Calibrated ensemble forecasts of the height of new snow using quantile regression forests and ensemble model output statistics

Abstract

Abstract. Height of new snow (HN) forecasts help to prevent critical failures of infrastructures in mountain areas, e.g. transport networks and ski resorts. The French national meteorological service, Météo-France, operates a probabilistic forecasting system based on ensemble meteorological forecasts and a detailed snowpack model to provide ensembles of HN forecasts. These forecasts are, however, biased and underdispersed. As for many weather variables, post-processing methods can be used to alleviate these drawbacks and obtain meaningful 1 to 4 d HN forecasts. In this paper, we compare the skill of two post-processing methods. The first approach is an ensemble model output statistics (EMOS) method, which can be described as a nonhomogeneous regression with a censored shifted Gamma distribution. The second approach is based on quantile regression forests, using different meteorological and snow predictors. Both approaches are evaluated using a 22 year reforecast. Thanks to a larger number of predictors, the quantile regression forest is shown to be a powerful alternative to EMOS for the post-processing of HN ensemble forecasts. The gain of performance is large in all situations but is particularly marked when raw forecasts completely miss the snow event. This type of situation happens when the rain–snow transition elevation is overestimated by the raw forecasts (rain instead of snow in the raw forecasts) or when there is no precipitation in the forecast. In that case, quantile regression forests improve the predictions using the other weather predictors (wind, temperature, and specific humidity).

An evaluation of variable selection methods using Southern Africa solar irradiation data

“…This excellent superiority can be attributed to the ability of QRF to learn any pattern of any response for any provided data (Dega et al, 2023). SI data is non-Gaussian and this study shows that the algorithm can be applied to both deterministic correction and probabilistic calibration of a skewed distribution of meteorological features, as claimed by Evin et al (2021). Apart from exhibiting strengths of both RFs and QR modelling, the hybrid algorithm demonstrated that it handled well the weaknesses in both QR and RF modelling separately.…”

Performance of precipitation phase partitioning methods and their impact on snowpack evolution in a humid continental climate

“…However, that study did not consider the uncertainties associated with phase partitioning. Evin et al (2021) highlighted the poor performance in snowpack accumulation simulations due to errors in phase partitioning. In our study, when focusing on all the accumulation events and the warm events, the PPM type greatly impacted SWE accumulation.…”

QRF4P‐NRT: Probabilistic Post‐Processing of Near‐Real‐Time Satellite Precipitation Estimates Using Quantile Regression Forests

“…Because although static experiments use fixed climatology as predictors, the samples are relatively evenly distributed (i.e., homogeneous). However, the homogeneity of Analog experiments was degrading heavily (Evin et al., 2021). Dynamic experiments exhibited the best performance and played an unbeatable role, which confirms the importance of near‐real‐time predictors.…”

“…Results indicated that the QRF model performs better than EMOS and can bring additional value to the human forecaster. Evin et al (2021) proposed using the QRF model to calibrate ensemble forecasts of the height of new snow, which also indicated that QRF could be applied to the correction of skewed distribution for variable similar to precipitation. In addition to the post-processing of MWPs, QRF has also been successfully applied to hydrological ensemble post-processing (Tyralis & Papacharalampous, 2021;Tyralis et al, 2019) and soil uncertainty mapping (Kasraei et al, 2021;Vaysse & Lagacherie, 2017).…”