Study on Forest Growth Predicting by Using MODIS-NDVI Time Series

In this paper, the southwest forest area of China is chosen for the study area. This paper uses the first and second channel of MODIS L1B data which is been preprocessed as scaling, geometry correction and so on, for computing NDVI. The pretreatments are aimed at removing the effects from atmosphere, sensor and so on. Then we build NDVI time series for years based on MVC arithmetic. After that we build polynomial forest growth predicting model and Gauss-Newton forest growth predicting model for predicting the forest growth of study area. Meanwhile, this paper experiments the two models with real data for predicting the forest growth of the study area. The result shows that using MODIS-NDVI time series and mathematical method to build predicting models of forest growth is feasible.