śrem and 

Konin. The water stored is also used for 

irrigation 

purposes and to preserve biological 

life below the dam. The Pątnów-Konin 

and Adamów power plants used the 

water from the reservoir for cooling. 

Other goal of the Jeziorsko performance 

is production of hydro-power, though, 

this is not signifi cant due to the small 

difference between upper and lower 

water levels in the dam (about 10 m). 

The reservoir is also used for recreation 

and inland fi shery. 

Since the very beginning of the 

reservoir performance the changes of 

the Warta river bed and fl oodplains are 

observed in the inlet part of the Jeziorsko. 

Due to decrease of fl ow velocities there 

the capacity of sediment transport is 

lower and material is deposited. The 

most important processes occur in the 

river reach below the bridge in Warta 

town, between km 503 and km 500 (see 

Fig. 1). 

The annual scheme of reservoir 

performance foster the vegetation 

growth in this area. It enabled to 

establish a nature reserve for birds in 

the upper part of the reservoir in 1998. 

Since this time the important reservoir 

performance purpose is conservation of 

water conditions suffi cient for biological 

life in this area. Due to the existence of 

the nature reserve and Environmental 

Protection Ministry act valid since 1998, 

the river engineering actions are not 

allowed in the reservoir. The purpose 

of this regulation was the protection of 

birds habitats and biological life in the 

upper part of the considered object. 

The regulation causes also some 

problems indicated in previous papers, 

e.g. Wicher (2004), Wicher-Dysarz and 

Przedwojski (2005). The water levels 

in the backwater part are gradually 

increasing. Long-term results of these 

processes may cause degradation of the 

river in the inlet part of the reservoir as 

well as the reservoir capacity decreasing. 

Second risk is related to destruction of 

fl ood protecting and backwater dikes 

by overtopping. The losses occurring in 

such case are diffi cult to overestimate 

but the probability of this danger seems 

to increase. 

However, the Ministry act enabled the 

presented research in some sense. In other 

conditions the authorities responsible 

for the reservoir effectiveness and safe 

use should take steps preventing from 

these processes. Now, it is possible to 

observe the hydrologic regime evolution 

in the pure form. The only infl uences 

are reservoir performance, sediment 

accumulation and vegetation growth. 

Hence, the chosen study area properly 

fi ts the research purposes. 

COLLECTED DATA 

The set of available data consists of (a) 

system geometry, (b) discharges and 

water stages in the system, (c) state of 

the vegetation cover.

The channel and fl oodplain elevations 

are observed for long period of time in the 

area under consideration. The most basic 

set of data consists of the river regulation 

design prepared in 1975 (Matan 1975). 

The regulated river seemed to be stable 

and was not subject to any changes until 

the reservoir started to perform in 1991. 

Hence, this set of data may be considered 

as the original terrain and river bottom 

without any signifi cant infl uence of the 

analyzed processes. The another sets of 

applied data consists of measurements 

done in 1997 and 2004. After a few years 

of reservoirs performance the Warta river 

changes were checked and assessed by 

the team from Department of Hydraulic 

Engineering, Agricultural University of 

Poznań. The measurements results were 

described by Wicher (2004) and Wicher- 

-Dysarz and Przedwojski (2005). Hence, 

the system geometry data consists of 

three sets of river bottom and fl oodplain 

measurements in 37 cross-sections 

covering the river reach from Sieradz 

gauge station (km 520+850) to Jeziorsko 

dam (km 486+500). These data were 

used to describe three stages of the river 

bottom and fl oodplain evolution. 

Other data used in our analyses are 

discharges and water stages observed 

in Sieradz gauge station as well as 

water levels measured at the Jeziorsko 

dam headwater. The set of data from 

Sieradz gauge station consists of daily 

measurements done in periods 1963– 

–1970, 1973–1983, 1993–1995 and 

1997–2001. These are 27 years of 

hydrological data. Some of the data were 

measured and published by Institute of 

Meteorology and Water Management 

(1963–1970, 1973–1983). Others were 

collected by Regional Board for Water 


Management in Poznań. The later data are 

made available under the circumstances 

of cooperation between Department of 

Hydraulic Engineering and Regional 

Board for Water Management. The data 

collected from Sieradz gauge station 

are used to describe the hydrologic 

condition in the inlet of the analyzed 

river reach which are not affected 

by reservoir performance. The water 

stages at Jeziorsko headwater were 

observed during the period of reservoir 

performance, from 1992 to 2001. This set 

of data was collected by Regional Board 

for Water Management. The data were 

used to design the average scenario of 

reservoir performance during the year. 

The impact of vegetation on the 

roughness coeffi cients was assessed on 

the basis of measurements including 

plants localization and vegetation 

cover density. The stalk thickness was 

estimated in the same way. The necessary 

measurements were done by researchers 

from Department of Hydraulic 

Engineering during 2004–2005. 

On the basis of collected data hydromorphological 

stages of the system 

evolution is classifi ed as four stages of 

river bottom. The stages of river bottom 

enable the design of fi ve stages of 

hydraulic conditions in the water system. 

First analyzed state of the system is the 

primary bottom described in Matan 

(1975). This geometry was considered 

as pre-dam conditions. The second 

geometry was prepared on the basis of 

measurements done in 1997 (Wicher 

2004). This stage of the system correspond 

to the system state a few years after the 

beginning of reservoir performance. 

The roughness in the particular crosssections 

is the same as in the 1975. It


does not describe the slight changes of 

vegetation cover in this period. Hence, 

the impact of pure sediment deposition 

may be assessed. The third and fourth 

geometry of the system was prepared on 

the basis of measurements done in 2004 

and corresponds to current conditions. 

However, the third scheme does not 

refl ex to the roughness changes due to 

the vegetation growth. This element is 

included in the fourth system geometry. 

Hence, the current state of the system 

may be assessed in two ways: with and 

without vegetation growth. The prepared 

stages of river bottom and hydraulic 

conditions are as follows: 

1) initial bottom refl ecting pre-dam 

conditions (1975) with free outfl ow 

in the outlet; 

2) initial bottom refl ecting pre-dam 

conditions (1975) with outfl ow 

corresponding to the average annual 

scheme of reservoir performance; 

3) bottom measured in 1997 with the 

initial roughness and outfl ow as 

above; 

4) bottom measured in 2004 with the 

initial roughness and outfl ow as 

above; 

5) bottom measured in 2004 with 

roughness refl ecting current state of 

vegetation and its seasonal variation, 

outfl ow corresponding to the 

average annual scheme of reservoir 

performance. 

In the fi fth stage the concept of 

roughness changes due to the vegetation 

growth is applied. The riparian vegetation 

impact on the fl ow conditions may be 

modeled as the changes of Manning 

roughness coeffi cients. Following 

Klopstra et al. (1997) concept of velocity 

changes due to the presence of non- 

submerged vegetation the roughness 

coeffi cient may be expressed as 

2 

3 md 

n= h Cw 

2g 

(1) 

where: 

g – acceleration of gravity, 

h – the depth, 

C w, m and d – the parameters describe the 

characteristics of riparian vegetation, 

C w – the coeffi cient refl ecting seasonal 

changes of leaves cover. 

Its values were taken as Cw = 1.05 

for winter period and Cw = 1.40 for 

summer (Armanini et al. 2005). m is 

average number of plants in one square 

meter and d is average diameter of 

bushes stalks. The presented approach 

was used to describe the changes of 

roughness in the fl oodplain. Parameters 

m and d were determined on the basis of 

measurements done in 2004 and 2005. 

The listed stages of river system were 

used for the simulation of steady and 

unsteady fl ow conditions along the reach 

under consideration. 

APPLIED METHODOLOGY 

The main basis for the assessment 

presented here is the set of collected 

data described in previous part of the 

paper. To estimate the hydrologic regime 

characteristics related to fi ve stages of 

river system several deterministic and 

stochastic techniques are used. The 

methodology applied consists of three 

main elements. These are (1) maximum 

fl ows probability of excedence, (2) 

hydrodynamic model simulations for

steady and unsteady conditions, (3) 

analysis of relative and cumulative 

frequencies of water stages in selected 

river cross-sections. The above elements 

are described shortly in the following 

parts of the paper. 

In the fi rst part of the analysis the 

maximum fl ows probability of excedence 

is elaborated. The curve is determined 

for the Sieradz gauge station located 

in the inlet of the system. The curve is 

shown in Figure 2. Crosses represent 

the annual maximum fl ows observed 

and probabilities of excedance assigned 

to them according to Weibull formula. 

The continuous line is the Pearson 

type III curve estimated by means of 

quintile method (Ozga-Zielińska and 

Brzeziński 1994). The estimated curve 

is used to determine the characteristic 

fl ows occurring in the system with some 

specifi ed probability of excedance. The 

chosen fl ows are commonly used for the 

assessment and design purposes. They 

are presented in Table 1. 

probability [%] 

100 

80 

60 

40 

20 

0 

 


observations 

Pearson III curve 

0 100 200 300 400 500 

maximum discharge [m 3 /s] 

FIGURE 2. Maximum fl ows probability of excedance 

for Sieradz gauge station 

TABLE 1. Selected fl ows with specifi c 

probabilities of excedance determined for the 

Sieradz gauge station 

Probability 

of excedance Denotation Value [m3 /s] 

10 % Q10 % 245.40 

5 % Q5 % 263.28 

1 % Q1 % 294.34 

0.5 % Q0.5 % 304.89 

Observed 

maximum 

Max Q 440.00 

The determined probable fl ows 

together with minimum and maximum 

water stages in the reservoir are used 

to simulate the steady fl ow conditions 

in the analyzed Warta river reach. The 

computations are done for each of fi ve 

river bottom stages listed in previous part 

of the paper. The hydrodynamic model 

is used in this purpose. The routines 

included in steady fl ow module of the 

classic HEC-RAS package are applied. 

The main basis of the used methodology 

is energy balance equation written for 

the compound channel. The complete 

description of HEC-RAS package may 

be found in Brunner (2002). 

The results of the simulations are 

changes of water levels in selected river 

cross-sections related to the maximum 

fl ows and stages of river bottom. They 

are presented and discussed in the next 

part of the paper. 

The second part of the performed 

analyses consists of infl ow scenarios, 

unsteady fl ow simulations and frequency 

analysis. The used infl ow scenarios are 

the observed discharges in the Sieradz 

gauge station. They are combined with 

the average annual scheme of water 

management in the Jeziorsko reservoir. 

The unsteady fl ow module included in


HEC-RAS package is used to simulate 

the behavior of the system related 

to fi ve stages of the river. Then the 

obtained results are analyzed by means 

of water stage relative and cumulative 

frequencies determined in several river 

cross-sections. The basic defi nitions 

of frequency curves are used. More 

information may be found in Chow et al. 

(1988). 

The selected cross-sections are 

located between km 503 and km 500, 

where the most intensive morphological 

and vegetation changes are observed. 

The results are presented and discussed 

in the next part of the paper. 

ANALYSIS OF SIMULATION 

RESULTS 

The described analyses are done for the all 

cross-sections in the river reach between 

the bridge in Warta town (Fig. 1) and the 

Jeziorsko reservoir. These are sections 

located at km 503+560, 503+380, 

503+140, 502+800, 502+300, 501+270, 

500+970, 500+510 and 499+890. The 

examples of obtained results are shown 

in Figures 3–10. Figures 3–6 presents the 

results related to the fi rst part of analyses 

done with the help of steady fl ow module. 

The rest of mentioned fi gures consists of 

relative and cumulative frequency curves 

elaborated on the basis of unsteady fl ow 

simulations. 

The Figures 3 and 4 present the water 

stages related to analyzed fl ows Q 1% and 

Max Q. The water stages are estimated 

for the pre-dam conditions (denoted as 

‘in 1975’) and current conditions (‘in 

2004’). These stages are drawn on the 

original cross-sections contours. The 

124 

123 

122 

121 

120 

119 

elevation 

[m a.s.l.] 

Q1% in 2004 

Q1% in 1975 

cross-section 

Max Q in 2004 

Max Q in 1975 

distance 

[m] 

0 100 200 300 400 500 600 

FIGURE 3. Cross-section shape and characteristic 

water stages in km 503+560 

124 

123 

122 

121 

120 

elevation 

[m a.s.l.] 

Max Q in 2004 

Q1% in 1975 

Q1% in 2004 

Max Q in 1975 

119 cross-section 

distance 

[m] 

118 

0 100 200 300 400 500 600 700 

FIGURE 4. Cross-section shape and characteristic 

water stages in km 503+140 

given examples present the results for 

the cross-sections located in km 503+560 

and km 503+140. It is clearly visible 

that the morphological and vegetation 

changes in the river system caused crucial 

increase of water stages in the system. 

The maximum observed fl ows should are 

close to the maximum dikes elevations 

in the system. It is important to indicate 

that further changes in this area may 

cause the dike break by overtopping.

The increase of water stages related to 

specifi c maximum fl ows is also presented 

in Figures 5 and 6. In this graphs the 

estimated changes of water stages related 

to the fl ows Q 10%, Q 1% and Max Q are 

plotted. The squares, circles and triangles 

represent values estimated for particular 

hydraulic conditions. Horizontal axis 

describes the time. The dotted lines 

and empty fi gures refl ects the changes 

of water stages without increase of 

vegetation cover. It is clearly visible that 

water stages are raising. The vegetation 

impact causes dramatic intensifi cation of 

these processes. 

123.5 

123.2 

122.9 

122.6 

122.3 

water stage 

[m a.s.l.] 

Max Q 

Q10% 

year 

122.0 

1975 1980 1985 1990 1995 2000 2005 

In the Figures presenting relative 

frequency functions (Figs. 7 and 9) the 

estimated initial stage of hydrologic 

regime is compared with the estimated 

current conditions. The initial stage 

is represented by Pearson type III 

probability density function fi tted to the 

results of simulation. In these fi gures 

the minimum fl oodplain elevations are 

denoted as dashed line. The Figures 8 

and 10 shows the gradual changes 

of cumulative frequency functions 


Q1% 

FIGURE 5. Changes of the water stages related to 

characteristic fl ows in cross-section km 503+560 

123.0 

122.7 

122.4 

122.1 

water stage 

[m a.s.l.] 

Max Q 

Q10% 

Q1% 

121.8 

year 

1975 1980 1985 1990 1995 2000 2005 

FIGURE 6. Changes of the water stages related to 

characteristic fl ows in cross-section km 503+140 

124 

123 

122 

121 

120 

water surface 

elevation 

[m a.s.l.] 

free outflow, 

bottom 1975 

effect of impacts, 

bottom 2004 

floodplain 

minimum 

relative 

frequency 

0.00 0.02 0.04 0.06 0.08 0.10 

FIGURE 7. The relative frequency function corresponding 

to the initial and current state of the 

hydrologi-cal regime in cross section located at 

km 503+560 

related to analyzed stages of the hydromorphological 

conditions in the river. 

Presented results shows two common 

trends in cross-sections located between 

km 503 and 501. The most frequent 

stages become higher and their relative 

frequency is greater. This means greater 

frequency of fl oodplain inundation, what 

favors the riparian vegetation growth.


124 

123 

122 

121 

water surface 

elevation 

[m a.s.l.] 

bottom 2004 

bottom 1997 

bottom 1975 

cumulative 

frequency 

120 

0.00 0.20 0.40 0.60 0.80 1.00 

FIGURE 8. The cumulative frequency function 

corresponding to the analyzed states of the hydrological 

regime in cross section located at km 

503+560 

123 

122 

water surface 

elevation 

[m a.s.l.] 

CONCLUSIONS 

bottom 2004 

+vegetation 

free outflow, bottom 1975 

effect of impacts, 

bottom 2004 

121 

floodplain 

minimum 

free outflow, 

bottom 1975 relative 

frequency 

120 

0.00 0.04 0.08 0.12 

FIGURE 9. The relative frequency function corresponding 

to the initial and current state of the 

hydrologi-cal regime in cross section located at 

km 503+14 

The obtained results shows that potential 

hydrologic regime in the analyzed area 

is signifi cantly changed. The impact 

of each single process as well as the 

cumulated impact of all processes have 

123 

122 

121 

water surface 

elevation 

[m a.s.l.] 

bottom 2004 

bottom 1997 

bottom 2004 

+vegetation 

free outflow, bottom 1975 

bottom 1975 

cumulative 

frequency 

120 

0.00 0.20 0.40 0.60 0.80 1.00 

FIGURE 10. The cumulative frequency function 

corresponding to the analyzed states of the hydrological 

regime in cross section located at km 

503+140 

one common trend. The higher water 

stages become more frequent and their 

frequency is greater. This is well visible 

in Figures 7–10. The changes are crucial 

even in the cross-section located in 

the longest distance from reservoir 

(km 503+560). The areas located along 

river reach were not fl ooded frequently 

there (Fig. 7). Now, the inundation is 

almost permanent. 

The risk related to the dike break by 

overtopping is raising, too. This aspect 

of the hydrologic regime changes is 

clearly visible in Figures 3–6. The 

channel capacity was huge enough to 

pass fl ows as Q 1% and Max Q in the past. 

Now, the changes of channel geometry 

and roughness cause huge resistance and 

increase of water stages. It is quite easy 

to predict, that further changes will result 

in natural catastrophe and damages. 

The presented results shows that the 

probability of such an event dramatically 

increases.

ACKNOWLEDGEMENT 

This work was supported by Polish 

Committee for Scientifi c Research under 

grant “Analysis of morphodynamic and 

vegetation impacts on the hydraulic 

conditions in backwater part of lowland 

reservoirs”, project no. 2 P06S 034 30. 

REFERENCES 

ABBOTT M.B. 1979: Computational 

Hydraulics, Elements of the Theory of 

Free Surface Flows. London: Pitman. 

ARMANINI A., RIGHETI M. & GRISENTI 

P. 2005: Direct measurement of vegetation 

resistance in prototype scale. Journal of 

Hydraulic Engineering, 43 (5): 481–487. 

BARKAU R.L. 1982. Simulation of the 

July 1981 fl ood along Salt River, Report 

for CE695BV, Special Problems in 

Hydraulics. Fort Collins, CO.: Department 

of Civil Engineering, Colorado State 

University. 

BRUNNER G.W. 2002: HEC-RAS, 

River Analysis System hydraulic 

reference manual, computer program 

documentation. Davis, CA.: US Army 

Corps of Engineers, Hydrologic 

Engineering Center. 

CHOW V.T., Maidment D.R. & Mays L.W. 

1988: Applied Hydrology. McGraw-Hill 

International Editions, Civil Engineering 

Series, McGraw-Hill Book Company. 

CUNGE J.A., HOLLY F.M.Jr & VERVEJ A. 

1980: Practical aspects of computational 

river hydraulics, Pitman Advanced 

Publishing Program. 

KEELAND B.D., CONNER W.H. & 

SHARITZ R.R. 1997: A comparison 

of wetland tree growth response to 

hydrologic regime in Louisiana and 

South Carolina. Forest Ecology and 

Management, 90: 237–250. 

KING S.L., ALLEN J.A. & McCOY J.W. 

1998: Long-term effects of a lock and 

dam and greentree reservoir management 


on a bottomland hardwood forest. Forest 

Ecology and Management, 112: 213–226. 

KITE G. 2001: Modelling the Mekong: 

hydrological simulation for environmental 

impact studies. Journal of Hydrology, 

253: 1–13. 

KLOPSTRA D., BARNEVELD H.J., 

NOORTWIJK J.M. von & VELZEN E.H. 

von 1997: Analytical model for hydraulic 

roughness of submerged vegetation. 

The 27 th Congress of the International 

Association for Hydraulic Research, 

Proceedings of Theme A, Managing 

Water: Coping with Scarcity and 

Abundance, San Francisco, California. 

LIGGETT M.B. & CUNGE J.A. 1975: 

Numerical methods of solution of the 

unsteady fl ow equations, in Mahmood 

and Yevjevich (1975). 

MAGILLIGAN F.J. & NISLOW K.H. 2005: 

Changes in hydrologic regime by dams. 

Geomorphology, 71: 61–78. 

MAHMOOD K. & YEVJEVICH V. (eds.) 

1975: Unsteady fl ow in open channels. 

Fort Collins, Colorado: Water Resources 

Publications. 

MAINGI J.K. & MARSH S.E. 2002: 

Quantifying hydrologic impacts following 

dam construction along the Tana River. 

Kenya. Journal of Arid Environments, 

50: 53–79. 

MATAN J. 1975: Jeziorsko reservoir in the 

Warta river, Hydro-engineering and land 

improvements structures in the reservoir 

backwater. Poznań: Hydroprojekt (in 

Polish). 

MUMBA M. & THOMPSON J.R. 2005: 

Hydrological and ecological impacts 

of dams on the Kafue Flats fl oodplain 

system, southern Zambia. Physics and 

Chemistry of the Earth, 30: 442–447. 

NILSSON C. & SVEDMARK M. 2002: Basic 

principles and ecological consequences 

of changing water regimes: riparian 

plant communities. Environmental 

Management, 30 (4): 468–480. 

OZGA-ZIELIŃSKA M., BRZEZIŃSKI 

J. 1994: Applied hydrology. Scientifi c 

Publishing PWN, Warsaw (in Polish).


PINTER N. & HEINE R.A. 2005: 

Hydrodynamic and morphodynamic 

response to river engineering 

documented by fi xed-discharge analysis, 

Lower Missouri River, USA. Journal of 

Hydrology, 302: 70–91. 

STROMBERG J.C. 2001: Restoration of 

riparian vegetation in the south-western 

United States: importance of fl ow 

regimes and fl uvial dynamism. Journal of 

Arid Environments, 49: 17–34. 

SURTEVANT B.R. 1998: A model of wetland 

vegetation dynamics in simulated beaver 

impoundments, Ecological Modeling, 

112: 195–225. 

WELLMEYER J.L., SLATTERY M.C. 

& PHILLIPS J.D. 2005: Quantifying 

downstream impacts of impoundment on 

fl ow regime and channel planform, lower 

Trinity River, Texas. Geomorphology, 

69: 1–13. 

WICHER J. 2004: Sediment accumulation 

in lowland reservoirs. Ph.D. thesis in the 

Agricultural University of Poznan (in 

Polish). 

WICHER-DYSARZ J., PRZEDWOJSKI B. 

2005: Modeling of sediment accumulation 

in the inlet part of Jeziorsko reservoir, 

Annual Reviews of Agricultural 

University of Poznan, 26: 483–493. 

Streszczenie: Ocena zmian warunków hydrologicznych 

wywołanych działaniem zapory Jeziorsko 

oraz procesami morfodynamicznymi 

w rzece Warcie. W artykule zostały opisane procesy 

morfodynamiczne zachodzące na wybranym 

odcinku rzeki Warty. Zjawiska te zostały wywołane 

działaniem zapory Jeziorsko wybudowanej 

w 1986 roku. Głównym celem badań jest oszacowanie 

wpływu tych procesów na aktualne warunki 

hydrologiczne panujące w rzece. Pierwsze 

napełnienie zbiornika do poziomu maksymalnego 

piętrzenia odbyło się w 1991 roku. W 1995 roku 

została uruchomiona elektrownia wodna „Jeziorsko”. 

W wyniku przerwania ciągłości transportu 

rumowiska, w górnej części zbiornika Jeziorsko 

nastąpiło akumulacja transportowanych przez 

rzekę cząstek rumowiska. W efekcie osadzania 

się rumowiska w górnej części zbiornika, zaczęła 

intensywnie rozwijać się roślinność oraz ptactwo 

wodne, które znalazło idealne warunki do rozwoju. 

