Spatial Analysis and Remote Sensing for Monitoring Systems of Oncomelania Nosophora Following the Eradication of Schistosomiasis

Jpn. J. Infect. Dis., 62, 125-132, 2009

Original Article Spatial Analysis and Remote Sensing for Monitoring Systems of Oncomelania nosophora Following the Eradication of Schistosomiasis Japonica in Yamanashi Prefecture, Japan Naoko Nihei*, Osamu Komagata, Mutsuo Kobayashi, Yasuhide Saitoh1, Kan-ichiro Mochizuki2, and Satoshi Nakamura3 Department of Medical Entomology, National Institute of Infectious Diseases, Tokyo 162-8640; 1Department of Parasitology, School of Veterinary Medicine Azabu University, Kanagawa 229-8501; 2Research and Development Center, PASCO Corporation, Tokyo 153-0043; and 3Department of Appropriate Technology Development and Transfer, Research Institute, International Medical Center, Tokyo 162-8655, Japan (Received August 18, 2008. Accepted January 19, 2009)

SUMMARY: In order to develop an inexpensive, simple, and accurate method of monitoring for the reemergence of schistosomiasis japonica in Yamanashi Prefecture, Japan, the distribution and habitation density of the inter- mediate host, Oncomelania nosophora, were spatially analyzed using geographic information systems. The 1967-1968 density distribution maps prepared by Yamanashi Prefecture and Nihei were digitized and geocoded. The habitats and population density of O. nosophora were estimated by referring to the data compiled by the Yamanashi Association for Schistosomiasis Control (1977). These earlier findings were compared with average population densities between 1996 and 2000 previously recorded (Nihei, N., Kajihara, N., Kirinoki, M., et al., Parasitol. Int., 52, 395-401, 2003 and Nihei, N., Kajihara, N., Kirinoki, M., et al., Parasitol, Int., 53, 199-205, 2004). A variance map was created to compare the spatial distribution maps of population density from each of the two periods of interest. The changes in distribution were remarkable and the map was found to be effective for future control. The most appropriate monitoring sites were chosen on the basis of the spatial population density maps and the variance map. Moreover, the paddy fields at risk were extracted using the normalized difference vegetation index value based on Advanced Land Observation Satellite images. The combination of this method with the global positioning system provides an inexpensive means of monitoring modern schistosomiasis endemic areas in Japan and also in China, the Philippines, and other countries as well, where the intermediate snail grows in paddy fields and marshlands under consistently wet conditions.

However, with the exception of those created by Iijima and INTRODUCTION Ishizaki (10) and Nihei and Asami (7), no detailed maps were The epidemic of schistosomiasis japonica in Yamanashi printed or preserved during this data collection process. Prefecture was officially declared to have ended in 1996, thus In this report, the data in the distribution map of 1967- 1968 completing the national eradication of this endemic disease. maintained by our group were compared with the results of However, populations of the intermediate snail host, namely, fixed-point observations recorded using geographic informa- Oncomelania nosophora, still exist. Hence, fixed-point ob- tion systems (GIS) from 1996 to 2000 (11,12). Furthermore, servations of O. nosophora habitats were conducted for 8 areas supportive of O. nosophora habitation were represented years spanning from 1996 to 2003, in accordance with deci- diagrammatically following analyses of aerial photographs sions made by the prefectural authorities, the Yamanashi and field surveys. The trends in habitat range and density Institute for Public Health (YIPH), and citizens who were were studied, enabling the identification of areas that still concerned about the reemergence of this disease (1,2). How- need to be continuously monitored. Snails were collected sev- ever, this series of observations was discontinued in 2003, eral times each year, and specimens were examined to identify since no infected snails had been identified during the study any cases of infection. If the schistosomiasis had indeed re- period. emerged, the condition of the snails could be immediately Various measures to eradicate the disease had been under- assessed. Here, we established a precise, efficient, inexpen- taken since 1913, when records maintenance began (3-9). sive, and rapid method of monitoring the risk of future out- Since the 1960s, information regarding measures taken to breaks of schistosomiasis japonica derived from imported target O. nosophora has been recorded; various factors have cases. It has been claimed that remote sensing and GIS have been included in these records, including distribution area, been actively adopted to monitor for outbreaks of schistoso- habitat density, development and dispersal of molluscicides, miasis japonica by considering factors such as climate change, and the positions and lengths of concrete irrigation channels. changes in land condition, and migration of patients in China and the Philippines (13,14). However, with few exceptions, *Corresponding author: Mailing address: Department of Medical there have been very few reports on the use of these techniques Entomology, National Institute of Infectious Diseases, Toyama (15-22). The results of the present observations could be used 1-23-1, Shinjuku-ku, Tokyo 162-8640, Japan. Tel: +81-3-5285- as a reference for monitoring the risk of schistosomiasis 1111, Fax: +81-3-5285-1147, E-mail: [email protected] japonica in countries other than Japan.

