天然林保护的神经网络分级规划 - 应用生态学报.PDF

2005 6 16 6 CHINESE JOURNAL OF APPLIED ECOLOGY, Jun. 2005, 16( 6)#1002~ 1006 * 1 1 1 1 2 罗传文 陈 焱 胡海清 沈海龙 范少辉 ( 1 , 150040; 2 , 100091) , . Arc/ Info , , , , . , . , , , ; , . 100 1- 9332( 2005) 06- 1002- 05 S7574 A 1 1 1 Neural network grade program of natural forest protection. LUO Chu nw en , CHEN Yi n , HU H iqing , 1 2 1 2 SHEN H ilong , FAN Sh ohui ( N or theast Forestr y Univ ersity , H ar bin 150040, China; I nstit ute of Forestry , A cademy of Ch inese For estry , Beij ing 10009 1, China) Ch in J A pp l Ecol . , 2005, 16( 6) : 1002~ 1006. In this p per , the implement steps of n tur l forest protection progr m gr ding ( NFPPG) with neur l network ( NN) were summ rized, nd the concepts of progr m illustr tion, p tch sign unific tion nd regress, nd inclining f ctor were set forth. Employ ing Arc/ Info GIS, t he tree species diversity nd r rit y, disturb nce degree, protection of ch nnel system, nd cl ssific tion m n gement in Moersh n N tion l Forest P rk w ere described, nd, used s the input f ctors of NN, the rel tionships between NFPPG nd bove f ctors w ere n lyzed. Through rt ifici lly determining tr ining s mples, the NFFPG of Moersh n N tion l Forest P rk w s built. T ested with ll p tches in the p rk, the gener liz tion of NFFP G w s s tisfied. NFPPG took both the cl ssific tion m n gement nd the protection of forest community types into ccount, s well s the ecologic l environments. T he ex cit t ion function of NFPPG w s not seriously s tur ted, indic t ing t