Czynniki te spowodowały spiętrzenie wody 

w tej części zbiornika. Dane historyczne oraz model 

hydrodynamiczny zostały wykorzystane, aby 

oszacować zmiany warunków hydrologicznych. 

Zestaw wykorzystanych danych zawiera: dane 

z pomiarów terenowych geometrii i morfologii 

koryta oraz pomiary stanów i przepływów z wybranych 

wodowskazów. Model hydrodynamiczny 

został wykorzystany do odtworzenia warunków 

przepływu w wybranych fazach zmian systemu 

wodnego. Analiza statystyczna wyników symulacji 

umożliwiła ocenę zmian charakterystyk hydrologicznych 

na wybranym odcinku rzeki Warty. 



Tomasz Dysarz, Joanna Wicher-Dysarz 

e-mail: dysarz@au.poznan.pl 

Akademia Rolnicza w Poznaniu 

ul. Wojska Polskiego 73a 

60-625 Poznań




Characteristics of the particulate matter PM10 concentration fi eld 

and an attempt to determine the sources of air pollution in the living 

district of Ursynów 

GRZEGORZ MAJEWSKI, WIESŁAWA PRZEWOŹNICZUK 

Warsaw Agricultural University – SGGW 

Department of Water Engineering and Environmental Restoration 

Abstract: Characteristics of the particulate 

matter PM10 concentration fi eld and an attempt 

to determine the sources of air pollution in the 

living district of Ursynów. The paper presents 

characteristics of the imission fi eld of PM10 

particulate matter with the use of basic statistical 

data such as mean values, variability ranges, 

numbers of exceedences of the allowable 

concentration, frequency distributions of 

concentration occurrence in individual periods, 

and time courses. Mean values of basic 

meteorological elements were calculated, circular 

graphs for percentiles of dust concentration were 

prepared, and percentiles of pollution plume rates 

were made in order to analyse meteorological 

conditions of dispersal and spreading out of air 

pollution. A synthetic WZ indicator of pollution 

has also been calculated that encompasses the 

joint impact of various meteorological elements 

on pollution concentration levels. 

Key words: particulate matter PM10, pollution 

plume rate, meteorological conditions. 

INTRODUCTION 

First, identifi cation of areas where 

allowable standards of pollution 

concentration are exceeded, and second, 

identifi cation of pollution sources 

being the threat to the environment 

are basic tasks for the atmospheric 

air quality monitoring system. The 

knowledge on the concentration fi eld 

and emission sources is a basis for 

setting correction programmes necessary 

for the improvement of environmental 

conditions. 

The order by Minister of Environment 

as of 5 July 2002 (DzU, No 115, pos. 

1003) settles in detail requirements 

which the air protection programs should 

answer to. It is said inside that act, 

among others, that apart of assessing the 

imission stated on a given area, the infl ow 

of pollution on that area should also be 

estimated. Discrimination between the 

emission infl uence and the impact of 

meteorological conditions on the rate 

of pollution concentration registered 

(Walczewski et al. 2000) is extremely 

important for proper execution of all 

tasks of environment protection services, 

because the increase or the decrease of 

pollution concentration values depend on 

both emission and dispersion conditions. 

The meteorological characteristics of the 

measurement period makes possible to 

evaluate properly the imission situation. 

Therefore, basic meteorological data 

should be taken into account at the 

interpretation of imission measurement 

results, and a detailed analysis of the 

wind rose from the measurement period 

would allow conclude on whether the 

measured concentrations came either 

from local emission sources or are rather

56 G. Majewski, W. Przewoźniczuk 

caused by an infl ow of pollution from 

outside the measuring network. 

The living quarter of Ursynow is 

one of the three districts of Warsaw, 

the Polish capital, for which the air 

protection correctional programs were 

made. That program was prepared 

because the allowable level of PM10 

particulate matter had been exceeded 

(Order No 62). 

The particulate matter is a commonly 

occurring pollution matter, the 

concentrations of which maintain in 

urban agglomerations at a rather high 

level exceeding frequently allowable 

values. Miniparticles dispersed in the 

air are considered to be one of more 

essential potential threats to human 

health as related to air pollution (i.a. 

Juda & Chróściel 1974, Stern 1994, 

Dockery 1996, Quarg 1996, Niecko 

et al. 1998, Warych 1999, Jabłońska 

2003). It is assessed that about one third 

of the population in Poland exposed to 

inhalation of dust pollution is bound to 

chronic diseases of the breathing system, 

that leads to circulatory system and 

cancer diseases (Warych 1999). 

Ursynów is a district of Warsaw, 

capital of Poland, situated almost entirely 

on the Warsaw Lowland, elevated 20–30 

m up of the level of water in the Vistula 

River, and it encompasses the southern 

regions of the Warsaw city. It covers the 

area of 44.6 km 2 constituting 8.6% of 

the Warsaw city total area, being at the 

third place as for the size within the city. 

Three nature reserves occur within the 

quarter mentioned; they are: the Natolin 

Wood, the Ursynów Slope – being the 

main topographic element shaping the 

local climate of the region mentioned, 

and the Kabaty Wood called the Stefan 

Starzyński Wood – a forest tract over 

900 ha in size lying within the Warsaw 

region; the latter wood is the most 

essential natural element of the southern 

aeration system for the capital. 

The industrial plants and structures 

equipped in their own boiler-rooms 

infl uencing the air quality on the area 

of the Ursynów quarter are as follows: 

Oncology Centre, Warsaw Plant of 

Aggregate Exploitation, Midas Polska 

Ltd, Volvo Auto Polska, Underground 

Standstill Station, PBM Południe Co, 

Radio Ceramics Plant, Stolbud Ltd, 

Geant Hypermarket, Unikon-Beton, 

Tesco Hypermarket, and others. The 

streets with the greatest intensity of car 

traffi c in this quarter are: Pulawska, 

KEN Avenue, Rodowicza Anody, 

Rosoła, Płaskowickiej, Ciszewskiego, 

Roentgena, and Poleczki street. The areas 

with the occurrence of low communalliving 

emission are as follows: ‘Green 

Ursynów’ along the Puławska Avenue, 

borders of the Kabacki Wood from the 

Zolna street up to the city boundary, 

the western side of the Puławska 

Avenue, the area encompassing the 

streets of Krasnopolska, Baletowa, 

Karczunkowska, Sarabandy, Farbiarska 

and Gawota, the area along the Ursynów 

Slope, and fi nally the area along the 

Prawdziwka street. 

MATERIAL AND METHODS 

An attempt has been undertaken 

in this paper to assess the imission 

situation in the Ursynow quarter, in the 

MzWarszUrsynow station area, as well as 

an attempt to point out the reasons for the 

existing state. The data for calculations

were taken from the automatic station 

of atmospheric air monitoring called 

the MzWarszUrsynow station (φN = 

= 52 o 09’39’’, λE=21 o 02’03’’, 102 m a.s.l.) 

working within the regional network of 

the Masovia Province (WIOS database). 

Measurements of SO 2, NO, NO 2, NO x, 

CO, and PM10 concentrations were made 

at that station, as well as measurements 

of basic meteorological elements. 

The calculations made from 

measurement data as averages in onehour 

intervals concerned the period 

Oct. 2003 – Sept. 2005 that was divided 

into winter half-years (the period from 

October to March, covering heating 

FIGURE 1. The Air Pollution Monitoring Stations in Ursynów 

Characteristics of the particulate matter PM10 concentration... 57 

seasons) and summer half-years (the 

period from April to September) . 

The MzWarszUrsynów measuring 

station under discussion is representative 

for the general urban ambience and it 

characterises well imission of pollution 

dust in the area of living quarters exposed 

to the impact of traffi c and communal 

and industrial emissions. Figure 1 

illustrates the location of the station. 

The concentration of PM10 particulate 

matter is measured there in continuously 

with MLU TEOM1400 analyser using 

the oscillation microweight method. The 

principle of the TEOM function consists 

in measuring the frequency of vibration


of a conical element on which the fi lter 

is fi xed. The increase of the fi lter mass 

due to deposition of dust particles on it 

causes the change in the frequency of 

measured vibration. This method ensures 

the detectability threshold below 0.06 

[μg·m –3 ] for mean one-hour values at 

the throughfl ow 3 [l⋅min –1 ]. The device 

has got the certifi cate of conformity with 

the PM10 measurement relay method 

according to the European standard 

(EN12341). 

The following calculations were the 

basis for characterising the concentration 

fi eld of PM10 particulate matter: 

1) mean concentration of dust in 

calculation periods, 

2) number of exceedences of the 

allowable concentration, 

3) variability ranges of concentration 

values in individual calculation 

periods, 

4) concentration histograms based 

on the Air Quality Scale rates 

recommended by WIOS, 

5) percentiles 25, 50, 70, and 98 of 

concentration values, 

6) mean annual 24-hour courses of dust 

concentration values. 

The data registered at 

MzWarszUrsynów and MzWarszSGGW 

stations were used for evaluation of 

meteorological conditions of pollution 

dispersal. Both those stations laid 800 

m apart from each other (correlation 

coeffi cient between temperatures 

measured on both stations is 0,99) 

are equipped in the automatic outfi t 

registering basic meteorological elements 

in 10-minute periods. Measurements 

of air temperature, wind speed, and air 

humidity at several levels are made at 

the MzWarszSGGW station. The air 

temperature values from heights 5 cm and 

22 m were used to calculate gradients. 

The analysis of meteorological 

conditions of pollution dispersal and 

spreading out covered the following 

points: 

1. Setting mean monthly values of basic 

meteorological elements (Table 1), 

2. Comparing the course of air 

temperature in calculation periods 

with the mean long term course, 

3. Making circular graphs for 

percentiles of particulate matter 

concentration as related to air infl ow 

directions, 

4. Making circular graphs for 

percentiles of pollution plume rate in 

order to get additional information 

on location of emission sources, 

5. 

Calculating synthetic WZ 

indicator of pollution for assessing 

meteorological conditions of 

pollution dispersal. 

The vicinity of the station was divided 

into equal sectors with the central α 

angle constituting 1/16 part of the full 

angle in order to prepare circular graphs 

for percentiles of particulate matter 

concentration. The percentiles of the 

order p = 25, 50, 70, 98 were calculated 

basing on the sets of dust concentration 

values occurring at the air infl ow from 

the direction related to the given sector 

α. Circular graphs were made for 

percentiles calculated for all sectors 

of wind directions; they are graphical 

presentations of concentration value 

distributions at various wind directions. 

Pollution fl ows were analysed for to 

get more information on directions of 

dust infl ows over the area under study 

(Klis, Matejczyk 2002). The pollution 

fl ow is a vector. The scalar size of this

TABLE 1. The selected statistical characteristics of PM10 concentration and meteorological elements from October 2003 to September 2005, station 

MzWarszUrsynów 

The 98th percentile [μg·m -3 ] 








Standard deviation [μg·m -3 ] 

Mean concentration PM10 [μg·m -3 ] 

Date of max. 

Max. S 24 [μg·m -3 ] 

Min. S 24 [μg·m -3 ] 

Precipitation [mm] 

Mean air humidity [%] 

Mean wind speed [m·s -1 ] 

Mean air temperature [ 0 C] 

Mean monthly cocentration of 

PM10 [μg·m -3 ] 

Month 

Calendar year 

Percentile of 1-hour 

Percentile of 24-hour 

The number of exceedance of D 24 

The number of concentrationS 24 

Winter, summer half-year period 

18,2 32,5 52,0 114,8 

19,5 29,1 42,4 93,9 

X 34,7 5,5 1,5 84,3 67,2 6 

XI 41,9 5,1 1,4 90,5 20,7 7 

XII 37,1 1,0 1,9 88,1 41,7 7 

181 46 5,2 156,4 29.01 38,3 22,7 21,2 35,6 50,5 97,7 

I 42,5 -4,9 1,7 88,5 30,9 10 

II 33,3 0,0 1,8 83,4 34,6 7 

III 40,5 3,7 1,6 78,4 40,9 9 

IV 42,0 8,9 1,2 63,7 52,2 10 

V 27,5 12,2 1,1 69,8 54,9 - 

VI 28,7 16,0 1,0 66,2 42,3 - 

182 18 9,9 75,2 16.04 33,3 12,3 24,0 31,2 41,7 62,0 

VII 28,5 18,1 1,1 70,3 61,3 1 

VIII 34,9 19,5 1,2 67,6 53,7 - 

IX 38,7 14,0 1,3 71,0 20,1 7 

X 38,4 10,2 1,5 81,5 39,7 7 

XI 29,1 3,4 1,7 89,1 62,1 2 

XII 41,2 1,7 1,6 90,5 11 7 

181 33 7,7 119,2 11.12 36,3 20,7 21,8 31,5 46,5 100,5 

I 29,7 0,8 1,9 84,5 33,1 4 

II 43,5 -3,1 1,7 84,8 27,2 7 

III 36,3 0,2 1,6 76,4 40 6 

IV 46,7 9,5 1,3 59,5 18,8 12 

V 28,1 14,0 1,2 66,6 56 4 

VI 28,5 16,4 1,1 64,6 45,3 - 

180 35 9,0 

98,0 03.04 35,7 17,2 23,3 30,6 44,5 79,9 

VII 29,0 20,8 1,0 58,8 4,7 - 

VIII 34,6 18,2 1,0 62,9 38,8 5 

IX 47,3 16,2 0,8 62,9 33,6 14 

winter half-year 

period 

2003 

summer half-year 

period 

2004 

18,9 29,8 46,7 116,2 

winter half-year 

period 

19,0 29,5 46,4 102,5 

summer half-year 

period 

2005


vector is equal to the amount of pollution 

fl owing in a time unit through the area 

unit perpendicular to the fl ow direction. 

This is the fl ow intensity [μg ·m –2 ·s –1 ], 

being the measure of the infl ow or outfl ow 

of a substance over an area unit situated 

within a territory (Stull 1995). Circular 

graphs were made for percentiles of the 

order p = 25, 50, 70, 98 after calculating 

the pollution fl ow intensities. 

A synthetic WZ indicator of pollution 

was calculated for discriminating 

between the impact of the emission and 

the infl uence of atmospheric conditions 

on the size of the imission registered. 

This indicator encompasses all impacts 

of various meteorological elements 

infl uencing the increase or decrease of 

air pollution during winter (Walczewski 

1997). It is expressed in the formula: 

WZ = [A(T) + B(v) + C(IN) + D(d)] · E(p) 

where: 

A(T) – component describing the 

temperature impact, 

B(v) – component describing the wind 

speed impact, 

C(IN) – component describing the 

inversion layer impact, 

D(d) – component describing the 

precipitation impact, 

E(p) – coeffi cient taking into account the 

atmospheric air pressure. 

The individual articles for the WZ 

indicator as adopted in the paper were 

defi ned in the following way: 

A(T) = 0 for T > 0 [°C], 1 for –5 < T ≤ 

[0°C], 2 for –10 < T ≤ –5 [°C], 3 for –15 

< T ≤ –10 [°C], 4 for T < –15 [°C], where 

T is the mean 24-hour temperature of air, 

B(v) = 0 for v > 2[m·s –1 ], 2 for 1< v ≤2 

[m·s –1 ], 3 for 0 < v ≤1 0[m·s –1 ], where 

v is the wind speed at the height of 22 

m above the ground level at the time 12 

UTC (10-minute mean), C(IN) = 0 when 

there is a lack of lower inversion layer at 

12 UTC, and 2 when the layer is present. 

The data necessary for verifying the 

existence of the inversion layer were got 

from measurements of air temperature 

at two levels on the mast installed in the 

MzWarszSGGW station. 

D(d) = 0 for d ≥ 4 [mm], 1 for 0.8 < d ≤4 

[mm], 2 for 0 < d ≤ 0.8[mm], 

E(p) = 1.5 when there is the anticyclone 

on a given day, and 1 in the rest of 

cases. 

The minimum WZ value may amount 

to 0, and the maximum is 16.5. The WZ 

index was used with success for the fi rst 

time by Walczewski (1997) in Kraków 

for identifi cation of the changes in the 

level of air pollution with SO 2. 

RESULTS 

The series of 1-hour concentration 

measurements of PM10 particulate matter 

gained in the period under study, i.e. 

1.09.2003–30.10.2005 contained 94% 

of correct results with variability range 

from 0.8 to 241.8 [μg·m –3 ]. One can 

notice the occurrence of a clear cycle of 

daily dust concentration variability, with 

two local peaks in morning and evening 

time – around 9h00 and after21h00, and 

the hours of their occurrence vary during 

the year, on the fi gure 2 presenting the 

mean hourly concentrations of PM10 

particulate within 24 hours. In the winter 

season the concentration maximums 

occur later in the morning and earlier 

in the evening as related to the summer 

season. A clear drop of PM10 particulate 

matter concentration is observed during

[μg•m -³ ] 

60 

50 

40 

30 

20 

10 

0 

00:00 

01:00 

02:00 

the day around 13 h 00 and in the night 

around 5 h 00. The occurrence of these 

relationships results from not only the 

daily course of pollution emission to the 

atmosphere (e.g. traffi c peaks) but also 

24 hour cycle of changes in the height 

of mixing layer and further on from 

the development and disappearing of 

convection processes, and general and 

local circulation. 

Daily concentrations of PM10 

particulate matter in the period under 

study (Tab. 1) acquired the values from 

the interval 5.2 – 156.4 [μg·m –3 ], that 

is from 10.4 to 312.8% respectively 

of mean daily allowable value (D 24 

= 50 [μg·m –3 ]) (Order by Minister of 

Environment of 6 June 2002). This value 

has been exceeded in 132 cases within 

the all measuring period, including 60% 

exceedences in heating seasons. 

The mean daily concentrations of 

dust acquired most frequently the values 

from 25 to 50 [μg·m –3 ] as it is shown 

in the frequency histogram of daily 

concentration occurrence (Fig. 3), based 

on intervals of the Air Quality Scale 

(WIOS) in both winter and summer 

seasons. The highest concentration 

03:00 

04:00 


w arm season cold season 

05:00 

06:00 

07:00 

08:00 

09:00 

10:00 

11:00 

FIGURE 2. Mean 1-hour concentrations of PM10 in the sequence of 24-hours in the station Mz- 

WarszUrsynów, 2004 

12:00 

13:00 

14:00 

15:00 

16:00 

17:00 

18:00 

19:00 

20:00 

values, exceeding D 24, occurred most 

frequently in winter months, this being 

probably linked with the dropping air 

temperature and with more intensive work 

of heating facilities. In summer seasons, 

the values from the intervals 50–75 and 

75–100 [μg·m –3 ] appeared as well, but the 

frequency of their occurrence was lesser 

than in winter seasons, but the values 

over 100 [μg·m –3 ] did not appear at all. 

In the warm season of year, exceedences 

of mean daily allowable value occurred 

most frequently in April and September 

(Tab. 1), but in the months V–VIII they 

appeared only sporadically. No one case 

of exceedence of the allowable standard 

occurred in the periods: V, VI, VIII of 

2004, and VII–VIII of 2005. 

The phenomenon of the occurrence 

of a considerable number of exceedences 

in April and September requires a deeper 

analysis. In the years under study, the 

reasons of such a state can be seen in low 

mean speed of wind, low mean humidity 

of air, low sums of precipitation, and 

low diversity in the infl ows of air masses 

(source: Daily Synoptic Bulletin of the 

IMGW 2003–2005). 

21:00 

22:00 

23:00


results [%] 

60 

50 

40 

30 

20 

10 

0 

35,1 

29,3% 

56,1% 

43,1% 

The mean monthly values of PM10 

concentration were from 27.5 [μg·m –3 ] 

(in May 2004) to 47.4 [μg·m –3 ] (in 

September 2005), while the values 

averaged for winter and summer year 

halves differed only very slightly from 

each other, and they were 38.3 [μg·m –3 ] 

and 36.3 [μg·m –3 ] in winter year halves 

and a bit less – 33.3 [μg·m –3 ] and 

35.7 [μg·m –3 ] in summer year halves 

respectively. 

Circular graphs of concentration 

percentiles related to individual sectors 

of wind rose were made and analysed 

in order to identify the directions of 

infl ows of particulate matter (Fig. 4a). 

This fi gure shows the difference in the 

distribution of percentiles in summer 

and winter seasons over individual 

sectors. They were more even in summer 

seasons, and this fact could evidence the 

impact of pollution sources located at 

various directions around the measuring 

point. The emission from traffi c sources 

prevailed in those periods because a 

thick network of streets with a high 

warm half-year cold half-year 

13,0% 

16,6% 

0–25 25–50 50–75 75–100 >100 

[μg•m 

–3 ] 

FIGURE 3. Frequency distributions in percents of 24-hours particulate matter concentration at Mz- 

WarszUrsynów, X 2003–IX 2005 

1,7% 

3,3% 

1,9% 

intensity of car traffi c occurred around 

the measuring point. 

In the summer 2004, when no 

overstandard values were observed, no 

dominant direction of pollution infl ow 

was seen. It results from the analysis 

of pollution plumes that the sources of 

emission intensities being the greatest in 

the period mentioned existed from the 

measuring station toward the South in 

the sectors 67–270 o N (Fig. 4b). 

A bit otherwise was in the summer 

2005. It is seen on Fig. 4a that the 

highest concentrations occurred at the air 

infl ow from NE direction (the value of 

percentile of 98% level in the NE sector 

was by 20% higher than in the remaining 

sectors), and this can suppose an impact 

of a large point source of emission. The 

analysis of pollution plumes has shown 

that the intensity of the emission was 

the greatest in the sectors 110–180 o N, 

and moreover in the sectors NE, N, 

WNW, where the following industrial 

plants were located: EC Siekierki 

Heating plant, Urban Construction

a) 

WNW 

W 

WSW 

WNW 

W 

WSW 

WNW 

W 

WSW 

WNW 

W 

WSW 

N 

NNW 

200 

NNE 

[μg•m 

NW 

150 

NE 

-3 Perecentiles concentrations of PM10 - 2003/2004 

] 

SW 

SSW 

100 

50 

0 

98% 

100 

50% 

70% 

50 

S 

0 

50% 

70% 


SSE 

SE 

ENE 

E 

ESE 

N 

NNW 

200 

[μg•m 

NNE 

NW 

150 

NE 

-3 Perecentiles concentrations of PM10 - 2004/2005 

] 

SW 

SSW 

98% 

S 

SSE 

SE 

Perecentiles concentrations of PM10 - 2004 

NW 

SW 

NNW 

SSW 

N 

200 

150 

100 

50 

0 

50% 

70% 

98% 

S 

[μg•m -3 ] 

NNE 

SSE 

NE 

SE 

Perecentiles concentrations of PM10 - 2005 

NW 

SW 

NNW 

SSW 

N 

200 

150 

100 

50 

0 

50% 

70% 

98% 

S 

[μg•m -3 ] 

NNE 

SSE 

NE 

SE 

ENE 

E 

ESE 

ENE 

E 

ESE 

ENE 

E 

ESE 

b) 

WNW 

W 

WSW 

N 

NNW 

250 

200 

NNE 

[μg•m 

NW 

NE 

-2 •s -1 Percentiles of PM10 pollution plume rate - 2003/2004 

] 

SW 

SSW 

150 

100 

50 

150 

100 

0 

50% 

70% 

50 

0 

98% 

S 

SSE 

SE 

N 

250 

NNW 

NNE 

[μg•m 

200 

NW 

NE 

-2 •s -1 Percentiles of PM10 pollution plume rate - 2004/2005 

] 

WNW 

W 

WSW 

WNW 

W 

WSW 

WNW 

W 

WSW 

SW 

SSW 

150 

100 

50 

0 

50% 

70% 

98% 

S 

SSE 

SE 

N 

250 

NNW 

NNE 

[μg•m 

200 

NW 

NE 

-2 •s -1 Percentiles of PM10 pollution plume rate - 2004 

] 

SW 

98% 

50% 

70% 

SE 

SSW 

SSE 

S 

Percentiles of PM10 pollution plume rate - 2005 

FIGURE 4. Circular graphs of percentiles concentrations of PM10 (a) and percentiles of PM10 pollution 

plume rate (b) in the cold and warm half-years from October 2003 to September 2005 at Mz- 

WarszUrsynów station 

NW 

SW 

NNW 

SSW 

N 

250 

200 

150 

100 

50 

0 

50% 

70% 

98% 

S 

[μg•m -2 •s -1 ] 

NNE 

SSE 

NE 

SE 

ENE 

E 

ESE 

ENE 

E 

ESE 

ENE 

E 

ESE 

ENE 

E 

ESE


Enterprise, Radio Ceramics Plant, and 

the Construction Woodworking Plant. A 

part of the pollution could also fl ow from 

the city centre. 

