The SMOS level 3 soil moisture was developed to take advantage of several revisits of the same point in 7 days. This multi-orbit retrieval was supposed to stabilize the VOD and so to improve the SM retrieval quality. Unfortunately, it was not the case leading to poor performances of the L3SM when compared to other soil moisture product.
This was not good so….
So we have worked on this issue for some time to identify the issue and correct it . We found the issue (Arnaud and Philippe) and the modification has been implemented in the operational processor since mid July 2019. The figure below presents the time series of level 3 SM and VOD L3 (in blue) for a site in Australia. Before the modification (left hand side of the vertical dotted line), one can noticed spurious peaks of SM (indicated by green circles) that are not present in other SM products (ESA level 2 in orange, SMOS IC in red). These poor retrievals are also noticeable in the VOD plot, where high values are reported.
Figure showing the temporal behaviour of soil moisture and VOD retrievals before and after the “bug” correction in July 2019.
With the correction, SMOS level 3 performs normally. A complete reprocessing with the corrected algorithm will be done in 2020, taking advantage of the general reprocessing with the new L1 and L2 processors.
Arnaud Mialon CESBIO