Assessment of inter-model variability and biases of the global water cycle in CMIP3 climate models

Abstract: Observed changes such as increasing global temperatures and the intensification of the global water cycle in the 20th century are also robust results of coupled general circulation models. In spite of this success model-to-model variability and biases that are small in first order climate responses however, have implications for climate predictability especially when multi-model means are used. We show that most climate simulations of 20th and 21st century A2 scenario performed with IPCC-AR4 models have deficiencies in simulating the global atmospheric moisture balance. Large biases of only a few models affect the multi-model mean global moisture budget and an imbalanced flux of -0.14 Sv exists whereas the multi-model median imbalance is only -0.02 Sv. For most models, the detected imbalances furthermore change over time. As a consequence, in 13 of the 18 IPCC-AR4 models examined, global annual mean precipitation exceeds global evaporation, indicating that there should be a "leaking" of moisture from the atmosphere whereas for the remaining 5 models a "flooding" is implied. The source / sink of moisture can also be regarded as atmospheric latent heating / cooling and hence as positive / negative perturbations of the atmospheric energy budget or non-radiative forcings in the range of -1 to +6 W/m2 (median is +0.1 W/m2). The global atmospheric moisture transport from oceans to land areas, which impacts the terrestrial water cycle also increases by about 5% with a model-to-model range from 1 to 13% in the 21st century. The variability is partially due to the different implementations of aerosol forcings in the models. The assessment of the pole wards shifts of dry zones in climate simulations show that the multi-model means of the two subsets of leaking and flooding models produce spatial shifts similar in size as expected from 21st century global warming simulations.