Spatiotemporal variation of potential evapotranspiration in the Qinling Mountain Areas and attribution analysis of typical watershed runoff

Mountain watersheds, as the critical regional water sources, are highly sensitive to climate change. It is a high demand to investigate the hydrological variations in mountain watersheds for the protection of regional ecological environments and sustainable water resources. This study aims to explore the spatiotemporal variation in the potential evapotranspiration and attribution analysis of typical watershed runoff in the Qinling Mountain Areas. Six typical watersheds were selected as the research objects. Meteorological data was daily collected from 79 meteorological stations from 1965 to 2019. Annual runoff data was from the hydrological control stations in the watersheds, together with the annual sunspot numbers, atmospheric circulation indices, and the Nino3.4 index, such as the Penman-Monteith equation. The modified Mann-Kendall trend test was employed to analyze the spatiotemporal variations in the potential evapotranspiration (PET) across the study area using cross-wavelet transforms and the time-varying Budyko framework. PET sensitivity was assessed on various climatic factors to explore the potential drivers of PET changes. Finally, the contributions of PET were quantified for the runoff variations across different periods. The results indicated: 1) There were some significant differences in spatial distribution and trends in the multi-year average PET and climate factors from 1965 to 2019. Multi-year average PET and sunlight hours (SH) exhibited a spatial pattern of higher in the northeast and lower in the southwest. While the maximum and minimum relative humidity (RHmax and RHmin) showed the opposite trends. Additionally, the temperature generally increased on an annual scale, while the SH and wind speed (WS) decreased across a broad area. The average annual WS of all vegetation types was below 1.8 m/s. The significant reduction in the annual WS was concentrated mainly in the vegetation types, such as the broad-leaved forest, cultivated vegetation, shrub, coniferous forest, and grass (P<0.05). It was also found that the differential modes were used to accurately simulate the annual variation of PET in the Qinling Mountain Area on an annual scale, with the coefficient of certainty R² (0.96) between the calculated and the actual value. 2) The annual PET and climatic factors exhibited different patterns of change across various vegetation types. Specifically, there was a significant decrease in the SH in the broadleaf forests and cultivated vegetation areas, with the spatially uneven distribution of relative humidity change trends (P<0.05). The sensitivity of annual PET to annual SH showed an overall pattern of low in the north and high in the east and west in spatial distribution, with a value range of -0.04~0.04. However, there was no spatial distribution pattern in the sensitivity of PET to RHmax. The highest sensitivity of annual PET to multiple climatic factors was found in the grassland vegetation. There was a great variation in the duration and scope of the influence of topographical factors, solar activity, and atmospheric circulation on the annual PET. The PET showed a highly significant decreasing trend with the increasing elevation and slope (P<0.001). Among them, 98.73% of meteorological stations shared a negative correlation between annual PET and the Arctic Oscillation index (AO index), and 20.25% shared a significant negative correlation (P<0.05). 3) Furthermore, the three northern foothill watersheds showed a decreasing trend in the annual runoff depth among the six typical watersheds. While two of the three southern foothill watersheds exhibited an increasing trend. There was a generally relatively small impact of PET on the runoff variation in the different typical watersheds and sliding window periods. The variation in the vegetation was the primary driving factor on the watershed runoff. The sustainable utilization of water resources in the mountain watersheds can greatly contribute to ecological protection and high-quality development in the Yellow River and Yangtze River Basins. Continuous monitoring and adaptive management were also highlighted in response to the future climates. The multiple types of data and reliable hydrological analysis were integrated to fully understand the complex interactions between climatic factors and hydrological processes in mountainous regions.