| The rapid economic development of mankind makes the content of CO2, CH4 and other gasses in air increased. It had brought many serious problems, such as the destruction of ecological system, greenhouse effect and many other serious problems. The carbon in the earth transferred from one subsystem to the other is called the global carbon cycle. Net Primary Productivity stands for the quantity converting from light to organic, and it plays a vital role in the terrestrial carbon circulation system.There are rich natural resource in the Midwest of Jilin province which is a science research region of grassland resource. The carbon cycle of this grassland is the most important physical and chemical processes of grassland biosphere. From the economic factors, the study of the carbon cycle of grassland is closely related to livestock and industry; From the perspective of sustainable development factors, the research of carbon cycle have a great significance on climate, water resources, controlling environmental pollution, etc.In this paper, we use the remote sensing data, the Meteorological data and the CASA model to carry out the research on NPP of the Midwest of Jilin province. The visual c++ programming language and GDAL library are used to read and display the raster data and get the vegetation value of NDVI. By combining with the interpolation module of meteorological data, we take the CASA model as a tool to carry out the NPP research. This software is a useful tool for a digital management of the grassland.The application program is development based on GDAL and MFC under the platform of VC6.0. MFC is used to implement the whole framework. GDAL is used to convert and read the raster data. The application program can be divided into three modules:the first one is the remote sensing image processing module, the function is to achieve fast reading of the large img file and compute each band albedo of remote sensing data, and then estimate NDVI normalized difference vegetation index and FPAR parameters. The second is meteorological data processing module, the function is to finish spatial interpolation of meteorological data, and then compute light energy conversion rateεof vegetation. The third is estimate NPP module, which puts the parameters obtained from above two modules into this module to get the result.After researching and developing the application program of the remote sensing NPP estimate based on CASA module, we got the following conclusions:(1) The information of the meteorological data in spatial is significant to estimate NPP in research regions, but the present meteorological data can be only received from some limited meteorological stations, and it requires interpolation into a continuous distribution spatial information. Only when the number of observation stations enough, the regional interpolation has higher accuracy. On the contrary, or in the area of surface changed much, it is difficult. Considering the special nature of interpolation in the sparse meteorological stations, we apply kriging interpolation algorithm in the new additional simulation station, and propose a novel spatial interpolation algorithm which is fit for sparse stations. With our algorithm, producing simulation station gets more simple under the condition of unaffecting the spatial variability of measured data. Also, the uniformity of stations can be ensure; kriging interpolation can be simple; computation speed can be improved. Finally, by cross testing, the experimental results show that error estimation of our method is less than existing ones.(2) After researching the NPP of grassland in Midwest of Jilin province, we can get that the NPP in the research area is consistent to the real area, so we can get the accurate results with our application program.(3) According to the characteristics of remote sensing images, this paper combine with meteorological data interpolation method to develop a software which is used to estimate NPP. The traditional software is non-open source or limited open source and professional of operators. Comparing with the traditional software needed to use remote sensing image processing software and GIS software together, our software is more convenient, and is easy to use. |