125 map. An isopleth map of O. nosophora density was created MATERIALS AND METHODS by superimposing the density isopleths onto this map. Data: (i) Data from YIPH (1967-1968): O. nosophora den- Creation of a density distribution map for O. nosophora sity distribution map by Nihei (no original map; extant map for 1996-2000: Snail habitat observations were conducted hand-drawn due to the lack of photocopying facilities at at 120 fixed points from 1996 to 2000, but due to budgetary a scale of 1:50,000 on topographic maps issued by the Geo- constrains, the number of fixed points was reduced to 60 be- graphical Survey Institute). (ii) 1970 statistical data for aver- tween the years of 2001 to 2003. For the present study, we age O. nosophora habitat density per municipality based on used the average populations for the first 5 years, when a data from YASC (1977) (5). (iii) 1996-2000: Data on O. higher number of fixed points were used. The number of speci- nosophora population density per fixed point calculated based mens collected was kept consistent within a 0.25 × 0.25-m on the results of a survey conducted in Yamanashi Prefecture frame at 2 water inlets per point. In order to compare these from 1996 to 2003 (12). (iv) Digital map of Yamanashi Prefec- data with those of the 1967-1968 GIS density distribution ture at a scale of 1:25,000 (PFM; Pasco Co., Tokyo, Japan). map, the averages were converted to equivalents collected (v) Aerial photos on a scale of approximately 1:20,000 (taken per square meter. The density distribution map per fixed point by the Japan Forestry Association, Tokyo, Japan in 2002). was displayed by classifying the data into 5 categories. Using (vi) Two advanced land observing satellite (ALOS) images these fixed points, we created a GIS spatial density distribu- taken in April and August 2007 (JAXA, Tokyo, Japan). (vii) tion map of the snail population for the period from 1996 to Data from the following GIS software: ArcGIS 9.1 + Spatial 2000 within a 250-m search area. Analyst Extension (ESRI, Redlands, Calif., USA). Creation of a spatial density variance map of O. nosophora Density distribution map of O. nosophora created on between 1967 -1968 and 1996 - 2000: GIS software, Arc View, the basis of surveys conducted in 1967 - 1968: The Yamanashi was used for dividing and displaying the density of each habi- Prefecture authorities had conducted a detailed survey of the tat of Oncomelania in 1967-1968. The habitat density at the density of snails in each habitat in each municipality for 2 investigated fixed points from 1996 to 2000 was similarly years. The distribution in the habitat was estimated as the divided and displayed as classes of 100. The values from the density of O. nosophora populations per square meter (m2) spatial distribution map of 1967-1968 were deducted from by dividing the habitat into 3 categories, namely, <10 snails/ those of the 1996-2000 map (green denotes positive values, m2 (low), 10-19 snails/m2 (medium), and ≥20 snails/m2 (high) white denotes no change, and red denotes negative values). each year, from which the density distribution map was As regards the range categories for habitat density at the 120 created. The coordinates were then added, and the resulting fixed points, maps were created ranging from 50 to –50. map was scanned. The snails mainly inhabited paddy fields Determination of the range of paddy fields with possible and associated irrigation channels; hence, habitat density is O. nosophora habitation baes on ALOS image data from represented as a linear, and not an aerial