Retrieval of grass canopy water content, leaf area index, and dry weight using TM/ETM+ data, a case study in a typical semi-arid steppe in northern china

In this research, three field surveys were carried out in a typical semi-arid steppe in northern China, respectively on early July, late July and late August in 2008. During each survey, 80 plots established according to the cyclic sampling design were investigated, and three bio-parameters including LAI (Leaf Area Index), CWC (Canopy Water Content) and DW (Dry Weight) were measured. Temporalspecific correlation analysis between TM spectral variables and bio-parameters revealed that the strength of the correlation changed dramatically across time. NDVI was more correlated with CWC and DW in late July, whereas LSWI was a better predictor for CWC and DW in late August. LAI was poorly correlated with TM spectral variables, except in late August. When the three survey datasets were analyzed together, TM band 5 and LSWI were the most suitable for bio-parameters estimation, highlighting the importance of SWIR band in ecological monitoring for semi-arid steppe.