The HIRLAM analysis of surface variables and parameterization of surface processes described in the companion paper is applied to a one year long assimilation experiment in order to evaluate the performance of the new package over an extended period covering a complete annual hydrological cycle. The time evolution of the slow evolving variables and the usage of diagnostic tools to assess the performance of the surface analysis are thoroughly discussed in the text. In particular, the usefulness of the screen level variables and soil moisture analyses to diagnose and identify model biases is highlighted. There are also sections devoted to discuss the behaviour of different tiles within a grid square and the convergence of the assimilation algorithm. The need of soil moisture analysis is justified by an additional experiment showing the appearance of long-term drift in near surface parameters when soil moisture analysis is switched off.