This paper and its companion report on the development of a sequential assimilation technique based upon optimum interpolation in order to initialize soil moisture in atmospheric models. A previous study by Mahfouf has demonstrated that it is possible to estimate soil moisture from the evolution of atmospheric temperature and relative humidity near the surface. The main purpose of this paper is to examine more precisely the dependence of atmospheric low-level parameters upon soil moisture and how this dependence is affected by various factors (soil characteristics, vegetation type, low-level wind). The sensitivity of atmospheric parameters to soil moisture is expressed as the statistical quantities of the optimal interpolation. The importance of observation errors, which define the relevance of the atmospheric parameters for the assimilation procedure, is also investigated. An analytical formulation of the optimal interpolation coefficients is proposed. Finally, the usefulness and limitations of ... Abstract This paper and its companion report on the development of a sequential assimilation technique based upon optimum interpolation in order to initialize soil moisture in atmospheric models. A previous study by Mahfouf has demonstrated that it is possible to estimate soil moisture from the evolution of atmospheric temperature and relative humidity near the surface. The main purpose of this paper is to examine more precisely the dependence of atmospheric low-level parameters upon soil moisture and how this dependence is affected by various factors (soil characteristics, vegetation type, low-level wind). The sensitivity of atmospheric parameters to soil moisture is expressed as the statistical quantities of the optimal interpolation. The importance of observation errors, which define the relevance of the atmospheric parameters for the assimilation procedure, is also investigated. An analytical formulation of the optimal interpolation coefficients is proposed. Finally, the usefulness and limitations of ...