Scholarlink

<p>In West Africa (WA), socioeconomic sectors such as agriculture, water management, electricity production, disaster risk management, and healthcare are heavily dependent on rainfall. It is therefore imperative to have reliable seasonal precipitation forecasts available sufficiently in advance to facilitate effective planning and informed decision-making across climate-sensitive sectors. The performance of 14 seasonal precipitation forecasts from the North American Multi-Model Ensemble (NMME) and European Centre for Medium-Range Weather Forecasts fifth-generation seasonal forecasting system (ECMWF-SEAS5) is evaluated for the June–September (JJAS) season. The Global Precipitation Climatology Centre (GPCC) and the African Rainfall Climatology Version 2 (ARC2) datasets are used as observational references. We began by verifying the ability of the forecasts to reproduce the climatology, and then assessed the performance of each model and the multi-model ensemble in predicting rainfall in WA at timescales ranging from 0 to 5&#xa0;months lead time. The NMME and SEAS5 models successfully capture the average JJAS seasonal precipitation, estimated at around 11&#xa0;mm/day in central and southeastern part of WA. Forecast skill is highest at shorter lead times but declines rapidly with increasing lead time, as reflected by the decreasing Heidke Skill Score values. Most models detect normal seasons with over 55% probability of Detection, but struggle to identify seasons above and below normal (probability of Detection &lt; 42%). The performance of the multi-model ensemble is not systematically superior to that of individual models, suggesting the need for more advanced weighting approaches. According to our results, the NMME and SEAS5 models represent a valuable tool during the first three lead times (i.e., leads 0–2) in WA, allowing for the anticipation of key seasonal features before the onset of the JJAS season, and thereby enhancing the integration of weather phenomena into decision-making processes.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Assessment of NMME and SEAS5 Forecasts for West African JJAS Rainfall

  • 期刊论文

  • 发表日期：

    2025年12月8日

查看全文
  • Armand Feudjio Tchinda,
  • Rodric Merimé Nonki,
  • Pascaline Liaken Dickmu,
  • Ossénatou Mamadou,
  • Stève Roméo Tanessong

摘要

In West Africa (WA), socioeconomic sectors such as agriculture, water management, electricity production, disaster risk management, and healthcare are heavily dependent on rainfall. It is therefore imperative to have reliable seasonal precipitation forecasts available sufficiently in advance to facilitate effective planning and informed decision-making across climate-sensitive sectors. The performance of 14 seasonal precipitation forecasts from the North American Multi-Model Ensemble (NMME) and European Centre for Medium-Range Weather Forecasts fifth-generation seasonal forecasting system (ECMWF-SEAS5) is evaluated for the June–September (JJAS) season. The Global Precipitation Climatology Centre (GPCC) and the African Rainfall Climatology Version 2 (ARC2) datasets are used as observational references. We began by verifying the ability of the forecasts to reproduce the climatology, and then assessed the performance of each model and the multi-model ensemble in predicting rainfall in WA at timescales ranging from 0 to 5 months lead time. The NMME and SEAS5 models successfully capture the average JJAS seasonal precipitation, estimated at around 11 mm/day in central and southeastern part of WA. Forecast skill is highest at shorter lead times but declines rapidly with increasing lead time, as reflected by the decreasing Heidke Skill Score values. Most models detect normal seasons with over 55% probability of Detection, but struggle to identify seasons above and below normal (probability of Detection < 42%). The performance of the multi-model ensemble is not systematically superior to that of individual models, suggesting the need for more advanced weighting approaches. According to our results, the NMME and SEAS5 models represent a valuable tool during the first three lead times (i.e., leads 0–2) in WA, allowing for the anticipation of key seasonal features before the onset of the JJAS season, and thereby enhancing the integration of weather phenomena into decision-making processes.