A fundamental issue when evaluating the simulation of precipitation is the difficulty of quantifying specific sources of errors and recognizing compensation of errors. We assess how well a large ensemble of high-resolution simulations represents the precipitation associated with strong cyclones. We propose a framework to breakdown precipitation errors according to different dynamical (vertical velocity) and thermodynamical (vertically integrated water vapor) regimes and the frequency and intensity of precipitation. This approach approximates the error in the total precipitation of each regime as the sum of three terms describing errors in the large-scale environmental conditions, the frequency of precipitation and its intensity. We show that simulations produce precipitation too often, that its intensity is too weak, that errors are larger for weak than for strong dynamical forcing and that biases in the vertically integrated water vapor can be large. Using the error breakdown presented above, we define four new error metrics differing on the degree to which they include the compensation of errors. We show that convection-permitting simulations consistently improve the simulation of precipitation compared to coarser-resolution simulations using parameterized convection, and that these improvements are revealed by our new approach but not by traditional metrics which can be affected by compensating errors. These results suggest that convection-permitting models are more likely to produce better results for the right reasons. We conclude that the novel decomposition and error metrics presented in this study give a useful framework that provides physical insights about the sources of errors and a reliable quantification of errors.
CITATION STYLE
Di Luca, A., Argüeso, D., Sherwood, S., & Evans, J. P. (2021). Evaluating Precipitation Errors Using the Environmentally Conditioned Intensity-Frequency Decomposition Method. Journal of Advances in Modeling Earth Systems, 13(7). https://doi.org/10.1029/2020MS002447
Mendeley helps you to discover research relevant for your work.