基于小型无人机与MODIS数据的草地植被覆盖度研究--以甘南州为例

Grassland fractional vegetation cover analysis using small UVAs and MODIS——A case study in Gannan Prefecture

以甘南州为研究区,采用Canon数码相机和小型无人机搭载相机在不同大小样方上获取草地植被覆盖度数码照片,结合2015年5月-10月的Terra/MODIS植被指数产品MOD13Q1,分析了增强型植被指数(EVI)和归一化植被指数(NDVI)与草地植被覆盖度之间的相关性,建立了研究区草地植被覆盖度的回归模型,并对模型进行了精度评价,筛选出甘南州草地植被覆盖度最优遥感反演模型,并对草地生长季时期覆盖度时空上的动态特征进行分析。结果表明,1)利用小型无人机搭载相机获取草地大样方植被数码照片的方法能用于地面草地覆盖度数据的获取;2)与NDVI相比,用EVI估算草地覆盖度更优,因此确定基于EVI构建的对数模型为甘南州草地植被覆盖度最优反演模型,模型精度可达88.00%;3)研究区2015年生长季草地植被覆盖度除了低平地草甸在8月达到最大值外,其它草地类型均在7月达到最大值;4)甘南州以中高植被覆盖度为主,主要分布在玛曲、碌曲、夏河以及合作四县市。整体而言,中西部和西南部区域草地覆盖度高于东部。通过精确草地植被覆盖度模型的建立,不仅有利于及时准确的了解草地植被覆盖度的时空分布状况和季节性动态变化,也有利于维护甘南州草地生态系统的持续稳定发展。

Gannan Prefecture was selected as the study area. Photos of different sized subplots of grasslands were taken with a camera （Canon） and small UAVs （unmanned aerial vehicles）. Terra/MODIS MOD13Q1 vegetation indices were obtained from May to October 2015. Correlation analyses were conducted between vegetation indices [enhanced vegetation index （EVI）, normalized difference vegetation index （NDVI）] and grass-land fractional vegetation cover （FVC）, and regression models were determined. Based on the accuracy of different regression models, the best estimation model for grassland FVC in Gannan Prefecture was selected, and grassland spatio-temporal characteristics of FVC during the growing season were analysed. The results showed that 1） small UAVs carrying cameras could be used to obtain digital photos containing data on grassland FVC; 2） EVI was more correlated with grassland FVC than NDVI, and the logarithmic model based on EVI was the most suitable model for monitoring grassland FVC in Gannan Prefecture, with model accuracy reaching 88.00% ; 3） grassland FVC of flat meadows in 2015 reached the maximum in August, and the remaining grassland types reached the maximum in July; 4） grasslands in Gannan Prefecture were dominated by high and midgrassland FVC levels, and these grasslands were mainly distributed in Maqu County, Luqu County, and Xiahe County and Hezuo City. Overall, the grassland FVC in the mid-western and south-western areas was better than that in the eastern area. Establishment of the best model for grassland FVC was beneficial not only in elu- cidating the spatio-temporal distribution and seasonal dynamic variation of grassland FVC, but also to the maintenance of stable development of grassland ecosystems in Gannan Prefecture.