Many components of the terrestrial energy and water budgets are not well observed at global scales yet are critical to understanding climate, climate change and the coupled Earth system. Operational analyses of global observations constrain models with the available observations; however, land fluxes (which all components are available) are generally produced from a model. The Coordinated Enhanced Observing Period (CEOP) has strived to compile many observation components and model data over a given period (2003-2004). Station data from many projects around the globe provide many flux components, but at point spatial scales. Satellite data provide observations for some fluxes components or closely related state variables. Additionally, seven operational analyses have been collected for CEOP from international forecast offices. While many assimilate the same observations the modeled fluxes can vary widely. Here, we will evaluate a new reanalysis system's land surface energy and water cycles compared to a multi-model analysis (including the variation among contributing models) and other available observations. We aim to characterize the uncertainty of the analyses, ultimately providing direction to future refinement of the models. Validation experiments for the GEOS5 data assimilation system being used for the Modern Era Retrospective-analysis for Research and Applications (MERRA) will be evaluated. The long term objectives of the work will also contribute to the development of observational based data sets of the land surface fluxes (e.g. GEWEX LandFlux project).