Assessment of ERA5 and ERA-Interim in Reproducing Mean and Extreme Climates over West Africa

Abstract

In situ data in West Africa are scarce, and reanalysis datasets could be an alternative source to alleviate the problem of data availability. Nevertheless, because of uncertainties in numerical prediction models and assimilation methods, among other things, existing reanalysis datasets can perform with various degrees of quality and accuracy. Therefore, a proper assessment of their shortcomings and strengths should be performed prior to their usage. In this study, we examine the performance of ERA5 and ERA-interim (ERAI) products in representing the mean and extreme climates over West Africa for the period 1981–2018 using observations from CRU and CHIRPS. The major conclusion is that ERA5 showed a considerable decrease in precipitation and temperature biases and an improved representation of inter-annual variability in much of western Africa. Also, the annual cycle is better captured by ERA5 in three of the region’s climatic zones; specifically, precipitation is well-reproduced in the Savannah and Guinea Coast, and temperature in the Sahel. In terms of extremes, the ERA5 performance is superior. Still, both reanalyses underestimate the intensity and frequency of heavy precipitations and overestimate the number of wet days, as the numerical models used in reanalyses tend to produce drizzle more often. While ERA5 performs better than ERAI, both datasets are less successful in capturing the observed long-term trends. Although ERA5 has achieved considerable progress compared to its predecessor, improved datasets with better resolution and accuracy continue to be needed in sectors like agriculture and water resources to enable climate impact assessment.

摘要

西非的现场数据稀少, 再分析数据集可以作为缓解数据可用性问题的替代来源。然而, 由于数值预测模型和同化方法自身的不确定性, 现有的再分析数据集具有不同程度的质量和精度表现力。因此, 有必要在使用前对再分析资料的优缺点进行适当评估。本研究我们利用CRU和CHIRPS的观测结果检验了ERA5和ERA-interim (ERAI) 产品对于1981-2018年西非平均气候和极端气候方面的表现力。主要结论是ERA5资料显示降水和气温的偏差显著减少, 西非大部分地区年际变率的代表性得以改善。此外, ERA5在该区域的三个气候带中能更好地捕捉到年循环, 特别是, 萨凡纳和几内亚海岸的降水量和萨赫勒地区的气温都得到了很好的再现。就气候极值而言, ERA5性能优越。尽管如此, 这两种再分析资料均低估了强降水的强度和频率, 并高估了雨天的数量, 原因在于再分析资料使用的数值模型往往更容易产生毛毛雨。尽管ERA5的再现能力优于ERAI, 但两个数据集在捕获观测到的长期趋势上均差强人意。尽管与前一代相比, ERA5取得了相当大的进展, 但农业和水资源等部门仍需要改进的具有较高分辨率和准确性的数据集, 以实现气候影响评估。