This paper evaluates the atmospheric controls contributing to actual evapotranspiration (ET) variability and trends across the ecological zones in the data-scarce region of West Africa.

Spatial and temporal ET dynamics in the region were analysed using three selected prognostic- and machine learning-based global ET over 1983 to 2012. In addition, the responses of annual actual ET to changes in annual rainfall, atmospheric evaporative demand (AED) and incident solar radiation were evaluated using joint time-frequency wavelet decomposition method and pixel-wise partial correlation analysis. Based on the Mann-Kendall statistics, results from the Global Land Evaporation Amsterdam Model (GLEAM) data revealed high variability (> 70 mm yr⁻¹) and significant rising trends (> 5 mm yr⁻¹) in annual ET in the Western Sahel. In contrast, results from CSIRO Penman-Monteith Leuning (PML) ET data showed declining trends by more than 4 mm yr⁻¹ in some parts of the Savannah zone during the study period.

Results also showed some similarities in the coupling between annual ET, rainfall and AED, particularly along the Sahelian belt time-frequency domain. This implies that these atmospheric variables largely explain the signals in ET variations across the zones. However, the atmospheric variables have different modulating powers on the ET variations in the region. Additionally, we find that other non-climatic factors may have also influenced the variations in ET during the study period. This could serve as a pointer to focusing on other non-climatic drivers, potentially affecting ET trend and variability in the study area. Therefore, the study highlights the importance of providing essential and robust information on ET variability and trend as affected by different atmospheric drivers and its potential impacts on water resources, particularly in the Sahelian zone where agricultural biomes are often threatened by water stress and drought.

本文评估了导致西非数据稀缺地区生态区实际蒸散量 (ET) 变化和趋势的大气控制。

使用 1983 年至 2012 年期间三个选定的基于预测和机器学习的全球 ET 分析了该地区的时空 ET 动态。此外，年度实际 ET 对年降雨量、大气蒸发需求 (AED) 和事件变化的响应使用联合时频小波分解方法和逐像素偏相关分析来评估太阳辐射。根据 Mann-Kendall 统计数据，全球土地蒸发阿姆斯特丹模型 (GLEAM) 数据的结果显示西萨赫勒地区的年度 ET具有高变异性（> 70 mm yr ^-1）和显着上升趋势（> 5 mm yr ^-1） . 相比之下，CSIRO Penman-Monteith Leuning (PML) ET 数据的结果显示下降趋势超过 4 mm yr ^-1在研究期间在萨凡纳地区的某些地区。

结果还显示了年 ET、降雨量和 AED 之间耦合的一些相似性，特别是在萨赫勒带时频域。这意味着这些大气变量在很大程度上解释了跨区域 ET 变化的信号。然而，大气变量对该地区的ET变化具有不同的调节能力。此外，我们发现其他非气候因素也可能影响了研究期间 ET 的变化。这可以作为关注其他非气候驱动因素的指针，可能会影响研究区域的 ET 趋势和变化。因此，该研究强调了提供有关受不同大气驱动因素影响的 ET 变异性和趋势及其对水资源的潜在影响的基本和可靠信息的重要性，