Satellite‐based evaluations of change in vegetation phenology have been explored extensively but land cover‐specific climate factors driving these anomalous changes are not fully understood in northern Sub‐Saharan Africa. In this study, we identified the climatic factors controlling the start of the season (SOS) extracted from GIMMS NDVI from 1988 to 2013 with onset of rainy season (ORS), annual mean temperature (Temp) and precipitation (PP) through the stepwise regression analysis. The results showed that the SOS shifted towards a late onset in a northward direction with distinct earlier and later trends in grassland and cropland, respectively. The stepwise regression has successfully built a model between SOS and its drivers in 46.0% of the total pixels, where its primary factor differed regionally across land covers. The ORS explained the local anomalous SOS change primarily at 44.7% of the pixels where the model was built. Although the ORS was the primary dominant factor in savannah and cropland, the Temp and PP were leading in grassland and shrubland, respectively, and all factors contributed evenly in evergreen forest. The difference of land cover‐specific primary factor implicates complex process in dependency of local vegetation phenology on physiological traits and climate regime across land cover in Sub‐Saharan Africa.

中文翻译：

---

1988年至2013年控制撒哈拉以南非洲北部植被物候发生的年际变化的气候因素

广泛研究了基于卫星的植被物候变化评估，但在撒哈拉以南非洲北部，导致这些异常变化的特定于土地覆盖的气候因素尚未得到充分理解。在本研究中，我们通过逐步回归确定了从1988年至2013年从GIMMS NDVI中提取的控制季节开始（SOS）的气候因子，包括雨季（ORS），年平均温度（Temp）和降水（PP）。分析。结果表明，SOS向北偏晚发生，草原和耕地分别具有明显的前后趋势。逐步回归已成功地在占总像素46.0％的SOS及其驱动程序之间建立了模型，其中主要因素在不同的土地覆盖范围内存在差异。ORS解释了局部异常SOS的变化主要是模型生成像素的44.7％。尽管ORS是稀树草原和农田的主要主导因素，但Temp和PP分别在草原和灌木丛中处于领先地位，所有因素在常绿森林中的贡献均均。特定于土地覆盖的主要因素的差异牵涉到复杂的过程，这取决于撒哈拉以南非洲各地土地覆盖范围内当地植被物候对生理特征和气候状况的依赖性。