Neighborhood Ensemble Copula Coupling: Smoother and sharper calibrated ensembles

Abstract

Abstract Ensemble Copula Coupling (Schefzik et al. 2013) is a widely used method to produce a calibrated ensemble from a calibrated probabilistic forecast. This process improves the statistical accuracy of the ensemble; in other words, the distribution of the calibrated ensemble members at each grid point more closely approximates the true expected distribution. However, the trade-off is that the individual members are often less physically realistic than the original ensemble: there is noisy variation among neighboring grid points, and, depending on the calibration method, extremes in the original ensemble are sometimes muted. We introduce Neighborhood Ensemble Copula Coupling (N-ECC), a simple modification of ECC designed to mitigate these problems. We show that, when used with the calibrated forecasts produced by Flowerdew’s (Flowerdew 2014) reliability calibration, N-ECC improves both the visual plausibility and the statistical properties of the forecast.