Evaluating the Impact of Demographic Transition in the Context of ...

jsnds.org

Evaluating the Impact of Demographic Transition in the Context of ...

EVALUATING THE IMPACT OF DEMOGRAPHIC TRANSITION IN THE CONTEXT OF THE TOKAI-TONANKAI-NANKAI EARTHQUAKE, JAPANThen, discrimination analysis to integrate the series ofregional classifications is undertaken, and to developdemographic transition patterns, a self-sufficiencyindex is developed. With mapping using the seismicintensity data of the Tokai-Tonankai-Nankai Earthquake,analysis is conducted using GIS to discuss theimpact of demographic transition. The details of eachpart are given as follows.1. Conducting regional classificationIt is noted that detailed aging population proportioncan provide more complete family structureinformation in the regions. [16] Cluster analysis is appliedto a great amount of 1-km 2 mesh-based populationdata to acquire a regional classification.Eighteen levels of 5-year population proportions(aged 0~4 to 80~84 and aged 85+) and the variablesof each mesh are collected from the population censusdata. Repeating the same application to various populationcensus time periods, a regional classification atevery particular time period is obtained.2. Developing demographic transitionpatternsDiscrimination analysis is conducted to adjustseveral sets of regional classifications and to integratethe adjusted sets into a demographic transition pattern.With the dendrogram graph of cluster analysis results,we can render clustered meshes in GIS. Based on thespatial distribution of regional classification, we canconclude which common regional types consistentlyexist. The meshes that remain stably classified at everytime period are collected as “sample meshes” andare used as samples for discrimination analysis.For one regional type, the meshes are ruled inmesh code and equally divided into the “experimentalgroup” and the “control group.” The experimentalgroup is made to be the discriminant. And the controlgroup is used to test the reliability of the discriminantusing a true positive rate test. And the discriminantwith the best performance is applied to the study areato adjust the regional classification. Furthermore,transformation of the demographic transition patternsis discussed regarding self-sufficiency by a referencestudy.Four methods of discrimination are applied here.The details and results of each are described in Chapter3.3. Application to place evaluationThe regional exposure characteristics and regionalself-sufficiency are integrated in GIS in orderto conduct spatial analysis in the affected area of theTokai-Tonankai-Nankai Earthquake.Regional exposure characteristics are determinedby the exposure scenario numbers and the maximumseismicity intensity among these appointed scenarios.And regional self-sufficiency is based on the previoussection. Discrimination analysis is applied in the latestPopulation Census 2005 to obtain the current situation.To reach a more accurate assessment, the aboveanalysis is conducted on a 1-km 2 scale.III. Building a classification of demographictransition patternsStudy areaOsaka Prefecture and Wakayama Prefecture areselected as study areas due to the variety of regionaldevelopment, from the center of Osaka City to themid-mountainous area of the Kii Peninsula. As themost aging metropolitan area, Osaka’s populationincrease has been slowing down since 1990. Meanwhile,population decline started to accelerate in Wakayama,one of the worst prefectures in 2005. Thus,the study area is representative of a near-future demographictransition. Another reason is that the studyarea is exposed to both the Nankai Earthquake and theTokai-Tonankai Earthquake (see fig. 3).OsakaWakayama TonankaiSeismicityM 4.5- 4.9M 5.0- 5.4NankaiM 5.5- 5.9 NankaiM 6.0- 6.4M6.5 -OsakaWakayama TonankaiSeismicityM 4.5- 4.9M 5.0- 5.4M 5.5- 5.9M 6.0- 6.4M6.5 -Figure 3. Osaka and Wakayama, both exposed in thescenario of the Nankai Earthquake (L) andthe Tokai-Tonakai Earthquake (R)23


