摘要
采用探索性数据分析技术——投影寻踪回归(PPR)模型,定量研究了干旱区农业土地生产力影响因素问题。以塔里木河流域40个县市作为研究单元,以近20a(分5个研究时段)的农业土地生产力(粮食产量)作为因变量,人类活动的5个因素(引水量、耕地、人力、化肥和机械力)作为自变量,通过建立PPR模型,研究了这一区域农业土地生产力增进中人类活动因素的贡献度。研究结果表明,耕地、引水量等资源性因素虽然在20世纪90年代后有一定程度的下降,但仍是这一区域生产力增进的主要因素,化肥增进作用显著。结果也表明,PPR模型适合于这类研究。
The projection pursuit regression (PPR), a new exploratory data analysis method, was used to make a quantitative evaluation on human factors in the increase of land agricultural productivity (LAP) in arid area. Taking the 40 counties in Tarim River basin as samples, using food output as dependent variable, and 5 factors of human activities (water, cultivated land, agricultural labor forces, fertilizer, mechanical power) as independent variables, 5 PPR models were set up at 5 time phases (1980,1985,1990,1995,2000) so as to study the contributing degrees of the factors of human activities in the LAP increasing. The computing results show that the contributing degrees of resource factors (water and land) have appeared some descend since 1990's but is still the dominant factors in LAP increasing. The increasing role of fertilizer is obvious. The results also prove that PPR is a practicable method for the similar studies.
出处
《干旱地区农业研究》
CSCD
北大核心
2004年第3期168-174,共7页
Agricultural Research in the Arid Areas
基金
国家重点基础研究发展规划项目(G1999043505)
关键词
干旱区
农业
土地生产力
人类活动
塔里木河流域
化学肥料
land agricultural productivity
human activities
Tarim River Basin
projection pursuit
dynamic analysis
arid area
