Roadmap for future ATBD development

Across CIMR bands, L-band has the highest sensitivity to soil moisture, but also the coarsest spatial resolution. At its native resolution, CIMR L-band meets the observations needs of hydroclimatology (<60 km, daily), yet a potential CIMR enhanced L-band is targeted to meet the needs of global hydrometeorology (~15 km, daily). The exploitation of CIMR oversampling by means of Backus-Gilbert or Scatterometer Image Reconstruction techniques applied at L-band will be instrumental to this end. This will be developed as part of a CIMR Remapper tool and is out of the scope of the DEVALGO project. In this ATBD, a prototype algorithm for CIMR L-band sharpening with C/X-bands to estimate an equivalent ~15 km L-band has been developed, based on the approach of [Santi, 2010]. The spatial enhancement of L-band brightness temperatures has been previously demonstrated based on SMAP/AMSR radiometry data [Santi et al., 2018, Zhang et al., 2025], which supports the conclusion that CIMR C-band can capture information for the spatial enhancement of CIMR L-band. The investigation of OZA correction at C/X-bands as well as sharpening approaches for L-band resolution enhancement will be central for the development of the CIMR multi-frequency soil moisture product.

A central part of this ATBD is the performance assessment of the prototype soil moisture algorithm. Soil moisture retrievals are conducted at two spatial scales. The first retrieval is based on the inversion of L-band simulated TBs at their native resolution (<60 km, 36 km EASE2 grid). The second retrieval is based on disaggregation of L-band simulated TBs to a finer resolution (~15 km, 9 km EASE2 grid) using C-band simulated TBs. To this end, the performance assessment has been focused on a synthetic reference dataset. Further steps require the application of the approach to satellite data and validation of the obtained SM estimates against in-situ and model reference data. Validation shall follow the soil moisture product validation good practices protocol [Montzka, 2020], potentially leveraging the Quality Assurance for Soil Moisture (QA4SM) service.

For the development of the CIMR soil moisture algorithm, one central task should be the global calibration of soil roughness and effective scattering albedo across L-, C-, and X-bands with satellite data. This new calibration shall be assessed against the proposed calibration in the current prototype, which is based on the SMOS-IC retrieval algorithm [Fernandez-Moran et al., 2017].

Another important aspect is the land surface temperature, which is needed as input for the SM retrieval algorithm. To this end, ERA5 skin and soil temperatures are proposed as primary temperature ancillary datasets, but the potential use of LST retrievals from CIMR’s Ka/Ku bands should be the subject of future investigations (e.g. [Holmes et al., 2009, Song et al., 2019, Prigent et al., 2016, Jiménez et al., 2017]). This approach will potentially allow for a reduced dependence on auxiliary data by exploiting CIMR multi-frequency capabilities.

For estimating microwave vegetation indices (MMVI), the CIMR SM product may serve as an initial guess to invert the tau-omega model and estimate VOD and ω from CIMR L-, C-, and X-bands. The retrieval of time-dynamic VOD and ω has been previously demonstrated at L-band [Baur et al., 2021]. These microwave vegetation parameters represent microwave attenuation and scattering properties at different canopy depths, which are deeper at lower frequencies. This allows capturing interactions between various wavelengths and canopy components such as branches, stems, and leaves, relating to significant vegetation characteristics like vegetation water content and above-ground biomass.