031428A / NC CT Region 3/24/11 Carolina Weekly email early due: 3/21/11 9.875” x 12.50” NAA OA02 LPI C: Randy D: Peter P: Darleneopens thursday,march 31 at 9:00 a.m.carolina pavilionIf your idea of a thrill is finding great fashionat huge savings, you’ll want to be there.Enter for a chance to win a $2,000‘Race Through the Rack’ shopping spree.*Arrive early to enter drawing. Winner will be announced at 8:45 a.m.,before the store opens. One lucky person will have 90 seconds to runthrough our store and select up to $2,000 worth of merchandise.In celebration of our opening, the first 1,000 customerswill receive a complimentary Nordstrom Rack reusabletote as a gift with purchase. One per customer.*No purchase necessary. Purchasing will not improve your chances of winning. Winner must be present at time of drawing. One shopping spree entry per person. Open to legal U.S. residents, age 18 or older. See store for official rules.Carolina Pavilion, 9563 South Blvd., Charlotte | 704.705.2161 | nordstromrack.com | facebook.com/nordstromracktwitter.com/nordstrom_rackPage 4 • March 25-31, 2011 • Matthews-Mint Hill Weekly031428A.NSO Rack Carolina Pavilion_ Charlotte_ NC.indd 1www.matthewsminthillweekly.com3/16/11 12:49 PMFinal


H. CHEN, N.MAKI, H.HAYASHIDespite 2005, the nearest neighborhood method constantlyhas a TP rate of over 99%, almost the same aslinear discrimination. Because the discriminant itselfis the population pyramid of one regional type, connectingto the idea of cohort analysis while examiningthe demographic transition from 4 pyramids (1990-2005) is clearly desirable.Yet the weighted distance is not found to improvethe TP rate of the nearest neighborhood methodin general. We believe that the weighted distancemight increase the difficulty in discriminating the lessaging areas.Above all, the pyramids of each regional type areapplied to the whole meshes of Osaka and WakayamaPrefecture to adjust the original cluster result (see fig.5) using the nearest neighborhood method. Figure 6shows the adjusted distribution of the 3 regional typesin 4 periods.Examining the adjusted resultWhile the rescheduled distance of 4 clusteranalysis is generated, another factor associated withthe difference between the 4 time periods (fig. 6) isthe transition of the study area itself. In this way, theactual urban development of Osaka and Wakayama isgiven in order to examine whether the regional typein the 4 periods can be regarded as one demographictransition.Comparing the adjusted result in 1990 with theresult of 2005 in fig. 6, we can collect meshes withdifferent regional types (see fig. 7). Remarkably, it isfound that the distribution of these meshes can correspondto certain events in urban development (seelegend in fig. 7).Those transformed from regional type B to A areinclined to experience urban renewal (central Osaka)or expanding local cities (Arida). By contrast, meshesfrom type A to B cover the first large-scale cohousing(Senri New Town) in Japan, decaying areas, urbanareas (Nishinari district) and certain aging local areas.And the distribution of meshes from type B toC is also found to correspond to the marginalizationof Wakayama’s mountainous areas such as the Kiimidistrict.Ultimately, it is proved that the 3 concluded demographictransition patterns can clearly correspondto one specific demographic transition. And these patternsconstantly exist. With the nearest neighborhoodmethod, the discrimination can effectively identifytransformation at the local scale.IV. examining self-sufficiency in demographictransition patternsBased on Chapter 3, the pyramids of 3 regionaltypes are collected to analyze the characteristics ofthe main demographic transition patterns. Referringto studies related to disaster recovery, demographictransition patterns are analyzed to discuss the selfsufficiencythat they represent.Pyramids of 1990, 1995, 2000, and 2005 areshown in table 3. Following the idea of cohort analysis,5-year aged population Pm in one pyramid and5-year older aged population Pm+1 in the pyramid 5years later are considered to be the same cohort. Also,the red line in table 3 marks the most aging cohort ofthe current population in 2030.1. Pattern A—Sustainable patternBaby boomers (those born in 1947-1949) andbaby boom juniors (those born in 1971-1974) lead thetransition in the pyramid of pattern A. Accompanyingthe aging of these generations, there is no apparentdecrease in the population of those 20-35 years old. Itimplies that, although the fertility decline since 1980has contributed to a general decrease in the youngpopulation, the young working population has stillconstantly emigrated. As life step theory [17] notes,sufficient employment and higher education are connectedto the emigration of the young labor population.Additionally, the fertility of the next generationcan be secured. That is, the young labor populationand new birth rate are stable enough to support thepopulation structure.As the white papers [13] and [14] note, with the twocritical factors of well-equipped facilities and sufficienthuman resources, pattern A presents the placeslikely to be sustainable in recovery.2. Pattern B—Dependent patternIn pattern B, it is found that a more juvenile populationoccupies the base of the pyramid in the 1990s.However, the abundant number of juveniles doesnot contribute to the growth of the young workingagepopulation in the next pyramid. This representsconstant and severe emigration of the young labor26