value. The coded April and August 2007: The land surrounding O. nosophora density data for the established 3 categories were joined for 2 habitats in the Kofu Basin consist of paddy fields, fallow lands, years per code, and a GIS layer was subsequently created. orchards, and other types of land. In order to determine the When the maps were overlaid and the lines were duplicated, range of paddy fields and fallow lands with possibile O. the higher numeric value was selected, and a density distri- nosophora habitation, the normalized difference vegetation bution map was created by defining 0.5 m as the line buffer index (NDVI) was computed using a near-infrared band and (hereafter referred to as the “1967-1968 GIS density distri- a visible-light red band of 2 ALOS images taken in April and bution map”). in August of 2007 (23). This index represents the existence Estimation of the average density of O. nosophora per and degree of vegetative activity, and shows the value nor- municipality and supplementary data for 1967-1968: In malized by –1 to 1. We set the threshold value of the NDVI order to prepare statistical data on the snail population den- to determine the area of flat land lacking vegetative cover, sity, density was divided into 3 categories that could be and we visually deciphered the ALOS images that had been depicted in the maps. Accordingly, the municipality map was taken in April. We then calculated the threshold value for overlaid with the 1967-1968 GIS density distribution map. determining the area of land with extensive rice cultivation To enable aggregation of the linear data per municipality, the and tracts of land with thick vegetation using the ALOS im- 3 categories of habitat density (i.e., “high,” “medium,” and age data that had been obtained in August. Furthermore, two “low”) were defined in 2 ways: (a) 20, 10, and 1/m2 and (b) areas were overlaid, namely, the area of flat land ranging from 25, 15, and 5/m2, following the general method of presuming –0.36 to –0.22 of the threshold value of the NDVI of the a numerical value. The snail density per municipality and the image data in April, and the area of green tracts of land with total populations were calculated (GIS density). The calcula- a threshold value of 0.28 or more of the NDVI in August. tion of density from the GIS supplementary data was per- The area of flat land on the image obtained in April that formed using “(a),” since these values were obtained when appeared as a tract of land with thick vegetation in the image the GIS density and the existing data were compared. A obtained in August was determined to be the area used for supplementary GIS habitat density map was then created on paddy cultivation. the basis of regression line. Creation of a spatial density distribution map for O. RESULTS nosophora for 1967-1968: The O. nosophora habitat ex- tended to paddy fields and fallow farmlands from irrigation Estimation and spatial analysis of the population den- channels during the rice cultivation period, and from spring sity of O. nosophora in Yamanashi Prefecture in 1967- to autumn. Considering such expansion, a GIS spatial density 1968: The 1967-1968 O. nosophora digital density distribu- distribution map was created by the addition of supplemen- tion map of each municipality shows that the population den- tary data from the search ranges of 100 m, 250 m, and 500 m sities of the snails differed significantly according to type of on ArcGIS using the supplementary GIS habitat density land examined (Fig. 1). Snail habitats were found not only