In the winter season 2003–2004 

the distribution of concentrations in 

individual sectors of the wind rose was 

almost evenly, with a slight overweight 

(under 10%) of concentration values 

occurring in the sectors 180–203 o N and 

338–360 o N. This shows an impact of 

some sources on the imission fi eld, the 

sources that were dispersed in various 

directions around the measuring station. 

It results from the analysis of pollution 

plumes that the greatest intensity of 

emissions came from the sources located 

in the sector 67–160 o N and in the sector 

315–360 o N where point emission 

sources existed as it was mentioned in 

the preceding paragraph. 

In the season 2004–2005 the highest 

concentrations occurred at the air infl ow 

from northern direction (Fig. 4a). The 

analysis of the plumes, similarly as 

for the preceding season, revealed the 

most intensive emission from the sector 

67–180 o N with local heating facilities 

and areas with low communal and 

living emissions (Fig. 1). Moreover, in 

that season one could note an infl ow of 

greater amounts of dust from the sector 

248–290 o N. The diagram of plumes 

showed the existence of sources with 

great concentrations of emission from 

that direction where pollution sources 

worked intensively. 

The calculation of the WZ pollution 

index described in the preceding section 

was the next stage of the analysis of 

meteorological conditions important 

for pollution dispersal. The calculations 

were made for both heating seasons 

mentioned. However, only the results 

from January 2004 and 2005 were put 

down in the paper due to its limited size. 

It is seen on Figure 5 presenting 

the monthly courses of mean daily 

PM10 concentrations compared to the 

analogous course of the WZ index that 

the maximums of concentration values 

cover the maximums of the index values. 

Sometimes a slight retardation of the 

concentration maximum as compared to 

the highest index value can be noted, this 

fact being well grounded because the 

concentration resulted from unfavourable 

meteorological conditions. 

In January 2004 meteorological 

conditions were less favourable for 

pollution spreading out than in January 

2005, because the mean WZ I.2004 

index amounts to 5.3 (max = 12), and 

WZ I.2005 is 3.5 (max 8.2), and the mean 

concentrations were: 42.5 [μg·m –3 ] 

and 29,7 [μg·m –3 ] respectively. A more 

detailed analysis of those winter months 

has shown that high air pressure areas 

shaped the weather in January 2004. 

Either frosty and dry polar continental air 

fl ew from NE and E or arctic air from the 

north. At that time an increased emission 

of pollution from heating sources could 

occur at very low air temperatures. Not 

only the pressure pattern but also a small 

amount of precipitation (7 days with 

precipitation over 0.8 mm) favoured 

increased concentrations of pollution. 

However in January 2005 the mean 

monthly air temperature was higher than 

in 2004 and that pattern could cause a 

bit lesser emission from heating sources. 

The following impacts: occurrence of 

low air pressure patterns and the fronts 

accompanying them, the infl ow of polar 

marine air, the increase of wind speed,

PM10 [μg·m -3 ] 

PM10 [μg·m -3 ] 

180 

160 

140 

120 

100 

80 

60 

40 

20 

0 

and the frequency of precipitation 

had contributed to better pollution 

spreading out and they could cause a 

drop in registered values of pollution 

concentration. 


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 

January 2004 D24 PM10 WZ 

- 24-hour permissible concentration 

70 

60 

50 

40 

30 

20 

10 

0 

D24 

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 

January 2005 

PM10 WZ 

FIGURE 5. Comparison of monthly variabilities of air pollution WZ indicator and PM10 concentration 

(24-hours) at MzWarszUrsynów station 

D24 

STATEMENTS AND 

CONCLUSIONS 

The high values of PM10 particulate 

concentrations occurring in the 

Ursynow living quarter were caused, at 

a considerable rate, by a low emission 

coming from main communication 

14 

12 

10 

8 

6 

4 

2 

0 

14 

12 

10 

8 

6 

4 

2 

0 

WZ indicator 

WZ indicator


avenues and from dispersed sources 

(local heating facilities, individual 

domestic fi re places) occurring in the 

sectors 67–180 o N and 245–360 o N. Low 

sources were of local importance but 

their impact decided on the scale of the 

threat from pollution. This threat results 

just from the fact that the low sources are 

located near human living areas; therefore 

dust aerosol breathed in is a ‘fresh’ one 

with a high share of miniparticles, and 

at the same time it is very dangerous 

for human health. This threat has been 

confi rmed by introductory results of the 

quality analysis of PM10 found at the 

MzWarszSGGW station that showed 

the presence of heavy metals in the dust, 

such as Ni, Pb, Cr, Zn, Fe, Cu (archival 

data of the Meteorology and Climatology 

Department of the Warsaw Agricultural 

University). 

Since the year 2003, i.e. from 

starting the measurements of PM10 

particulate matter concentration at the 

MzWarszUrsynow station, a decreasing 

tendency of this pollution is observed, 

although the number of allowable 

exceedences of mean daily values 

maintains still at a high level (mean daily 

allowable value of the particulate matter 

concentration was exceeded in 132 cases 

in the whole measuring period). 

The differences in seasonal 

concentrations were caused by 

emissions of particulate matter from 

burning out fuels in low sources being 

greater in winter seasons at relatively 

worse conditions of aeration and air 

self-cleaning that occur in the cold 

season. Meteorological conditions were 

the factor deciding on the imission 

registered in all winter months under 

study. It was impossible to found for the 

study period that a pollution emission 

decrease caused by implementation of 

the air quality improvement program 

had been the reason for the betterment of 

the imission situation in the area under 

analysis. According to the opinion of 

the authors, the decrease of particulate 

matter amount registered in the period 

2003–2005 was caused by the betterment 

of meteorological conditions. 

REFERENCES 

DOCKERY D.W., CUNNINGHAM J., 

DAMOKOSH A.I., NEAS L.M., 

SPENGLER J.D., KOUTRAKIS P., 

WARE J.H., RAIZENNE M, SPEIZER 

F.E. 1996: Health effects of acid aerosols 

on North American children: respiratory 

symptoms. Environ. Health Perspect. 

104, 500–505. 

EN12341 – Determination of the PM10 

fraction of suspended particulate matter – 

Reference method and fi eld test procedure 

to demonstrate reference equivalence of 

measurement methods. November 1998. 

JABŁOŃSKA M. 2003: Skład fazowy pyłów 

atmosferycznych w wybranych miejscowościach 

Górnośląskiego Okręgu Przemysłowego. 

Wydawnictwo Uniwersytetu 

Śląskiego. Katowice. Juda J., Chróściel 

S. 1974: Ochrona powietrza atmosferycznego. 

NIEĆKO J., NIEĆKO M., NIEĆKO U. 1998: 

Charakterystyka pyłu zawieszonego 

i jego wpływ na organizm ludzki. Air 

protection in theory and application. 

Section I, s. 29-49. 

KLIŚ CZ., MATEJCZYK M. 2002: Ocena 

wpływu źródeł na jakość powietrza 

w świetle ustawy „Prawo ochrony 

środowiska”. Ochrona Powietrza i Problemy 

Odpadów, vol. 36, nr 3: 95–98. 

Rozporządzenie Nr 62 Wojewody Mazowieckiego 

z dnia 8 grudnia 2003.

Rozporządzenie Ministra Środowiska z dnia 

6 czerwca 2002 r. w sprawie dopuszczalnych 

poziomów niektórych substancji 

w powietrzu oraz marginesów Tolerancji 

dla dopuszczalnych poziomów niektórych 

substancji (DzU z 2002r. nr 87 poz. 

798). 

STERN A.C. et al. 1994: Fundamentals of air 

pollution. Academic Press, San Diego. 

STULL R.B. 1995: Meteorology Today For 

Scientists and Engineers, West Publishing 

Comp. New York. 

Quarg, 1996: Airborne Particulate Matter in 

the United Kingdom. Third Report of the 

Quality. 

WALCZEWSKI J. et al. 2000: Wykorzystanie 

danych meteorologicznych w monitoringu 

jakości powietrza. Biblioteka 

Monitoringu Środowiska, Warszawa. 

Walczewski J. 1997: Wskaźnik meteorologiczny 

określający prawdopodobieństwo 

wzrostu zanieczyszczenia powietrza 

w okresie zimowym. Wiad. IMiGW, t. 

XX (XLI). 

WARYCH J. 1999: Zanieczyszczenie powietrza 

cząstkami aerozolowymi i wynikające 

stąd problemy. Ochrona powietrza 

i Problemy Odpadów. Vol. 33, nr 3, majczerwiec. 


Streszczenie: Charakterystyka pola imisji oraz 

próba określenia źródeł zanieczyszczenia powietrza 

pyłem zawieszonym PM10 na obszarze 

dzielnicy mieszkaniowej Ursynów na podstawie 

danych pomiarowych z automatycznej stacji 

monitoringu atmosfery. W pracy przedstawiono 

charakterystykę pola imisji pyłu zawieszonego 

PM10 za pomocą podstawowych statystyk, takich 

jak wartości średnie, zakresy zmienności, liczby 

przekroczeń dopuszczalnego stężenia, rozkłady 

częstości występowania stężeń w poszczególnych 

przedziałach, przebiegi czasowe. W celu przeanalizowania 

warunków meteorologicznych dyspersji 

i rozprzestrzeniania się zanieczyszczeń, obliczono 

średnie wartości podstawowych elementów 

meteorologicznych, sporządzono wykresy kołowe 

percentyli stężenia pyłu, oraz percentyli natężenia 

strumieni zanieczyszczenia. Obliczono również 

syntetyczny wskaźnik zanieczyszczenia WZ, ujmujący 

łączne oddziaływanie różnych elementów 

meteorologicznych na poziom stężeń. 

MS. received July 2006 


Wydział Inżynierii i Kształtowania Środowiska, 

SGGW, 

02-787 Warszawa, ul. Nowoursynowska 159




Runoff volume and slope gradient relationship – laboratory 

investigations 

ANNA BARYŁA 

Department of Environmental Improvement, Warsaw Agricultural University – SGGW 

Abstract: Runoff volume and slope gradient 

relationship – laboratory investigations. The 

objective of the research was to analyze infl uence 

of the slope gradient on the volume of surface 

runoff. Runoff is the primary driving variable in 

the water – induced erosion process. The overland 

fl ow process is strongly affected by the slopes 

because of the effective rainfall rate at the surface. 

The laboratory physical model diameters were 

120 cm long, 15 cm wide and 120 cm deep. The 

experiment the measurements of water erosion 

was done for different slopes: 5%, 10%, and 

15%. The rainfall was artifi cial simulated and his 

intensity for experimental separate sets. Based on 

performed research it can be concluded that the 

volume of the runoff depends on the gradient of 

the slope. As the gradient increase the volume of 

the runoff also increase. 

Key words: rainfall simulation, physical model, 

experiment. 

INTRODUCTION 

The runoff of rainfall water on the slope 

depends on several factors such as: 

decline, form, length and plant cover 

of the slope, amount ant intensity of the 

rainfall as well as the water properties of 

the soil (Carson and Kirbky 1972, Słupik 

1981, Morgan 1986, Józefaciuk 1996, 

Kinnell 1997). 

Among physical soil properties it is 

the soil infi ltration ability that infl uences 

the runoff the most. The impact of 

infi ltration on the possibility of surface 

runoff occurrence is revealed particularly 

when the surface lay is hardly permeable. 

The structure and the porosity of the soil 

have a decisive impact on the amount of 

infi ltration (Józefaciuk A. and Józefaciuk 

C. 1996). Basic causes of variation 

of infi ltration speed are changes of 

retentive capacity in the ground, the type 

of the gradient and the primary capacity 

value, treated as moisture in a time given 

(Słupik 1981, Schmidt 1997). 

Theoretical base of infi ltration was 

formed by Horton (1993); according to 

him the surface runoff forms on a slope 

when the intensity of rainfall exceeds the 

infi ltration capacity of the soil. In view 

of gradual fi lling of the soil pores with 

the water, the infi ltration capacity of the 

soil diminishes as the rainfall lasts. When 

the intensity of the rainfall exceeds the 

infi ltration process, a so called superinfi 

ltration runoff occurs. However, 

when the surface ground lay is totally 

saturated with the water, a so called 

saturated zone runoff is formed (Słupik 

1981, Kinnell 1997, Józefaciuk A. and 

Józefaciuk C. 1996, Rejman and Usowicz 

1999). A particular type of it is a runoff 

in the initial phase of intense rainfall 

formed thanks to a dried-off and hardly 

permeable shell or a blockage laying 

on the surface against the infi ltrating 

water which is created by the air bound 

in the soil. This type of runoff is called 

a pre-infi ltration runoff (Schmidt 1997,

70 A. Baryła 

Kinnell 1997). In this case the surface 

runoff lasts until the soil is softened 

and the soil moisture is complemented. 

This type of runoff occurs the most often 

on heavy soils, thus short and of little 

intensity rainfalls very often can entirely 

constitute the runoff (Kinnell 1997). The 

process of forming the surface runoff 

depends in a big measure on, among 

others, the degree of covering of the soil 

surface. The practice has shown that it is 

more intense on open soils. 

FIGURE 1. Schema of laboratory experiment 

In natural conditions because of 

their big differentiation, it is extremely 

diffi cult to estimate precisely the impact 

of each of the factors mentioned above on 

the course of the runoff and its amount. 

The aim of the research was to determine 

the impact of the initial soil moisture, the 

decrease and the intensity of the rainfall 

on the value of surface runoff. The results 

of laboratory experiments are presented 

in the paper.

MATERIALS AND METHODS 

The research was carried out on a ground 

model with the following diameters: 136 

cm of length, 120 cm of altitude, 15 

cm of width (Fig. 1). The inside oh the 

model was fi lled with clay sand fi rmly 

poured layer by layer and alternately 

compacted. 

In order to enable the observation 

of water fl ow in the soil the front wall 

of the model was made of a transparent 

polymethacrylate board. Inside the model 

16 probes TDR were installed in order 

to measure the soil moisture. 0.8 mm 

of diameter each of them, the distance 

between them was 5 mm. They were 

connected to the TDR (Malicki 1996) 

set ant to the computer. All the time 

during the experiment (in ten-minutes 

intervals) changes of soil moisture were 

registered. The rainfall was simulated by 

microsprinklers; the intensity of outfl ow 

depended on the water pressure, which 

enabled modeling of rainfalls of different 

intensity. 

The surface runoff water was collected 

at the end of the model by a trough 

having an outlet directly connected to 

the measuring vessel. The measurements 

of outfl ow and moisture were made in 

the same intervals. 12 experiments were 

carried out on the model with three slopes 

of the terrain surface – 5, 10 and 15%, 

TABLE 1. Physical properties of the soil used in 

experiment 

Depth [m] Density 

[g·cm –3 ] 

Porous 

[%] 

Filtration 

coeffi cient 

[cm·s –1 ] 

0.00–0.15 1.68 34.1 15.2*10–4 

0.15–0.45 1.62 35.6 7.84*10–4 

0.45–0.80 1.58 36.8 4.2*10–4 

Runoff volume and slope gradient relationship... 71 

and different rainfall intensity: 41 mm 

h –1 , 46 mm h –1 , 52 mm h –1 and 54 mm 

h –1 . The duration of every experiment 

was 1 hour. 

Surface runoff in the time from the 

physical model was collected at the end by 

trough, from of which one accompanied 

to dishes measuring. Physical properties 

of the soil are set together in the Table 1. 

Lesser and lesser values of fi ltration 

coeffi cient in subsequent layers may 

evidence about more compacted lower 

parts of the ground in the model. 

RESULTS AND DISCUSSION 

Knowledge and proper determination 

of interdependencies between surface 

runoffs and factors determining their 

value, intensity and frequency of 

occurrence is indispensable in the 

modeling of the value of surface runoff 

on a slope. On the Figure 2a are shown 

the results of surface runoff for the 

slope of 5%. The analyze of the results 

obtained can lead to the conclusion that 

at the rainfall intensity of 41 mm h –1 and 

a slope of 5% the summary runoff was 

0.8 mm h –1 , instead at a slope of 15% the 

summary runoff was 5.6 mm h –1 . When 

the rainfall intensity was increased by 5 

mm h –1 (46 mm h –1 ) at a slope 5% the 

summary surface runoff amounted to 1.2 

mm h –1 at 15 mm h –1 – 11.8 mm h –1 . 

The time of starting-up the surface 

runoff at the intensity of 41 mm h –1 was 

22 minutes and was the longest time of 

starting-up the surface runoff at a slope 

10%. The shortest time was observed at 

the intensity of rainfall of 54 mm h –1 and 

then the runoff started 10 minutes after 

the experiment had started.

72 A. Baryła 

Surface runoff [mm] 



12 

10 

8 

6 

4 

2 

0 

12 

10 

8 

6 

4 

2 

0 

12 

10 

8 

6 

4 

2 

0 

a) 

41 mm/h 

46 mm/h 

52 mm/h 

54 mm/h 

0 10 20 30 40 50 60 

b) 

0 10 20 30 40 50 60 

c) 

5% 

10% 

15% 

0 10 20 30 40 50 60 

Time [min] 

FIGURE 2. Dynamics of runoff surface with rain 

intensity and different slope (a – 5%, b – 10%, 

c – 15%). 

As at the 5 and 10% slopes, the highest 

value of surface runoff during 1 hour was 

obtained at the rainfall intensity of 54 

mm and equaled to 10.86 mm; this value 

was twice higher than surface runoff 

obtained at the intensity of 41 mm (3.68 

mm). At the intensity of rainfall equal 41 

mm h –1 the surface runoff increased after 

10 minutes, and at the intensity of 54 mm 

h –1 – 2 minutes. The research also proved 

that in the initial rainfall phase there is an 

intensifi ed runoff volume until a balance 

is obtained between the amount and the 

energy of runoff. The surface runoff 

speed stabilisation depends mainly on 

the rainfall intensity, which makes a very 

thin layer on the soil surface. For that 

reason the runoff value remains at the 

similar level, although the surface layer 

is almost fully saturated. 

The performed research confi rmed 

that there is an important impact of the 

land slopes on the course and value of 

surface runoffs (Fig. 3). At a slope of 

5% surface runoffs’ value was more 

than fi ve times less than at a slope of 

15% at the same rainfall intensity. The 

research performed by Szafrański and 

others (1998) revealed that on grounds 

at a slope 3–6% surface runoff value is 

low and therefore the erosion danger 

is very slight. A severe water erosion 

danger may occur only on the lands at 

a slope of more than 10%. None the less 

during violent summer high-intensity 

thunderstorms the phenomenon of linear 

erosion can occur even at slopes of less 

than 6% (Rejman 1999). 

Coefficients of runoff C [%] 

30 

20 

10 

0 

slope 5% 

slope 10% 

slope 15% 

C=0,36466*P; C=0,36466*P; R=99 R=99 

C=0,230434*P; C=0,230434*P; R=96 R=96 

C=0,0669003*P; C=0,0669003*P; R=91 R=91 

0 10 20 30 40 50 60 

Rainfall P [mm] 

FIGURE 3. Relationship between rainfall intensity 

and coeffi cients of runoff

The Figure 3 presents a relationship 

between rainfall intensity and coeffi cients 

of runoff for the results of the research. 

The determination coeffi cient amounted 

to respectively: for a slope 5%–99, 

for a slope 10%–96 and for 15% –91. The 

research performed revealed that when 

the rainfall intensity volume increases, 

surface runoff increases too. 

CONCLUSIONS 

1. 

2. 

The research performed on a ground 

model confi rmed an important role 

of land slopes played in a course 

and value of surface runoffs. At a 

slope equal 5% the runoffs’ value 

was almost fi ve times less than at 

slopes equal 15%. It was ascertained 

that there is a rectilinear dependency 

between surface runoffs volume and 

land slopes. 

Surface runoffs volume described 

by the runoff coeffi cient was 0,02 to 

0,05 at a land slope 5% and 0,13 to 

0,21 at a land slope equal 15%. 

REFERENCES 

CARSON M.A., KIRBKY M.J. 1972: 

Hillslope form and process. University 

Press, Cambridge. 

JÓZEFACIUK A., JÓZEFACIUK C. 1996: 

Mechanizm i wskazówki metodyczne 

badania procesów erozji. Biblioteka 

Monitoringu Środowiska w Warszawie. 

KINNELL P.I.A. 1997: Runoff ratio as 

a factor in the empirical modeling of 

soil erosion by individual rainstorms. 

Australian Journal of Soil Research No 

25, p 1–6. 

MALICKI M.A., PLAGGE R., ROTH 

C.H. 1996: Improving the calibration of 

dielectric TDR soil moisture determination 

Runoff volume and slope gradient relationship... 73 

taking into account the solid soil. European 

J. Soil Sci. 47, 57–366. 

MORGAN R.P.C. 1986: Soil erosion and 

conservation. Longman Scientifi c & 

Technical, Essex, UK. 

REJMAN J., USOWICZ B. 1999: Quantitative 

description of water ad soil transport in 

process of water erosion. Acta Agrophysica, 

no 23, p. 143–148 (in Polish). 

SCHMID B. 1997: Critical rainfall for overland 

fl ow from an infi ltrating plane surface. 

Journal of Hydrology 193, 45–60. 

SŁUPIK J. 1981: Rola stoku w kształtowaniu 

odpływu w Karpatach Fliszowych. Prace 

Geografi czne, 142, 1–94. 

SZAFRAŃSKI C., FIEDLER M., STASIK 

R. 1998: Rola zabiegów melioracyjnych 

w ochronie przeciwerozyjnej gleb 

terenów bogato urzeźbionych Bibliotheca 

Fragmenta Agronomica tom 4B/98. 

WISCHMEIER W.H., SMITH D.D. 1978: 

Predicting rainfall erosion losses. USDA 

Agric. Handb. 537. U.S. Gov. Print. 

Offi ce, Washington DC, p. 1–58. 

VAN DIJK A.I.J.M., BRIJNZEEL L.A. 

Modeling runoff and soil loss from bench 

terraced hillslopes in the volcanic uplands 

of West Java, Indonesia. 

