Estimation Of Forest Net Primary Productivity In Heilongjiang Province Based On Remote Sensing

Forest net primary productivity(NPP) is one of the basic quantitative characters of forest ecosystem.NPP represents the biologic and systematic properties for a specific kind of forest ecosystem;besides,it also reflects the impacts of various environmental factors on forest growth.NPP acts not only as the driving force of the carbon cycle,but also as a primary factor to investigate the carbon source and sink,and to adjust the ecosystem processes.Using the ETM+ remote sensing data,DEM data and meteorological data,and developing the NPP estimate system of forest vegetation that based on the C-FIX estimate model to study the forest resources of Heilongjiang Province by GIS and RS,the NPP of forest vegetation of Heilongjiang Province from May to September in 2000 was estimated, and the spatial pattern of forest NPP in Heilongjiang Province and the relationship between NPP and site type factors were quantitative analyzed,which provided a theoretical foundation for further prediction of carbon sink ability of terrestrial ecosystem in Heilongjiang Province.The primary conclusions were listed as below:(1) The gross NPP of forest in Heilongjiang Province from May to September in 2000 was 69.75×10¹²gC,the maximum NPP value for month appeared in June;shared 27.23% of the gross forest NPP from May to September.June,July and August were the major months for NPP to accumulate.In May,high levels of NPP distributed among the northern of Great Xingan Mountains,the southern and north-western of Xiao xing'an Mountains, and the southern of Zhangguangcai Range and Laoye Range.In September,high level of NPP mainly centered around southem of Xiao xing'an Mountains and Zhangguangcai Range.Forest land owned gross forest NPP of 60.99×10¹²gC,which accounted for 87.44%of total forest NPP from May to September,while shrub land 10.41%,open forest land 1.63%,and other types of land accounted for 0.52%of the gross forest NPP.(2) Got the forest NPP that respectively corresponding to the six categories soil by mask processing the gross NPP of Heilongjiang Province from May to September in 2000. It showed that the order of Vegetation NPP level from high to low was dark brown forest soil,brown taiga soil,swamp,meadow soil,planosol and black soil,meanwhile,the average forest NPP from high to low was dark brown forest soil,brown taiga soil,black soil,meadow soil,swamp and planosol.The order of average NPP basically reflected the integrated soil fertility of these six soil types in Heilongjiang Province.(3) The monthly NPP estimation model of Heilongjiang Province from May to September in 2000 was build by analyzing the relationship between vegetation NPP of forest and the latitude and longitude,elevation,temperature,solar radiation,slope gradient, slope direction and soil type.It concluded that the forest NPP in Heilongjiang was closely related to the elevation and solar radiation.(4) Based on the C-FIX model,taking ERDAS IMAGE as the platform for the development,the quantitative estimate system of forest vegetation NPP in Heilongjiang Province was realized through spatial Modeling Language(SML) and the ERDAS Macro Language(EML).The system could accomplish the evaluation and extraction of NPP data for various soil types according to the classified remote sensing image data,and achieve the goal of estimating forest NPP quantitatively and automatically.The innovation of this dissertation:(1) The forest vegetation NPP from May to September was quantitatively estimated using TM images in the scale of Heilongjiang Province,which provided an idea and method to the NPP study of province scale.(2) A monthly forest NPP estimation model was built using multivariate linear regression method,which analysed the relationship of NPP and the latitude and longitude, elevation,temperature,solar radiation,slope gradient,slope direction and soil type in Heilongjiang Province from May to September in 2000.(3) The quantitative estimate system of forest vegetation NPP in Heilongjiang Province was developed and realized,taking ERDAS IMAGE as a platform.This estimate system realized the quantification and automation of forest vegetation NPP estimation that based on the classified remote sensing image data and related factors.