Assessment of the relationship between thermal-infrared-based temperature-vegetation dryness index and microwave satellite-derived soil moisture

Soil moisture (SM) is an important parameter in terrestrial ecological and hydrological processes, particularly in arid and semi-arid regions. However, a highly accurate SM grid data set, which is used as a reference for the data quality, is not really suitable for the validation of other SM products. Thus, a more effective method may be necessary for evaluation of SM grid data. The temperature-vegetation dryness index (TVDI), which is estimated by the relationship between land surface temperature and vegetation index data, has been developed to assess regional water stress. Based on previous studies, we assumed a negative linear relationship between SM and the TVDI to establish the evaluation method of SM grid data. Although a highly accurate measure of SM obtained by use of microwave sensors may not always exhibit a negative linear correlation with the TVDI, the pixels of strong negative linear correlation between them signifies at least a higher accuracy of the two data at that position. The negative relationships between microwave satellite sensor-derived SM and the TVDI were tested by application of 16-day scaled satellite data in the Sahel. We determined that the correlations differ spatially according to vegetation distribution. That is, when compared with a lower correlation in the arid Sahara to the north, a higher correlation (-0.9 < r < -0.7) was observed in the savannas, shrublands, and grasslands to the south. Our comparison results will be useful in developing validation methodologies for SM grid data in an alternative way under conditions of insufficient in situ measurements.