A global analysis of variability and trends of soil moisture

As slowly evolving state variables of the Earth surface, both the sea surface temperature (SST) and soil moisture have longer "memory" of past states than does the atmosphere. The variability and trends of these two variables are thus very useful for detecting global climate variability and climate change. However, the variability and change of soil moisture is much less well studied than that of SST, largely due to the paucity of soil moisture observations on a global scale. Thus we must turn to model products to estimates of soil moisture variations. The interannual variability and trend of soil moisture is examined in this paper using five global soil moisture products, three produced by land surface model calculations, and two from coupled land-atmosphere model reanalyses. Several recent studies show that a multi-model analysis of the land surface outperforms individual models in nearly all circumstances. Thus, the characteristics in the multi-model average of soil moisture are evaluated, and are validated by comparing to available in situ measurements. Also, the relationship between the interannual variability of soil moisture and SST is explored.