Toward improvement of cloud representation in the Eurasian continental interior in CPS

Abstract

 In February 2022, the Japan Meteorological Agency started operating its third-generation seasonal ensemble prediction system (CPS3; Coupled Prediction System version 3; Hirahara et al. 2023). CPS3 is based on an atmosphere/ocean/land/sea-ice-coupled model; the atmosphere model has a resolution of TL319 (approx. 55 km) L100 (up to 0.01 hPa) and the ocean model has 60 layers with a global resolution of 0.25° x 0.25°. Along with the widely improved performance of each model, prediction skills of CPS3 are significantly better than those of the previous system. There is, however, still room for improvement, including a positive IOD-like SST bias in the Indian Ocean, a positive surface air temperature bias over the interior of Eurasia, a negative cloud cover bias in subtropical ocean stratocumulus regions, and an arctic radiation flux bias. Cumulus and cloud parameterization errors are some of the primary causes of these biases. Focusing on the Eurasian continental interior, predicted cloud cover is fewer than observed amount throughout the year, and diurnal and seasonal variations are insufficiently represented. To represent cloud condensation, CPS3 has a cloud parametrization using a diagnostic method based on Smith (1990), where spatial fluctuations in cloud water variables that cannot be resolved on a grid scale are represented by assuming a specific Probability Density Function (PDF). In CPS3, a symmetric top-hat PDF (JMA 2022) is used. Although detrain from cumulus is included in cloud water content, fluctuations in cloud water variables due to cumulus are not fully considered, which is one of the possible causes of the negative bias in cloud cover. To address this issue, representation of asymmetries of the fluctuation in cloud water variables due to cumulus in the PDF can be adapted. At this meeting, we will report on the above topics and the progress of development for the next CPS.