YU B., SOMBATPANIT S., ROSE C.W., 

CIESIOLKA, C.A.A. and COUGHLAN 

K.J. 2000: Characteristics and modeling 

of Runoff Hydrographs for Different 

Tillage Treatments. Soil Science Society of 

American Journal No 64, p. 1763–1770. 

Streszczenie: Zależność pomiędzy spływem powierzchniowym 

a spadkiem terenu – doświadczenie 

laboratoryjne. Spływ powierzchniowy jest 

procesem złożonym, a czynniki od których zależy 

są zmienne zarówno w czasie, jak i w przestrzeni. 

Poznanie i właściwe określenie zależności pomiędzy 

spływami powierzchniowymi a czynnikami 

wpływającymi na ich wielkość, natężenie i częstotliwość 

występowania jest niezbędne m.in. dla 

oceny istniejących modeli spływu powierzchniowego. 

Badania laboratoryjne wykonano na modelu 

gruntowym o wymiarach: długość 136 cm, wysokość 

120 cm, szerokość 15 cm. Wnętrze modelu 

wypełnione zostało piaskiem gliniastym mocnym

74 A. Baryła 

zagęszczanym warstwowo. Badania dowiodły, że 

w początkowej fazie opadu następuje zwiększenie 

wielkości spływu do momentu ustalenia się stanu 

równowagi pomiędzy intensywnością i energią 

opadu a wielkością i energią spływu. Prędkość 

ustabilizowania się spływu powierzchniowego 

zależy głównie od intensywności opadu, który 

powoduje tworzenie się na powierzchni gleby 

cienkiej warstewki. Wartość odpływu utrzymuje 

się przez to na zbliżonym poziomie, mimo że 

wierzchnia warstwa osiągnęła prawie pełny poziom 

nasycenia. 


Author’s address: 

Katedra Kształtowania Środowiska – SGGW 

ul. Nowoursynowska 166 

02-787 Warszawa 

e-mail: anna_baryla@sggw.pl




Infl uence of sprinkling irrigation and nitrogen fertilization on health 

status of potato grown on a sandy soil 

DARIUSZ PAŃKA1 , ROMAN ROLBIECKI2 , CZESŁAW RZEKANOWSKI2 1Department of Phytopathology, University of Technology and Agriculture, Bydgoszcz, Poland 

2Department of Land Reclamation and Agrometeorology, University of Technology and Agriculture, 

Bydgoszcz, Poland 

Abstract: Infl uence of sprinkling irrigation and 

nitrogen fertilization on health status of potato 

grown on a sandy soil. The fi eld experiment was 

carried out in the years 2001–2003 at Kruszyn 

Krajeński near Bydgoszcz on a sandy soil 

(Typic Hapludolls). The fi eld experiment was 

done using the split-plot method, in a dependent 

system with two variable factors (sprinkler 

irrigation and nitrogen fertilization) and three 

replications. ‘Drop’ early potato cultivar was 

taken into consideration. It is the most frequent 

cultivated cultivar in Kujawy-Pomerania region. 

The following levels of experimental factors were 

used: water: O – without sprinkler irrigation and 

W – sprinkler irrigation according to tensiometer, 

nitrogen fertilization: N1 – 75 and N2 – 125 kg 

N ha –1 , respectively. Potato tubers were analyzed 

at once after harvest from the point of view of 

occurrence of common scab, black scurf, potatotuber 

dry-rot and potato-tuber soft-rot symptoms. 

Estimation of degree of infestation was carried 

out on 50 tubers taken at random from the each 

plot. Nine-degree scale was used (0–8°), where: 

0° means lack of infestation symptoms (sound 

tubers) and 8° means over 50% of a tuber surface 

with symptoms of common scab or black scurf. 

In case of other diseases a percentage of tubers 

with symptoms of infestation was estimated. On 

the basis of the investigation results it was found 

that sprinkler irrigation had a signifi cant infl uence 

on the occurrence of common scab symptoms. 

Stronger infestation was noted on irrigated 

treatments. A strong infl uence of the rainfall 

course in the vegetation period of the potato on 

the infestation of tubers by Streptomyces sp. was 

observed. An infl uence of nitrogen fertilization on 

the occurrence of common scab and black scurf 

symptoms was not found. The cultivar tested was 

characterized by a higher resistance to infestation 

by Rhizoctonia solani than Streptomyces sp. 

Key words: potato, sprinkling irrigation, fertilization, 

nitrogen, Streptomyces sp., Rhizoctonia solani, 

common scab, black scurf, sandy soil. 

INTRODUCTION 

High quality yield is at present the main 

aim of the table potato production. 

The quality of potato is affected most 

strongly by such the factors like cultivar, 

course of weather conditions during the 

vegetation period, mineral fertilization, 

soil cultivation and plant protection 

against agrophages [3], [7], [9], [16]. 

Good quality yield can be produced when 

optimal conditions for the potato growth 

and development are created. In such the 

conditions the pressure of agrophages is 

limited to minimum. 

Pathogens developing on tubers are 

included to the most dangerous pathogens. 

They often decreased the suitability of 

potato for food processing. In addition 

these pathogens decreased the quality 

of seed-potatoes and they can cause 

considerable storage losses [1], [6], [15]. 

Common scab, rhizoctoniose, tuber dryrot 

and tuber soft rot are included to the 

most dangerous diseases in cultivation 

of potato. They occur commonly. Their

76 D. Pańka, R. Rolbiecki, Cz. Rzekanowski 

development are infl uenced by a number 

of such the factors like for example 

moisture, temperature or mineral 

fertilization [8], [10], [11], [14]. In 

Poland cultivation of potato is conducted 

mostly on light soils characterized by a 

low fertility and limited water-holding 

capacity. Because of this, higher doses 

of mineral fertilization and irrigation 

of potato plantations should be used, 

especially in the regions characterized 

by periodical rainfall defi cits. All of 

this with connection of other factors 

can infl uence the development of potato 

pathogens. 

The aim of the study was to determine 

the infl uence of nitrogen fertilization and 

sprinkler irrigation on the potato tuber 

health of ‘Drop’ cultivar, grown on the 

light soil. 

MATERIAL AND METHODS 

The fi eld experiment was carried out 

in the years 2001–2003 at Kruszyn 

Krajeński near Bydgoszcz on a sandy soil 

(Typic Hapludolls). The water reserve to 

1 m depth of soil at fi eld capacity was 

87 mm and the available water 67 mm. 

The fi eld experiment was done using the 

split-plot method, in a dependent system 

with two variable factors (sprinkler 

irrigation and nitrogen fertilization) and 

three replications. ‘Drop’ early potato 

cultivar was taken into consideration. It is 

the most frequent cultivated in Kujawy- 

-Pomerania region. 

The following levels of experimental 

factors were used: 

– water: O – without sprinkler 

irrigation and W – sprinkler irrigation 

according to tensiometer, 

nitrogen fertilization: N1 – 75 and 

N2 – 125 kg N ha –1 , respectively. 

The plot area for harvest was 15 m 2 . 

The mineral fertilization was applied 

at following doses: 80 kg P2O5 ha –1 

(superphosphate) and 140 kg K2O ha –1 

– 

(potash salt). Fertilization was used presowing 

in early spring. 

Chemical protection against potato 

late blight, colorado potato beetle was 

conducted in particular years of the 

study. Potato tubers were analyzed at 

once after harvest from the point of view 

of occurrence of common scab, black 

scurf, potato late blight, potato-tuber dryrot 

and potato-tuber soft-rot symptoms. 

Estimation of degree of infestation was 

carried out on 50 tubers taken at random 

from the each plot. 9-degree scale was 

used (0–8°), where: 0° means lack of 

infestation symptoms (sound tubers) and 

8° means over 50% of a tuber surface 

with symptoms of common scab or 

black scurf. In case of other diseases 

a percentage of tubers with symptoms 

of infestation was estimated. Infection 

degrees were transformed into infection 

indexes (II) according to Townsend and 

Heuberger formula [17]. Obtained data 

were statistically analyzed using analysis 

of variance. Mean values were verifi ed 

with Tukey’s test. 


In all of the study years, symptoms 

of infestation by Streptomyces sp. and 

Rhizoctonia solani, occurred on potato 

tubers most often. Other pathogens were 

noted sporadically and they amounted 

less than 1% tubers with disease 

symptoms.

Signifi cant infl uence of sprinkler 

irrigation on the occurrence of common 

scab symptoms on tubers was observed 

(Tab. 1). Tubers from irrigated plots 

were signifi cantly higher infected 

by Streptomyces sp. every year. The 

strongest symptoms of infestation 

occurred in 2001, on plots fertilized with 

the higher nitrogen dose. The lowest 

infestation by this pathogen occurred in 

exceptional dry year 2003 (Tab. 2). In this 

year the infection index on irrigated plots 

was lower than 24,5%. Similar results 

obtained Sadowski et al. [10], [11]. 

They observed a decrease of common 

scab symptoms occurrence on the very 

light soils, in years characterized by 

small amounts of rainfall. In such the 

Infl uence of sprinkling irrigation and nitrogen... 77 

conditions, irrigation was favourable 

to the development of Streptomyces sp. 

An infestation increase of tubers by this 

pathogen on irrigated plots, in years with 

lower rainfall, noted also Gładysiak and 

Czajka [5] as well as Borówczak and 

Gładysiak [2]. 

In years with higher rainfall amounts 

they observed a decrease of the common 

scab occurrence. The fi rst two years 

of the study (2001 and 2002) were 

characterized by higher rainfall amounts, 

especially year 2001 (Tab. 2). In spite of 

this, the infection increase of tubers by 

Streptomyces sp. on irrigated treatments 

was obtained. Reason of this phenomenon 

can be a low rainfall in June in these 

years. Amounts of rainfall in this month 

TABLE 1. Effect of nitrogen fertilization and irrigation on cv. ‘Drop’ tubers infection with Streptomyces 

sp. (Infection index in %). Kruszyn Krajeński 2001–2003 

Years Water variant 

Nitrogen fertilization 

75 kg·ha 

Mean 

–1 125 kg·ha –1 

Non irrigated 

18.7 a1 

a 

18.5 a 

a 18.6 a 

2001 

Sprinkler irrigation 

52.8 a 

b 

53.3 a 

b 53.1 b 

Mean 35.8 a 35.9 a 

Non irrigated 

15.7 a 

a 

8.7 a 

a 12.2 a 

2002 


43.4 a 

b 

50.1 a 

b 46.8 b 

Mean 29.6 a 29.4 a 

Non irrigated 

1.4 a 

a 

2.5 a 

a 1.9 a 

2003 


21.3 a 

b 

24.5 a 

b 22.9 b 

Mean 11.3 a 13.5 a 

Non irrigated 

11.9 a 

a 

9.9 a 

a 10.9 a 

2001–2003 


39.2 a 

b 

42.6 a 

b 40.9 b 

Mean 25.6 a 26.3 a 

1Mean values followed by the same letter in columns and rows are not signifi cantly different at α = 0.05, 

according to Tukey’s test.


TABLE 2. Rainfall conditions recorded over vegetation period 

Period 

April May 

Rainfall (mm) 

June July August Total 

1951–2000 26 40 56 70 48 240 

2001 45 30 49 106 27 257 

2002 13 50 44 108 41 256 

2003 18 18 30 106 18 190 

2001–2003 25 33 41 107 29 234 

were, like in 2003, lower than the manyyear 

average. According to Sawicka [13] 

potato is the most susceptible to infection 

in the period from the beginning of tuber 

formation to the stage when diameters of 

tubers achieve 1,5–2 cm. Small amounts 

of rainfalls in this period are favourable to 

the development of common scab. ‘Drop’ 

is the very early cultivar, characterized 

by a short vegetation period, and 

because of this its largest susceptibility 

to infection occurred in the years, 

exactly in the period with low rainfall. 

Sprinkler irrigation was then favourable 

to the pathogen development. According 

to Szutkowska [14], the occurrence of 

common scab is mostly affected by the 

soil moisture and the soil temperature in 

the period of the early growth of tubers. 

The highest infection was observed when 

the soil temperature was about 21°C, 

and its moisture around 6%. Favourable 

arrangement of these factors in the years 

2001 and 2002 in the remaining months 

of vegetation was probably the reason 

of the stronger infection of tubers on 

irrigated treatments as compared to the 

year 2003. 

Signifi cant infl uence of the nitrogen 

fertilization dose on the common scab 

occurrence on tubers was not observed. 

Czajka et al. [4] observed an increase 

of common scab infection between 

treatments with doses 0 and 80 kg N ha –1 . 

Doses which were increased above 80 kg 

N ha –1 had no larger effect. In the own 

experiments the occurrence of common 

scab was also on a similar level in case 

of both the doses tested. 

Signifi cant infl uence of sprinkling 

irrigation on infection of tubers by 

Rhizoctonia solani was observed 

only in 2003 (Tab. 3). This year was 

characterized by the lowest amount of 

rainfall during the vegetation period 

of potato. Sprinkler irrigation in 

such the condition was favourable to 

the development of the pathogen as 

compared to control treatments (without 

irrigation). According to Weber [16], 

the high moisture of soil is favourable 

to sclerotia formation on tubers. This 

dependence was observed also in our 

earlier investigations [8]. Similar results 

were obtained also by Sadowski et al. 

[12]. In the remaining years of the study, 

with higher rainfall during the period of 

the potato growth, no differences in black 

scurf occurrence between irrigation and 

control treatments were observed. 

A signifi cant effect of the nitrogen 

fertilization dose on black scurf of

potatoes was not observed, too. In 

investigation of Czajka et al. [4] the 

signifi cantly highest level of black scurf 

was noted on treatment with 80 kg 

N ha –1 , and the signifi cantly lowest level 

of black scurf of potatoes – on treatments 

with 0 and 160 kg N ha –1 , in the years 

with average amount of rainfall. Higher 

nitrogen fertilization decreased the 

pathogen occurrence in warm and dry 

years. 

CONCLUSIONS 

On the basis of the investigation results 

it was found that sprinkler irrigation had 

a signifi cant infl uence on the occurrence 

of common scab symptoms. Stronger 

infestation was noted on irrigated 

treatments. 


TABLE 3. Effect of nitrogen fertilization and irrigation on cv. ‘Drop’ tubers infection with Rhizoctonia 

solani (Infection index in %). Kruszyn Krajeński 2001–2003 

Years Water variant 

2001 

2002 

2003 

2001–2003 

Nitrogen fertilization 

75 kg·ha –1 

125 kg·ha –1 

A strong infl uence of the rainfall 

course in the vegetation period of the 

potato on the infestation of tubers by 

Streptomyces sp. was observed. An 

infl uence of nitrogen fertilization on 

the occurrence of common scab and 

black scurf symptoms was not found. 

The cultivar tested was characterized 

by a higher resistance to infestation by 

Rhizoctonia solani than Streptomyces 

sp. 

REFERENCES 

Mean 

Non irrigated 21.81 24.2 23.0 

Sprinkler irrigation 20.1 20.8 20.5 

Mean 21.0 22.5 



Mean 14.4 16.1 

Non irrigated 

12.0 a 

a 

10.9 a 

a 11.5 a 


20.6 a 

b 

25.2 a 

b 22.9 b 

Mean 16.3 a 18.1 a 



Mean 17.2 18.9 

1 Mean values without letters or followed by the same letter in columns and rows are not signifi cantly 

different at α = 0.05, according to Tukey’s test. 

1. BOLIGŁOWA E., ŁABZA T., GLEŃ K., 

PUŁA J. 1999: Infection of potato tubers 

by storage diseases depending on organic 

fertilization. Progres in Plant Protection, 

Vol. 39 (2): 898–901. 

2. BORÓWCZAK F., GŁADYSIAK S. 

1999: Disease infestation of potato tubers


depending on irrigation and cultivation 

system. Progress in Plant Protection/ 

Postępy w Ochronie Roślin, Vol. 39 (2): 

786–788. 

3. CZAJKA W. 1988: Badania nad 

występowaniem ważniejszych 

bakteryjnych i grzybowych chorób 

ziemniaka na tle wybranych czynników 

agrotechnicznych oraz zabiegów 

chemicznych. Acta Acad. Agricult., 

Tech. Olst. Agricultura Supplement, C 

44: 1–58. 

4. CZAJKA W., CWALINA B., CZAJKA 

M., FABISIEWICZ M. 1999: Porażenie 

bulw ziemniaka patogenami w zależności 

od nawożenia mineralnego. Progress in 

Plant Protection/ Postępy w Ochronie 

Roślin, 39 (2): 852–855. 

5. GŁADYSIAK S., CZAJKA M. 1996: 

Sprinkling infl uence on potato cultivars 

infestation by diseases. Nowe Kierunki 

w Fitopatologii, Materiały z Sympozjum, 

Kraków: 217–221. 

6. JABŁOŃSKI K., CZERKO Z. 1995: Zbiór 

i przechowywanie ziemniaków. Fundacja 

Rozwój SGGW, Warszawa: 62–69. 

7. MATKOWSKI K., PROŚBA-BIAŁCZYK 

U., PLĄSKOWSKA E. 2002: Infl uence 

of cultivar mixture of potato and chemical 

control against Phytophthora infestans 

(Mont.) De Bary on the health status of 

tubers. Advances of Agricultural Sciences 

Problem Issues, 489: 261–268. 

8. PAŃKA D., PIŃSKA M. 2004: Infl uence 

of potassium fertilization on health status 

of Barycz and Triada potato cultivars. 

Prace Komisji Nauk Rolniczych i Biologicznych, 

BTN, 39, Seria B, 52: 261–269. 

9. RUDKIEWICZ F., SIKORSKI L., 

ŚLĄZAK I. 1983: Wpływ rodzaju gleby, 

nawożenia i zwalczania Phytophthora 

infestans na rozwój niektórych chorób na 

roślinach i bulwach ziemniaka. Biul. Inst. 

Ziemn., 30: 157–170. 

10. SADOWSKI CZ., GRABARCZYK S., 

RZEKANOWSKI CZ. 1988: Wpływ 

nawadniania na występowanie Streptomyces 

scabies (Taxter) i Rhizoctonia solani Kühn 

na bulwach ziemniaków uprawianych 

na glebie bardzo lekkiej. Acta Acad. 

Agricult. Techn. Olst., Agricultura 47: 

45–54. 

11. SADOWSKI CZ., PESZEK J., 

RZEKANOWSKI CZ., SOBKOWIAK 

S. 1996: Effect of irrigation and 

different nitrogen fertilization rates on 

the occurrence of Streptomyces scabies 

(Taxter) on potato cultivated on very 

light soil. Plant Breed. and Seed Sci. 40, 

No. 1–2: 45–49. 

12. SADOWSKI CZ., KORPAL W., LENC 

L., KAWALEC A. 2003: Health status of 

tubers of potato cultivated under organic 

and integrated conditions. In: Obieg 

Pierwiastków w Przyrodzie, Gworek, B. 

and Misiak, J. (eds). Monografi a tom II: 

682–686. 

13. SAWICKA B. 1995: The infl uence of 

selected elements of meteorological 

conditions on potato tubers with common 

scab. Zesz. Probl. Post. Nauk Roln. 419: 

89–93. 

14. SZUTKOWSKA M. 1999: Czy można 

ograniczyć porażenie bulw parchem 

zwykłym. Ziemn. Pol. 3: 5–9. 

15. ŚNIEG L. 1992: Wpływ nawadniania 

i nawożenia azotem na występowanie 

niektórych chorób bulw ziemniaka i ich 

odporność na uszkodzenia mechaniczne. 

Fragmenta Agronomica, 3/92: 49–57. 

16. WEBER Z. 1976: Wpływ przedplonu 

i innych czynników na występowanie 

rizoktoniozy ziemniaka (Rhizoctonia 

solani Kuhn). Rocz. Nauk Roln., Ser. E, 

2: 45–65. 

17. WENZEL H. 1948: Zur Erfassung des 

Schadenausmasses in fl anzenschutzversuchen. 

Pfl anzenschutzberichte 15: 

81–84. 

Streszczenie: Wpływ deszczowania i nawożenia 

azotowego na zdrowotność ziemniaka uprawianego 

na glebie bardzo lekkiej. Ziemniak jest 

bardzo często atakowany przez liczne patogeny, 

mogące powodować znaczne straty plonu. Do 

najgroźniejszych chorób należą: Phytophthora 

infestans, Rhizoctonia solani, Streptomyces sp., 

Erwinia carotovora subsp. carotovora i Fusarium

spp. Ich rozwój, a tym samym szkodliwość, zależy 

od wielu czynników. W największym stopniu 

na przebieg procesu chorobowego wpływają 

opady, temperatura, dostępność składników pokarmowych 

i stosowane zabiegi pielęgnacyjne. 

Celem przeprowadzonych badań było określenie 

wpływu nawadniania i zróżnicowanego nawożenia 

azotowego na porażenie bulw ziemniaka 

przez najgroźniejsze patogeny. Doświadczenie 

z uprawą ziemniaka odmiany ‘Drop’ wykonano 

w latach 2001–2003 jako dwuczynnikowe, na 

glebie lekkiej, należącej do kompleksu żytniego, 

w trzech powtórzeniach. Pierwszym czynnikiem 

było nawadnianie deszczowniane stosowane 

w zależności od potrzeb roślin oraz występujących 

niedostatków wody, a drugim nawożenie azotowe 

na poziomie 75 i 125 kg N ha –1 . Bezpośrednio po 

zbiorach analizowano bulwy pod kątem występowania 

objawów parcha zwykłego, rizoktoniozy, 

zarazy oraz suchej i mokrej zgnilizny. 

W doświadczeniu zaobserwowano w kolejnych 

latach trwania badań duże nasilenie występowania 

na bulwach parcha zwykłego i rizoktoniozy. 

Deszczowanie powodowało wzrost porażenia 

bulw przez S. scabies i spadek zanieczyszczenia 

bulw sklerotami R. solani. Stopień porażenia 

bulw różnicowało także nawożenie azotowe. 

Pozostałe choroby notowano sporadycznie, a ich 

nasilenie zależało od przebiegu warunków atmosferycznych 

w poszczególnych latach trwania 

eksperymentu. 




Dariusz Pańka 

Department of Phytopathology, 

University of Technology and Agriculture, 


Kordeckiego 20, 85-225 Bydgoszcz, Poland 

e-mail: panka@atr.bydgoszcz.pl 

Roman Rolbiecki 

Department of Land Reclamation 

and Agrometeorology, 



Bernardyńska 6, 85-029 Bydgoszcz, Poland 

e-mail: rolbr@atr.bydgoszcz.pl 

Czesław Rzekanowski 

Department of Land Reclamation 

and Agrometeorology, 



Bernardyńska 6, 85-029 Bydgoszcz, Poland 

e-mail: rzekan@atr.bydgoszcz.pl




Heat balance and climatic water balance in vegetation period 

of spring wheat 

ELŻBIETA MUSIAŁ 1, JOANNA BUBNOWSKA 1 , EDWARD GĄSIOREK 1 , 

LESZEK ŁABĘDZKI2 , 

1 

Agricultural University of Wrocław, Departament of Mathematics 

2 IMUZ – Bydgoszcz 

Abstract: Heat balance and climatic water 

balance in vegetation period of spring wheat. This 

paper characterizes the changes of heat and climatic 

water balance structure during the growing season 

of spring wheat in four observatories: Wrocław- 

Swojec 1964–2000, Bydgoszcz 1946–2004, 

Gorzów Wielkopolski 1970–1995 and Łódź 1954– 

–1995. Study of changes and trends contains the 

following elements of heat balance: net radiation, 

latent heat fl ux,sensible heat fl ux, soil heat fl ux 

and their contribution in net radiation. Climatic 

water balance is defi ned as a difference between 

precipitation and potential evapotranspiration 

reckoned by the Penmann method. 

