A Proactive Quality Control (PQC) scheme based on Ensemble Forecast Sensitivity to Observations (EFSO)

Tse-Chun Chen* and Daisuke Hotta and Eugenia Kalnay
University of Maryland

Operational numerical weather prediction (NWP) systems occasionally exhibit “forecast skill dropouts” in which the forecast skill drops to an abnormally low level, due in part to the assimilation of detrimental observational data. Recent studies have shown that a diagnostic technique called Ensemble Forecast Sensitivity to Observations (EFSO) can detect such observations (Kalnay et.al 2012; Ota et al. 2013, Tellus A). Based on this technique, a new Quality Control (QC) scheme called Proactive QC (PQC) has been proposed which detects detrimental observations using EFSO after 6 hours from the analysis when the analysis at the next cycle becomes available for verification and then repeats the analysis and forecast without using the detected observations (Hotta 2014).

In Hotta (2014), PQC has been shown, with NCEP’s quasi-operational system, that it can reduce the 24-hour forecast error from the detected skill dropout events. Showing such great potential, operational implementation is highly desired. We will show that PQC correction can significantly reduce forecast errors up to 5 days, and the reduction magnitude and the benefitted areal coverage can grow with synoptic weather disturbances through out 3- to 4-day forecasts. In addition, the operational center imposes very tight schedule in order to deliver the products on time, thus reduction of computational cost is necessary before PQC can be implemented. To avoid performing the analysis twist, the most expensive part of PQC, we tested the accuracy of constant-K approximation proposed by Hotta (2014), which assumes the Kalman gain K doesn’t change much given the fact that only a small subset of observation is rejected. We also presented another option for implementation that repeating analysis only within the range of impact of the detrimental observations, which is only a small part of the globe. In this presentation, we will demonstrate the performance and feasibility of PQC implementation.



*email: tcchen@umd.edu
*Preference: Poster