Evaluation of brightness temperatures observations from SMOS microwave satellite in comparison with L-MEB model simulations data

The European Space Agency (ESA’s) Soil Moisture and Ocean Salinity (SMOS) satellite mission was launched in November 2009. SMOS carries the first L-band (1.4 GHz) 2-D synthetic aperture microwave radiometer that produces multi-angular dual polarized (or fully polarized) brightness temperature. The objective of SMOS mission is to provide global surface soil moisture maps over the land surfaces with an accuracy of 0.04 m3m−3. The SMOS soil moisture retrieval algorithm was developed, which processes Level 1C products (multi-angular brightness temperatures) to Level 2 SM products (soil moisture maps). This algorithm is based on the comparison between the brightness temperatures from SMOS and the simulated brightness temperatures data (simulated TB) using L-MEB model. Thus, the evaluation of SMOS brightness temperatures is a necessary step before using of Level 2 Soil Moisture products. Therefore, the objective of this research is to evaluate the horizontal and vertical full polarized brightness temperatures data (TBh, TBv) from the SMOS MIR_SCLF1C products at the five meteorological stations in the west and southwest of Iran. Evaluation of SMOS brightness temperature data (SMOS TB) was done through a comparison between the SMOS TB and simulated TB from the L-MEB model. The SMOS MIR_SCLF1C (Level 1C Full Polarization Land Science measurements) products, which were provided through the ESA, contains the multi-angular brightness temperatures at the top of the atmosphere in the antenna polarization reference frame. In this study, the MIR_SCLF1C products, version 505 for the period 2012-2013 were evaluated. The ESA’s SMOS Matlab codes on Linux was used to reading and deriving TB, Incidence angles, Geometric and Faraday rotations and other required data from MIR_SCLF1C products.
The L-MEB (L-band Microwave Emission of the Biosphere) model is the radiative transfer model, which has been specifically developed to simulate the L-band microwave emission (brightness temperature) over land surfaces. In this research, the simulation of TB (TBh, TBv) at the five meteorological stations was carried out using L-MEB model (MATLAB function) and ground-based measurements. The model was simulated TB at the Earth’s surface reference. Therefore, SMOS TB data was projected from the antenna reference frame to the Earth’s surface reference frame using an algorithm provided by the CESBIO (Centre d’Etudes Spatiales de la BIOsphére) team. Four statistical metrics and Taylor diagram were used for the evaluation of results; the Root Mean Squared Difference (RMSD), the centered Root Mean Square Difference (cRMSD), the Mean Bias Error or bias and the correlation coefficient (R).The Taylor diagrams are used to represent three statistical metrics (R, cRMSD and standard deviation) on two dimensional plots to graphically describing how closely SMOS TB matches simulated TB.
Based on the research algorithm, the Evaluation model for the SMOS brightness temperatures data (TBh, TBv) was obtained. The Evaluation model was run for five metrological stations and simulated TB data from L-MEB model and SMOS BT from the MIR_SCLF1C product was saved as the output of the model to evaluation.
The results of the comparison between the SMOS TBh, TBv data and simulated TBh, TBv show that SMOS TB have an underestimation at Ahvaz, Sararod, Sarableh stations, whereas an overestimation of the SMOS BT was detected at Darab, Ekbatan stations. According to RMSD results, the SMOS TBh data at Ahvaz, Sarableh stations and the SMOS TBv data at Ahvaz, Darab, Sarableh stations have the highest accuracy.
The Taylor diagrams shows the strong correlation (RTBh = 0.8-0.9 and RTBv =0.81-0.93) between the SMOS TB and simulated TB data at all stations. Besides, the lowest value of the cRMSD of the SMOS TB data was obtained at Ahvaz (TBh =5.34, TBv = 5.67 K) and Darab stations (TBh =8.54, TBv = 5.9 K). In addition, these diagrams indicate that the standard deviation of SMOS TBh data at Sarableh, Ahvaz, Sararod stations and SMOS TBv data at Sarableh, Darab, Ahvaz stations are closer to the simulated TB data than other stations. Overall, the findings of this paper give valuable information about the uncertainties and errors of SMOS brightness temperatures data (MIR_SCLF1C) in the study area. Therefore, this research could be as a reference for using the SMOS soil moisture products (Level 2 Soil Moisture) in hydrology and meteorology studies in Iran.