A Study Based Upon MODIS DATA Of Hulunbeier Steppe ANPP Dynamic Variation

At present,along with technical of RS,GIS and GPS,there are many methods of steppe monitoring.In order to adapt to the new situation of steppe monitoring, information technology must be applied to large areas of steppe resources survey and monitoring. We have access to a variety of remote sensing data, and remote sensing image content is also very rich. Although remote sensing is applied to Above-ground Net Primary Productivity estimation,research based on use of ground-spectroscopy and remote sensing data to model predicted grassland biomass is little. Therefore,research based on the relationship between the remote sensing data and ground-spectroscopy data is important to Monitoring of Aboveground Net Primary Productivity.In order to improve the research and application of hyperspectral remote sensing to estimate the grassland biomass, ASD Fieldspec3 Spectroradiometer is used for the spectral measurements of arid steppe on Hulunbeier Plateau. The models are built for expressing the relationship between MODIS NDVI and NDVI from ground experiments by the one-sample linear regression method, the nonlinear regression method and the stepwise analysis method. Relationship model between MODIS NDVI and NDVI from ground experiments can be basis of large-scale using of high-altitude remote sensing geographic.we also bulid models for estimating Above-ground Net Primary Productivity (ANPP) with Normalized Difference Vegetation Index (NDVI). Based on the above two models, we can use the MODIS NDVI estimate Above-ground Net Primary Productivity(ANPP). We estimated the Hulunbeier steppe ANPP production which is under the influence of recipitation from 2000 to 2009. The main research results are as follows:(1)The results shows that it is linear relationship between MODIS NDVI and F-NDVI which is from ground experiments.It can be expressed by the equation:MODIS NDVI=0.808F-NDVI+0.106(R2=0.785, P<0.001);the relationship between AVHRR NDVI and F-NDVI is also linear,it can be expressed by the equation: AVHRR NDVI=0.584F-NDVI+0.043(R2=0.65, p<0.001).After taking all factors into consideration, MODIS data is selected and put to use.(2)The ground spectral models are built for estimating Above-ground Net Primary Productivity (ANPP) with Normalized Difference Vegetation Index (NDVI) by the one-sample linear regression method, the nonlinear regression method and the stepwise analysis method.Exponential model is selected and put to use, it can be expressed by the equation:ANPP=12.631e3.8972F-NDVI(R=0.85, p<0.001).(3)therefore,the equation used for ANPP estimation is as follow: ANPP=12.631e3.8972*[(MODIS NDVI-0.106)/0.808](4)Based on the study results,MODIS NDVI can be used estimating Above-ground Net Primary Productivity (ANPP). We estimated the Hulunbeier grassland ANPP production which is under the influence of recipitation from 2000 to 2009.The results show that both production of ANPP and precipitation are all getting smaller in the last decade.But they have no strong correlations.Because ANPP is affected not only by precipitation but also by evaporation, temperature and other factors. We must study the ANPP which is under the influence of multiple factors.