EVALUATING THE IMPACT OF DEMOGRAPHIC TRANSITION IN THE CONTEXT OF THE TOKAI-TONANKAI-NANKAI EARTHQUAKE, JAPANpopulation. Eventually, the decrease in the populationof those aged 20-35 has contributed to an apparentdecrease in new births (bottom of the pyramid) since2000.Accompanying population aging and the decreasein the juvenile population, population agingis expected to accelerate gradually. Insufficient advancedservice and a lower proportion of the workingagepopulation reveal less capability of the 2 notedfactors in disaster recovery. [13], [14] In contrast to thesustainable pattern, pattern B shows a dependentcharacteristic.3. Pattern C—Marginal patternIn pattern C, an extremely aging and unstablepyramid is shown in table 3. The collective amountin every pyramid is 100%. In this case, while a greatamount of the most senior cohorts gradually disappears,the rest of the cohorts, the middle part of thepyramid, should become wider in the next pyramid.In contrast, no increase in the proportion of juvenilesand working-age population is found. It clearly representsa social decrease in population in the non-agingpopulation. Following cohort analysis, a more unstablepyramid with a wider top and a narrower bottomis formed.Compared with the dependent pattern, a low proportionof working people with social decrease showsthat not only advanced services but also basic servicesfor living might hardly be provided. The pattern refersto regions that have less capability to support themselvesand that are more likely to become isolated ina disaster. [12] Also, it implies that limited human resourcesare available for recovery. Accompanying thedemographic transition, a more unstable populationpyramid is expected. Based on the severe populationdecline and decaying services, this marginal demographictransition represents places where energymight be insufficient to continue their recovery.Above all, the characteristics of the 3 demographictransition patterns are observed in the contextof disaster recovery. Based on the self-sufficiency thateach pattern represents, this index is provided to analyzethe demographic transition of exposed areas.Application to the exposed areas of TTNN Eq.Applying the 2005 pyramid of demographictransition patterns to the mesh-based Population Census2005, discrimination analysis using the nearestneighborhood method is conducted in 151,558 meshesof the whole of Japan. Figure 8 shows the result ofTTNN Earthquake-affected areas (over M4.5).TABLE III. CHARACTERISTICS OF DEMOGRAPHIC TRANSITION PATTERNS ASSOCIATED WITH SELF-SUFFICIENCYPatterns 1990 1995 2000 2005 Characteristics of Self-sufficiencySustainablePatternDependentPatternMarginalPatternBaby boom generationJunior baby boom generationMain generationYouth (20-35)• Young working population and fertility of thenext generation are secured.• There are well-equipped facilities and sufficienthuman resources for recovery.• Ongoing population aging and emigration ofthe young working are expected.• Unsatisfactory advanced services and fewerhuman resources weaken region capability.• More unstable pyramid• Basic services might hardly be maintainedand sufficiency in recovery is unlikely.• Human resources and service from outside arerequired.27