126 Table 1. Estimation of population number of O. nosophora, and population density calculated per municipality for both cases, namely, (a) an estimate of smaller populations and (b) an estimate of larger populations Mean population Estimated GIS supplementary No. of habitats GIS habitat density Municipality density No. of snails habitat density 1970 data 1970 data/m2 (a) 20, 10, 1/m2 (a) 20, 10, 1/m2 (b) 25, 15, 5/m2 (a) 20, 10, 1/m2 (b) 25, 15, 5/m2 1 Nirasaki City 25 12 187,317 5 9 7 12 2 Shirane Town 6 22 123,419 11 16 13 17 3 Hatta Village 6 35 150,082 9 14 19 24 4 Wakakusa Town 5 19 69,451 13 17 11 16 Western area subtotal 42 22 530,269 10 14 10 17 5 Futaba Town 15 5 10,549 2 6 4 8 6 Shikishima Town 10 5 66,594 9 13 4 8 Northern area subtotal 25 5 77,143 6 10 4 8 7 Ryuo Town 10 8 258,192 8 13 6 11 8 Showa Town 11 2 1,863 2 6 2 6 9 Tamaho Town 14 2 14,538 4 8 2 6 10 Kofu City 11 3 56,148 7 12 3 7 Central area subtotal 46 4 330,741 5 10 3 8 11 Isawa Town 19 1 155 1 5 2 6 12 Tatomi Town 10 4 307 1 5 3 8 13 Nakamichi Town 7 2 860 1 5 2 6 14 Sakaigawa Town 7 1 396 1 5 2 6 15 Yatsushiro Town 6 1 1,147 1 5 2 6 16 Misaka Town 4 1 1,665 1 5 2 6 17 Ichinomiya Town 8 1 133 1 5 2 6 18 Kushigata Town 0 0 454 1 5 1 5 Eastern area subtotal 61 1 5,117 1 5 2 6 Total 174 7 943,270 4 9 5 9 along the Midai River on the western side of the Kofu Basin municipalities on the basis of the existing data; the differ- and in swampy areas where the water level was high and prone ences between values observed in different municipalities to flooding downstream of the confluence of the Kamanashi were significant. The smaller range per municipality was used and Midai Rivers, but snail habitats were also identified in to estimate (a) small and (b) large types of population (1-13 paddy fields in the alluvial fan around the Kofu Basin. During snails/m2 in the former case, and 5-17 snails/m2 in the latter). the study period, some scattered snail habitats were observed The graph in Fig. 2 shows a comparison between the average in paddy fields within Kofu City, the Yamanashi Prefecture density corrected by GIS and the average density based on capital. the existing data (Table 1; GIS supplementary average den- Using smaller estimates for the density distribution (Fig. sity). Since the data in this figure demonstrate a close corre- 1), the total O. nosophora population was calculated per lation, the expression y = 0.519x + 1.315 (line a), was used to municipality for the first time (Table 1). The Kofu Basin is for the data correction. The corrected value was calculated usually divided into 4 areas (western, northern, central, and with the correlation coefficient, save for (H) and (K) points eastern) in order to assess epidemiological differences in schis- of the exception value from all endemic village data points tosomiasis. (line b). The smaller estimates were closer to the existing The total O. nosophora population in the Kofu Basin was data (3). Hence, the map for the estimated snail populations estimated to be 943,270. The largest population (258,192) was created using line widths of 1 m and the 3 density code for a municipality unit was observed Ryuo Town on the categories of 20, 10, and 1 snails/m2. Kamanashi alluvial fan in the central area. In the western Expansion of the snail habitat during 1967 - 1968 was area, the largest O. nosophora population (530,269) was ob- expected during the wet stage of paddy rice cultivation. In served on gently sloping grounds of the alluvial fan (Nirasaki order to estimate the spatial distribution of the O. nosophora City: 187,317; Hatta Village: 150,082; Shirane City: 123,419; population, the GIS spatial distribution maps were created etc.). The total population of O. nosophora in the low-lying with search areas at 500 m, 250 m, and 100 m from the den- swampy zone at the center of the Kofu Basin was estimated sity line (figures are omitted). A search area of 500 m was at 330,741. With the exception of in Ryuo Town, where the considered appropriate for the spatial analysis in order to snail population was the largest, in other towns in this area, accurately represent O. nosophora population distribution, the number of specimens was lower (e.g., in Tamaho and including that in low-density areas. However, this approach Showa). Further, there were few snail specimens in the area produced some false positives, implying that even in 1967- surrounding the Fuefuki River from the southern to the east- 1968, the snail populations were unlikely to inhabit areas ern area of the Kofu Basin (5,117). such as hilly terrain, uplands, and the upper reaches of allu- Based on the data provided above, the average density of vial fans. Our experience suggests that a map with a search the snail population was calculated for each municipality in area of 250 m could be used to adequately approximate the both of two cases, namely, (a) an estimate of smaller popula- distribution status, while at the same time highlight the high- tions (20, 10, and 1 snails/m2), and (b) an estimate of larger density areas relatively well. In general, it would not have populations (25, 15, and 5 snails/m2) using the arithmetic been easy to identify high-density habitats using a map with function in ArcView (Table 1). The average densities of the a buffer area of 100 m, because using this approach, the snail population were determined to be 0-35 snails/m2 in 18 distribution area would be less than the actual area.

