Subseasonal Precipitation Prediction for Africa: Forecast Evaluation and Sources of Predictability

Abstract

This paper evaluates sub-seasonal precipitation forecasts for Africa using hindcasts from three models (ECMWF, UKMO, and NCEP) participating in the Subseasonal to Seasonal (S2S) prediction project. A variety of verification metrics are employed to assess weekly precipitation forecast quality at lead times of one to four weeks ahead (Weeks 1-4) during different seasons. Overall, forecast evaluation indicates more skilful predictions for ECMWF over other models and for East Africa over other regions. Deterministic forecasts show substantial skill reduction in Weeks 3-4 linked to lower association and larger underestimation of predicted variance compared to Weeks 1-2. Tercile-based probabilistic forecasts reveal similar characteristics for extreme categories and low quality in the near-normal category. Although discrimination is low in Weeks 3-4, probabilistic forecasts still have reasonable skill, especially in wet regions during particular rainy seasons. Forecasts are found to be over-confident for all weeks, indicating the need to apply calibration for more reliable predictions. Forecast quality within the ECMWF model is also linked to the strength of climate drivers’ teleconnections, namely El Niño-Southern Oscillation, Indian Ocean Dipole, and the Madden-Julian Oscillation. The impact of removing all driver-related precipitation regression patterns from observations and hindcasts shows reduction of forecast quality compared to including all drivers’ signals, with more robust effects in regions where the driver strongly relates to precipitation variability. Calibrating forecasts by adding observed regression patterns to hindcasts provides improved forecast associations particularly linked to the Madden-Julian Oscillation. Results from this study can be used to guide decision-makers and forecasters in disseminating valuable forecasting information for different societal activities in Africa.

Progress and Challenges of Demand-Led Co-Produced Sub-Seasonal-to-Seasonal (S2S) Climate Forecasts in Nigeria

“…In addition, Vellinga et al (2013) and Batté et al (2018) have also shown that the ECMWF-S2S climate forecast has skill in representing weather and sub-seasonal climate forecasts over Nigeria's neighboring countries. Furthermore, de Andrade et al (2021) who utilized several verification metrics to assess the ECMWF-S2S' weekly precipitation forecast quality at lead times of one to four weeks ahead (weeks 1-4) during different seasons over Africa, and in comparison to other models, found that predictions from the ECMWF-S2S model are more skillful than those from other models, especially in the first two weeks. de Andrade et al (2021) also found that the ECMWF-S2S model's forecast quality is linked to the strength of climate drivers, vis-à-vis teleconnections such as Indian Ocean dipole, El Niño-Southern Oscillation, and the Madden-Julian oscillation.…”

“…Furthermore, de Andrade et al (2021) who utilized several verification metrics to assess the ECMWF-S2S' weekly precipitation forecast quality at lead times of one to four weeks ahead (weeks 1-4) during different seasons over Africa, and in comparison to other models, found that predictions from the ECMWF-S2S model are more skillful than those from other models, especially in the first two weeks. de Andrade et al (2021) also found that the ECMWF-S2S model's forecast quality is linked to the strength of climate drivers, vis-à-vis teleconnections such as Indian Ocean dipole, El Niño-Southern Oscillation, and the Madden-Julian oscillation. ECMWF-S2S dataset has 51 ensemble members thereby allowing analyses of the range of possible climate events, uncertainties in climate events and probabilities of occurrence of climate events across all meteorological variables.…”

“…Furthermore, some works have documented the skill and reliability of the ECMWF-S2S forecasts, e.g., Tompkins and Feudale (2010), Vellinga et al (2013), Batté et al (2018), Olaniyan et al (2018), de Andrade et al (2021, etc. They found that ECMWF-S2S climate forecast data has skill in representing weather and sub-seasonal climate simulations and forecasts over Nigeria, and where applicable over West Africa.…”

Flow dependence of wintertime subseasonal prediction skill over Europe

“…In this study, we follow a frequently used convention in the S2S community to define weekly lead times (e.g. Vitart, 2004;Specq et al, 2020;de Andrade et al, 2021) : week 1 goes from day 5 to day 11, week 2 from day 12 to 18, week 3 from day 19 to 25 and week 4 from day 26 to 32.…”

Drivers of Subseasonal Forecast Errors of the East African Short Rains

“…The lack of subseasonal forecast provision is unfortunate, as compared to most of the globe East Africa is a hotspot of forecast skill (de Andrade et al., 2020; Vigaud et al., 2019). The strong predictability arises in large part from the Madden‐Julian Oscillation (MJO), which shows a strong teleconnection with regional rainfall: models with skillful MJO predictions also capture the teleconnection to rainfall over East Africa (MacLeod et al., 2021).…”