The Effect of Vegetation on the Remotely Sensed Soil Thermal Inertia and a Two-Source Normalized Soil Thermal Inertia Model for Vegetated Surfaces

Abstract

The estimation of soil thermal inertia (STI) (P) on the vegetated surface is a challenging task due to the difficulty in acquiring soil temperature under vegetation. In most cases, mixed surface temperature (T) is used to replace soil temperature ( ${\rm T}_{\rm s}$ ) to estimate P. Inevitably, errors are introduced because of the effect of vegetation. In this paper, on the basis of a simplified STI concept and an operational algorithm of surface temperature separation, the differences of STI estimated from ${\rm T}_{\rm s}$ , vegetation temperature ( ${\rm T}_{\rm v}$ ) and T were quantified. When there is large difference between T and ${\rm T}_{\rm s}$ (as much as 10 K), the mean absolute percentage difference (MAPD) between STI estimated from ${\rm T}_{\rm s}$ ( ${\rm STI}_{{\rm S}}$ ) and STI estimated from T ( ${\rm STI}_{{\rm M}}$ ) can reach 60%. A normalized STI ( ${\rm STI}_{{\rm N}}$ ) to account for the vegetated surface was proposed in terms of the linear mixing theory, which can be used to estimate the relative soil water (SW) content. Under the condition that the wilting point of the soil moisture and the saturated soil moisture are known for an area, SW content can then be calculated from ${\rm STI}_{{\rm N}}$ . The comparisons of the relationships between soil moisture from the advanced microwave scanning radiometer-earth observing system and ${\rm STI}_{{\rm S}},{\rm STI}_{{\rm V}},{\rm STI}_{{\rm M}}$ , and soil moisture estimated from ${\rm STI}_{{\rm N}}$ show that ${\rm STI}_{{\rm N}}$ is the best indicator of soil moisture, with the highest correlation coefficient $({\rm R}^{2})$ of 0.64 and 0.75 for the two validation domains.