127 Fig. 1. The 1967-1968 digital density distribution map of O. nosophora per governmental region of Yamanashi Prefecture. This map shows the distribution areas by dividing the density per square meter of O. nosophora into 3 categories, namely, less than 10/m2, 10-19/m2, 20/m2 or more. Coded density data for 3 categories were overlaid with municipality map in order to estimate the population density of O. nosophora per municipality.

spatial density distribution map for 1996-2000 were not speci- fied in Kofu City, nor in other areas to the east (Fig. 4). This was because there were no habitats for O. nosophora due to advanced urbanization, which had brought about changes in the use of surrounding paddy fields and orchards. O. nosophora no longer inhabits the center of the Kofu Basin. In the western part of the Kofu Basin, high-density snail populations were observed in areas with paddy fields and orchards situated in complicated topographic terrains and alluvial fans (8). On the basis of these fixed points, we applied a buffer area of 250 m (Fig. 4) and created a map similar to that created for Fig. 2. Relationship between the average densities corrected by GIS 1967-1968 (Fig. 5). A variance map was created by overlay- and the average density based on the existing data. Line (a), correlation ing the 2 data periods (Fig. 6). The areas that required moni- except 2 points (Hatta Village and Kushigata Town) from all endemic toring were magnified and were overlaid with the fixed-point municipalities. Line (b), correlation of all endemic municipalities, (H): Hatta Village, (K): Kushigata Town. distribution map for 1996-2000 (Fig. 7). Figures 6 and 7 display the density difference per area observed for each period with respect to gradation. In the western region, at the The more accurate estimation of the snail habitat spatial confluence of the Midai and Kamanashi Rivers, the O. density distribution was compared to the current fixed-point nosophora habitat was large and the population density was observation results. In order to apply the data to future field high as well. Although this habitat had decreased in area surveys, an isoline of the density, namely, 1 specimen/m2, was between the years 1996 to 2000, high-density areas were prepared for the 1967-1968 density distribution map (Fig. nonetheless observed. On the variance map, high-density snail 3A). To this end, a density distribution map with a buffer area populations were observed in the western section during both of 500 m was created as the background, part of which was periods, although the higher value was observed in more magnified (Fig. 3B). The isoline shape of the Kofu Basin recent years, i.e., the study revealed the presence of a high- resembles the density distribution map for a search area of density snail population during the period 1996-2000 (green). 500 m. It was easy to identify areas on the map with a popu- Areas on the map that did not undergo any change in the lation density ranging from high to low, without relying on snail population density for 40 years were represented in color separation. When this map was overlaid on the map white. with a buffer area of 500 m, the snail population distribution There are currently few habitats at the center of the Kofu was estimated on the basis of the high-density areas, and popu- Basin, the central region of which is colored pink. At present, lation status could be accurately identified. Once this map urbanization has precluded snail habitation in some of these was expanded, it could be used for field surveys. The areas areas, although there remain some paddy fields in which this lacking a snail habitation were indicated as shch on the map. type of snail can survive. A clearly delineated boundary was Changes in the spatial density distribution of O. nosophora observed between the green and pink zones on the map. The between 1967-1968 and 1996-2000: Fixed points for the pink area, i.e., the northern portion of the Midai River, sug-

128 Fig. 3. (A) The 1967-1968 GIS spatial density distribution map with 500 m range and population density isolind map of O. nosophora in Yamanashi Prefetcture. (B) Magnified isoline map of O. nosophora (Fig. 3A) in northwestern part of Yamanashi Prefecture on spatial map with 500 m range. Red lines show the isoline of population density. Lighter blue zones between isolines show high density areas and darker purple zones show low density areas. Red, green, and blue short lines show habitats of O. nosophora.

Fig. 4. Spatial density distribution of O. nosophora in 1996-2000 cal- Fig. 6. Population variance map between 1967-1968 and 1996-2000. culated from the fixed-point observation (12). Dark color shows high population density of O. nosophora. areas, i.e., areas at risk, were also identified as determined based on aerial photographs (11). When the isodensity map for 1967-1968 was superimposed onto these aerial photo- graphs, the high-density snail habitation areas were more clearly visible. When the paddy fields and areas of fallow land, i.e., those areas that had been clearly identified as suit- able habitats for snail populations, were examined using the NDVI data of the 2 ALOS images taken in 2007 (Fig. 8), the need to survey the paddy fields in risk areas became evident.

