Including observation error correlation for ensemble radar radial wind assimilation and its impact on heavy rainfall prediction

Abstract

An assumption of uncorrelated observation errors is commonly adopted in conventional data assimilation. For this reason, high‐resolution data are resampled with strategies such as superobbing or data thinning. These strategies diminish the advantages of high temporal and spatial resolutions that can provide essential details in convection development. However, assimilating high‐resolution data, such as radar radial wind, without considering observation error correlations can lead to overfitting and thus degrade the performance of data assimilation and forecasts. This study uses a radar ensemble data assimilation system that combines the Weather Research and Forecasting model and Local Ensemble Transform Kalman Filter (WRF‐LETKF) to assimilate radar radial wind and reflectivity data. We present a strategy to include the error correlation of the Doppler radar radial wind in the WRF‐LETKF radar assimilation system and examine its impact on the accuracy of short‐term precipitation predictions based on a heavy rainfall event on June 2, 2017 in Taiwan. For radial wind, the horizontal error correlation scale is approximately 25 km according to the innovation statistics. The introduction of observation error correlation for radar radial wind assimilation produces more small‐scale features in wind analysis corrections compared to the experiment using an independent observation assumption. Consequently, the modification of wind corrections leads to stronger convergence accompanied by higher water vapor content, which enhances local convections. This results in more accurate simulations of reflectivity and short‐term precipitation. In particular, this advantage is identified for extreme heavy rainfall thresholds at small scales according to probability quantitative precipitation forecasts and fractions skill score.

A Novel Localized Fast Multipole Method for Computations With Spatially Correlated Observation Error Statistics in Data Assimilation

“…By directly including correlated error statistics in the assimilation scheme, we are able to use a shorter observation‐thinning distance or make full use of the information from high‐resolution observations (Bell et al., 2020; Fowler et al., 2018; Rainwater et al., 2015; Waller, Simonin, et al., 2016). This has been shown to improve both analysis accuracy and forecast skill in idealized data assimilation systems (e.g., Healy & White, 2005; Stewart et al., 2008, 2013) as well as operational systems (e.g., Fujita et al., 2020; Simonin et al., 2019; Yeh et al., 2022). In addition, assimilation of dense observations (e.g., Doppler radar observations) without thinning is important for convection‐permitting NWPs as they require information on small scales (Waller et al., 2021).…”

Sampling and misspecification errors in the estimation of observation‐error covariance matrices using observation‐minus‐background and observation‐minus‐analysis statistics

“…We refer to this approach as the Desroziers et al method, which is essentially a sampling approach that uses finite samples of the O$$$ - $$$B and O$$$ - $$$A differences to estimate observation‐error covariance matrices. The Desroziers et al method has been successfully used in operational data assimilation systems to estimate spatial error correlations for observations such as geostationary satellite observations (e.g., Michel, 2018; Waller et al ., 2016a) and Doppler radial winds (e.g., Waller et al ., 2016c, 2019; Yeh et al ., 2022), and to estimate satellite interchannel error covariances (e.g., Bormann et al ., 2016; Campbell et al ., 2017; Stewart et al ., 2014; Weston et al ., 2014).…”