Comparison of X-Band and L-Band Soil Moisture Retrievals for Land Data Assimilation

This paper explores for the first time assimilation of the X-band soil moisture retrievals by the advanced microwave scanning radiometer-Earth observing system and the advanced microwave scanning radiometer 2 in Environment Canada's standalone Modelisation Environmentale Surface et Hydrologie model over the Great Lakes basin, in comparison with the assimilation of L-band soil moisture retrievals from the soil moisture and ocean salinity mission. A priori rescaling on satellite retrievals is performed by matching their cumulative distribution function (CDF) to the model surface soil moisture's CDF, in order to reduce the satellite-model bias in the assimilation system. The satellite retrievals, the open-loop model soil moisture (no assimilation), and the assimilation soil moisture estimates are validated against point-scale in situ measurements, in terms of the daily-spaced anomaly time series correlation coefficient R (soil moisture skill). Results show that assimilating X-band retrievals can improve the model soil moisture skill for both surface and root zone soil layers. The assimilation of L-band retrievals results in greater soil moisture skill improvement Δ R A-M (the assimilation skill minus the skill for the open loop model) than the assimilation of X-band products does, although the sensitivity of the assimilation to the satellite retrieval capability may become progressively weaker as the open-loop skill increases. The joint assimilation of X-band and L-band retrievals does not necessarily yield the greatest skill improvement. Overall, Δ R A-M exhibits a strong dependence upon the difference between the satellite retrieval skill and the open-loop surface soil moisture skill.