Development And Application Of Fertilization Decision Supporting Systems Based On ComGIS

The key problem of scale change in application of fertilization model is how to calculate the fertilization rate for the plots in the target region with the fertilization model developed at field plot. This paper is object to construct a fertilization decision support system (FDSS) at county scale based on the fertilization model, which takes into account spatial change of soil fertility. By choosing the proper models, the system can recommend fertilization from point to area scale for different crops and fertilizers. The FDSS was established for Yujiang County, Jiangxi Province, which consists of 3 research contents, i.e. Building red soil upland fertilization model in hilly area, evaluating the environment effect by studying nitrogen transfer in soil profile of upland red soil, integrating fertilization model with GIS and application i.e. in YuJiang county.Nitrogen transfer in upland. Nitrogen transfer in agro-ecosystem is a key process affecting soil fertility and environment quality. The profile distribution of NO₃-N, NH₄-N in upland red soils with oil rape (Brassica campestris L)-peanut rotation (Arachis L.) system in hilly region was investigated by field experiment, then the effect of soil parent material, crop growth, rain, and N application rate was analyzed. The results showed that distribution of NO3-N in soil profile varied with the soil parent materials. Nitrate in soil layer at 60-80cm depth changed at a range of 20.38³2.94 mg.kg^-1 for red soil derived from red sandstone during the whole rotation period, showing a continuous nitrate transfer process. Nitrate content increased with the depth in soil profile after continuous precipitation, indicating a driven force of precipitation for nitrate moving. N application rate mainly affected the change of soil NO₃-N content in the surface layer (0-20cm), the storage of NO₃-N in the whole profile (0-80cm) increased with N application rate at the end of rotation period.Fertilization model in hilly upland. The field experiments including single N, Ptreatments were carried out in two different soils derived from red sand stone and Quaternary red clay for oil rape and peanut rotation system. Results showed that oil rape yield (Y, kg.hm^-2) was the function of soil classify (AP_class) and N fertilizer (N, kg.hm^-2) with the equation of Y=266.114×AP_class +2.866×N+393.26. Considering economical profit, N fertilizer use efficiency and NO3-N storage in 0-80cm soil profile, 69kgN.hm-2 is a proper amount for peanut. The phosphorus use efficiency decreased with the increasing P2O5 fertilizer for both crops. In oil rape season the phosphorus use efficiency was as high as 10% in the treatment of 90kg.hm-2 P2O5. For peanut the phosphorus use efficiency in red soil derived from red sandstone with higher available phosphorus content is significantly higher (p<0.05) than that from red clay. Economic benefit of phosphorus in the whole rotation was higher for P treatment with lower P application rate. The relationship equation is as follows: P = -0.0463X2 + 2.8253X + 12763, in which P is the phosphorus profit (yuan.hm-2), X is the amount of application P2O5 (kg.hm-2). The investigation of Yujiang County in 2003 showed the available phosphorus in surface soil was increased to a higher level. Therefore decreasing P application is necessary.The integration of fertilization model and GIS. The fertilization model can be embedded in the component geographic information system by using the visual program language, which chooses the open soft structure. This model reduces the development cost; shorten the period of development and integration. The fertilization decision supporting systems was developed for Yujiang county Jiangxi province, by inserting MapObjects into the Visual Basic development environment and combining the techniques of database, ComGIS, decision supporting systems, based on the theory of balanced fertilization. Spatial Database and Attribute Database are the mostly important components of the decision support system. 1:50000 Spatial Database was structured...