Applying scaled vegetation greenness metrics to constrain simulated transpiration anomalies: A study over Australia.

The feasibility of using vegetation greenness metrics as a proxy for transpiration variability over Australia is demonstrated. Several global evapotranspiration datasets, one of which provides transpiration data and is constructed independently of the vegetation greenness measurements, are compared to four satellite-based observations representative of the state of the vegetation over several regions in Australia. Further estimates of the transpiration are obtained by decomposing the evapotranspiration datasets using an ensemble of land surface model simulations. On monthly time scales, the greenness anomaly metrics show a near one-to-one relationship with the transpiration estimates when the time series are appropriately scaled by the mean. The authors demonstrate that anomalous vegetation greenness metrics, when properly scaled, provide a tool for evaluating transpiration variability simulated by land surface models and observation-based evapotranspiration datasets that include transpiration. These methods provide a new test to help constrain the dynamic behavior of the land surface in climate model simulations.

FIG. 2. Kendall tau between the monthly anomalies of the (left) GLEAM ET and (a) MODIS LAI, (c) MODIS EVI, (e) AVHRR NDVI, and (g) MODIS-derived PVF. Also shown is Kt between the monthly (right) J2010 ET anomalies and (b) MODIS LAI, (d) MODIS EVI, (f) AVHRR NDVI, and (h) MODIS-based PVF. Only values of Kt that are significant at the 95% level are included.