Key words: heat balance, spring wheat, sensible 

heat fl ux, latent heat fl ux, climatic water balance. 

INTRODUCTION 

Energy and water relations in different 

ecosystems are best described by water 

and heat balance structure. Heat and 

water balance are connected through 

a streamvapour, that transports a huge 

quantity of energy to the atmosphere. 

Because of it any change in water balance 

infl uences heat balance and vice versa. 

Climate changes with their 

consequences are connected with heat 

exchange between active surface and 

atmosphere. Therefore, searching for 

climate variability concentrates on 

studying heat and water balance of 

different ecosystems. Different surfaces 

transform this energy in a particular way 

[Bubnowska, Gąsiorek, Łabędzki, Musiał 

2005], [Bubnowska, Gąsiorek, Łabędzki, 

Musiał, Rojek 2005], [Kędziora, Olejnik, 

Kapuściński 1989], [Kapuściński 2000], 

[Olejnik 1996]. The knowledge of heat 

balance for various kinds of surface 

allows characterizing changes in fl ux 

distribution. 

METHODS 

The calculation of components of active 

surface heat balance was carried aut by 

using the BMC model, worked out by 

Olejnik and Kędziora [1991, 1999] 

The heat balance of ecosystems is 

described by the following equation 

[Kapuściński 2000; Kędziora 1989]: 

Rn + LE + H + G = 0 

where: 

Rn – net radiation [Wm –2 ] 

G – soil heat fl ux [Wm –2 ] 

H – sensible heat fl ux [Wm –2 ] 

LE – latent heat fl ux [Wm –2 ] 

The exact description of applied 

calculation method could be found in 

a previous paper by the same authors 

[Musiał 2001; Bubnowska, Gąsiorek

84 E. Musiał et al. 

Łabędzki, Musiał, 2005] and in [Bowen 

1926], [Karliński, Kędziora 1968], 

[Shuttleworth, Wallace 1985]. 

Water balance in this paper is defi ned 

as a difference between precipitation and 

potential evapotranspiration reckoned by 

Penmann method. Penman showed that 

the latent heat fl ux could be expressed as: 

Δ 

( Rn + G) + Ea 

γ 

LE = 

⎛ Δ ⎞ 

⎜1+ γ 

⎟ 

⎝ ⎠ 

where: 

E a – ability of air evapotranspiration 

[Wm –s2 ], 

E a = 7,44(1 + 0,54v)d 

v – wind speed at 2 m height [ms –1 ], 

d – vapour pressure defi cit [hPa], 

γ – psychrometric constant 

γ = 0,655 [hPaK –1 ], 

∆ – mean rate of change of saturated vapour 

pressure with temperature [hPaK –1 ], 

LE, Rn, G – like above. 

A simple relationship exists between 

evapotranspiration ETP expressed in 

[mm] and latent heat fl ux LE expressed 

in [Wm –2 ]: 

LE 

ETP = n 

28,34 

Where n – number of days in decade, in 

month. 

RESULTS 

The study was carried out on data 

from the following meteorological 

observatories: Wrocław-Swojec (1964- 

–2000), Bydgoszcz (1946-2003), Łódź 

(1954-1995) and Gorzów Wielkopolski 

(1975–1995). 

The variability of heat and water 

balance components was analyzed in 

given perennials. 

HEAT BALANCE IN VEGETATION 

PERIOD OF SPRING WHEAT 

Mean values of net radiation calculated 

for vegetation period of a spring wheat 

(IV–VIII) in ssubsequent years in 

Gorzów Wielkopolski fl uctuated between 

94 and 112 Wm –2 . The lowest value 

was obtained in 1962 when the sum of 

precipitation in vegetation period was 

also the lowest. Latent heat fl ux absorbed 

from 61 to 68% of net radiation, and 

sensible heat fl ux – from 24 to 61%. The 

soil heat fl ux accounted for around 8% of 

net radiation. 

Values of heat balance components in 

other observatories had similar variability. 

Net radiation in Łódź changed from 

90W/m 2 in 1960, when the precipitation 

was the highest (476.7 mm), to 115 W/m 2 

in 1983, when the precipitation was the 

lowest (171.7 mm). In Wrocław and 

Bydgoszcz net radiation fl uctuated from 

89 to 112 [Wm –2 ] and from 89 to 109 

[Wm –2 ] respectively. Latent heat fl ux 

absorbed from 58% in Bydgoszcz to 

69% of net radiation in Łódź. 

The highest values of net radiation 

during the growing season were observed 

in Łódź and Gorzów Wielkopolski, 

whereas in Wrocław and Bydgoszcz the 

values of Rn were lower. 

Latent heat fl ux values were the 

highest in Łódź, where net radiation was 

the highest. This may be due to the fact 

that in Łódź more energy was supplied 

and therefore, more of it could be used 

for evapotranspiration.

Regarding the Figure 3, the following 

conclusion could be drawn: in Łódź and 

Gorzów Wielkopolski more energy was 

transferred to the atmosphere from the 

active surface of spring wheat than in 

Wrocław and Bydgoszcz. In 1982–1995 

perennial, common for all observatories, 

there was a distinct increasing tendency 

Heat balance and climatic water balance... 85 

TABLE.1. Components of heat balance during the growing season of spring wheat (IV–VIII) in 

Gorzów Wielkopolski (1970–1995) 

Rok Rn -LE -G -H -LE/Rn -H/Rn -G/Rn P 

1970 101 66 8 27 0,65 0,27 0,08 215,3 

1971 105 64 8 33 0,61 0,31 0,08 256,4 

1972 97 62 7 28 0,64 0,29 0,07 306,2 

1973 105 66 8 31 0,63 0,30 0,07 273,7 

1974 94 59 8 27 0,63 0,29 0,08 330,6 

1975 110 71 8 31 0,65 0,28 0,07 233,0 

1976 110 72 8 30 0,65 0,28 0,07 172,7 

1977 98 65 8 25 0,66 0,26 0,08 460,0 

1978 101 64 8 29 0,63 0,29 0,08 279,0 

1979 102 65 8 29 0,63 0,29 0,08 225,5 

1980 95 59 8 28 0,62 0,30 0,08 320,1 

1981 104 63 9 32 0,61 0,31 0,08 306,9 

1982 112 70 9 33 0,63 0,30 0,07 153,6 

1983 109 71 8 30 0,65 0,28 0,07 214,4 

1984 95 60 7 28 0,63 0,29 0,08 353,1 

1985 103 65 8 30 0,63 0,29 0,08 229,6 

1986 106 68 8 30 0,64 0,28 0,08 255,2 

1987 94 62 7 25 0,66 0,27 0,07 365,3 

1988 95 59 8 28 0,62 0,30 0,08 259,7 

1989 107 70 8 29 0,65 0,27 0,08 190,9 

1990 108 66 9 33 0,61 0,31 0,08 285,4 

1991 102 69 8 25 0,68 0,24 0,08 221,0 

1992 108 72 8 28 0,67 0,26 0,07 170,0 

1993 100 64 8 28 0,64 0,28 0,08 268,0 

1994 107 71 8 28 0,66 0,26 0,08 259,0 

1995 108 69 8 31 0,64 0,29 0,07 320,0 

Rn – net radiation [Wm –2 ] , LE – latent heat fl ux [Wm –2 ], G – soil heat fl ux [Wm –2 ], H – sensible heat 

fl ux [Wm –2 ], P – sum of precipitation in period IV–VIII [mm] 

for sensible heat fl ux. Thus, the amount 

of energy used for heating atmosphere is 

growing. 

Due to the fact that the values of heat 

balance components depend on the net 

radiation value, it is worthy looking into 

contribution of each component to net 

radiation (Rn).


FIGURE 1. Variation of mean ten-days values of net radiation (Rn) during the growing season 

of spring wheat in Wrocław – Swojec, Bydgoszcz, Gorzów Wielkopolski and Łódź 

FIGURE 2. Variation of mean ten-days values of latent heat fl ux (LE) during the growing season 

of spring wheat in Wrocław – Swojec, Bydgoszcz, Gorzów Wielkopolski and Łódź 

FIGURE 3. Variation of mean ten-days values of sensible heat fl ux (H) during the growing season 

of spring wheat in Wrocław – Swojec, Bydgoszcz, Gorzów Wielkopolski and Łódź

Looking at the course of mean values 

of latent heat fl ux contribution in net 

radiation, distinct decreasing tendency 

for this contribution in the last 20 years 

is seen. Therefore, less and less energy 

is used for evapotranspiration, especially 

in Wrocław and Bydgoszcz. 

Increasing share of sensible heat fl ux 

in net radiation in all observatories stress 

that more and more energy in all regions 

is used for heating atmosphere. 

The above mentioned results are 

concordant with those by Musiał [Musiał, 

Gąsiorek, Rojek 2004], Trepińska[1997], 


FIGURE 4. Variation of mean ratios of latent heat fl ux and net radiation (LE/Rn) during the growing 

season of spring wheat in Wrocław – Swojec, Bydgoszcz, Gorzów Wielkopolski and Łódź 

FIGURE 5. Variation of mean ratio values of sensible heat fl ux and net radiation (H/Rn) during the 

growing season of spring wheat in Wrocław – Swojec, Bydgoszcz, Gorzów Wielkopolski and Łódź 

[Ryszkowski Kędziora, 1995] and 

[Kożuchowski 2004]. 

The temperature increase is a 

consequence of the growing contribution of 

latent heat fl ux to net radiation. Regression 

equations in Table 2 confi rm this. 

CLIMATIC WATER BALANCE 

The net climatic water balance determines 

conditions of plant vegetation [Bac, 

Rojek, 1979, 1982]. This index may be 

the source of information on climate 

change effects.


TABLE 2. Mean parennial air temperature values for periods IV–VIII and IV–IX with regression 

equations 

Obserwatory 

T weg 

[°C] 

T pc 

[°C] 

s sr 

[°C] 

Linear regression equation 

Tendency 

[°C/10 years] 

Bydgoszcz 15.22 15.0 0.92 y = 0.0176x + 14.7 0.18* 

Gorzów 

14.57 14.4 0.85 y = 0.0477x + 13.9 0.48* 

Wielkopolski 

Łódź 14.32 14.1 0.88 y = 0.0129x + 14.0 0.13** 

Wrocław 14.93 14.7 0.83 y = 0.0347x + 14.3 0.35* 

Tweg – mean seasonal yearly air temperature (IV–VIII), 

Tpc – mean seasonal, yearly air temperature (IV–IX), 

ssr – standard devation of temperature 

*) – statistically signifi cant for α = 0.05 

**) – statistically signifi cant for α = 0.3 

FIGURE 6. Variation of climatic water balance during the growing season of spring wheat (IV–VIII) 

in Bydgoszcz (1946–2003) 

FIGURE 7. Variation of precipitation (P) and climatic water balance (CWB) during the growing season 

of spring wheat (IV–VIII) in Wrocław (1964–2000)

In Bydgoszcz climatic water balance 

was positive in two years during the 

period 1946–2004. Those two years 

were chracterized by high sums of 

precipitation: 599,1 mm in 1980 and 

506,6 mm in 1985. 

In years 1980 and 1985 the spring 

wheat vegetation period was characterized 

by rather low (390,56 and 419,35 mm, 

respectively) evapotranspiration values, 

whereas mean evapotranspiration value 

for years 1946–2004 was 456,5 mm. 


In Wrocław climatic water balance 

was positive in 1980 (11,3 mm) and in 

1986 (9,2 mm). 

During the spring wheat vegetation 

period in 1980, the precipitation 

value was 438 mm and the potential 

evapotranspiration reached the level of 

426,7 mm. The year 1997 is noteworthy 

due to the fl ood in Wrocław. During 

that spring wheat vegetation period 

the precipitation sum (494 mm) was 

52% higher than the mean value. In 

spite of such high precipitation value, 


of spring wheat (IV–VIII) in Łódź (1954–1995) 


of spring wheat (IV–VIII) in Gorzów Wielkopolski (1970–1995)


TABLE 3. Evapotranspiration tendencies in 4 examined regions 

Obserwatory 

P 

[mm] 

ETP 

[mm] 

the net climatic water balance in the 

1997 growing season remained negative 

(–31 mm). 

In Łódź, during the 1954–1995 

perennial, there were 3 years with 

positive net climatic water balance: 1960 

(precipitation 477 mm, climatic water 

balance 20 mm), 1980 (458 mm and 

38,9 mm, respectively) and 1985 (464 

mm and 17 mm, respectively). Mean 

precipitation sum in the growing season 

for spring wheat in Łódź was 312 mm. 

In Gorzów Wielkopolski, the climatic 

water balance was positive only once and 

reached the value of 48,2 mm in 1977, 

when the precipitation sum in vegetation 

period was the highest (460 mm). 

All analyzed regions were 

characterized by continuous water 

shortage in vegetation period. During 

the period 1970–1995 the worst water 

conditions were observed in Gorzów 

Wielkopolski (the highest water shortage 

sum – 5474 mm), whereas the best terms 

were seen in Wrocław (water shortage 

sum of 4640 mm was 15% lower than 

that in Gorzów). During the wheat spring 

growing season the values of climatic 

water balance were positive only in years 

with the highest precipitation sums. 

Linear regression equation 

CONCLUSIONS 

The enlargement of negative net climatic 

water balance, along with increasing 

potential evapotranspiration, confi rms 

diminishing atmospheric precipitation. 

The increase of temperature during 

the spring wheat growing season in all 

discussed regions is a consequence of 

enlarging sensible heat fl ux, used for 

heating atmosphere. 

The growing temperature values 

enlarge the saturation defi ciency, thus 

allowing more water to evapotranspirate 

(PET increase). Consequently, growing 

precipitation enlarge the negative net 

climatic water balance. 

The tendencies observed among heat 

balance components are concordant with 

changes seen among the components of 

climatic water balance. 

The research supported by KBN grant 

in years 2004–2007. 

REFERENCES 

Tendency 

[°C/10 years] 

Bydgoszcz 281 457 y = 0.29x + 443.4 2,9 

Gorzów Wielkopolski 266 477 y = 4.91x + 466.7 49,1* 

Łódź 312 505 y = 2.0x + 470.6 20** 

Wrocław 326 500 y = 2.0x + 468.2 20* 

P – mean seasonal, yearly precipitation (IV–VIII) 

ETP – mean seasonal, yearly evapotranspiration (IV–VIII), 


**) - statistically signifi cant for α = 0.01 

BAC S., ROJEK M. 1979: Klimatyczny 

bilans wodny a odpływy w Polsce, Przegl. 

Geofi z. 24(3) 293–297.

BAC S., ROJEK M. 1977: Metodyka oceny 

stosunków wodnych obszarów rolniczych 

na podstawie danych klimatycznych, 

Zesz. nauk ART Olszt. Nr 21, 13–24. 

BAC S., ROJEK M. 1982: Klimatyczne 

bilanse wodne w Polsce [w:] Bac S. 

(red.) Agrometeorologiczne podstawy 

melioracji wodnych w Polsce. PWRiL 

Warszawa. 

BOWEN I.S. 1926: The ratio of heat losses 

by conduction and by evaporation from 

any water surface. Phys. Rev., 27, p. 

779–787. 

BUBNOWSKA J., GĄSIOREK E., 

ŁABĘDZKI L., MUSIAŁ E., ROJEK 

M. 2005: Bilans cieplny lasu iglastego 

w latach o ekstremalnych opadach i jego 

wieloletnie zmiany w rejonie Bydgoszczy 

i Wrocławia. Woda-Środowisko-Obszary 

Wiejskie t. 5, z. spec. (14) s. 69–82. 

BUBNOWSKA J., GĄSIOREK E., ŁA- 

BĘDZKI L., MUSIAŁ E. 2005: Struktura 

bilansu cieplnego łanów w ekstremalnych 

warunkach opadowych na tle 

wielolecia. Woda-Środowisko-Obszary 

Wiejskie, t. 5 z. 2(15) s. 31–52. 

ŁABĘDZKI L., BĄK B. 2004: Standaryzowany 

klimatyczny bilans wodny jako 

wskaźnik suszy. Acta Agrophys. Vol. 

3(1) s. 117–124. 

KAPUŚCIŃSKI J. 2000: Struktura bilansu 

cieplnego powierzchni czynnej na tle 

warunków klimatycznych środkowozachodniej 

Polski, Rocz. Nauk. AR Pozn. 

Rozpr. Nauk. AR Poznań, 303, s. 250. 

KARLIŃSKI M., KĘDZIORA A. 1968: 

Rozważania metodyczne przy układaniu 

kalendarza przyrody dla województwa 

szczecińskiego, Pr. Kom. Nauk Roln. 

Kom. Nauk Leśn. PTPN 24, s. 159–175. 

KĘDZIORA A. 1999: Podstawy 

agrometeorologii, Poznań, PWRiL. 

KĘDZIORA A., OLEJNIK J., KAPUŚCIŃ- 

SKI J. 1989: Impact of landscape structure 

on heat and water balance. Ecol. 

Intern. Bull. 17 ss. 1–17. 

KOŻUCHOWSKI K. 2004: Skala, uwarunkowania 

i perspektywy współczesnych 


zmian klimatycznych w Polsce, Łódź, 

Biblioteka s. 170. 

MUSIAŁ E. 2001: Modelowanie procesu 

ewapotranspiracji rzeczywistej i prognozowanie 

jego tendencji. Zesz. Nauk. AR 

Wroc., nr 412, Rozpr. 182, ss. 116. 

MUSIAŁ. E. , GĄSIOREK E., ROJEK M.S. 

2004: Zmienność temperatury powietrza 

w obserwatorium Wrocław–Swojec w 

latach 1964-2001. Acta Agroph. nr 105 

Vol. 3(2) s. 333–342. 

OLEJNIK J., KĘDZIORA A. 1991: A model 

for heat and water balance estimation and 

its application to land use and climate 

variation., Earth Surface Processes 

Landforms Vol.16, ss. 601–617. 

OLEJNIK J. 1996: Modelowe badania 

struktury bilansu cieplnego i wodnego 

zlewni w obecnych i przyszłych 

warunkach klimatycznych, Rocz. Nauk. 

AR Poznań z. 268, ss. 125 

RYSZKOWSKI L., KĘDZIORA A. 1995: 

Modifi cation of the effects of global 

climate change by plant cover structure 

in an agricultural landscape. Geogr. Pol. 

Vol. 65, s. 5–34. 

RYSZKOWSKI L., KĘDZIORA A. 1993: 

Rolnictwo a efekt szklarniowy. Kosmos 

42 s. 123–149. 

SHUTTLEWORTH W.J., WALLACE J.S. 

1985: Evaporation from sparse crops-an 

energy combination theory, Quart. J. R. 

Met. Soc. 111, pp. 839–895. 

TREPIŃSKA J. 1997: Wahania klimatu 

w Krakowie (1792–1995). Kraków Inst. 

Geogr. UJ, s. 1–204. 

Streszczenie: Bilans cieplny i klimatyczny bilans 

wodny dla okresu wegetacji pszenicy jarej. Praca 

charakteryzuje zmiany struktury bilansu cieplnego 

i klimatycznego bilansu wodnego w czasie okresu 

wegetacji pszenicy jarej dla czterech stacji: 

Wrocław-Swojec 1964–2000, Bydgoszcz 1946– 

2004, Gorzów Wielkopolski 1970–1995 i Łódź 

1954–1995. W pracy przeanalizowano zmiany 

i trendy takich składowych bilansu cieplnego, 

jak: saldo promieniowania i strumienie ciepła 

jawnego, ciepła utajonego przeznaczonego na 

parowanie, ciepła wymienianego z podłożem oraz 

ich udział w saldzie promieniowania.




E. Musiał, J. Bubnowska, E. Gąsiorek 

Agricultural University of Wrocław 

Departament of Mathematics 

ul. Grunwaldzka 53 

50 357 Wrocław 

Poland (071)3205659 

Leszek Łabędzki 

ul. Glinki 60 

85 174 Bydgoszcz 

Poland 

musial@ozi.ar.wroc.pl




Climatic and agricultural water balance for grasslands in Poland 

using the Penman-Monteith method 

WIESŁAWA KASPERSKA-WOŁOWICZ, LESZEK ŁABĘDZKI 

Institute for Land Reclamation and Grassland Farming, 

Regional Research Centre in Bydgoszcz 

Abstract: Climatic and agricultural water 

balance for grasslands in Poland using the 

Penman-Monteith method. The aim of the paper 

was to estimate the spatial variability of climatic 

and agricultural water balance for two-cut 

meadows (for two hay yield: 5 and 7 Mg ha-1) 

in different agro-climatic regions in Poland. It let 

to estimate meadow water needs under different 


The climatic water balance is a difference 

between precipitation and reference evapotranspiration 

calculated according to the Penman- 

Monteith formula. The agricultural water balance 

is a difference between precipitation and potential 

crop evapotranspiration. Potential grassland 

evapotranspiration is the result of multiplication 

crop coeffi cient and reference evapotranspiration 

according to the Penman-Monteith formula. The 

crop coeffi cient values were calculated on the 

basis of multiannual lysimeter experiments of 

meadow evapotranspiration. 

The analysis was carried out in the growing season 

(from April to September) in the years 1970–1995 

for 17 meteorological stations located in different 

regions of Poland. The balances were done for 

ten-day periods, months and the whole growing 

season. 

The climatic water balance indicates the risk 

of agro-climatic drought. The agricultural water 

balance indicates water defi cit and irrigation water 

needs of two-cut meadows in different regions of 

Poland. The negative values of this balance mean 

the scarcity of water for grasslands. The values 

of agricultural water balance are shown on the 

maps. 

Key words: climatic and agricultural water 

balance, two-cut meadow. 

INTRODUCTION 

The estimation of drought hazard and 

irrigation needs for plants can be estimated 

using crop coeffi cients method according 

to the Penman-Monteith formula. 

This method for calculation reference 

evapotranspiration ET o is commonly 

recommended and used in the word To 

estimate crop water requirements, one 

can relate potential evapotranspiration 

from the cropped soil with an optimum 

water supply under consideration to an 

estimated reference evapotranspiration 

by means of crop coeffi cient [1, 3]. 

Evaluation of crop water demands, 

agricultural droughts and irrigation 

requirements can be made with the 

help of climatic water balance and 

agricultural water balance. The climatic 

water balance can be interpreted as the 

indicator of climate dryness [7]. The 

agricultural water balance indicates 

the values of irrigation water needs for 

plants. These two balances are needed to 

the preliminary estimation of irrigation 

needs.

