Forecast Skill of Minimum and Maximum Temperatures on Subseasonal‐to‐Seasonal Timescales Over South Africa

Abstract

Forecast skill of three subseasonal‐to‐seasonal models and their ensemble mean outputs are evaluated in predicting the surface minimum and maximum temperatures at subseasonal timescales over South Africa. Three skill scores (correlation of anomaly, root‐mean‐square error, and Taylor diagrams) are used to evaluate the models. It is established that the subseasonal‐to‐seasonal models considered here have skill in predicting both minimum and maximum temperatures at subseasonal timescales. The correlation of anomaly indicates that the multimodel ensemble outperforms the individual models in predicting both minimum and maximum temperatures for the day 1–14, day 11–30, and full calendar month timescales during December months. The Taylor diagrams suggest that the European Centre for Medium‐Range Weather Forecasts model and MM performs better for the day 11–30 timescale for both minimum and maximum temperatures. In general, the models perform better for minimum than maximum temperatures in terms of root‐mean‐square error. In fact, the skill difference in terms of correlation of anomalies (CORA) is small.

Probabilistic skill of statistically downscaled ECMWF S2S forecasts of maximum and minimum temperatures for weeks 1–4 over South Africa

“…The week 1, week 2, week 3 and week 4 averages for December, January and February are computed and averaged to form weeks 1-4 DJF seasons. The ECMWF model is used because in our previous study, it performed better than the other S2S models in terms of skill over southern Africa (Engelbrecht et al, 2021;Phakula et al, 2020).…”

“…This study seeks to determine the probabilistic forecast skill level of statistically downscaled European Centre for Medium-Range Weather Forecasts (ECMWF) S2S forecasts in predicting maximum and minimum temperatures for weeks 1-4 lead times during 20-year December-January-February (DJF) seasons over South Africa. In our previous work, we tested the deterministic skill of S2S forecasts in predicting surface temperature over South Africa, without any downscaling (Phakula et al, 2020). Therefore, in this study, the focus is to determine if statistical downscaling of model forecasts improves skill or not.…”

“…The South African Weather Service (SAWS) has the mandate to provide weather and climate information to the public of South Africa. Hence, SAWS issues temperature and rainfall weather forecasts on a daily basis and seasonal climate predictions once a month (e.g., Phakula et al, 2018; Phakula et al, 2020). There is a ‘gap’ between weather forecasting and seasonal climate predictions, referred to as subseasonal‐to‐seasonal (S2S), a timescale from 2 weeks up to 2 months (de Andrade et al, 2021; Klingaman et al, 2021; Mariotti et al, 2018; Moron & Robertson, 2020; Mundhenk et al, 2017; Wang & Robertson, 2018; White et al, 2017).…”

Literature survey of subseasonal‐to‐seasonal predictions in the southern hemisphere

“…The skill of the ECMWF, UKMO, and Centre National de Recherches Meteorologiques (CNRM; Voldoire et al, 2013) S2S project models and their MME was investigated in predicting minimum and maximum temperatures for days 1–14 (week 1 + 2), days 11–30, and days 1–30 (full month calendar) over South Africa (Phakula et al, 2020). Higher skill was found for week 1 + 2 in predicting both minimum and maximum temperatures, with the MME outperforming the individual models.…”

Using the Residual Network Module to Correct the Sub-Seasonal High Temperature Forecast

“…As early as 2010, the World Weather Research Programme (WWRP) and the World Climate Research Programme (WCRP) identified the need for in-depth research on the sub-seasonal to seasonal (S2S) scale to improve the prediction of extreme weather events, such as droughts, floods, and cyclones (Brunet et al, 2010). As the research continues, researchers have repeatedly argued the value of sub-seasonal forecasting studies as a key component of completing seamless forecasts (Pegion et al, 2019;Xiang et al, 2019;Phakula et al, 2020;Vijverberg et al, 2020;Manrique-Sun e´n et al, 2020;Merryfield et al, 2020). However, this timescale is hard to forecast due to the fact that the lead time is sufficiently long that the memory of the atmospheric initial conditions is lost and it is too short a time range for the variability of the ocean to have a strong influence on the atmosphere.…”