DISCUSSION

Fig. 5. Spatial density distribution of O. nosophora in 1967-1968 dis- GIS can be a rather useful tool for analyzing the distribu- played by same methods in 1996-2000. tion and population density of the intermediate host of schis- tosomiasis japonica, as well as for identifying spatial changes gests a particular pattern of change, as this area used to be a in the distribution of this host. All figures generated for high-density habitat, and has since been converted into an this study were geo-coordinated, thus enabling the overlay industrial complex. of global positioning system (GPS) monitoring data (11). A Determination of areas at risk using NDVI values from high-quality, yet simple and inexpensive monitoring system the ALOS image data: Possible O. nosophora habitation was thus established on the basis of data acquired by mobile

129 Fig. 7. Results of snail monitoring survey using camera with GPS on detail aerial photographs and population isoline map around the habitats of O. nosophora in Yamanashi Prefecture, Japan.

(indicative of the schistosomiasis japonica risk) and national spatial data that include population migration. This helps guarantee the safety of local citizens concerned about such risks. Furthermore, such an approach can be used for eradi- cation measures that take the environment into consideration, and the addresses demands for appropriate surveys and moni- toring. Such a monitoring system should be established in the future, as eradication measures based on variations in the geographical distribution of O. nosophora have long been considered essential. Such eradication measures were initially delayed in Japan due to the confusion prevalent during the post-World War II period. Once the snail habitat range reached its peak, a distribution map was created to provide data for the concretization of irrigation channels, to gain a better un- derstanding of the current status of the host, and to generate Fig. 8. Monitoring paddy fields in risk area in Yamanashi Prefectuer measures for O. nosophora eradication. Prior to the emer- using the NDVI data of the 2 ALOS images taken in 2007. Paddy gence of the current GIS technology, different types of maps fields are areas of –0.36 to –0.22 of the threshold value of the NDVI and data had been prepared, and these previous sources were of ALOS image data in April, and the area of green tract of land with used in the present GIS analysis. Researchers at the YIPH a threshold value of 0.28 or more of the NDVI in August. and Nihei in the 1960s created maps replying primarily on their expertise in cartography at a time when maps on a scale GIS (12). Digital density distribution maps composed of 3 of 1:25,000 of the complete habitat range for O. nosophora population density layers (“high,” “medium,” and “low”) were unavailable. enabled the visualization of distinct areas identified as having A spatial density map was created here for the entire Kofu different population density ranks. The population density per Basin. In order to compare this map with the distribution maps square meter (m2) value, which is the standard index, was es- that had been drafted in the past, numerical data was neces- timated based on the results of population surveys conducted sary. Accordingly, a GIS analysis was conducted using the during different years and by using different survey methods available data on the population densities in each municipal- applied in the Kofu Basin in Yamanashi Prefecture. Recently, ity, other estimates (statistical data), and the geographical data the importance of GIS in the fields of parasitology and tropi- supplemented by preexisting distribution maps; the calculated cal medicine has been emphasized in countries other than data were confirmed to be consistent with the statistical data. Japan. However, to date, the few studies that have been For the 1996-2003 population survey (11,12), fixed-point conducted have only involved the visualization of data and observations were conducted using GPS accurate to within the analysis of distribution elements via overlay with other 50 cm; these previous results not only contributed to the find- factors. The present case study demonstrates that GIS tech- ings of the present analysis, they are expected to remain use- nology in Japan can be effectively used for related future ful for future reference. studies. The map of density differences highlighted any possible Emergency measures for quickly dealing with a variety of spatial and numerical discrepancies with respect to the find- problems have been developed by techniques involving the ings of conventional ecological science. This map was used to overlay of a digital map detailing O. nosophora distribution compare quality requirements between the 2 maps (Figs. 4

