关键词: 改进单纯形, 神经网络, 自疏规律, BP-MSM混合算法

Abstract: From May 2004 to February 2005,an investigation was made on the avian communities in four typical habitats,i.e.,village and farmland,urban woodlot,floodplain,and riverine wetland in Wuhu City,with the related diversity indices analyzed.A total of 117 species belonging to 15 orders and 31 families was recorded.The avian species and G-F index in village and farmland,urban woodlot,and floodplain were higher in summer than in winter,while the species number,ShannonWiener index and Pielou index in riverine wetland were significantly higher in winter than in summer.The avian species and G-F index in urban woodlot were the highest,and the Shannon-Wiener index and Pielou index in floodplain were higher than those in other three habitats.The highest density was observed in village and farmland in summer and in urban woodlot in winter,being 31.1140 and 42.9160 ind·hm^-2,respectively.The avian species,density,Shannon-Wiener index and G-F index in riverine wetland were the lowest.Avian diversity could be enhanced by increasing spatial heterogeneity and reducing human disturbance.

Key words: Modified simplex method, Artificial neural network, self-thinning, BP-MSM mixed algorithms