RS,GPS和GIS集成系统在新疆北部天然草地估产技术中的应用进展

ADVANCES IN STUDY ON THE REMOTE SENSING TECHN OLOGY AND GPS AND GIS INTEGRATION SYSTEMS IN ESTIMATING GRASSLAND YIELD APPLICATIONS IN THE NORTH OF XINJIANG,CHINA

利用１９９１～１９９６年在新疆天山北坡不同草地类型上观测的草地可食产量，环境与遥感资料等，使用ＲＳ技术、ＧＰＳ和ＧＩＳ集成系统进行了多重相关分析和遥感估产技术的深入研究，并在图象处理、信息提取、信息应用和ＲＳ－ＧＰＳ－ＧＩＳ一体化估产方法及遥感知识与草原生态专业知识结合方面获得一定研究进展。研究结果表明，４个草地类型的可食鲜干草产量与两种遥感绿度值间存在着极显著相关性（Ｐ＜０．０１），ｒ值均在０．６７９以上，且通过Ｆ检查和精度分析。一般在类型Ⅱ、Ⅲ和Ⅳ，是鲜草产量与ＲＶＩ相关性好于ＮＤＶＩ，而在类型Ⅰ则相反。进而从６种数学方程式中选优，建立了地学、光学和非线性遥感估产模型，并在实际估产中加以应用、检验和给出了生态学解释及机理分析，使大面积草地可食牧草遥感估产精度达到７５．８％以上，实现了遥感大面积估产目标和草地生态学意义及ＲＳ－ＧＰＳ－ＧＩＳ与草地专家系统一体化集成的应用。

Using the green herbage yields,environment and the remote sensing(RS)data observed in different grassland types in the northern slope of Tianshan,Xinjinag Province from 1991 to 1996,the correlation analysis and the estimating yield methods were studied using the 3S technology integration.The methods of processing images,and collecting and using information,and lining RS data and grassland data were explored in the paper.The results showed that there existed an obvious correlation between the fresh herbage yields and RVI and NDVI ( P <0 01)in grasslands of four types,the correlation coefficients( r )were above 0 679.The study also found that the correlation between fresh or hay yields and RVI was better than NDVI in Ⅱ,Ⅲ and Ⅳ types.The opposite was found in I types.The finest nonlinear geographical,the optical and satellite remote sensing estimating yield models were made from 6 curve types.The effects were checked by the observing yields and statistics.The ecological interpretations and mechanism analysis of the results were given.The total accuracy of the RS estimating green yields was above 75 8% in large areas.The objectives were realized by the RS,GPS and GIS and grassland expert system integrations.