摘要
提高净初级生产力(NPP)的估算精度,需要充分认识不同空间分辨率的遥感数据对NPP估算结果的影响差异,即NPP的空间尺度效应问题。本文借鉴了多种成熟的光能利用率NPP模型的优点,同时充分考虑了研究区生态环境的典型特点,建立了针对研究区域的基于光能利用率原理的植被净初级生产力遥感估算模型。选取了具有空间尺度代表性的4种遥感数据作为NPP遥感估算模型的输入参数,估算了4种空间分辨率的NPP模拟值,对比分析了这4种分辨率的土地覆盖类型空间格局的变化和NPP分布情况。结果表明:随着遥感数据空间分辨率的降低,NPP的模拟值呈逐渐增加的趋势,但其变化的程度差异较大。其中,分辨率由30m降低为1km时,NPP模拟值变化程度最大,增加了15.7%。
Scale effect of net primary productivity（NPP）models refers to the effects on NPP estimates from remote sensing data at different spatial resolutions.To improve the reliability of NPP estimates,the scale effect of the NPP model needs to be further studied.In this paper,a vegetation NPP model was proposed based on light use efficiency theory and remote sensing data.Meanwhile,many models based on light use efficiency were made use of,with accounting for typical characteristics of the environment in the study area.Four types of remote sensing data representing different scales were used as inputs for the NPP models,and the NPP at the four spatial resolutions was calculated in succession.Furthermore,the estimated NPP results and land cover maps derived at the four spatial resolutions were analyzed in detail.The study was conducted in the middle regions of Dari County in Qinghai Province,situated between 98°51＇E and 99°44＇E and between 33°18＇N and 33°43＇N.Alpine meadow dominates the study area.Remote sensing data used encompassed MODIS,Landsat TM,CBERS and NOAA/AVHRR data.A new classification method integrating supervisor classification,non-supervisor classification and decision tree classification was adopted to derive land cover types with 14 classes from remote sensing images,involving city,water body,ice,desert,gobi,sparse shrub,sparse grassland,grassland mixed with cropland,cropland,dense shrub,dense grassland,needle-leaved forest,needle-leaved and broad-leaved mixed forest,and broad-leaved forest.In order to analyze variations in spatial patterns for different resolution images,number of patches,fractal dimension index,dominance index,fragmentation index,and Shannon diversity index were calculated at 20 m resolution,30 m resolution,1 km resolution and 8 km resolution,respectively.Results indicated that in July 2006,mean NPP values in the study area were 49 gC/m2 at 20 m resolution,51 gC/m2 at 30 m resolution,59 gC/m2 at 1 km resolution,and 63 gC/m2 at 8 km resolution,respectively.This demonstrates
出处
《资源科学》
CSSCI
CSCD
北大核心
2010年第9期1783-1791,共9页
Resources Science
基金
教育部人文社会科学重点研究基地项目基金(编号:08JJD790142)
辽宁省教育厅创新团队项目(编号:2007T095)
辽宁省教育厅科学研究一般项目(编号:L2010226)
关键词
净初级生产力
光能利用率模型
多源遥感数据
尺度效应
Net primary productivity
Light use efficiency model
Multi-source remote sensing data
Scale effect
