Seasonal and inter-annual dynamics of water vapor flux based on five-year eddy covariance measurements over an alpine grassland in arid Central Asia

Abstract

Extensive arid and semi-arid ecosystems in Central Asia are threatened by aridification and desertification owing to intensified evaporation and extreme climates. To understand the mechanisms of water vapor (H₂O) transformations and assess the water balance under climate change in Central Asian grasslands, knowledge of H₂O ecosystem-scale flux and its seasonal and interannual dynamics is important. Based on the eddy covariance technique, this study measured the five-year H₂O flux over the Bayinbuluk Grassland in Central Asia and investigated its environmental controls. The results showed that the grassland was a net source of H₂O flux, emitting 1432 ± 93 mm y⁻¹ from 2017 to 2021, with a mean annual precipitation of 237 ± 69 mm. Seasonal changes in H₂O fluxes were higher during the growing season than that during the non-growing season, with annual maxima generally occurring from July to August. A clear unimodal diurnal pattern in the H₂O flux was observed during both seasons from 2018 to 2021, with peak values appearing at approximately 14:30. We further conducted a wavelet analysis on this long-term quasi-continuous H₂O flux time series and investigated its temporal variability and wavelet coherence with environmental variables. Daily periodicity in H₂O fluxes was detected during most of the growing season. The variations in H₂O fluxes were in phase with changes in air temperature and solar radiation on a daily timescale, with relative humidity showing a negative correlation with H₂O flux. Changes in precipitation, air temperature, soil temperature, and photosynthetically active radiation exhibited stronger positive correlations with H₂O fluxes at monthly and annual timescales than daily timescales. In addition, annual evapotranspiration increased at a rate of ∼100 mm y⁻¹ during the study period, despite precipitation and air temperature showing no apparent increasing trends. Grassland ecosystems in arid Central Asia are expected to emit more H₂O under a warming climate, leading to greater water scarcity and heightened aridity. Our study highlights the importance of conducting long-term continuous eddy covariance and time-series analyses to enhance our understanding of the temporal variability in grassland H₂O exchanges.