Results from the Diurnal land/atmosphere Coupling Experiment (DICE)

Many studies of boundary layer parametrizations (such as those under GABLS) treat the surface as a fixed lower boundary condition, either specifying the surface fluxes or near surface temperatures. In a similar way, land surface models are widely developed and tested in isolation, driven by observed near-surface atmospheric variables. Whilst simplifying the system in this way has undoubtedly delivered many improvements to models on both sides, the applicability of such studies to the real world is readily called into question because feedbacks between the atmosphere and land are being suppressed. As an example, results from the GLACE experiment highlighted the areas of the globe that were “hotspots” for land/atmosphere coupling, with stronger coupling strength between soil moisture and precipitation. However, the results from this experiment also showed that there was wide variation in the strength of coupling at the hotspot regions between the participating models. Further analysis following the protocol of the GLACE experiment suggested that interactions between the atmospheric parametrisations were responsible for the land/atmosphere coupling strength in some models, but our understanding of these interactions within models is limited.

DICE is an international experiment designed to identify and understand the interactions and feedbacks between the land and atmospheric boundary layer. The GABLS2 (CASES99) experiment has been re-visited, but with both boundary layer and land surface communities included within the project development and analysis. The result is a multi-stage project with ensembles of simulations that allows the sensitivity characteristics of each component (land and atmosphere) to be assessed and compared with the characteristics of the coupled versus uncoupled simulations. In this presentation we will give details of the experimental protocol, including the justification for each of the stages. Initial results from the experiment will be presented along with conclusions about characteristics of the integrated system responsible for the coupling strength in the models.