Inference of the potential predictability of seasonal land-surface climate from AMIP ensemble integrations

• Main Authors: Phillips, T. J., Santer, B. D.
• Other Authors: United States. Department of Energy.
• Format: Article in Journal/Newspaper
• Language: English
• Published: Lawrence Livermore National Laboratory 1995
• Subjects: Forecasting; Seasonal Variations; 54 Environmental Sciences; Data; Climate Models; Climates; Sea ice
• Online Access: https://digital.library.unt.edu/ark:/67531/metadc671892/

A number of recent studies of the potential predictability of seasonal climate have utilized AGCM ensemble integrations--i.e., experiments where the atmospheric model is driven by the same ocean boundary conditions and radiative forcings, but is started from different initial states. However, only a few variables of direct relevance to the climate of the land surface have been examined. In this study, the authors infer the potential predictability of 11 climate variables that are indicative of the energetics, dynamics, and hydrology of the land surface. They used a T42Ll9 ECMWF (cycle 36) AGCM having a land-surface scheme with prognostic temperature and moisture of 2 layers occupying the topmost 0.50 meters of soil, but with monthly climatological values of these fields prescribed below. Six model realizations of decadal climate (for the period 1979--1988) were considered. In each experiment, the SSTs and sea ice extents were those specified for the Atmospheric Model Intercomparison Project (AMIP), and some radiative parameters were prescribed as well. However, the initial conditions of the model atmosphere and land surface were different: the first two simulations were initialized from ECMWF analyses, while the initial states of subsequent realizations were assigned values that were the same as those at the last time step of the preceding integration.