Improving the precipitation forecast over the Eastern Mediterranean using a smoothed time‐lagged ensemble

Abstract

In the absence of convection permitting numerical weather prediction (NWP) ensembles, the most recent deterministic NWP precipitation forecast is usually addressed. However, the exact intensity, location and timing of a deterministic precipitation forecast is not always reliable because of the chaotic nature and complexity of precipitation formation mechanisms. This study examines a way to optimize the use of precipitation forecasts for deterministic NWP models. More specifically, it suggests using a spatially smoothed time‐lagged ensemble (TLE) to obtain more reliable precipitation forecasts. A global NWP model (integrated forecast system—IFS) and a regional convection permitting NWP model (COSMO) over the Eastern Mediterranean during the period 2016–2018 were used for the analysis. First, the paper defines light, light–moderate and moderate intensities for 6 hr accumulated precipitation (6hAP) and investigates the corresponding definitions for 1 hr accumulated precipitation (1hAP). Next, fractional skill score (FSS) are used to estimate the optimal spatial smoothing scale of a deterministic precipitation forecast for the three intensity categories for the 6hAP and 1hAP. The FSS is also used to compare COSMO and IFS deterministic precipitation forecasts, and to analyse the skill degradation with the forecast range. It is quantitatively shown that the useful scale of precipitation forecasts is smaller for larger accumulation time intervals. Finally, precipitation forecasts for TLE are formed from successive smoothed deterministic forecasts and compared with the most recent deterministic forecast. It is found that, on average, TLEs have better skills for both 6hAP and 1hAP. The reason for this improvement in skill is illustrated using a case study as an example.

The set‐up and evaluation of fine‐scale data assimilation for the urban climate of Amsterdam

“…To verify the spatial representation of precipitation, the fractional skill score (FSS) was calculated for daily accumulated precipitation of the consecutive data assimilation runs for domain 2 that covers The Netherlands. The FSS is a neighbourhood‐based verification method for evaluating the spatial accuracy of precipitation (Roberts and Lean, 2008) and has been applied in real case verification studies (Ebert, 2009; Mittermaier et al ., 2013; Skok and Roberts, 2016; Imhoff et al ., 2020; Khain et al ., 2020). In order to use it, the gauge‐adjusted radar data (Overeem et al ., 2009a) were transformed with an average weighted interpolation scheme (Met Office, 2018) to match the model resolution.…”

Hydrometeorological analysis and forecasting of a 3-day flash-flood-triggering desert rainstorm

“…However, as already suggested, the forecasted rainfall might be erroneously placed in space and time (Ben Bouallègue and Theis, 2014;Collier, 2007). To account for this source of uncertainty, a simple and costeffective forecast-shifting approach was applied: shifting the last two available COSMO runs closest to the April 26 th flashflood occurrence (i.e., Apr 25, 21:30, and Apr 26, 09:30) within a reasonable spatial error range of 20 km (Armon et al, 2020;Khain et al, 2019;Fig. 9) in 1-km intervals.…”

“…Global weatherprediction models are routinely used by meteorological agencies worldwide, but their spatiotemporal scales are too coarse for flash-flood applications (Sene, 2013). In recent years, convection-permitting models with spatial resolution of ≤3 km have enabled explicit representation of the convective process, providing better representation of rainfall and better forecast skills on the flash-flood scale (Armon et al, 2020;Clark et al, 2016;Khain et al, 2019;Prein et al, 2015). However, the finer scale increases the sensitivity of these models to initial conditions, leading to spatial uncertainties in their output (Bartsotas et al, 2016;Ben Bouallègue and Theis, 2014;Collier, 2007;Sivakumar, 2017).…”

Precipitation forecast in China based on reservoir computing

“…On the basis of the models, Clark et al used Convection-Permitting numerical weather prediction models to forecast the precipitation over the UK and the Alpine region, respectively, in 2016 [ 4 ]. In the absence of Convection-Permitting numerical weather prediction ensembles, Khain et al used a smoothed time-lagged ensemble method to forecast the precipitation in 2019 [ 5 ]. In addition, the mechanism of precipitation and the circulation of water resources are also important.…”