94 W. Kasperska-Wołowicz, L. Łabędzki 

The aim of the paper was to estimate 

the spatial variability of climatic and 

agricultural water balance for grasslands 

in different agro-climatic regions in 

Poland and to show these values on 

the maps. These maps let to estimate 

water needs for 2-cut meadow with the 

different yield in the years with different 


METHODS AND MATERIAL 

The climatic water balance CWB is the 

difference between precipitation sum P 

and reference evapotranspiration ETo calculated according to the Penman- 

Monteith formula [1, 6]: 

CWB = P – ETo (1) 

TABLE 1. The crop coeffi cients k c for the Penman- 

Monteith equation for 2-cut meadow in Poland 

Month 

Ten-day 

period 

k c coeffi cient for hay 

yield: 

7 Mg⋅ha –1 5 Mg⋅ha –1 

IV 1 0.50 0.45 

2 0.75 0.70 

3 0.95 0.80 

V 1 1.00 0.90 

2 1.15 1.00 

3 1.20 1.10 

VI 1 1.30 1.20 

2 0.55 0.45 

3 0.65 0.55 

VII 1 0.80 0.70 

2 0.90 0.80 

3 1.10 1.00 

VIII 1 1.30 1.15 

2 1.20 1.10 

3 1.35 1.25 

IX 1 1.10 1.10 

2 1.10 1.10 

3 1.10 1.10 

The agricultural water balance AWB is 

the difference between precipitation sum 

P and potential crop evapotranspiration 

ETP: AWB = P – ETP (2) 

where: 

ETP – is evapotranspiration of wellwatered 

crop calculated from the 

formula: 

ETP = ETo⋅kc (3) 

where: 

kc – crop coeffi cient, 

ETo – reference evapotranspiration 

(mm). 

The kc coeffi cients values were 

calculated on the basis of multiannual 

lysimeter experiments and depended on 

the hay yield (Tab. 1) [4, 6, 8]. 

The analysis was carried out in 

the growing season (from April to 

September) in the years 1970–1995 

for 17 meteorological stations. The 

chosen stations are situated in different 

agroclimatic regions of Poland (Tab. 2). 

According to Bac, Koźmiński, Rojek 

[2] taking into account water conditions 

one can distinguish regions: A – dry, 

B – moderately wet, C – wet. Taking 

into account thermal and solar energy 

conditions one can distinguish regions: 1 

– hot and sunny, 2 – hot and moderately 

sunny, 3 – hot and cloudy, 4 – moderately 

hot and sunny, 5 – moderately hot and 

moderately sunny, 6 – moderately hot 

and cloudy, 8 – cold and moderately 

sunny. 

The agricultural water balance AWB 

was calculated for 2-cut meadows with the 

hay yield of: 7 Mg⋅ha –1 and 5 Mg⋅ha –1 . 

The negative values (–) of climate 

and agricultural water balance showed

TABLE 2. Meteorological stations and agroclimatic 

regions according to Bac [2] 

Number of Name of the Agro-climatic 

the station station region 

1 Koszalin C8 

2 Bielnik B4 

3 Biebrza B8 

4 Chojnice B8 

5 Szczecin A6 

6 Frydrychowo A2 

7 Poznań A3 

8 Warszawa A2 

9 Zielona Góra A1 

10 Łódź A5 

11 Sosnowica A1 

12 Wrocław B2 

13 Opole B3 

14 Częstochowa B6 

15 Zamość B5 

16 Kraków C5 

17 Rzeszów C8 

the drought risk and water defi cit for 

grasslands irrigation, the positive values 

(+) – showed the excess of water. 

The calculation of the balances have 

been done for ten-day periods, months 

and the whole growing season (from 

April to September). The probability 

distributions of climatic water balance 

and agricultural water balance were 

determined. It was assumed that the 

examined period was classifi ed as wet 

at the probability of exceedence equal to 

25%, mean – 50% and dry – 75%. 

RESULTS 

Climatic water balance 

The climatic water balance CWB is 

different in Poland and depends on 

Climatic and agricultural water balance... 95 

agrometeorological conditions in the 

examined regions. The mean multiannual 

values of this balance (at the p = 50%) 

for growing season (April to September) 

were negative in all examined stations. 

In the north (stations 1–3) and south 

(stations 14, 16, 17) parts of the country 

the values of climatic water balance 

were negative but higher than –40 mm. 

The values of CWB less than –100 mm 

were in the central Poland, in the regions 

represented by the stations numbered 

5–10. The lowest values of the balance 

(less than –150 mm) were in the region 

between the Odra and left side of Wisła 

rivers, especially represented by the 

stations 5–7 and 9–10. The highest risk 

of agro-climatic drought in Poland was 

observed on the area from the north-west 

to the central part of Poland (Fig. 1). 

The lowest monthly values of CWB 

in central Poland were in May, June 

and July. It resulted from the monthly 

distribution of precipitation and reference 

evapotranspiration. For instance in 

Poznań countryside (station 7), the 

multiannual mean monthly sum of ET o 

reaches the highest values in July (about 

100 mm), a little less in June, May and 

August (about 80 mm) [5]. 

The higher values of CWB in the 

average year were found in the south 

(stations 14, 16, 17) part of Poland, 

the region situated near the Karpaty 

mountains and in the north part of Poland 

– near Bałtyk sea (station 1) and in the 

Biebrza river valley (station 3) – in the 

north-east part of the country. 

Agricultural water balance 

In the wet growing season (at 25% 

probability) agricultural water defi cit 

did not occur in the north and south 

parts of Poland. The negative values of


FIGURE 1. Mean multiannual precipitation (1) and the climatic water balance (2) in the growing season 

(IV–IX) in different regions of Poland 

agricultural water balance for low and 

high yield meadows were observed in 

the Warta river basin (stations 6, 7, 10), 

in the west part of Poland (stations 5, 

9) and in the Polesie Lubelskie region 

(station 11). It means that even in the wet 

year the big area of Poland is threaten by 

droughts. The highest water defi cit for 

meadow was in the Szczecin countryside 

(Fig. 2a). 

In the mean growing season (at 50% 

probability) the values of agricultural 

water balance were negative for 7 

Mg⋅ha –1 hay yield meadow in the whole 

area of Poland. The high water defi cits 

for grasslands (more than 100 mm) were 

observed on the area from the northwest 

to the central part of Poland. The 

highest scarcity of water (more then 150 

mm) was observed in Warta and middledown 

Odra river basins. In mean year, in 

all regions of Poland, water defi cit for 5 

Mg⋅ha –1 hay yield meadow was 40–47 

mm less than for 7 Mg⋅ha –1 hay yield 

meadow (Fig. 2b). 

In the dry growing season (at 75% 

probability) agricultural water defi cit was 

observed in the whole area of Poland. 

The highest values of water defi cit were 

in the area between Odra and the left side 

of Wisła and they exceeded 200 mm for 

7 Mg⋅ha –1 yield meadow and 160 mm for 

5 Mg⋅ha –1 yield meadow. Water defi cit 

for high yield meadow was less than 

100 mm only in south part of Poland in 

Kraków countryside (Fig. 2c). 

The values of climatic water balance 

(CWB) and agricultural water balance 

(AWB) for the 7 Mg⋅ha –1 hay yield 

meadow did not differ very much in the 

whole mean growing season (at 50% 

probability). These values differed in 

particular months. They depended on 

agrometeorological conditions and the 

time of fi rst and second cut of meadow. 

The monthly distribution of climatic

a 

FIGURE 2. Agricultural water balance for 2-cut 

meadow at the yield 7 Mg·ha –1 (upper number, 

mm) and 5 Mg·ha –1 (lower number, mm) in different 

agroclimatic regions of Poland in the growing 

season (IV–IX) a) wet, b) mean, c) dry 

• 01 – number of meteorological station, 01 – Koszalin 

and agricultural water balance is shown 

on the example of Wielkopolska region 

(dry and hot) – represented by Poznań 

agrometeorology station (7) and the 

Biebrza river valley (moderately wet and 

cold) – represented by station 3. (Fig. 3). 

The highest water defi cits for 

meadow were in May and September 

– the months of intensive grass growth. 

In Wielkopolska region, in these months, 

they exceed 50 mm and in Biebrza river 

catchment – 30 mm. In the other months 

the values of climatic water balance 

were higher than values of agricultural 

b 

c 


water balance. In Biebrza region the 

values of climatic and agricultural water 

balance were positive in September. In 

every month they were higher than in 

Wielkopolska region. 

CONCLUSIONS 

1. 

The climatic water balance CWB is 

different in Poland and depends on 

agrometeorological conditions. The 

highest risk of agro-climatic drought 

in Poland was observed in the region


a 

b 

FIGURE 3. Mean monthly multiannual climatic water balance (1) and agricultural water balance for 

7 Mg·ha –1 yield 2-cut meadow (2) in the growing season (IV–IX) in a) Poznań, b) Biebrza

2. 

3. 

4. 

5. 

between the Odra and left side of 

Wisła rivers. 

In the wet year the big area of Poland 

is threaten by droughts. The negative 

values of agricultural water balance 

for low and high yield meadows 

were observed in the west, central 

and east part of Poland. 

In the mean year the values of 

agricultural water balance were 

negative for 7 Mg⋅ha –1 hay yield 

meadow in the whole area of Poland. 

The highest scarcity of water (more 

than 150 mm) was observed in Warta 

and middle-down Odra river basins. 

Water defi cit for 5 Mg⋅ha -1 hay yield 

meadow was 40–47 mm less than for 

7 Mg⋅ha –1 hay yield meadow. 

In dry year meadow water defi cit was 

observed in the whole area of Poland. 

The highest water needs were in the 

area between Odra and left side of 

Wisła and they exceeded 200 mm 

for 7 Mg⋅ha –1 yield meadow and 160 

mm for 5 Mg⋅ha –1 yield meadow. 

Water defi cits should be satisfi ed 

by effi ciently performing irrigation 

systems. 

REFERENCES 

1. ALLEN R.G., PEREIRA L.S., RAES D. & 

SMITH M. 1998: Crop evapotranspiration 

– Guidelines for computing crop water 

requirements. FAO Irrigation and 

drainage no 56, pp. 300. 

2. BAC S., KOŹMIŃSKI C., ROJEK M. 1993: 

Agrometeorologia. [Agrometeorology] 

Warszawa. PWN, pp. 250 (in Polish). 

3. FEDDES R.A., LENSELINK K.J. 1994: 

Evapotranspiration. ILRI Publication 16. 

Drainage Principles and Applications. 

Wageningen, p. 145–173. 


4. KACA E., ŁABĘDZKI L., CHRZA- 

NOWSKI, S,. CZAPLAK I., KASPER- 

SKA-WOŁOWICZ W. 2003: Gospodarowanie 

zapasami wody użytecznej 

gleb torfowo-murszowych w warunkach 

regulowanego odpływu w różnych regionach 

agroklimatycznych Polski. Woda, 

Środowisko, Obszary Wiejskie. Rozprawy 

naukowe i mono-grafi e nr 9, ss. 

118. Falenty, Wydaw. IMUZ. [Managing 

useful water resources in peat-moorsh 

soils at a regulated outfl ow in different 

agro-climatic regions of Poland. Water- 

-Environment-Rural Areas, Treatises & 

Monographs 9. Falenty: IMUZ, Poland, 

pp 108] (in Polish, summary in English). 

5. KASPERSKA-WOŁOWICZ W., ŁA- 

BĘDZKI L. 2004: Porównanie ewapotranspiracji 

wskaźnikowej według Penmana 

i Penmana-Monteitha w różnych 

regionach Polski. Woda, Środowisko, 

Obszary Wiejskie t. 4 z. 2a (11) s. 123– 

–136. [A comparison of reference evapotranspiration 

according to Penman and 

Penman-Monteith in various regions of 

Poland. Water-Environment-Rural Areas 

4(2a). Falenty: IMUZ, Poland, p. 123– 

136] (in Polish, summary in English). 

6. ŁABĘDZKI L. 1997: Potrzeby nawadniania 

użytków zielonych – uwarunkowania 

przyrodnicze i prognozowanie. 

Rozpr. Habil. Falenty: Wydaw. IMUZ ss. 

120. [Grasslands water requirements – 

nature conditions and forecasting. Falenty: 

IMUZ, Poland, pp. 120] (in Polish). 

7. ŁYKOWSKI B. 1986: Wskaźnik 

suchości klimatu P – E w nowym ujęciu. 

Problematyka melioracji w nauczaniu 

i badaniach naukowych. Warszawa: 

Wydaw. SGGW s. 59–68. [Indicator of 

climate dryness P – E in a new approach. 

Reclamation problems in teaching and 

research studies. Warszawa: SGGW, p. 

59–68] (in Polish, summary in English). 

8. ROGUSKI W., SARNACKA S., DRUPKA 

S. 1988: Instrukcja wyznaczania potrzeb 

i niedoborów wodnych roślin uprawnych 

i użytków zielonych. Mat. instruktażowe 

nr 66, Falenty: IMUZ ss. 90. [Guidelines


for estimating water needs and water 

defi cits for cultivated plants and 

grasslands. Instruction Materials no 66, 

Falenty: IMUZ, pp. 90] (in Polish). 

Streszczenie: Klimatyczny i rolniczy bilans wodny 

dla użytków zielonych w Polsce z wykorzystaniem 

metody Penmana-Monteitha. Celem pracy 

była ocena przestrzennego zróżnicowania wodnych 

bilansów klimatycznego i rolniczego dla 

dwukośnych użytków zielonych (dla plonów 5 

i 7 Mg ha –1 ). Autorzy przeprowadzili analizę bilansów 

sporządzonych w okresie dekady, miesiąca 

oraz wegetacji dla 17 stacji meteorologicznych 

w różnych regionach Polski. W przeciętnym okresie 

wegetacji IV–IX wartość rolniczego bilansu 

wodnego jest ujemna dla plonów w wysokości 

7 Mg ha –1 w całym kraju. Największe niedobory 

(powyżej 150 mm) obserwuje się w dorzeczu 

Warty. Niedobór dla plonów wysokości 5 Mg ha –1 

jest o 40–47 mm mniejszy w porównaniu z plonami 

użytków zielonych w wysokości 7 Mg ha –1 . 

MS. received December 2006 


Wiesława Kasperska-Wołowicz 

Leszek Łabędzki 

IMUZ 

Glinki 60, 85-174 Bydgoszcz 

Poland 

e-mail: imuzbyd@by.onet.pl




Variation of climatic water balance and heat balance for various 

ecosystems in Wrocław in the years 1964–2000 

ELŻBIETA MUSIAŁ, JOANNA BUBNOWSKA, EDWARD GĄSIOREK 

Agricultural University of Wrocław, Department of Mathematics 

Abstract: Variation of climatic water balance and 

heat balance for various ecosystems in Wrocław 

in the years 1964–2000. The study contains a 

description of water relations, characterized by 

cumulative climatic water balance, in the region 

of Wrocław-Swojec. Heat relations during the 

vegetation period were described by heat balance 

of various active surfaces. The evaluation of 

climatic water balance variation was based on 

monthly sums of precipitation and potential 

evapotranspiration. The latter was calculated 

according to the Penman’s model for the warm 

half of the year (IV–IX) and according to the 

Tichomirow’s and Iwanow’s model for the cold 

half of the year (X–IX). The values of climatic 

water balance were assessed for the period from 

X to IX in every year, as well as for all years 

in common during the period 1964–2000. The 

perennial 1964–2000 in Wrocław-Swojec was 

characterized by potential evapotranspiration 

increase and a decreasing trend in yearly sums of 

atmospheric precipitation. The calculated values 

of cumulative climatic water balance proved the 

deepening water shortage in this area. Due to 

aggravating water defi cit in Wrocław latent heat 

fl ux decreases and sensible heat fl ux, used for 

heating the atmosphere, increases, thus causing 

the warming up in this region. 

Key words: potential climatic water balance, 

potential evapotranspiration, heat balance, latent 

heat fl ux, sensible heat fl ux, Bowen’s ratio. 

INTRODUCTION 

Investigations on variation of heat 

balance components for various 

ecosystems, performed in the Wrocław- 

Swojec region, made the authors think 

about changes of climatic water balance 

in this region. The importance of climatic 

water balance in meteorology is due to 

the fact that net climatic water balance 

defi nes conditions for plant vegetation, 

infl uences changes of outfl ow indexes 

in river cachments and retention. 

The investigation was performed 

in years 1964–2000. Every analyzed 

year started in October and ended in 

September. Creation of cumulative 

climatic water balance from October 

to September gave the possibility to 

evaluate water overfl ow or defi cit after 

autumn and winter, at the beginning 

of vegetation period and during this 

period until October. Due to the fact that 

climatic water balance is connected with 

heat balance through the stream vapour, 

the values of balance components were 

evaluated for coniferous forest, spring 

wheat and potatoes. The main aim of 

the study was to describe variations of 

water overfl ow and defi cit after autumn 

and winter in subsequent years and 

consequences of water defi cits in heat 

balance component values for examined 

ecosystems. 

METHODS 

The term of climatic water balance 

was fi rst used in Polish literature by 

Bac and Rojek in 1970 [Bac, Rojek

102 E. Musiał, J. Bubnowska, E. Gąsiorek 

1977, 1979, 1982]. The authors 

defi ned climatic water balance as a 

difference between precipitation and 

indicatory evapotranspiration [Rojek 

1994], [Rojek, Wiercioch 1994, 1995]. 

The name climatic water balance is 

mainly connected with net balance 

assessed for the warm period (IV–IX). 

Contradictory to the above mentioned 

investigations, the authors of this study 

created a term of potential climatic water 

balance (PCWB). PCWB is defi ned as 

a difference between precipitation and 

potential evapotranspiration calculated by 

Penman, assessed in the warm half of the 

year (IV–IX). When the cold half of the 

year (X–III) is taken into account, PCWB 

is a difference between precipitation and 

potential evapotranspiration calculated 

by Iwanow and Tichomirow [Kędziora 

1999], [Olejnik, Kędziora 1991]. 

Penman [Penman 1948, 1950, 1956, 

1963] assumed that the density of latent 

heat fl ux used for evapotranspiration can 

be calculated in the following way: 

⎡ 

⎤ 

LE = Rn G Ea 

+ 

⎢ + + ⎥ 

⎣ 

⎦ 

Rn G Ea 

= 

γ Δ 

( ) 

Δ γ γ 

Δ 

( + ) + 

γ 

= 

⎡ Δ ⎤ 

⎢1 

+ ⎥ 

⎣ γ ⎦ 

where : 

Rn – net radiation [Wm –2 ], 

G – soil heat fl ux [Wm –2 ], 

H – sensible heat fl ux [Wm –2 ], 

LE – latent heat fl ux [Wm –2 ], 

Ea – ability of air evapotranspiration 

[Wm –2 ], 

Ea = 7.44(1 + 0.54v)d, 

v – wind speed at 2 m height [ms –1 ], 

d – vapour pressure defi cit [hPa], 

∆ – mean rate of change of saturated 

vapour pressure with temperature 

[hPaK –1 ], 

γ – psychrometric constant γ = 0.655 

[hPaK –1 ]. 

A simple relationship exists between 

potential evapotranspiration ETP 

expressed in [mm] and latent heat fl ux LE 

expressed in [Wm –2 ]: ETP = nLE/28.34, 

where n is a number of days in decade, 

in month. 

Climatic water balance, through 

the stream vapour transporting huge 

amount of energy to the atmosphere, 

is connected with active surface heat 

balance. The components of heat balance 

are not independent due to the rule of 

evapotranspiration priority in nature. 

Therefore, both defi cit and excess in 

water, defi ned by cumulative potential 

climatic water balance at the beginning 

and during the vegetation period, decide 

on values of various heat balance streams. 

Thus, heat balance should be the next 

step of investigation. Heat balance of 

any active surface may be expressed by 

the following equation: 

Rn + LE + H + G = 0 

where H – sensible heat fl ux [Wm –2 ], 

Rn, LE, G – like above. 

Heat balance components for various 

ecosystems were calculated according 

to the MBC [Kędziora 1999], [Olejnik, 

Kędziora 1991]. 


Seasonality of precipitation and potential 

evapotranspiration in the perennial 

1964–2000 in Wrocław-Swojec is

shown in Figure 1 as sums of potential 

evapotranspiration and precipitation 

in the following months of the year. In 

our study we assume that one season 

is the period from October, when 

vegetation period ends, until September. 

Characteristics of precipitation and 

potential evapotranspiration seasonality 

is as follows: the highest sums of 

atmospheric precipitation and potential 

evapotranspiration in Wrocław-Swojec 

during the period 1964–2000 were seen 

in June, July and August. Moreover, apart 

from January, February, October and 

November, atmospheric precipitation was 

Variation of climatic water balance... 103 

lower than potential evapotranspiration. 

The analysis of mean sums of potential 

climatic water balance in this perennial 

revealed that net PCWB was negative in 

March, April, May, June, July, August and 

September. This may be due to the fact 

that the vegetation period in the above 

mentioned perennial was characterized 

by continuous water defi cits. 

The description of 1964–2000 

perennial in Wrocław-Swojec by 

yearly precipitation and potential 

evapotranspiration sums show that the 

last two decades of the XXth century were 

characterized by increasing potential 

FIGURE 1. Sums of precipitation (P) and potential evapotranspiration (ETP) in Wrocław-Swojec 

(1964–2000) 

FIGURE 2. Mean monthly sums of potential climatic water balance (PCWB) in Wrocław-Swojec 

(1964–2000)


evapotranspiration and diminishing 

atmospheric precipitation (Fig. 3). 

Precipitation decrease and potential 

evapotranspiration increase resulted in 

enlarging net potential climatic water 

balance during the years 1964–2000 

(Fig. 4). 

Tendencies of yearly changes 

in precipitation and potential 

evapotranspiration sums as well as 

basic statistical characteristics are 

shown in Table 1. In the 37-year period 

of observation in Wrocław-Swojec, 

potential evapotranspiration has shown 

a growing tendency both during the 

whole year and during the period from 

X–III. The last two decades of the XXth 

century revealed quicker increase of 

potential evapotranspiration than in the 

earlier period. However, the potential 

evapotranspiration rise in the period 

X–III was less pronounced than in the 

preceding years. Decreasing trends of 

yearly atmospheric precipitation sums 

and the sums in the period X–III, suggest, 

the aggravating atmospheric precipitation 

shortage in the last 20 years. 

The next step was the characteristics 

of the period 1964–2000 in Wrocław- 

Swojec, conducted in the perennial for 

every season from X to IX separately, by 

assessing cumulative potential climatic 

water balance. This characteristics 

looks as follows: in years 1964–1986 

net PCWB, depending on precipitation, 

was alternately positive and negative 

FIGURE 3. Yearly sums of potential evapotranspiration and precipitation in Wrocław-Swojec (1964– 

–2000) 

FIGURE 4. Yearly sums of potential climatic water balance in Wrocław-Swojec (1964–2000)

throughout the period X–IX. It was often 

seen that within the analyzed season net 

PCWB was positive until April, which 

means that precipitation covered the 

needs of potential evapotranspiration. 

Contrarily, since May net PCWB 

became negative, which was the sign 

of deepening water defi cit. The detailed 

analysis starts in the year 1964/1965. 

