Answers are grounded in this paper and cite their sources.
This work aims at improving the accuracy of ensemble precipitation forecasts for large events (return period of 20 years or larger for 24-h precipitation) generated from the Global Ensemble Forecast System (GEFS) ensemble mean forecast in support of the Hydrologic Ensemble Forecast Service of the U.S. National Weather Service (NWS). The proposed prototype application, the conditional bias-penalized regression (CBPR)-aided Meteorological Ensemble Forecast Processor (MEFP), is comparatively evaluated with the MEFP currently used operationally at the River Forecast Centers (RFCs). For evaluation, hindcasting experiments were carried out for dependent and independent validation of mean areal precipitation and streamflow hindcasts at 82 and 48 locations in 13 RFCs, respectively. It is shown that, for 24-h precipitation, the CBPR-aided MEFP improves forecast skill up to 10 days of lead time over the current MEFP. The average margin of improvement is larger for shorter lead times and exceeds 20% for days 1 and 2. For mean daily streamflow, the CBPR-aided MEFP improves skill up to 2 weeks of lead time. The average margin of improvement is over 10% for days 1–7 and exceeds 15% for days 3–4. That the improvement is larger for heavier precipitation is expected from the type I versus type II error trade-off employed in CBPR. This work empirically validates that the above performance characteristic holds in the CBPR-aided MEFP ensemble forecasts as well.
No takes yet. Share an insight, caveat, or question.
Kim et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: