Crop Information Systems for Soybean and Azuki Bean in Hokkaido, JAPAN

ANCHA SRINIVASAN

Abstract

ARCVIEW was used to examine relationships among soybean and azuki bean crop distribution, yield, quality, soil and climatic conditions in Hokkaido. Several variables affecting yield and quality were mapped and new relationships among them analyzed. Such an analysis was useful to even determine new areas for introducing these crops. It is planned to make this data widely available through the Hokkaido Regional Agricultural Information System, which is based on the World Wide Web and other Internet tools. The system should facilitate sound decision making by agricultural policy makers, private input agencies, and farmers in Hokkaido.

___________________________________________________________________________

Introduction

Up to date information on soils, climate, crops and varieties is essential for making appropriate farm management decisions. In most cases, however, farmers are not provided with information in a timely manner due to logistic and other reasons. Further the data concerning a crop or cropping system is usually available in various places such as research stations, extension offices, etc. If an information system relevant to a particular crop is available and accessible via Internet, farmers can readily use such information for managing farm resources more effectively. In such system, availability, quality, standardization, coherence, and accessibility of data are all important. Geographic Information System (GIS) forms the core component of this system. Because GIS permits integration of large sets of data in a spatial framework and enables more in-depth data analyses than non-spatial methods, the interactions between biological (genetic), physical (soil, climate), and environmental (management, socio-economic) factors pertaining to a crop can be identified, derived, evaluated and analyzed more accurately (Srinivasan, 1997). Moreover, GIS technology supports visualization by offering refined techniques of altering complex data into clear and interesting display. The ability of a GIS to integrate spatially based information from a variety of data sources in various forms (tables, maps, imagery, etc.) makes it particularly attractive for use in crop management.

Soybean and azuki bean have been grown in Hokkaido for many decades. However, because of relatively high labor requirement for their cultivation, unstable yields, and reduced profitability especially in case of soybean, farmers have recently reduced acreage under cultivation. In 1997, Hokkaido produced 28,200 t of soybean, and 57,200 t of Azuki bean accounting for 20 and 79% of national production respectively (Anonymous, 1998). The annual demand for soybean and azuki bean, however, is very high in Japan and is estimated at 4.6 and 0.15 million tons respectively (Hokkaido Govt., 1994, 1997). The wide deficit between domestic supply and demand is met with imports from other countries including USA, China and others. The development of crop information systems, which provide information on various aspects including the effects of climate and soil on productivity, is expected to enhance opportunities for growing and managing these crops more efficiently in existing regions and in extending their cultivation to new regions within Hokkaido. An enormous amount of data on adaptation of these crops in various geographic regions was already gathered and kept on paper files at various research centers, and experimental stations throughout Hokkaido. For instance, the multi-location multi-year field trials on hundreds of soybean and azuki bean germplasm alone produced invaluable data that, until now, were difficult to interpret and use. Compilation, integration, and analysis of such data in a GIS are bound to yield valuable information. The main objective of crop information system, therefore, is to develop a geo-referenced data recording and information exchange system relevant to soybean and azuki bean within Hokkaido. Upon full development, the system will serve as a quick reference tool for scientists, technicians, field support staff, farmers and policy makers providing up-to-date information on performance of these legumes and their varieties under specific soil and climatic conditions, crop production systems, etc. It will also permit further data analysis and serve as an information dissemination system to allow easy retrieval and supply of regularly updated data to a large number of institutions and users within Hokkaido and elsewhere. The target users of crop information systems are officials of agricultural cooperatives, researchers in agricultural experiment stations, and planners of the prefecture and city governments. In addition, farmers and farmer-based organizations, crop consultants, seed dealers, sales representatives of seed companies, grain marketers, crop insurance agents, and dealers of crop inputs such as fertilizers, pesticides, and herbicides also find these systems useful.

Accessibility is as important as compilation and analysis of data. In order to make the data and maps generated from crop information systems widely available to public, Hokkaido Agricultural and Environmental Information System (HAEIS) is being developed on the lines of the Great Lakes Regional Environmental Information System (GLREIS). GLREIS is a well-known initiative undertaken in USA by the Consortium for International Earth Science Information Network (CIESIN). The goal here is to create a region-wide computer-based information system that provides soil, climate, crop, and market information matched to needs of farmers in Hokkaido. HAEIS is designed to use a variety of capabilities provided by traditional database management technology, GIS, Global Positioning Systems (GPS), and image analysis (remote sensing). HAEIS is now in initial stages of development and further refinements are necessary for its successful launch on to the Internet.

Data and Methodology

The project’s aim is to make use of the existing data and update the current databases rather than generating new data. During the first year of this 5-year project (1997-2002), collection and integration of data relating to crop distribution, yield, quality, and soil and climatic conditions in various regions of Hokkaido is being done with help from researchers in National Agriculture Research Center, and Hokkaido National Agricultural Experiment Station. Initially, the collaborating researchers were asked to emphasize problems specific to soybean and azuki bean that required solving or upstream research that would contribute to the expansion of area and production of these legumes in Hokkaido. Both site suitability and profitability of crop were considered in determining potential areas for expansion and/or intensification of their cultivation.

Insofar as data handling is concerned, the methodology being followed is similar to that adopted for the International Wheat Information System (Fox et al., 1996). The main difference is that soybean and azuki bean information systems will be integral parts of the Hokkaido Agricultural and Environmental System (HAEIS), which will be accessible via Internet and provide many users with a set of data and flexible means to achieve numerous goals. When fully developed, the system would provide quick user-friendly options for managing and searching entire data related to soybean and azuki bean in Hokkaido. Because the system would gather data from genebanks, laboratories, and field tests, it would help researchers avoid duplication of expensive evaluations. Users will be guided through the available databases by a graphical interface prompting selections of crop, varieties, test locations, and yield and quality variables for which data are available. The crop information system can be used to demonstrate genotypic comparison for a particular trial spot, display yield/quality stability of a selected variety, show genotypic comparisons within and between environments, provide genotypic information, and give test location data and management practices information via a map interface. It will allow users to explore crop production data interactively in real-time without specialized and expensive software.

Initially, the base maps showing the administrative boundaries of villages and towns, and soil type in Hokkaido were digitized and converted into shape files using ARCSHAPE command. Attribute data for each geographic region (crop yield, variety, previous quality ratings, cultural treatments including input use, etc.) are being compiled from several PC-based RDBMS including Microsoft Access, and FoxPro. ARCVIEW Spatial Analyst was used for preparing thematic maps and other analyses that permit relating the current intensity of the distribution of the crop and its yield to the physical environment. Both spatial and temporal analyses of area, production, and yield of soybean and azuki bean were conducted. By linking historical information stored in the current databases to spatial data sets, an analysis of cultural treatments will be conducted to provide new insights into crop response in various geographic regions to management treatments. The design of crop information system also accommodates future expansion of functionality and data.

Salient findings

The nature of potential and constraints confronting soybean and azuki bean in Hokkaido differ in various production environments. During the past 3 decades, the annual mean temperature in Hokkaido varied between 7 and 11oC. In contrast, mean temperature during the growing season from May to October ranged from 12 to 19oC. The temperature variation by day or season was also large. Precipitation varied from 900 to 1500 mm (less than the rest of the nation) but the hours of sunlight (around 1100) were similar to those of other prefectures.

The major constraints for intensive cultivation of soybean and azuki bean in Hokkaido are low temperature, limited duration of frost-free period, high probability of rainfall at harvest, etc. Low temperature stress was the main abiotic constraint accounting for large temporal variation in soybean and azuki bean yields. For example, mean temperatures during post-flowering phase were lowest in 1993 and the mean productivity of soybean (1.1 t ha-1) and azuki bean (1.0 t ha-1) was also the lowest in the entire decade. The low temperatures during flowering reduce seed quality as well, as evidenced by an increased seed discoloration and cracking (Srinivasan and Arihara, 1994). In 1993, yields of both soybean and azuki bean were very low in all but Ishikari and Sorachi districts. The low yields in Sorachi, Soya, Abashiri, and Rumoi districts seem to be related to the presence of Psuedogleys (poorly drained soils that have developed from clayey diluvial sediments), which have poor physical properties such as fine texture, high bulk density and low water holding capacity. However, further research into these factors is necessary before a definite conclusion can be drawn.

In 1997, soybean productivity increased to 2.2 t ha-1 as against the long-term average of 1.9 t ha-1. On the other hand, azuki bean productivity was 1.8 t ha-1 as against the long-term average of 1.7 t ha-1. Data on crop distribution of soybean and azuki bean in various regions indicated that cultivation of both crops was largely confined to the south, central and western regions of Hokkaido. Nearly 31% of soybean and 43% of azuki bean is grown in Tokachi district alone. Soybean and azuki bean are not grown at all in as many as 62 and 54 out of 212 towns/villages in Hokkaido respectively. The Eastern and northernmost regions (Kushiro, Nemuro, and Soya districts), because of low temperatures and freezing soils, are unsuitable for both crops. The mean temperatures between June and September were below 15oC in these regions. The Western Sea of Japan coast enjoys relatively high temperatures in summer because of Tushima ocean current. In contrast, the Pacific Ocean Region to the east and the Sea of Okhotsk region to the northeast have relatively low temperatures due to the influence of the cold Kurile current. The study also revealed that soybean and azuki bean cultivation had disappeared in many parts of Oshima, Shirbeshi and Abashiri districts during the past 2 decades. Spatial analysis was thus useful in identifying shifts in soybean and azuki bean cultivation.

An analysis of the effects of variation in area and yield (production per unit area) on variation in total production showed that area accounts for around 27.5% and yield, 45% of yearly variation in soybean production in Hokkaido, and the two together around 91% variation. However, multiple regression analysis showed a dominant effect of yield (67.8% of yield variation) and a relatively low effect of area (24.2%). Several geographic regions with consistently low yields could be identified. They include Mikasa, Bifuka, Otoineppu, Nakagawa, Kumaishi, Kamoenai, Shakotan, Taisei, Furubira, and Akaigawa for soybean, and Nayoro, Shimokawa, Otoineppu, Nakagawa, Kamishihoro, Sarabetsu, Churui, Taiki, Hiro, Toyokoro, Urahoro, Ikutahara, and Engaru for azuki bean. These areas need further improvement in terms of improved drainage, and use of cold and waterlogging tolerant genotypes, etc. On the other hand, productivity of both legumes was consistently high in Ishikari and Sorachi districts, despite high yield variation within each district. The reasons for such variation in productivity in each district are now under investigation.

The hands-on exercises on soil, climate and crop characterization of soybean and azuki bean cropping regions using ARCVIEW Spatial Analyst has brought out some relevant points that need immediate attention. The points include identification of scale of mapping, proper agroclimatic zoning and development of suitable model/expert system depending on soil, climatic and crop variability. Once these are accomplished, spatial analysis would help us identify regions where soybean and/or azuki bean cultivation can be most profitable in terms of probability for production of high quality seeds.

Several maps on yield and quality of soybean and azuki bean in various regions of Hokkaido are in preparation. Initially, maps on duration of frost-free period, probability for rainfall occurrence, mean temperature during the growing season, mean temperature during reproductive growth, number of hours below 15oC during day, yield variation in different regions, soils prone to waterlogging, distribution of dominant genotypes in each geographic region, etc. have been prepared. In addition, a series of graphs depicting changes in soybean and azuki bean productivity during the past two decades is prepared and the relationships among various components are being examined. These maps will soon be available for public use through Internet as a component of the Hokkaido Agricultural and Environmental Information System.

Integration of growth simulation models into GIS is planned for the future to allow users to develop a spatial decision support system for legume crop management. It is expected that such a spatial decision support system would be a forerunner of several other initiatives for efficient crop management in Hokkaido. The development of crop information systems is particularly timely, especially when one considers the enormous growth of information technology and the increasing numbers of growers and crop consultants with Internet access. It is hoped that by utilizing crop information systems, growers can quickly evaluate numerous scenarios and arrive at the optimal mix of crop management practices. It is also anticipated that as the accuracy and capabilities of the crop information systems improve, and as more farmers adopt GIS as a tool for managing spatially based information, the need to depend on narrative information would decrease. Farmers will then start to make better decisions about inputs for crops and management practices within and between fields. Farmers may also be able to evaluate trade-off of quantity over quality of the produce, and select management strategies that optimize growers’ objectives. With all this information, it should be possible for farmers to turn obstacles into opportunities.

This exercise on development of information systems for soybean and azuki bean in Hokkaido region has prepared the way for similar work on other crops and it should be a relatively simple matter to extend these studies to other legumes and possibly to cereals. In future, it should be feasible to employ the data from the study of constituent crops to evaluate the sustainability and stability of specific cropping systems, and better understand how the systems respond to environmental changes.

Acknowledgments

I thank Ms. Y. Komata (Regional Science Institute), Dr. Joji Arihara and Mr. Motoki Takahashi (National Agriculture Research Center, Tsukuba), and Mr. H. Hamaguchi, (Hokkaido National Agricultural Experiment Station, Memuro) for their valuable suggestions and help.

References

Anonymous. 1998. Crop Statistics for Hokkaido. Office of Statistical Information, Sapporo, Hokkaido.

Fox, P.N., Lopez, C., Skovmand, B., Sanchez, H., Herrera, R., White, J.W., Duveiller, E. and van Ginkel, M. 1996. International Wheat Information System (IWIS), Version 1. Mexico, D.F.: CIMMYT. On compact disk.

Hokkaido Government. 1994. Regional Policy for Bean Production in Hokkaido (In Japanese). March 1994. Department of Agriculture, Hokkaido Government.

Hokkaido Government. 1997. Hokkaido Profile, March 1997. International Relations Division, Office of the Governor, Hokkaido Government.

Srinivasan, A. 1997. Crop Information Systems for Food Legumes in Hokkaido, Japan. In: United Nations Environment Programme (1997) GIS in Agricultural Research: Awareness Package. (Denisov, N., Heberlein, C., Czaran, L. and Simonett, O. Eds.). UNEP/DEIA/TR.97-9. Case Study No. 19.

Srinivasan, A. and Arihara, J. 1994. Soybean seed discoloration and cracking in response to low temperatures during early reproductive growth. Crop Sci., 34:1611-1617.

___________________________________________________________________________

Ancha Srinivasan, Ph.D.

Senior Researcher, Regional Science Institute

4-13, Kita 24 Nishi 2, Kita-ku, Sapporo 001-0024 Japan

___________________________________________________________________________

Last revised January 15, 1999

1