Toward a new cloud analysis and prediction system

The International Cloud analysis Workshop, held on September 1-3, 2009, at Boulder, Colorado, reviewed current and recent cloud analysis efforts and evaluating the state of the science, synergies, and collaborations in modeling clouds. Various experts in cloud modeling, observations, and data assimilation met to move decisively toward a realization of cloud analysis systems for operational use. An outcome of discussions was a recommendation to develop specialized one-dimensional variational data assimilation (1D-Var) cloud retrievals that match the data assimilation system to ensure consistency. It was found that the ensemble Kalman filter (EnKF) allows direct use of nondifferentiable observation operators, such as those simulating abrupt cloud geometrical properties, for example, cloud-top and cloud-base vertical levels. It was also mentioned that background error modeling could be improved through the use of better balance relationships, either by using statistical regressions or by using more sophisticated nonlinear balance relationships.