基于Sentinel-2数据的典型草原光合/非光合植被覆盖度估算

Estimating fractional cover of photosynthetic/non-photosynthetic vegetation in a typical steppe region based on Sentinel-2 data

及时、准确地掌握草原光合植被覆盖度(fractional cover of photosynthetic vegetation,fPV)和非光合植被覆盖度(fractional cover of non-photosynthetic vegetation,fNPV)信息对草地资源的合理利用以及草原生态系统的监测和保护具有重要意义。本研究选取锡林郭勒典型草原为研究区,以Sentinel-2A MSI多光谱影像为数据源,基于光合植被(photosynthetic vegetation,PV)-非光合植被(non-photosynthetic vegetation,NPV)-裸土(bare soil,BS)模型,利用多端元光谱混合分析(multiple endmember spectral mixture analysis,MESMA)方法估算研究区的fPV、fNPV,并对感兴趣区内fPV、fNPV的时空动态变化进行分析。结果表明,对于锡林郭勒典型草原来说,PV、NPV以及BS光谱库内端元存在明显的光谱异质性特征,确定最优的端元组合是成功估算fPV和fNPV的关键;MESMA可较好地估算fPV和fNPV,fPV估算的RMSE为5.52%(R2=0.72),fNPV估算的RMSE为9.37%(R2=0.61);利用MESMA估算fPV和fNPV的季节变化符合草原植被物候发育特征,可应用于监测典型草原地区fPV和fNPV的时空动态变化。

It is important to timely and accurately understand the information on fractional cover of photosynthetic vegetation(fPV)and non-photosynthetic vegetation(fNPV)for the rational use of grasslands,as well as monitoring and protecting grassland ecosystems.In this study,we used the Xilingol typical steppe region as the study area and used the Sentinel-2A MSI multispectral imagery data as the data source.Based on the model for photosynthetic vegetation(PV),non-photosynthetic vegetation(NPV),and bare soil(BS),the multiple endmember spectral mixture analysis(MESMA)method was used to estimate fPV and fNPVin the study area.Then,we used MESMA to analyze the spatiotemporal dynamic changes in fPV and fNPVin the region of interest.The research results showed:For the Xilingol typical steppe region,the PV,NPV,and BS endmember libraries had an obvious intra-variability feature.Therefore,determining the optimal endmember combinations is a key process to successfully estimating fPV and fNPV;MESMA could estimate fPV and fNPV well,with an RMSE of5.52%for fPV(R^2=0.72)and an RMSE of9.37%for fNPV(R^2=0.61);Estimated seasonal variation in fPV and fNPVusing MESMA was consistent with the developmental characteristics of pasture phenology,and MESMA can be used to monitor spatiotemporal dynamic changes in fPV and fNPVin the typical steppe region.