GIS-Based Forage Adaptability In Decision Support System Of Returning Cropland Back To Grassland

GIS-based Forage Adaptability in Decision Support System of Returning Cropland back to Grassland Ph.D. Candidate: Sun JuanSupervisor: Jiang Wenlan, ProfessorForage adaptability (FA) is the main problem of returning cropland back to grassland in Gansu province. The present research attempts to computerize FA and design a decision support system called GIS-based FA in Decision Support System of Returning Cropland back to Grassland (GFADSSCG), and hopes that it can provide support for the work in Gansu province. Several studies have already been developed to computerize FA. Most of them concentrated on quantitative part of FA including how the climatic factors meet forage growth and how forag fits in with ecological factors of their growth conditions. First, the main factors affecting FA were ascertained according to theories of grassland integrative sequence taxonomy, the eco-climate suitability degree to forage growth and forage biological traits. Second, membership functions of these two aspects were ascertained. Membership functions of the former aspect were studied several years ago, and this study combined it with fuzzy logic in order to gain the forage distribution range. And we built membership functions of the other aspect with fuzzy logic and Analytic Hierarchy Process (AHP). Thirdly, the reasoning mechanism for FA was built based on these membership functions. Finally, it was combined with GIS through computer programs and the collected knowledge on FA. GFADSSCG was built based on GIS. GFADSSCG employs the power and abilities of the Decision Support System (DSS) in solving complex problems to help the users select suitable forage for a target site of returned cropland. The results are as follows:Quantitative definition of FA is decided by the matching degree between forage and ecological factors of their growth conditions, including how the climatic factors meet forage growth and how forage fits in with ecological factors of their growth conditions. The former aspect reflects FA's spatial continuous distribution, and the latter reflects the suitable forage of a target site. The main affecting factors of FA are annual sunlight hours (I), annual average temperature (T), relative humidity (R), annual accumulative temperature (＞0℃) (θ), annual rainfall (P), frost-free days (D), elevation (A), maximal temperature (H), minimal temperature (L), up to 9 factors, in all.In regard to membership functions of affecting factors for FA, the above 9 factors were used to study two aspects of FA. (1)annual sunlight hours (I), annual average temperature (T), relative humidity (R), annual accumulative temperature (＞0℃) (θ), annual rainfall (P), frost-free days (D) were used to study the former aspect, and their membership functions refer to membership functions of climatic fitting degrees of forage growth. (2)annual accumulative temperature (＞0℃) (θ), annual rainfall (P), frost-free days (D), elevation (A), maximal temperature (H), minimal temperature (L) were used to study the latter aspect, and their membership functions were formulated by fuzzy logic and the weights between 6 factors were ascertained through AHP, the total membership function is T=0.3Tθ +0.2Tp+0.2TD+0.1TA+0.1TH+0.1TL, thereinto, Tθ ,Tp,TD,TA,TH and TL are membership function of annual accumulative temperature (＞0℃) (θ), annual rainfall (P), frost-free days (D), elevation (A), maximal temperature (H), minimal temperature (L), respectively. Reasoning mechanism for FA was built based on fuzzy logic and the collected knowledge of FA.Reasoning mechanism consists of two aspects: judgments of whether forage species can grow at a target site and what forage specie(s) can grow on a target site. The first aspect was used to describe special distributions of forage species, and the second aspect was used to choose suitable forage specie(s) for a target site.22 GIS-based FA distribution maps were formed through combining the reasoning mechanism for FA with GIS by computer programs, and the reasoning mechanism for Deci...