Net potential climatic water balance 

(PCWB) was positive for the whole 

period from X to IX (precipitation 661 

mm). Other seasons with positive net 

PCWB were as follows: 1967/1968 

(precipitation 706 mm), 1970/1971 with 

693.5 mm of precipitation and 1976/1977 

(precipitation 722.6 mm). In the above 

mentioned years precipitation was higher 

than the mean value in the perennial 

(571 mm). The following seasons had 

negative net potential climatic water 

balance 1969/1970 with 525.6 mm of 

precipitation, 1972/1973 with 425 mm of 

precipitation, 1973/1974 with 431.4 mm 


TABLE 1. Basic statistical charateristics of evapotranspiration and precipitation in Wrocław-Swojec 

(1964–2000) 

Years 

1964– 

–2000 

1980– 

–2000 

season 

P 

(mm) 

σ p 

Linear regression 

equation for 

precipitation 

Tendency 

mm/ 

/10 year 

ETP 

(mm) 

Linear regression 

equation for 

evapotranspiration 

Tendency 

mm/10 year 

X–IX 571 92 y = –1.8x+604 –18* 746 y = 4x+670 40** 69 

I–XII 570 95 y = –1.9x+606 –19* 745 y = 3,9x+672 39** 73 

X–III 200 55 y = –0,5x+210 –5 180 y = 2,8x+128 28** 45 

X–IX 550 83 y = –2.3x+580 –23* 777 y = 5,7x+670 57** 73 

I–XII 555 104 y = –3.6x+600 –36* 772 y = 6,9x+640 69** 79 

X–III 196 52 y = –0,7x+205 –7 204 y = 1,4x+190 14** 38 

P – mean precipitation, σ p – standard deviation for P, ETP-mean evapotranspiration 

σ ETP – standard deviation for ETP 

**) – statistically signifi cant for α = 0.01 


σ ETP 

of precipitation, 1977/1978 with 599.2 

mm of precipitation and 1983/1984 

with 544.3 mm of precipitation. In other 

seasons of the perennial 1964-1986 

potential climatic water balance was 

positive until April, whereas after that 

period became negative. Since 1986/1987, 

when net potential climatic water balance 

was positive, in every coming year water 

defi cit was growing. Even in 1997, the 

year of fl ood in Wrocław, precipitation in 

July did not manage the water shortage 

to vanish. Deepening water defi cits are 

clearly seen in the last two decades of the 

XXth century (Fig. 5a, b, c, d, e, f). 

Water defi cit or excess during 

vegetation period have signifi cant 

infl uence on the values of heat balance 

components. It follows from the rule 

of evapotranspiration priority, that if 

there is suffi cient amount of water in the 

substrate, the excess of energy is used 

for evapotranspiration prior to heating 

air and soil. Therefore, this rule connects


a 

b 

c 

d 

150 

100 

50 

0 

-50 

100 

50 

0 

-50 

-100 

-150 

-200 

-250 

0 

-100 

-200 

-300 

-400 

-500 

0 

-50 

-100 

-150 

-200 

-250 

1986/1987 

X X-XI X-XII X-I X-II X-III X-IV X-V X-VI X-VII X-VIII X-IX 

1987/1988 


1988/1989 


1996/1997 

X X-XI X-XII X-I X-II X-III X-IV X-V X-VI X-VII X-VIII X-IX

e 

f 

100 

0 

-100 

-200 

-300 

50 

0 

-50 

-100 

-150 

-200 

-250 

-300 

climatic water balance and heat balance. 

Thus, in the following investigations 

heat balance components and their 

contribution in net radiation were 

calculated for ecosystems of coniferous 

forest, potatoes and spring wheat (Fig. 

6, 7, 8). The consequences of deepening 

water defi cits in Wrocław-Swojec during 

the last 20 years are as follows : 

1. Decreasing latent heat fl ux used for 

evapotranspiration; 

2. Slow increase of sensible heat fl ux 

used for heating atmosphere. 

Increasing values of Bowen’s ratio 

(H/LE) [Bowen 1926] for coniferous 

forest, potatoes and spring wheat prove 

sensible heat fl ux enlargement in the last 

20 years (Fig. 9). The indicated above 

tendencies suggest the climate warming 

up in this area. 

1997/1998 



1999/2000 


FIGURE 5. Changes of potential climatic water balance (PCWB) in Wrocław-Swojec for the period 

X–IX in years 1986–2000 

CONCLUSIONS 

1. 

2. 

3. 

The 1964–2000 perennial in Wrocław- 

-Swojec is characterized by increased 

yearly sums of potential 

evapotranspiration and decreasing 

trend in yearly precipitation sums. 

The analysis of a yearly course of 

potential cumulative climatic water 

balance in the season from October 

to September revealed growing 

water defi cits in this region, more 

evident in the last decades of the 

XXth century. 

Growing water defi cits in this area 

are the cause of increasing sensible 

heat fl ux in the vegetation period 

for coniferous forest, spring wheat 

and potatoes, as well as decreasing 

tendency for latent heat fl ux.


FIGURE 6. Variation of mean ten-days values of sensible heat fl ux (H) during the growing season four 

coniferous forest, potatoes and spring wheat in Wrocław-Swojec (1964–2000) 

FIGURE 7. Variation of mean values of sensible heat fl ux and net radiation ratio (H/Rn) during 

the growing season for coniferous forest, potatoes and spring wheat in Wrocław-Swojec (1964–2000)


FIGURE 8. Variation of mean values of latent heat fl ux and net radiation ratio (LE/Rn) during the growing 

season for coniferous forest, potatoes and spring wheat in Wrocław-Swojec (1964–2000) 

FIGURE 9. Variation of mean ten-days values of the Bowen’s ratio (H/LE) during the growing season 

for coniferous forest, potatoes and spring wheat in Wrocław-Swojec (1964–2000)


4. 

Due to increasing values of sensible 

heat fl ux used for heating the 

atmosphere, the air temperature in 

this region rose 0,03 o C per year, 

which means warming up in this 

area. 

REFERENCES 

BAC S., ROJEK M. 1977: Metodyka oceny 

stosunków wodnych obszarów rolniczych 

na podstawie danych klimatycznych, 

Zesz. nauk ART Olszt. Nr 21: 13–24. 

BAC S., ROJEK M. 1979: Klimatyczny 

bilans wodny a odpływy w Polsce, Przegl. 

Geofi z. 24(3): 293–297. 

BAC S., ROJEK M. 1982: Klimatyczne 

bilanse wodne w Polsce. [w:] Bac S. (red.) 

Agroklimatyczne podstawy melioracji 

wodnych w Polsce. PWRiL, Warszawa. 

BOWEN I.S. 1926: The ratio of heat losses 

by conduction and by evaporation from 

any water surface. Phys. Rev., 27, p. 

779–787. 

KĘDZIORA A. 1999: Podstawy agrometeorologii, 

PWRiL, Poznań. 

OLEJNIK J., KĘDZIORA A. 1991: A model 

for heat and water balance estimation and 

its application to land use and climate 

variation., Earth Surface Processes 

Landforms vol. 16 ss. 601–617. 

PENMAN H.L. 1948: Natural evaporation 

from open water, bare soil and grass. 

London: Proc.Royal Soc. Vol. 193, 120– 

146. 

PENMAN H.L. 1950: Evaporationover the 

BritishIsles, Q.J. Roy. Met. Soc., 76 

372–83. 

PENMAN H.L. 1956: Evaporation: an 

introductory survey, Netherlands J. Agric. 

Sci. 4, –29. 

PENMAN H.L. 1963: Vegetation and 

Hydrology, Tech. Comm. Nr 53, Comm. 

Bur. of Soils, Harpenden. 

ROJEK M.M., WIERCIOCH T. 1995: 

Zmienność czasowa i przestrzenna 

parowania wskaźnikowego , ewapotranspiracji 

aktualnej i niedoborów opadowych 

w Polsce nizinnej w okresie 1951– 

–1990, ZNAR Nr 268, Monografi e VI. 

ROJEK M. 1994: Time variablity of climatic 

water balances in selected meteorological 

stations in Poland. Zesz. Probl. Post. 

Nauk Rol., 405, 147–153. 

ROJEK M., WIERCIOCH T. 1994: Indicatory 

evapotranspiration os summer half-year 

in various long-term periods. Zesz. Probl. 

Post. Nauk Rol., 405, 155–161. 

Streszczenie: Zmienność klimatycznego bilansu 

wodnego i bilansu cieplnego różnych ekosystemów 

dla Wrocławia w latach 1964–2000. Praca zawiera 

charakterystykę składowych bilansu klimatycznego 

bilansu wodnego we Wrocławiu w latach 1964– 

–2000. W analizowanym okresie obserwuje się 

trend wzrostu rocznych sum ewapotranspiracji 

potencjalnej oraz spadku rocznych sum opadów 

atmosferycznych. W związku ze wzrostem strumienia 

ciepła jawnego, ogrzewającego atmosferę 

obserwuje się wzrost średniej rocznej temperatury 

powietrza o 0,03°C. Ten wzrost temperatury 

związany jest ze wzrostem defi cytu wody dla lasu 

liściastego, pszenicy jarej i ziemniaków oraz rosnącym 

strumieniem ciepła jawnego przeznaczonego 

na ogrzanie atmosfery. 

MS received December 2006 

Authors address: 

Elżbieta Musiał, Joanna Bubnowska, 

Edward Gąsiorek 

Department of Mathematics 

Agricultural University of Wrocław 

ul. Grunwaldzka 53, PL-50357 Wrocław 

Polska 

The research supported by KBN grant in 

years 2004–2007.




Investigation for biological nitrogen removal from wastewater using 

simultaneous nitrifi cation/denitrifi cation technology 

GIEDRÉ VABOLIENÉ 

Department of Water Supply and Management, Vilnius Gediminas Technical University, Lithuania 

Abstract: Investigation for biological nitrogen 

removal from wastewater using simultaneous 

nitrifi cation/denitrifi cation technology. Biological 

nitrogen removal from the wastewater is based on 

the nitrifi cation and denitrifi cation processes at 

the biological treatment plant with the activated 

sludge. Different technological schemes can 

be used for the above mentioned processes. 

“BioBalance” technology as the newest way 

of nitrogen and phosphorus removal has been 

applied at Utena Wastewater Treatment Plant. 

“BioBalance Symbio” technology for the nitrogen 

removal is based on the active sludge technology 

with simultaneous nitrifi cation/denitrifi cation. 

The aeration zone and the anoxic zone are in 

one tank. The nitrifi cation and denitrifi cation are 

carried out during the aeration switching on and 

off. Nitrifi cation and denitrifi cation processes 

have been estimated during fi ve experiments in 

the aeration tanks, when durations of aeration 

and low aeration were from 120 to 180 min. The 

impact of aeration regime on biological nitrogen 

removal has been estimated during this scientifi c 

work. 

Key words: biological nitrogen removal, 

nitrifi cation, denitrifi cation, nitrate, ammonium 

nitrogen, biological active potential (BPA). 

INTRODUCTION 

Biological nitrogen removal from the 

wastewater is based on the nitrifi cation 

and denitrifi cation processes at the 

biological treatment plant with the 

activated sludge (Berţinskienë 1999). 

The nitrifi cation is carry out by the 

nitrifi cation bacteria, which are divided 

into two groups. The fi rst group 

bacteria (Nitrosomonas, Nitrosospira, 

Nitrosococcus, Nitrosolobus genera) 

oxidise the ammonium hydrate to nitrite. 

The second group bacteria (Nitrobacteria, 

Nitrospira, Nitrococcus genera) oxidise 

nitrite into nitrate (Bitton 1994). The 

energy is produced during both stages 

of the nitrifi cation. Nitrifi cation bacteria 

consume the energy for CO 2 assimilation. 

Thus nitrifi cation bacteria are autotrophic 

to the carbon. Nitrifi cation bacteria grow 

signifi cantly slower than other bacteria 

during the intensive nitrifi cation process 

(Droste 1997). 

The nitrogen cycle ends by its return 

back to the atmosphere in nature i.e. by 

denitrifi cation. Denitrifi cation is one of 

anaerobic respiration variation (nitrate 

respiration), when NO 3 ions are being 

used as a fi nal electrons acceptor for 

the oxidation of organic substances. 

The denitrifi cation is oxygen split from 

the nitrate, after that – from nitrite by 

denitrifi cation bacteria. The denitifi cation 

is being carried out by heterotrophic 

bacteria: pseudomonas, spirillium, 

tiobacillus, alkaligenes, bacillus. They 

use organic pollution, existing in the 

wastewater as carbon source. There are 

two main stages in the denitrifi cation. 

None of the stages will take place till 

there will be any suffi cient amount of the 

dissolved oxygen. The fi rst is carrying out

112 G. Vaboliené 

nitrate reduction up to nitrite; the second 

is reducing nitrite up to nitrogen gas. 

During the second stage nitrites change 

nitrate as electrons acceptors, and the 

nitrogen gas changes nitrite as a product 

to fi nish the denitrifi cation reaction. 

Different technological schemes, 

following the source of carbon oxidation 

during denitrifi cation processes, can be 

used for the above mentioned processes 

(Henze et al. 1995). The 1st scheme, 

i.e. carbon source – carbon of raw 

wastewater. The scheme is composed of 

the denitrifi cation tank, nitrifi cation tank, 

and secondary clarifi ers. The 2nd scheme, 

i.e. carbon sources – additionally added 

easily biodegradable organic substances 

(methanol, ethanol, etc.). The scheme 

is composed of the nitrifi cation tank, 

denitrifi cation tank, additional aeration 

tank, and secondary clarifi ers. The 3rd 

scheme, i.e. carbon source – carbon 

of raw wastewater. According to this 

scheme the aeration zone and the anoxic 

zone are in one tank. The nitrifi cation 

and the denitrifi cation is carry out 

during the aeration switching on and off 

(Matuzevičius et al. 1998). 

Usually schemes of the nitrogen 

removal are combined with phosphorus 

removal schemes. Different new 

technologies are used to decrease 

phosphorus and nitrogen quantity in 

the wastewater. The mostly advanced is 

“BioBalance” technology, where nitrogen 

is removed following the 3rd nitrogen 

removal scheme, and the anaerobic 

zone is equipped before the nitrifi cation/ 

denitrifi cation tank the phosphorus 

biological removal. The nitrogen removal 

according to “BioBalance Symbio” 

technology carried out in a simultaneous 

way of nitrifi cation and denitrifi cation; 

controlled the oxygen supply by NADH 

sensor according to sludge activity. 

Commonly, there exist co- 

-ferments that function as suppliers 

of hydrogen and electrons and are 

nicotinamideadeninedinucleotide NAD + 

and its phosphoresced compound 

derivative NADP + . Reduced NAD(P)H 

form, can be oxidised once more during 

the formation of ATP. During the 

processes of metabolism co-ferments 

NADH and NADPH are produced in 

all microorganisms. NAD(P)H quantity 

in the microorganisms produced 

depend on their activity. The abovementioned 

activity depends on existing 

sludge loads, supplied in the form of 

nutritious matters. This means that, 

for example, NAD(P)H quantity will 

remain stable under the conditions of 

constant quantity of nutritious matters 

supplied to the active sludge system, 

the constant load and sludge activity. 

The sludge activity and at the same time 

NAD(P)H production increase gradually 

at the same time with an increase of the 

sludge load because of a more intensive 

supply of nutritious matters. Differently, 

NAD(P)H production decreases during a 

decrease of the nutritious matter supply. 

During research this was determined 

by linear dependences (Norgard et 

al. 1996). “BioBalance” measuring 

equipment NADH fl uorescensor is 

based on the process mentioned above. 

The fl uorecsensor fi xes only NADH 

that gives information on how much 

energy the microorganisms possess. The 

information on the energy is defi ned as the 

biological active potential (BPA). NADH 

fl uorescensor controls denitrifi cation 

and nitrifi cation process according to 

the sludge activity in a single aeration

tank. Ammonia is oxidised to nitrates by 

nitrifying bacteria during the nitrifi cation 

phase. The nitrates are reduced to 

molecular ni trogen (N 2) by denitrifying 

bacteria during the denitrifi ca tion 

phase. These processes are carried out 

periodically. Organic matter is oxi dised 

bacterially during both the nitrifi cation 

and denitrifi cation phases, with oxygen 

and nitrates, respectively, as the oxidising 

agents and phosphorus is absorbable with 

special microorganisms. 

The aim of the work was to evaluate the 

impact of aeration regime on biological 

nitrogen removal from wastewater using 

simultaneous nitrifi cation/denitrifi cation 

technology. 

MATERIALS AND METHODS 

Researches have been carrying out 

during period (July–November, 2005) 

at Utena Wastewater Treatment Plant 

in Lithuania. Two aeration tanks have 

been used for the biological wastewater 

treatment in Utena wastewater treatment 

plant. The nitrogen removal according to 

“BioBalance Symbio” technology carried 

out in a simultaneous way of nitrifi cation 

and denitrifi cation; controlled the oxygen 

supply by NADH sensor according to 

sludge activity. 

Investigation for biological nitrogen removal... 113 

Two aeration tanks worked at the same 

time. Wastewater-fl ow after mechanical 

treatment was divided into both aeration 

tanks simultaneously. However, the 

regime of aeration in each aeration tank 

was different. Aeration can be carried 

out in one aeration tank, in the meantime 

low aeration can be carried out in the 

other aeration tank. The term of aeration 

and low aeration can be different in both 

aeration tanks. Because of the abovementioned 

reasons, this research’s 

results were analysed of both aeration 

tanks. Low oxygen concentration 

0.1–0.5 mg O 2 l –1 has been held until the 

denitrifi cation carries out and the higher 

oxygen concentration 0.5–1 mg O 2 l –1 

has been maintained for the nitrifi cation 

processes carry out periodically. 

Conditions for the nitrifi cation process 

were held from 120 to 180 min, and 

after that 120–180 min were spent for 

the denitrifi cation. The duration of the 

nitrifi cation and the denitrifi cation had 

been held in fi ve different regimes during 

experiments (Tab. 1). The researches 

from 2 to 9 occasions in each aeration 

tank have been carrying out during 

different aeration regime. 

The total nitrogen concentration 

in wastewater after the mechanical 

treatment, ammonium nitrogen and 

nitrate concentration in the beginning of 

aeration and by the end of aeration rate 

TABLE 1. The duration of the nitrifi cation and the denitrifi cation during experiments 

Number of experiment The duration of aeration rate, min The duration of low aeration rate, min 

1 150 150 

2 180 150 

3 

180 

120 

120 

120 

4 120 180 

5 150 180


in the both aeration tanks, total nitrogen 

concentration in wastewater after 

biological treatment had been measured 

during all experiments. The active sludge 

concentration and the volatile suspended 

solids of active sludge in the aeration tank 

had been measured. The average fl ow in 

the aeration tank had been fi xed. The 

effi ciency of total nitrogen removal and 

the active sludge load in the aeration tank 

had been estimated. Total nitrogen had 

been estimated adding Kjeldahl nitrogen 

to nitrites and nitrates nitrogen. All 

analysis was carried out using standard 

methods (LST ISO 5815:1989). 

RESULTS 

The total nitrogen concentration in 

wastewater after the mechanical treatment 

fl uctuated from 37 to 52 mg/l, average 44 

mg/l during fi ve experiments (Fig. 1). 

The fi rst experiment carries out 21 

day during period (July 19 – August 8). 

Conditions for the nitrifi cation carried 

Total nitrogen, [mgN/l] 

55 

50 

45 

40 

35 

30 

out were held 150 min, and 150 min – for 

the denitrifi cation. The researches were 

carrying out 9 times in the both aeration 

tanks. The gathered results proved that 

the effi ciency of total nitrogen removal 

fl uctuated from 87 to 91%, average 88%. 

At start of experiment, the concentration 

of ammonium nitrogen in the beginning 

of aeration rate fl uctuated from 0.84 to 1.4 

mg/l, by the end of aeration rate – from 

0.47 to 0.91 mg/l (Fig. 2). However, the 

concentration of ammonium nitrogen 1 

mg/l was estimated in one aeration tank 

by the end of aeration rate after 7 day of 

experiment start. 

After that, aeration tanks worked at 

mentioned above aeration regime yet one 

week. Then it was estimated that partial 

nitrifi cation was carried out in the both 

aeration tanks by the end of aeration rate. 

The concentration of ammonium nitrogen 

by the end of aeration rate only increased 

from 1.4 to 2 mg/l till experiment was 

fi nished. Following experimental results 

can be seen, that the duration of aeration 

was not enough. The effi ciency of total 

1 2 3 4 5 6 7 8 9 

aeration and reduced aeration rate 150 min 

aeration rate 180 min, reduced aeration rate 150 min 

aeration rate 180, 120 min, reduced aeration rate 120 min 



FIGURE 1. Total nitrogen concentration in wastewater after the mechanical treatmen

a 

b 

N-NH4, N-NO3,[mg/l] 

N-NH4, N-NO3[mg/l] 

2,5 

1,5 

0,5 

nitrogen removal decreased from 91% at 

experiment start to 87% at the experiment 

fi nish. The concentration of nitrate was 

changing in small interval during fi rst 

experiment: in the beginning of aeration 

rate 0.1÷0.6 mg/l, by the end of aeration 

rate 0.45÷1.2 mg/l (Fig. 2). A low 

concentration of nitrate in the beginning 

of aeration rate provided that 150 min. 

reduced aeration rate was enough for 

complete denitrifi cation carry out. 

3 

2 

1 

0 

2,5 

2 

1,5 

1 

0,5 


1 2 3 4 5 6 7 8 9 10 

N-NH4 in the beginning of aeration rate 

N-NO3 in the beginning of aeration rate 

N-NH4 by the end of aeration rate 

N-NO3 by the end of aeration rate 

0 

1 2 3 4 5 6 7 8 9 





FIGURE 2. The concentration of ammonium nitrogen and nitrate in the beginning and by the end of 

aeration rate at fi rst (a) and second (b) aeration tank when aeration regime: aeration and reduced aeration 

rate – 150 min 

Consequently results of fi rst 

experiment indicated the condition 

of second experiment. The second 

experiment carries out 11 day during 

period (August 9–19). Conditions 

for the nitrifi cation carried out were 

held 180 min, and 150 min – for the 

denitrifi cation. It was extended duration 

of aeration for complete nitrifi cation. 

The researches were carrying out 8 times 

in the both aeration tanks. The obtained 

results proved that the effi ciency of


total nitrogen removal fl uctuated from 

83 to 89%, average 85%. During two 

days of the second experiment the 

concentration of ammonium nitrogen in 

the beginning of aeration rate decreased 

from 3,1 to 0.8 mg/l (Fig. 3). Process 

of nitrifi cation carries out completely. 

Duration of aeration was enough till the 

end of experiment. The concentration of 

nitrate changed more than during fi rst 

experiment. The concentration of nitrate 

by the end of aeration rate exceeded 1 

mg/l several time. The effi ciency of 

a 

b 

N-NH4, N-NO3, [mg/l] 


3,5 

3 

2,5 

2 

1,5 

1 

0,5 

0 

2,5 

2 

1,5 

1 

0,5 

0 

total nitrogen removal decreased from 

89% at experiment start to 83% at the 

experiment fi nish. 

The third experiment carries out 

during 7 days period (August 22–28). 

Aeration tanks worked in follow aeration 

regime: aeration rate – 180 min, reduced 

aeration rate – 120 min fi rst 5 day of 

experiment. Later aeration regime 

was changed: aeration rate – 120 min, 

reduced aeration rate – 120 min. The 

researches were carrying out 5 times in 

fi rst aeration regime and 2 times in second 

1 2 3 4 5 6 7 8 





1 2 3 4 5 6 7 8 






aeration rate at fi rst (a) and second (b) aeration tank, when aeration regime: aeration rate 180 min, reduced 

aeration rate – 150 min

aeration regime in the both aeration 

tanks. The gathered results proved that 

the effi ciency of total nitrogen removal 

fl uctuated from 73 to 80%, average 77%. 

Though the concentration of ammonium 

nitrogen in the beginning of aeration rate 

fl uctuated from 0.13 to 2.2 mg/l, but it 

already decreased and fl uctuated from 

0.01 to 0.96 mg/l by the end of aeration 

rate (Fig. 4). However nitrate increased 

signally by the end of aeration rate. The 

concentration of nitrate was estimated 

from 0.47 to 11 mg/l in the beginning 

of aeration rate and 1.2÷17 mg/l by 

a 

N-NH4, N-NO3 , [mg/l] 

18 

16 

14 

12 

10 

8 

6 

4 

2 

0 


the end of aeration rate. The reduced 

aeration rate was preferred for complete 

denitrifi cation, so the effi ciency of total 

nitrogen removal decreased to 73%. 