H. CHEN, N.MAKI, H.HAYASHIFIGURE 8. Distribution of the sustainable pattern (blue), dependent pattern (brown), and marginal pattern (green):The combined meshes (black) are the extremely lowly populated areas where the 5-year aged populationfrom the census data is unvailable. According to the population census of 2005, metropolitan citiesare marked in red. Also, the captical cities of the prefecture are marked in orange.It is found that the sustainable-patterned meshes(blue) comprise wide areas in 3 major metropolitancities in the Kinki, Chubu, and capital regions. Andthe metropolitan cities (e.g., Fukuoka, Kita Kyushu,Hiroshima, Okayama, Hamamatsu, and Shizuoka)are also covered by sustainable-patterned meshes.Meanwhile, besides the areas surrounding themajormetropolitan cities, a great amount of dependentpatternedmeshes (brown) cover vast areas in theChugoku region, Shikoku region, and Shikoku region,even most of the area of the prefectural capital cities(e.g., Miyazaki, Oita, and Kochi). And the marginalpatterned(green) and combined meshes (black) sharea similarity of distribution. Most of these meshes applyto part of the coastline and mountainous areas inthe Kii Peninsula, Shikoku region, Chugoku region,and Kyushu region.V. Place evaluationScenario settingsThe Tokai-Tonankai-Nankai Earthquake (seetable 4; fig. 9) includes several scenarios as stated bythe Central Disaster Prevention Council. [] Additionally,a month’s time delay between the Tokai-TonankaiEarthquake and the Nankai Earthquake is likelyto occur and result in severe damage. [18] Consideringthe areas affected by the Nankai (N), Tokai-Tonankai(TTN), and Tokai-Tonankai-Nankai Earthquake(TTNN), and days of delay (Delay), 4 scenarios areexamined in this paper.TABLE IV. SETTING THE SCENARIO OF THE TTNNEARTHQUAKEScenario N TN TTN TTNN DelayMechanism Individual Individual Linkage Linkage LinkageMagnitude M8.4 M8.1 M8.4 M8.5 M8.4;M8.4Possibility 1) 50% 60-70% No Data No Data No DataExposure 2) 50,413 36,970 48,171 89,315 84,536Sea Area N TN T+TN T+TN+N T+TN;N1) Probability in the coming 30 yrs (Headquarters for EarthquakeResearch Promotion, 2009)2) Exposure: Areas (km 2 ) exposed above M4.5Regional seismicity exposureTo analyze the characteristics of regional exposure,it is essential to consider the intensity as wellas the probability of each scenario at the same time.The meshes are collected using the rule of maximumintensity (intensity scale) and exposure scenarios(scenario numbers). Note that the number of exposurescenarios here is not representative of the times of occurrencebut of the concept of potential.The TTNN Earthquake will affect the entire plateof the Nankai Trough. Yet some of the affected areasare exposed to stronger seismicity than individually inthe N or TTN Earthquake. [19] In this case, it is necessaryto take another count of TTNN, apart from theN and TTN Earthquakes. As for the scenario of “De-28


H. CHEN, N.MAKI, H.HAYASHIAlso, there are 32,781 meshes exposed to 2 scenarioswith a maximum intensity of M4.5~5.4. And846 meshes are only exposed to the TTNN Earthquakewith an intensity of M4.5~5.4. These wideareas are spread along the coastline of the Kyushuregion and the Chugoku region, the northern Shikokuregion, inner land of the Chubu region, and part of thecapital region.Integrative place evaluationTo integrate the above 5 regional exposure characteristicswith the 3 demographic transition patternsassociated with self-sufficiency, fig. 8 and fig. 9 aremapped in order to conduct a place assessment. Insteadof a complicated legend, we collect mesh numbersclassified into patterns (sustainable, independent,and marginal; uninhabited and combined meshes) aswell as the exposure characteristics (scenario numberswith maximum intensity) in table 5.It is found that more than 500 km 2 (around19,000 inhabitants) spread over the Kii Peninsula areexposed to a severe intensity of every scenario, butare constructed by the marginal-patterned and combinedmeshes. Based on their demographic characteristicsand the exposure, these meshes are consideredto be the most vulnerable regions. More human resourceand service support from outside is necessaryin disaster recovery.In the Chubu region, the vast industrial districtalong the coast is mostly classified as having a sustainablepattern. Yet these areas are also exposed tostrong intensity in the TTN and TTNN Earthquake.Although sufficient social services and a large working-agepopulation are highly cumulative and beneficialto recovery, devastating damage is also expected.In this sense, a damage evaluation is suggested.Southern Shikoku shares similar characteristicsof seismicity exposure to the Chubu region. But fewsustainable-patterned meshes are found in this region.Apart from Kochi City, dependent- and marginalpatternedmeshes occupy the long coastal areas andthe mountainous inner land. Exposed to the strong intensityof the N and TTNN Earthquakes, it is believedthat, as an issue affecting the Kii Peninsula, wide-areamanagement is necessary in disaster recovery.In the Kinki region, the core areas and suburbanareas are both exposed to all scenarios with seismicityof M4.5~5.4. Also, large sustainable-patternedmeshes are found in this area. Generally speaking,severe direct damage of casualties or physical damageto buildings is rarely of concern. Yet, as the secondarymajor metropolitan area, indirect damage is expectedwhen the TTN and the N Earthquake occur individuallywith a time delay. Thus, business continuity planning(BCP) is very important in relieving the impactand adjusting effectively in recovery.Unlike the capital region, the coastal areas ofKyushu and Chugoku, Northern Shikoku, and the innerland of Chubu Region are mostly covered by sustainable-and dependent-patterned meshes. Vast areasare exposed to either the N or the TTN Earthquakewith seismicity of M4.5~5.4. Excluding certain sustainable-patternedmeshes in prefectural cities, manylocal areas (around 15,000 km 2 , 30% of the affectedareas) are covered by independent transition-patternedmeshes. This represents the necessity of reinforcingthe capability of local government.VI. ConclusionIn this study, we aimed to build a self-sufficiencyindex based on demographic transition patterns in orderto examine the impact of demographic transitionin the context of the Tokai-Tonankai-Nankai Earthquake,in contrast to previous concerns regarding exposedpopulations.This study observed the past 15 years of demographictransition and analyzed a great amountof mesh-based population census data of 1990-2005to develop 3 demographic transition patterns basedon cluster analysis and a series of discriminationanalyses. And the transition of each pattern revealsthe capability of social services and human resourcesconcerning the self-sufficiency. Applying the patternsdeveloped to entire affected areas, an integrativeplace evaluation with scenario setting is presented toreinforce disaster management in the Tokai-Tonankai-Nankai Earthquake. The findings are as follows.1) Sustainable pattern: This was constructed by thebaby boom and baby boom junior generation. Aconstant young working population and stablefertility of the next generation result in wellequippedsocial services and sufficient humanresources that are beneficial to disaster recovery.2) Dependent pattern: Ongoing population agingand the emigration of young working people are30


