Study region: The Akesu National Station of Observation and Research for Oasis Agro-ecosystem in the Tarim River Basin, a typical arid region in northwest China. Study focus: Groundwater evapotranspiration (ETg) is a key discharge pathway in arid regions, influenced by hydrothermal conditions and vegetation water use, complicating accurate estimation. Using an automatic water level-compensating lysimeter, ETg was monitored hourly for P euphratica, T ramosissima, and bare land at groundwater depths of 0.7, 1.1, and 1.5 m in August 2023. Variations at hourly and daily scales were analyzed, and a modified White method was applied to estimate ETg. Principal component analysis and a random forest model were used to assess meteorological influences. New hydrological insights for the region: Deeper groundwater reduced ETg from bare land but increased ETg from P. euphratica and T. ramosissima. Vegetation transpiration as a proportion of total evapotranspiration increased with groundwater depth. The average errors in estimating P. euphratica and T. ramosissima ETg using White (20-4) were 30 % and 32 %, and the average error in estimating bare land ETg using White (20-8) was 23 %, which is a reduction of 5, 18, and 4 %, respectively, compared to the traditional White method. Ground temperature was the primary meteorological factor influencing ETg, followed by relative humidity, air temperature, and solar radiation, collectively accounting for > 70 %. These findings enhance understanding of ETg dynamics and improve White methods, critical for groundwater assessment in arid regions.