Study On The Theory And Method On Dynamic Forescating For Forest Resources Based On Two Dimensional Kalman Filtering

The article established the two dimensional coordinate system and studied the dynamic of forest stock volume and the area of forest trees by using kalman filtering and the material of sample plot.The main results are al follows:The author builded two dimensional coordinate system by using the horizontal,y-coordinate of 481 sample plots of forest resource inventory from 1999 to 2002 in Lishui,Zhejiang Province.The author studied the dynamic for forest resource by taking forest resources in a sample plot as state variable.It show that two dimensional kalman filtering can not only be used in predicting the microscopic dynamic change for forest resources but also in predicting the macrospic dynamic change; The sample plots are divided into pine,Cunninghamia lanceolata and broad-leaved tree species according to their dominant tree species and their tree ages.The biggest error is only 4.98%.It show the forecasting model and the method are good.Land in Lishui will be divided into timber protective forest , economic forest , young stand, middle-aged stand, near-mature forest, mature forest, over-mature stand and so on according to land classification technological standard.The author studyed the dynamic for forest land.It show that the error in area prediction and economic forest is small because there are more sampling units and more area in these classes.But there are bigger error in special stand and unsucceed attorestation sites.It will need to enlarge sampling unit to make more accuration.But the error in predicting result of mature stand and over-mature stand is bigger because there are only a few sampling unit and a bigger coefficient of variation. We will not get better answer unless we enlarge more sampling unit...