Overview of applications of Remote Sensing Data Records and Reanalysis for the study of surface processes

The EUMETSAT Land Surface Analysis Satellite Application Facility (LSA-SAF) now offers a wide range of satellite-derived products for land surface monitoring. The catalogue comprises variables quantifying different terms of the surface energy balance (land surface temperature – LST - and emissivity, downwelling radiative fluxes and turbulent fluxes), as well as several vegetation-related indicators, such as the Leaf Area Index, Fraction of Vegetation Cover, Evapotranspiration, Net Primary Production and Fire Radiative Power. The availability of these datasets, especially taking into account that the time series now span nearly two decades,  already allows many interesting applications, overviewed in this presentation.

Comparisons of remote sensing data for land surfaces with corresponding model data have already been useful: the standard L2 (clear sky) LST has been used to diagnose a systematic cold bias of ERA5 skin temperature over the Iberian Peninsula. Offline simulations using H-TESSEL revealed that the bias could be alleviated using a more realistic representation of vegetation than what is currently used in ERA5. A recently developed product by LSA SAF allows LST retrievals for all-weather conditions, using a surface energy balance model to provide estimates under cloudy pixels. This product is compared to ERA5-Land skin temperature, showing that despite the increased level of detail of the latter (with respect to ERA5), it is still not representing the former correctly. ERA5 Land skin temperature shows large biases (of more than 10 K) and phase errors (with the satellite LST warming up prior to ERA-Land during the morning and cooling down earlier in the late afternoon). Comparisons of the different terms of the surface energy balance from ERA5-Land and LSA SAF are currently in progress to identify causes of the biases.

Another interesting application of LSA SAF products is the study of vegetation recovery over wild fire scars. Five wild fire events over Portugal were analyzed in terms of the long term anomalies introduced by the fire in 3 variables: LST, Albedo and Fraction of Vegetation Cover (all provided by LSA SAF). Results suggest that albedo returns to close-to-normal conditions in less than a year, while LST anomalies last much longer.  

Finally, trends in the land-ocean thermal contrast were evaluated over Western Iberia and Northwest Africa (due to its importance in generating coastal mesoscale circulations). The study used long time series from 1) satellite – LST from CM-SAF and SST from GHRSST; 2) ERA5 global reanalysis and 3) UERRA regional reanalysis. The results strongly depend on the used dataset and sub-region, with UERRA showing a sharp decrease of the thermal contrast over Iberia, while ERA5 shows a positive trend.

These results emphasize the need to improve the representation of surface processes in numerical models, particularly over land surfaces. This presentation shows that datasets such as the ones provided by the LSA SAF are key to such improvements.