EVALUATING THE IMPACT OF DEMOGRAPHIC TRANSITION IN THE CONTEXT OF THE TOKAI-TONANKAI-NANKAI EARTHQUAKE, JAPANexpected to accelerate. Unsatisfactory advancedservices and fewer human resources will likelyweaken region capability.3) Marginal pattern: Severe natural and socialdecrease in population result in a highly agingpopulation pyramid. In contrast to the dependentpattern, even basic services are barely maintained.More support from outside is required.4) Kii Peninsula: The fewer human resources andpoor social services are unable to cope with disasterrecovery in the severely and most multiexposeddistrict.5) Chubu region: Vast sustainable-patterned areasimply stably working immigrants in recovery.Yet the devastating damage from the TTN andTTNN Earthquake reveals the necessity of damageevaluation.6) Southern Shikoku region: Under the strong intensityof the N and the TTNN Earthquake, outsidethe only prefectural capital city, the distributionof vast dependent - and marginal-patternedmeshes stresses the importance of wide-areamanagement.7) Kinki region: This area is exposed to all scenariosbut with a seismicity scale under M5.5.Severe physical damage and casualties are notexpected, yet BCP is suggested to relieve theimpact while in the time-delay scenario and foreffective adjustment.8) Coastal areas of the Kyushu region and Chugokuregion, northern Shikoku region, and inner landof the Chubu region: These areas are exposed toeither the N or the TTN Earthquake with seismicityof M4.5~5.4. However, vast independenttransition-patterned meshes imply the necessityof reinforcing the capability of local government.Additional work could usefully infer these findingsfor further comprehensive place evaluation withother information required such as infrastructure.AcknowledgmentsThis research was financially supported by theMinistry of Education, Culture, Sports, Science andTechnology of Japan as Tokai Tonankai Nankai EarthquakeCorrelation Evaluation Research. Please notethat the article is reconstructed from a conference paperpublished in IIASA-DPRI, 2009. [20][][][][][][][][][][10][11]ReferencesHeadquarters for Earthquake Research Promotion(2009, Jul), “National Seismicity Estimation Map,”retrieved Jul, 2009, Official Homepage, Japanese,.Central Disaster Prevention Council (2003, Mar), “Abstractof Disaster Management against the Tonankai—Nankai Earthquake,” retrieved Nov, 2008, CabinetOffice Homepage, Tonankai Nankai Earthquake ExpertsInvestigation Committee, Japanese, .National Institute of Population and Social SecurityResearch (IPSS) (2007, May), “Population Projectionsfor Prefectures of Japan,” retrieved Oct, 2008, IPSSHomepage, Japanese, .N. Maki, H. L. Chen, and S. Suzuki, “Response to PossibleEarthquake Disasters in the Tokai, Tonankai, andNankai Areas, and Their Restoration/ReconstructionStrategies,” Journal of Disaster Research, vol.4, no.2,pp.142-150, 2009.H. Blanco, “Pre-event Disaster Planning: TowardsMore Sustainable Communities,” Composite Journal,AIJ, vol. 6, pp. 117-121, 2008.N. Maki and H. Hayashi, “The Efficiency of BuildingCodes for Earthquake Risk Reduction Disaster Managementfor Housing,” Journal of Social Safety Science,vol. 2, pp. 243-250, 2001.S. Cutter, B. Boruff, and W. L. Shirley, “Social Vulnerabilityto Environmental Hazards,” Social ScienceQuarterly, vol. 84, pp. 242-261, 2003.D. S. Mileti, Disaster by Design: A Reassessment ofNatural Hazards in the United States, Washington D.C.: Joseph Henry Press, 1999.S. L. Cutter, T. M. Jerry, and M. S. Scott, “Revealingthe Vulnerability of People and Places: A Case Studyof Georgetown County,” South Carolina, Annals of theAssociation of American Geographers, vol. 90, no. 4,pp. 713-737, 2000.M. L. Carreño, O. D. Cardona, and A. H. Barbat, “UrbanSeismic Risk Evaluation: A Holistic Approach,”Natural Hazards, vol. 40, pp. 137-172, 2007.N. Nojima, M. Kuse, and M. Sugito, “Population Exposureto Seismic Intensity by Recent Earthquakes(2000-2005) in Japan and its Correlation with BuildingDamage and Human Casualty,” Journal of Natural DisasterScience, vol. 25, no. 2, pp. 165-182, 2006.31