130 and 5). On the basis of the 1967-1968 data for O. nosophora, maps with buffer areas of 500 m, 250 m, and 100 m were ACKNOWLEDGMENTS created. It was assumed that each buffer area represented a The authors are grateful to Professor Nobuo Ohta, Shaolong Lu, and different distribution. To create the GIS density distribution Tianping Wang for their helpful discussion. This work was supported in part by a Grant-in-Aid for Scientific Re- map, the data and the numerical results shown on the graph search (B) (310, 18406012, 2006) of the Ministry of Education, Culture, were analyzed. Based on these maps, the habitat density Sports, Science and Technology, and in part by the sustainable Co-existence results were considered to be suitable for comparison by speci- of Humans, Nature and the Earth project of Research Revolution 2002 fying 0.5 m on one side (set to 1 m) as the linear data, and an (RR2002) of the Ministry of Education, Culture, Sports, Science and Tech- empirical buffer area of 250 m was set. Clear differences were nology of Japan. We received the grant of joint developments of ALOS from JAXA. observed when these maps were overlaid onto the 1967-1968 spatial density distribution map with a buffer area of 250 m. Methods to accurately reveal habitats and changes in the REFERENCES population density of O. nosophora have been developed. 1. Kajihara, N., Sawanobori, I. and Aikawa, H. (2001): A change of the distribution and the density of Oncomelania snails after the closing of However, the habitat range will still need to be adjusted in the prevalence of schistosomiasis in Yamanashi Prefecture. Annu. Rep. accord with future O. nosophora population reductions or Yamanashi Inst. Public Health, 45, 15-19 (in Japanese). expansions. These methods can be applied in future studies 2. Minai, M., Hosaka, Y. and Ohta, N. (2003): Historical view of schisto- to monitor regions for the potential growth and expansion of somiasis japonica in Japan: implementation and evaluation of disease- O. nosophora populations. control strategies in Yamanashi Prefecture. Parasitol. Int., 52, 321-326. 3. Yamanashi Association for Schistosomiasis Control (YASC) (1977): It was essential to assess environmental factors such as Fight against schistosomiasis in Yamanashi. Heiwa Print, Kofu (in Japa- actual topography, groundwater, urbanization, and agricul- nese). tural land use in order to interpret geographical differences 4. Yamanashi Association for Schistosomiasis Control (YASC) (1981): between the 2 maps (Figs. 4 and 5). Although some correla- Fight against schistosomiasis in Yamanashi. Heiwa Print, Kofu (in Japa- tion was observed with the overlay of different data, our in- nese). 5. Komiya, Y. (1964): The prevention of schistosomiasis japonica. p. 243- terpretation of the correlation was facilitated by having a good 274. In K. Morishita, Y. Komiya, H. Matsubayashi (ed), Progress of understanding of the snail’s habitat conditions and elements Medical Parasitology in Japan. vol. 1. Meguro Parasitological Museum, related to the actual decline in the snail population. It was Tokyo. also deemed necessary to consider the potential for false posi- 6. Okabe, K. (1964): Biology and epidemiology of Schistosoma japonicum tives and false negatives in this study. and schistosomiasis. p. 185-211. In K. Morishita, Y. Komiya, H. Matsubayashi (ed), Progress of Medical Parasitology in Japan. vol. 1. Isoline maps of population density that were created on Meguro Parasitological Museum, Tokyo. the basis of the 1967-1968 population distribution map could 7. Nihei, N. and Asami, S. (1972): Medical geographical study on schisto- potentially be used during field research for gaining a more somiasis japonica--on the topography and soil as controlling factors of detailed understanding of the relationship between density the distribution of Oncomelania hupensis nosophora (1). Geographical Rev. Jpn., 45, 391-409 (in Japanese with English abstract). and location. It is important to consider the availability of 8. Nihei, N., Asami, S. and Tanaka, H. (1981): Geographical factors influ- data and to create user manuals regarding that data, so that encing the population numbers and distribution of Oncomelania such maps can be effectively used and a management sup- nosophora and the subsequent effect on the control