摘要
利用Landset_5 TM遥感影像数据,经几何校正、大气校正、辐射校正等前期处理后,计算比值指数(RVI)和再归一化植被指数(RNDVI),结合野外调查获得的自然草地生物量数据,在GIS中反演祖历河流域自然草地生物量,建立回归反演模型,并对两种植被指数的反演精度进行比较;利用RNDVI建立的模型,在Arc GIS的Grid模块中重采样计算,获得流域自然草地生物量的空间分布地理信息数据层。结果表明:RNDVI比RVI反演精度更高,相关系数R2=0.939(P=0.00)。从RNDVI和反演的生物量空间分布规律看,祖厉河流域自然植被生物量空间植被分布极不均匀,流域南部植被生物量明显高于北部地区,高海拔地区明显大于低海拔地区。从畜牧业可利用性分析,流域不可利用草地面积占总草地面积的53.16%,限制性可利用草地面积占总草地面积的28.67%,可利用草地面积仅占总草地面积的18.16%。因此,流域自然草地畜牧业养殖潜力有限,需结合家庭养殖发展畜牧业。
The re-normalized differential vegetation index （RNDVI） and ratio vegetation index （RVI） were calcu- lated by using the Landsat 5 TM image data processed with geometric, atmospheric and radiometric calibrations. The natural grassland productivity GIS regression retrieval model was then established by using the above calculated indexes combined with the natural grassland biomass data from field work in the Zulihe River Basin. The accuracy of the indexes were assessed through the model. The results showed the RNDVI had a higher accuracy than RVI, with a correlation coefficient R2=0. 939 （ P = 0.00）. The spatial distribution of the natural grassland biomass was uneven across the Zulihe River Basin according to the RNDVI and the retrieved distribution. The vegetation biomass was apparently higher in the southern region than that in the north part. From the perspective of forage availability analysis, the unusable grassland accounted for 53.16% of the total grassland in the Basin while the usable grass- land accounted for only 18.16%. The rest of the grassland animal husbandry would be limited by the natural grassland , accounted for 28.67%, was restrictive usable. The productivity and need to rely on livestock breeding to improve the population.
出处
《干旱区研究》
CSCD
北大核心
2014年第6期1147-1152,共6页
Arid Zone Research
基金
"973"预研项目"草地与农牧交错带生态系统生产力形成机制"(G2000028603)