H. CHEN, N.MAKI, H.HAYASHI[12][13][14][15][16]K. Ota, Y. Kataie, S. Sakaguchi, S. Nakase, M.Sawada, S. Kondo, K. Fukutome, and C. Watanabe,“An Examination of a Method for Evaluating DisasterMitigation Ability with Attention to Isolation and Selfsupportof Villages in Intermediate and MountainousAreas, Kii Peninsula,” Composite Journal, ArchitecturalInstitute of Japan, vol. 6, pp. 49-56, 2008.Cabinet Office (2007, Jun), “White Paper on DisasterManagement 2007,” retrieved Jul, 2009, Cabinet OfficeHomepage, Japanese, .Cabinet Office (2009, Jun), “White Paper on DisasterManagement 2009,” retrieved Jul 1, 2009, Cabinet OfficeHomepage, Japanese, .T. Mori, “The Tendency of Aging Population at theHousing Estates in theLocal Cities—A Case Study ofOkayama Urban Area,” Journal of The City Planningof Japan, vol. 40, no. 3, pp. 757-762, 2005.H. Chen, N. Maki, and H. Hayashi, “Population Exposureto Tonankai-Nankai Earthquake under the Considerationof Population Transition in 2030,” in: InternationalSymposium on City Planning 2009, Tainan,Taiwan, pp. 289-299, 26-28 Aug, 2009.[17][18][19][20]M. Okumura, “Estimation Method of the Future AgeStructure of Small Area based on the Mesh-basedPopulation Census Data,” Journal of the City PlanningInstitute of Japan, vol. 40, no. 3, pp. 193-198, 2005.K. Terumoto, S. Suzuki, Yoshikawa, K. Inagaki, S.Beniya, H. Tabata, and A. Ouno, “On the Severityof Problems and Necessity of Measures during theInterval between the Tokai, Tonankai, and NankaiEarthquakes—A Case Study in Tanabe City,” Journalof Social Safety Science, vol. 10, pp. 416-426, 2008.H. Chen, N. Maki, and H. Hayashi, “Examining theRegional Exposure Characteristics in the Tokai, Tonankai,and Nankai Earthquake under the Considerationof Future Population Decline—A Preliminary Studyin Future Population Exposure and Infrastructure Exposure,”Journal of Japan Society for Natural DisasterScience (accepted, in Japanese).H. Chen, N. Maki, and H. Hayashi,” The Impact ofDemographic Transition in Pre-disaster Managementof Tokai-Tonankai-Nankai Earthquake, Japan,” in: The9th IIASA-DPRI Conference on Integrated DisasterRisk Management, The Young Scientist Session, Kyoto,Japan, pp. 60-67, 12-16 Oct, 2010.32

More magazines by this user
Similar magazines