A parameterization of SEVIRI and MODIS daily surface albedo with soil moisture: Calibration and validation over southwestern France

- We estimate daily MODIS soil albedo.
- We calibrate MODIS soil albedo with soil moisture measured in situ.
- We built a prognostic albedo from soil moisture estimates.

In climate models, it is important to simulate the partitioning of the surface albedo into soil and vegetation albedo components because these latter address different processes that are time-scale dependent. Vegetation albedo primarily varies along with the growing season while soil albedo shows day-to-day variations caused by rainfall events. In this study, the objective is to disentangle soil albedo from surface albedo within the visible (VIS) and near infrared (NIR) spectral bands of MODIS and SEVIRI sensors in order to yield a calibration with surface soil moisture (SSM). In a first step, we derive global static maps of soil and vegetation albedos from MODIS products at the resolution of 0.05° over a 4-year period. In a second step, we estimate a daily MODIS white-sky albedo (WSA) by combining 8-day BRDF and TERRA/AQUA reflectance values for each orbit pass. These MODIS products are then projected on the SEVIRI grid of 4 km for further comparison. Then, a physical method is presented to get a daily soil albedo from both MODIS and SEVIRI data. Good correlations are obtained between satellite time series for 2007 and 2008 of retrieved soil albedo and in situ SSM measurements at the 12 SMOSMANIA stations located in southwestern France. A function of an exponential type between albedo and SSM, as supported by the theory, could be verified. As in the VIS domain, the goodness of fit r2 is about 0.42 and 0.54 for Lahas and Condom, respectively. The Chi-Square test indicates that the relationship is significant with p-value < 0.01. Finally, it is shown that most anomalies of soil albedo correlate with anomalies of soil wetness, which yields crucial importance for studying the energy budget.