of schistosomiasis port system can be developed while emergency surveys are japonica in Japan. Soc. Sci. Med., 15D, 149-159. conducted for projects such as reallocation of land for a 9. Tanaka, H. and Tsuji, M. (1997): From discovery to eradication of schis- tosomiasis in Japan: 1847-1996. Int. J. Parasitol., 27, 1465-1480. different use and future monitoring. We encourage the use of 10. Iijima, T., Ito, Y. and Ishizaki, T. (1965): Epidemiological study of schis- such data in tandem with a standardized GIS accessed via tosomiasis japonica for school children in Futaba Town, Kitakoma Gun, a server, which will aid in the development of response proto- Yamanashi Prefecture, Japan. Ann. Rep. Yamanashi Pref. Hyg. Lab., cols. 10, 22-30 (in Japanese). The aerial photographs, the fixed-point survey, and the iso- 11. Nihei, N., Kajihara, N., Kirinoki, M., et al. (2003): Fixed-point observa- tion of Oncomelania nosophora in Kofu Basin--establishment of monitor- line map for the 1967-1968 population density indicated ing system of schistosomiasis japonica in Japan. Parasitol. Int., 52, 395- whether or not O. nosophora populations were found at the 401.Corrected and republished in: Parasitol. Int. (2004), 53, 199-205. survey points. This combination of aerial photographic and 12. Nihei, N., Kajihara, N., Kirinoki, M., et al. (2006): Establishment of a GPS data collection provides an easily applied, precise, and GIS monitoring system for schistosomiasis japonica in Kofu, Japan. advantageous survey method, which could also be used in Anna. Trop. Med. Parasitol., 100, 143-153. 13. Zhou, X.N., Yang, G.J., Yang, K.H., et al. (2008): Potential impact of developing countries. climate change on schistosomiasis transmission in China. Am. J. Trop. Intermediate snail hosts cannot be monitored directly with Med. Hyg., 78, 188-194. ALOS images. However, an accurate and effective method 14. King, C.T. (2008): Where snails no longer fear to trend. Am. J. Trop. of identifying and forecasting relevant conditions within a Med. Hyg., 78, 185. changing environment includes the deciphering of images to 15. Seto, E., Liang, S., Qiu, D., et al. (2001): A protocol for geographically randomized snail surveys in schistosomiasis fieldwork using the global identify potential habitats via consideration of ground-surface positioning system. Am. J. Trop. Med. Hyg., 64, 98-99. studies, land usage in regions where disease is prevalent, topog- 16. Seto, E., Wu, W., Qiu, D., et al. (2002): Impact of soil chemistry on the raphical characteristics, and water utilization characteristics. distribution of Oncomelania hupensis (Gastropoda: Pomatiopsidae) in The method described here could also be used to survey vari- China. Malacologia, 44, 319-343. 17. Seto, E., Xu, B., Liang, S., et al. (2002): The use of remote sensing ous types of snail that serve as intermediate hosts in other for predictive modeling of schistosomiasis in China. Photogram. Eng. countries (e.g., in mountainous or hilly terrain in China and Rem. Sens., 68, 167-174. the Philippines). In general, using satellite images instead of 18. Spear, R., Gong, P., Seto, E., et al. (1999): Remote sensing and GIS a distribution map as a background is expected to be helpful for schistosomiasis control in mountainous areas in Sichuan, China. Geogr. in related endeavors. Inf. Syst., 4, 14-22. 19. Ohmae, H., Sinoun, M., Kirinoki, M., et al. (2004): Schistosomiasis mekongi: from discovery to control. Parasitol. Int., 53, 135-142. 20. Yang, G.I., Vounatsou, P., Zhou, X.N., et al. (2005): A review of geographic information system and remote sensing with applications to

131 the epidemiology and control of schistosomiasis in China. Acta Trop., 22. Zhang, Z.Y., Xu, D.Z., Zhou, X.N., et al. (2005): Remote sensing and 96, 117-129. spatial statistical analysis to predict the distribution of Oncomelania 21. Guo, H.G., Vounatsou, P., Cao, C.L., et al. (2005): A geographic informa- hupensis in the marshlands of Vhina. Acta Trop., 96, 205-212. tion and remote sensing based model for prediction of Oncomelania 23. Baker, J.R., Muller, R., Rollinson, D., et al. (2000): Advances in Parasit- hupensis habitats in the Poyang Lake area, China. Acta Trop. 96, 213- ology. Remote sensing and Geographical Information Systems in Epide- 222. miology. Academic Press, London.

132