Climatic Constraints of Spring Phenology and Its Variability on the Mongolian Plateau From 1982 to 2021

The start of vegetation growing season (SOS) plays an important role in the energy cycle between the land and atmosphere. Due to the limited temporal span of a single satellite sensor through time, the continuous variation of the SOS over 40 years has not been adequately quantified. Using the overlapping periods (2001-2015) between the Global Inventory Modeling and Mapping Studies (GIMMS) (1982-2015) and Moderate Resolution Imaging Spectroradiometer (MODIS) normalized difference vegetation index (NDVI) (2001-2021) data sets, we construct an NDVI data set covering the period 1982-2021 on the Mongolian Plateau and further construct a map of relative climatic constraint on the SOS (divided into "temperature-constrained," "precipitation-constrained," and "other" regions) for quantifying SOS variability. We show that the constructed NDVI data set has high consistency and continuity with earlier GIMMS NDVI data. Regions with the SOS constrained by temperature account for 55.3% of the plateau and are located in northwestern and northeastern cold areas, while regions with the SOS constrained by precipitation constitute over 34.7% and are located in central and southwestern drier regions. Importantly, the temperature-constrained SOS has continuously and significantly advanced, with a total advance of 4.8 days over 40 years. In contrast, the precipitation-constrained SOS reversed from advancing to delaying in 2005. This suggests that differentiating the climatic constraint on the SOS might be a practical treatment for reducing the uncertainties in the SOS trends in previous studies. Interestingly, the precipitation-constrained SOS does not significantly correlate to both the chilling and forcing temperatures, indicating less dependency of the SOS on chilling, which may not have been well considered previously. Plain Language Summary Shifts in the spring phenology of plants have profound effects on ecosystem structure and function. Trends in spring phenology have become increasingly controversial as observational periods have extended. To reduce the uncertainties in the controversial changes of spring phenology and explore the trends of the start of the growing season (SOS) over 40 years on the Mongolian Plateau, we divided the plateau into three subregions and quantified the main climatic constraints on spring phenology. Based on these divisions and quantifications, we observed different trends of spring phenology in different subregions. We can infer that warming could continuously advance spring phenology in the future in areas where the spring phenology is constrained by temperature, but the magnitude of the advance may be smaller than that estimated before the 1990s. However, the advance of spring phenology in regions where the spring phenology is constrained by precipitation will be delayed with continuous warming. Due to none of the single remotely sensed normalized difference vegetation index (NDVI) data sets covering sufficiently long periods to predict future trends of spring phenology, the construction of a consistent and continuous NDVI time series using the overlapping periods of different available sensors, especially for regions with an annual NDVI > 0.1, is encouraged.