Then the aeration regime was changed 

and aeration tanks worked in other 

regime (aeration rate – 120 min, reduced 

aeration rate – 120 min) last 2 day of 

experiment for reduction of nitrate. Then 

aeration rate was changed from 180 

min to 120 min concentration of nitrate 

decreased to 0.75 mg/l in the beginning 

and by the end of aeration rate during 2 

1 2 3 4 5 6 7 





b 

18 

16 

14 

12 

10 

8 

6 

4 

2 

0 

1 2 3 4 5 6 7 






aeration rate at fi rst (a) and second (b) aeration tank, when aeration regime: aeration rate 180, 120 min, 

reduced aeration rate – 120 min 

N-NH4, N-NO3, [mg/l]


day of experiment. The effi ciency of total 

nitrogen removal increased to 73%. 

Next experiment was carrying 

out interchanging aeration regime. 

Aeration rate reduced from 180 min. 

to 120 min., and reduced aeration rate 

prolonged from 120 min. to 180 min. 

for reduction of nitrate. The fourth 

experiment carries out 7 day during 

period (August 29–September 4). The 

researches were carrying out 7 times in 

the both aeration tanks. The gathered 

results proved that the effi ciency of total 

nitrogen removal fl uctuated from 73 to 

a 

b 



7 

6 

5 

4 

3 

2 

1 

0 

10 

8 

6 

4 

2 

0 

88%, average 83%. The nitrifi cation 

carried out completely fi rst fi ve day 

of experiment. The concentration of 

ammonium nitrogen fl uctuated from 

0.31 to 1.4 mg/l, 0.02–0.7 mg/l in the 

beginning and by the end of aeration rate 

respectively (Fig. 5). However later (two 

last day of experiment) the concentration 

of ammonium nitrogen increased in the 

beginning and by the end of aeration 

rate. Only partial nitrifi cation carried out. 

Consequently concentration of nitrate 

decreased progressively and decreased 

until 0.5, 0.31 mg/l in the beginning of 

1 2 3 4 5 6 7 





1 2 3 4 5 6 7 








aeration rate at the each aeration tanks 

respectively and 0.89 and 0.33 mg/l by the 

end of aeration rate at the each aeration 

tanks respectively during third day of 

experiment. Enough duration of reduced 

aeration for completely denitrifi cation 

had positive impact on effi ciency of total 

nitrogen removal, which increased to 

88% later decreased to 82%. 

The fi fth experiment carries out 

74 day during period (September 5 

–November 18). Conditions for the 

nitrifi cation carried out were held 150 

a 

b 



5 

4 

3 

2 

1 

0 

4 

3,5 

3 

2,5 

2 

1,5 

1 

0,5 

0 


min, and 180 min – for the denitrifi cation. 

The researches were carrying out 7 times 

in the both aeration tanks. The gathered 

results proved that the effi ciency of 

total nitrogen removal fl uctuated from 

80 to 89%, average 85%. Following 

concentration of ammonium nitrogen 

0.44–4.6 mg/l and 0.06÷0.65 mg/l in 

the beginning and by the end of aeration 

rate respectively, nitrifi cation process 

carry out completely (Fig. 6). However 

concentration of nitrates (3.8 mg/l by 

the end of aeration rate) showed partial 

1 2 3 4 5 6 7 





1 2 3 4 5 6 7 









denitrifi cation process. It is proved that 

the duration of reduced aeration was not 

enough for denitrifi cation. The effi ciency 

of total nitrogen removal decreased from 

89% at experiment start to 80% at the 

experiment fi nish. 

The effi ciency of total nitrogen 

removal fl uctuated from 73 to 91% 

during fi ve experiment when aeration 

tank worked in different aeration regime 

(Fig. 7). The total nitrogen concentration 

in wastewater after the biological 

treatment fl uctuated from 3.9 to 12 mgN/ 

/l during fi ve experiments. 

CONCLUSIONS 

1. 

2. 

Using biological nitrogen removal 

technologies, when nitrifi cation 

and denitrifi cation are in process 

changing aeration intensiveness, it 

is very important properly evaluate 

aeration and low aeration duration. 

When total nitrogen concentrations 

of mechanically treated wastewater 

ranged from 37 till 52 mgN/l, it 

will be proper to use two aeration 

Efficiency of nitrogen removal,% 

95 

90 

85 

80 

75 

3. 

4. 

regimes: 150 minutes aeration and 

150 minutes low aeration rates, 

similarly 180 minutes aeration and 

150 minutes low aeration rates. 

During 120 minutes conditional 

short aeration rate and prolonged low 

aeration rate, presently ammonium 

nitrogen increase, following partial 

denitrifi cation outcome. Working in 

this regime it is possible to obtain 

completed nitrifi cation. 

During experimental research it was 

obtained, when aeration and low 

aeration rates are similar or slightly 

different approximately 30 minutes, 

nitrogen removal effectiveness is the 

biggest. 

ACKNOWLEDGEMENT 

This research study is dedicated to EC 

fi nanced FP6 project MAPO „Enhancing 

Research and Development Projects 

to fi nd Solutions to Struggle against 

various Marine Pollutions”. The author is 

involved into deliverable on innovation 

roadmap for technologies based on 

70 

1 2 3 4 5 6 7 8 9 

aeration and reduced aeration rate 150 min 


aeration rate 180, 120 min, reduced aeration rate 120 min 



FIGURE 7. The effi ciency of total nitrogen removal when aeration tank worked in different aeration 

regime

gaps within current state-of-the-art. 

I’m gratefully to all Small Medium 

Enterprises participated into MAPO 

project activities; more Small Medium 

Enterprises working into related to 

marine pollutions areas are welcome to 

join future activities. 

REFERENCES 

BERŢINSKIENË J. 1999: Water 

microbiology. Textbook Vilnius: 

Technika, p. 144. 

HENZE M., HARREMOES P., JANSEN C., 

ARVIN E. 1995: Wasterwater Treatment. 

Biological and Chemical Processes. 

Springer-Verlag, p. 383. 

BITTON G. 1994: Wastewater microbiology. 

Wiley-Liss, New York, p. 456. 

DROSTE R. 1997: Theory and practice of 

water and wastewater treatment, p. 954. 

MATUZEVIČIUS A., PAULAUSKIENË 

Z. 1998: Experimental researches of 

phosphorus and nitrogen consumption 

and nitrogen removal from the 

wastewater in the biological treatment 

plant. 3rd International Conference. 

Cities Engineering and Environment, 

VGTU, “Technika”, p. 137–142. 

NORGARD P., HELMO K., SORENSER E. 

1996: Purifi cation process for nitrogen 

removal controlled by NADH. Vand og 

Jord, Danish, vol 3, p. 126–129. 

TSAI M.W., WENZEL M.C., EKAMA G.A. 

2003: The effect of residual ammonia 

concentration under aerobic conditions 

on the growth of Microthrix parvicella in 

biological nutrient removal plants. Water 

Research 37, p.p. 3009–3015. 

HATZICONTINOU G.J., ANDREADKIS A. 

2002: Differences in nitrifi cation potential 

between fully aerobic and nitrogen removal 

activated sludge systems. Water Science & 

Technology. Vol. 46, No 1/2, p. 297–189. 

JENICEK P., SVEHLA P., ZABRANSKA J., 

DOHAYOS M. 2004: Factors affecting 

nitrogen removal by nitrifi cation/ 


/denitrifi cation. Water Science & 

Technology. Vol. 49, No 5/6, p. 73–79. 

Lithuanian Ministry of Environment Protection. 

1994: Unifi ed methods of the wastewater 

and surface water quality researches. 

Chemical analysis methods. Part I. 224 p. 

Lithuanian Ministry of Environment 

Protection, 2002: LAND 47-1:2002. 

(ISO 5815:1989), 17 p. 


Protection, 2000: LAND 32-2000.10 p. 


Protection, 2005: LAND 66-2005. 7 p. 


Protection, 2005: LAND 58:2003. 24 p. 

Streszczenie: Badania biologicznego usuwania 

azotu ze ścieków za pomocą technologii jednoczesnej 

nitryfi kacji i denitryfi kacji. Metoda biologicznego 

usuwania azotu ze ścieków oczyszczanych 

w komorach osadu czynnego opiera się na 

procesach nitryfi kacji i denitryfi kacji. Do wyżej 

wymienionych procesów można stosować różne 

ciągi technologiczne. Technologia „BioBalance” 

zaliczana do najnowszych metod usuwania azotu 

i fosforu została wdrożona w oczyszczalni ścieków 

w Utenie (Litwa). Technologia „BioBalance 

Symbio” do usuwania azotu wykorzystuje metodę 

osadu czynnego z jednoczesnym prowadzeniem 

procesów nitryfi kacji i denitryfi kacji. Strefy tlenowa 

i beztlenowa znajdują się w jednej komorze. 

Procesy nitryfi kacji i denitryfi kacji są prowadzone 

poprzez naprzemienne włączanie i wyłączanie 

napowietrzania komory. W pracy przedstawiono 

wyniki badań przebiegu procesów nitryfi kacji 

i denitryfi kacji podczas pięciu cykli badań w komorze 

napowietrzania w warunkach intensywnego 

i wolnego napowietrzania. W pracy oszacowano 

wpływ warunków napowietrzania na skuteczność 

usuwania biologicznego azotu. 

MS. received July 2006 

Author’s address: 

Giedré Vaboliené 

Vilnius Gediminas Technical University, 

Dept. of Water Supply and Management, 

Saulëtekio al. 11, AIF, LT-10228, Vilnius-40, 

Lithuania. 

e-mail: giedre.v@freemail.lt




Development of technologies used in agricultural engineering work 

on an example of selected stages of land consolidation process 

URSZULA LITWIN*, JAROSŁAW JANUS*, MARIUSZ ZYGMUNT** 

*Department of the Geodesical Arrangement of Rural Settlements 

**Department of Geodesy, Agricultural University of Cracow 

Abstract: Development of technologies used in 

agricultural engineering work on an example of 

selected stages of land consolidation process. 

The land consolidation process is one of the most 

time-consuming and complicated agricultural 

engineering tasks performed by geodesists. Their 

automation and computerisation is not as advanced 

as for other types of geodetic works. The reason 

for that is a limited number of works that have 

been performed in this range since the mid 1980s, 

i.e. during the period of intensive IT (Information 

Technology) implementation in geodesy. At 

present one can observe again the growth in 

interest in performance of land consolidations 

processes. The purpose of this elaboration is to 

introduce the latest solutions IT applied in land 

consolidation process on an example of the land 

consolidation in Wojków village. 

Three characteristic stages of the process are 

presented herein: creation of assessment register 

before consolidation, preliminary selection of plots 

for designed blocks and designing. Proposal for a 

method for automation of the above-mentioned 

activities using a digital map developed by means 

of Bentley MicroStation software, retaining the 

requirements imposed by land consolidation 

instruction, was discussed. 

The above-mentioned stages are illustrated 

by examples from the land consolidation carried 

out in Wojków village in years 2001–2004. The 

work was performed on these premises using the 

technology presented herein. It allowed reduction 

of the time of land consolidation process by 

approx. 30%. 

Key words: spatial structure of village, land 

consolidation. 

INTRODUCTION 

The effi cient use of rural areas has 

become one of the key problems in 

the age of Poland’s membership of the 

European Union. To use the potential 

of rural areas effectively, among other 

things a signifi cant improvement in the 

structure of the areas is required. One of 

the problems to be solved is an excessive 

dispersion and fragmentation of farms 

[Urban 1981, Hopfer and Urban 1984, 

Harasimowicz 1995]. The stagnation in 

performance of the land consolidation 

process that occurred in the 1980s 

has intensifi ed the problem. Only the 

increased interest in land consolidation 

process and related prospects for 

signifi cant growth in the number of 

consolidations resulted in appearance 

of the fi rst IT tools to automate some of 

the process’s stages [Janus and Zygmunt 

2005]. This project describes selected 

functionalities of the system supporting 

modern land consolidation process, 

which is the only system in Poland to 

treat this process in a comprehensive 

manner.

124 U. Litwin, J. Janus, M. Zygmunt 

CREATION OF DESIGNED 

BLOCKS 

The land consolidation process the new 

system was tested on was carried in 

Wojków village, Mielec district in years 

2001–2004. The area of the consolidated 

land was 670 ha and comprised of 1503 

plots. There was no digital map for the 

area, therefore the fi rst thing was to 

convert the contents of analogue map 

into the digital form. Bentley Systems 

software was used for this purpose. The 

digital map was checked for correctness 

by the software the purpose of which was 

to pick up any topological errors. Another 

stage of the check was to compare a set 

of plots on the map to the plots included 

in the land register survey. The checks 

ensured that register of lands before 

consolidation was created properly. To 

facilitate the work, the software had the 

functionality of semi-automatic creation 

of maps with comparative assessment. 

The borderlines of assessment 

contours occurred as a result of automatic 

FIGURE 1. Division of object into 9 parts 

selection of appropriate borders of plots 

and facilities, such as: roads, rivers and 

railway lines. For a digital map prepared 

like that the software automatically 

generates register of lands before 

consolidation. The register is prepared 

so that it can be printed in A3 format 

[Litwin and Zygmunt 2005]. 

To enable multi-station operation 

the object was divided into 9 designed 

areas. The software has functionalities 

allowing this operation to be performed 

automatically (Fig. 1). 

Each of the newly created designed 

areas was then divided into designed 

blocks (Fig. 2). Division into blocks may 

also be carried out automatically. 

Information on how the lines of 

assessment contours run was used to defi ne 

boundaries of designed blocks. Relevant 

function of the software defi nes a new 

boundary upon the assessment contour 

number is indicated. When designing 

the technology, complete latitude in 

defi nition of the boundaries of designed 

blocks was adopted. The only condition

to be met by the designed block’s 

boundary is topological correctness. The 

topological correctness is controlled at 

the operator’s request and any errors are 

immediately communicated with their 

precise location. 

PRELIMINARY SELECTION OF 

PLOTS FOR DESIGNED BLOCKS 

Another stage of a land consolidation 

process that has been automated in a 

comprehensive manner is the preliminary 

selection of plots for designed blocks. 

This procedure is one of more important 

stages of a land consolidation process. It 

is performed after the land assessment 

register is approved and before the stage 

of designing a new land arrangement is 

commenced. The purpose of the stage 

is to make the initial assignment of 

lands in individual farms to dedicated 

designed blocks. The number and 

sizes of the blocks most often result 

from the run of boundaries of designed 

invariants, such as boundaries of roads, 

Development of technologies used in agricultural engineering... 125 

FIGURE 2. Division of designed blocks within area no 01-Wojków 

watercourses, administrative boundaries 

and boundaries of building sites. The 

essential problem that occurs during this 

operation is the necessity of considering 

all participants’ requests regarding the 

location of the newly divided lands. The 

appropriate performance of preliminary 

selection activities is impeded by the 

following factors: 

– large number of participants 

– 

in the process (in case of land 

consolidations carried out over large 

areas it may reach a few thousands), 

most often exceeding the number of 

separated designed blocks by several 

dozen times; 

necessity of considering diversifi 

cation in designed blocks of classes 

and arable lands when collecting 

requests; 

– necessity of as even collection of 

requests for individual blocks as 

possible so that further design of a 

new land arrangement in accordance 

with this selection would be 

feasible.


To facilitate this operation the process 

participants submit their requests in 

more than one version to increase the 

likelihood of proposing such a land 

arrangement that the expectations of 

most participants could be come up. 

In the time of domination of analogue 

technologies, this stage of the process 

was extremely time-consuming and 

any errors made during the stage were 

decisive for any other operations carried 

out during the whole land consolidation 

process. In particular, making changes 

to the already submitted requests and 

control of acquired data took a lot of 

time. 

The purpose of the solutions designed 

is a comprehensive support for the initial 

selection stage, in particular: 

– supporting the negotiations with 

parties to the land consolidation 

process by visualisation of the 

farm’s lands against the demarcated 

boundaries of designed blocks; 

– ensuring that designed blocks are 

fi lled by value assignments as 

proposed by participants in the land 

consolidation process; 

– ensuring that the data being entered 

are verifi ed for conformity to 

requirements of the consolidation 

instruction and the land consolidation 

and exchange act; 

– immediate delivery of any 

information on the registration 

unit and designed blocks that may 

be useful on this stage of the land 

consolidation process, including 

– 

data from the assessment register 

before consolidation; 

preparation of fi nal summary of 

equivalents to be designed for every 

farm and for every block. 

The following items are used as input 

data for designed solutions: 

– data from assessment register before 

consolidation; 

– geometrical data to defi ne the 

boundaries of separated designed 

blocks; 

– layer of former condition plots; 

– layer of classifi cation contours; 

– layer of designed blocks. 

From the user’s point of view, the main 

element of the system to support the 

initial selection for designed blocks is 

the specially designed form. 

The most important components of 

this form are: 

– information panel containing, among 

other things, data on the unit, former 

condition plots included in it, data 

on area and value of the unit, and a 

lot of other information required by 

the consolidation instruction; 

– table containing information on 

collected requests for the unit; 

– fi gures and graphical information 

showing the current fi lling status of 

individual designed blocks; 

– graphical window showing the 

arrangement of former condition 

plots, lands of active farm, boundaries 

of designed blocks and classifi cation 

contours. 

The following method for collecting 

requests for a single farm was proposed: 

– the fi rst step is to select the active 

registration unit, which results in 

fi lling relevant information fi elds 

with data to describe the unit; 

– the next step is to select the value 

for which a request in its basic and 

alternative version is submitted; 

– 

for every version the appropriate 

designed block is selected;

CONCLUSIONS 

The approach to the land consolidation 

process presented herein allows 

achievement of signifi cant economic 

advantages resulting from, among 

other things, reduction in duration of 

individual operations. Choosing Bentley 

MicroStation software as a system 

platform enabled using a rich set of CAD 

Development of technologies used in agricultural engineering... 127 

– the last step is confi rmation of any tools and allowed signifi cant acceleration 

decisions made. 

of work performance. 

It is possible to cancel each of the The functions allowing quick redesign 

decisions for any farm at any time, of the plot arrangement made it possible 

which makes the specifi c land value to transfer a lot of design variants to the 

return to the appropriate designed block. land consolidation process participants, 

The fi nal effect of this stage of a land which effected in lack of complaints, 

consolidation process is generating a and thus completing the process on 

number of summaries that are required time. Simultaneously with reduction in 

further in the design work carried out duration of individual stages, conditions 

in the digital map environment and the were achieved for improvement in 

purpose of which is to design the target quality of a land consolidation process as 

land arrangement. 

regards its adjustment to the participants’ 

requests and through automatic detailed 

DESIGN FOR A PRESET VALUE 

verifi cation of the project for conformity 

to requirements of the consolidation 

instruction and the land consolidation 

Designing starts with block selection. 

and exchange act. The implementation 

It requires the number of block to 

of the discussed IT technologies for land 

be indicated by the mouse. Next the 

consolidation process in Wojków village 

designing direction is defi ned and score 

confi rmed the practical usefulness and 

and number of designed plot are entered. 

correctness of designed solutions. It also 

Data are entered using a special form. 

needs to be mentioned that individual 

Coordinates of the newly designed 

elements of the proposed method for 

plot are calculated then and boundaries 

carrying out this stage are subject to 

and number are inserted into the block. 

constant modifi cations as the described 

The designing direction may be altered 

technology is implemented in successive 

at any time. The designed boundaries 

land consolidation processes in Poland. 

may be deleted and the block that occurs 

after such a boundary is deleted may be 

subject to another division. 

REFERENCE 

HARASIMOWICZ S. 1995: Wpływ cech 

działki i gospodarstwa na wartość 

dochodową gruntów. (Infl uence of plot 

and farm features on revenue value of 

lands). Zesz. Nauk. AR w Krakowie, ser. 

Geodezja No 16, 77–86 (in Polish). 

HOPFER A., URBAN M. 1984: Geodezyjne 

urządzanie terenów rolnych. PWN. 

Warszawa (Geodetic development of 

agricultural areas) PWN, Warsawa (in 

Polish) 

JANUS J., ZYGMUNT M. 2005: Technologia 

kompleksowej automatyzacji prac


scaleniowych (Technology of comprehensive 

automation of land consolidation 

process). Materials from the 17th Scientifi 

c and Technical Session in the series 

“Current issues in geodesy and cartography”, 

Nowy Sącz (in Polish). 

LITWIN U., ZYGMUNT M. 2005: Nowa 

technologia generowania rejestru gruntów 

przed scaleniem (New technology 

for generation of land register before 

land consolidation). Zesz. Nauk. AR 

w Krakowie, ser. Geodezja No 21 (in 

Polish). 

URBAN M. 1981: Ekonomika i organizacja 

gospodarstw rolnych (Economics and 

organisation of agricutural farms), PWN, 

Warszawa (in Polish). 

Streszczenie: Rozwój technologii wykorzystywanych 

w pracach urządzeniowo-rolnych na przykładzie 

wybranych etapów scalenia gruntów. Prace 

scaleniowe należą do najbardziej czasochłonnych 

i skomplikowanych zabiegów urządzenioworolnych 

wykonywanych przez geodetów. Proces 

ich automatyzacji oraz informatyzacji nie jest tak 

zaawansowany jak w przypadku innych rodzajów 

prac geodezyjnych. Przyczyną takiego stanu 

jest ograniczenie ilości prac z tego asortymentu 

wykonywanych od połowy lat osiemdziesiątych, 

czyli w okresie intensywnego wdrażania technik 

informatycznych w geodezji. Obecnie można 

zaobserwować ponowny wzrost zainteresowania 

wykonywaniem scaleń gruntów. Celem autorów 

było przybliżenie najnowszych rozwiązań informatycznych 

stosowanych w procesie scalenia 

gruntów, na przykładzie obiektu scaleniowego 

Wojków. 

W pracy przedstawiono trzy charakterystyczne 

etapy procesu scalenia: tworzenie rejestru 

szacunkowego przed scaleniem, etap wstępnego 

naboru do kompleksów projektowych oraz 

projektowanie. Omówiono propozycję sposobu 

automatyzacji wymienionych czynności z wykorzystaniem 

mapy numerycznej opracowanej przy 

pomocy programu Bentley MicroStation, z zachowaniem 

wymogów narzucanych przez instrukcję 

scaleniową. 

Wymienione etapy zilustrowano przykładami 

pochodzącymi ze scalenia przeprowadzonego 

w latach 2001–2004 na gruntach wsi Wojków. 

Prace na tym obiekcie wykonano z zastosowanej 

przedstawionej technologii. Pozwoliło to skrócić 

czas scalenia o około 30%. 



Urszula Litwin, Jarosław Janus, 

Katedra Geodezyjnego Urządzania Terenów 

Wiejskich 

Akademia Rolnicza w Krakowie 

30-059 Kraków 

Al. Mickiewicza 24/28 

jarek@cracow.pl 

Mariusz Zygmunt 

Katedra Geodezji 

Akademia Rolnicza 

30-059 Kraków 

A. Mickiewicza 24/28 

m.zygmunt@geodezy.com.pl

Land Reclam. 37.indb - Annals of Warsaw University of Life ...

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?