柴达木盆地蒸散发遥感估算与耗水有效性评价

Remote sensing estimation of evapotranspiration and efficiency evaluation of water consumption in Qaidam Basin

陆面蒸散发(ET)是自然生态系统水分耗散的主要方式,准确把握其时空变化特征,对于区域水资源合理利用与生态环境保护具有重要意义。针对我国西北干旱内陆区实测资料匮乏的现状,基于MODIS产品,构建了具有时空二维属性的地表温度-植被指数特征空间,实现了柴达木盆地陆面蒸散发的时空连续估算;在此基础上,进一步分离土壤蒸发与植被蒸腾,开展研究区自然生态系统耗水有效性评价。结果表明:(1)柴达木盆地近九年多年平均ET为188.75 mm,受降水空间分布格局影响,ET具有明显的从东南向西北减少趋势;(2)土壤蒸发与植被蒸腾多年平均值为171.06 mm和14.26 mm,后者年内峰值出现时间总体比前者晚1个月,具有滞后效应;(3)盆地陆地生态系统多年平均耗水总量为430.94亿m^(3),其中高效、中效和低效耗水的占比分别为6.55%、52.57%和40.88%。在区域尺度以时空连续的方式揭示了ET水分消耗的有效性,可为基于ET水资源管理目标的实现提供重要科学支撑。

Terrestrial evapotranspiration(ET), composed of soil evaporation and plant transpiration, is the main mode of water dissipation in the natural ecosystems. In general, ET returns about two thirds of the annually terrestrial precipitation to the atmosphere on the global scale, and the ratio can reach as high as 80% in arid areas. As a result, accurate knowledge of ET is crucial to understand the water budget as well as its effect on ecosystems. This is particularly important for water resources management and eco-environmental protection in arid regions. Compared with ground-based ET measurements recorded by sparse flux towers, satellite remote sensing(RS) is a suitable tool to map regional ET patterns in a temporally and spatially consistent manner, yielding a series of RS-based ET models with varying mechanisms and degrees of complexity. Considering the sparse distribution of meteorological observations in the Northwest of China, we adopted the surface temperature-vegetation index(TVX) method in this study to estimate daily ET in Qaidam Basin from 2011 to 2019. Uniquely, the theoretical boundaries of the TVX feature space were determined pixel by pixel based on the surface energy balance principle, so ET could be mapped for regions with complex topography in a spatially consistent manner. Then a simple dual-source ET framework was used to partition evapotranspiration into soil evaporation(E) and plant transpiration(T), and finally the water consumption efficiency of natural ecosystems was evaluated by further partitioning E into soil evaporation over bare land and soil evaporation beneath the canopy. All the above schemes were demonstrated with Terra Moderate Resolution Imaging Spectroradiometer(MODIS) products. Results show that the annual ET of Qaidam Basin from 2011 to 2019 decreased at first and then increased, with an average value of 188.75 mm. As for seasonal variations, ET from April to September accounted for 80% of the annual ET with the peak value arising in June or July. Influenced by the spatial patterns of precipitation, the ET presents an obviously decreasing trend from southeast to northwest and from the surrounding mountains to the interior of the basin. Local high ET values in the interior of the basin are mainly distributed around the rivers, lakes, and springs. The annual average E and T in the recent nine years is 171.06 mm and 14.26 mm, respectively. Although the seasonal variations of both E and T follow a trend similar to sinusoidal curves with the peak values observed in summer, the peak of the latter appears one month later than that of the former. In terms of spatial distribution, soil evaporation shares patterns consistent with total ET, but plant transpiration is mainly observed in the southeast of the basin, showing a semi-ring structure. The annually total water consumption of terrestrial ecosystems in the basin is 43.094 billion m^(3) on average. The corresponding proportion of high-efficiency, medium-efficiency and low-efficiency water consumption is 6.55%, 52.57% and 40.88%, respectively. Due to the extremely low fractional vegetation cover of Qaidam Basin, the amount of water consumption with high-efficiency is much lower than the other two. However, when considered water consumed by bare land and dense canopy, the water consumed by dense canopy per unit area is 297.83 mm, which is much higher than that consumed by bare land. In conclusion, this study reveals the water consumption efficiency of ET in the whole Qaidam Basin in a temporally and spatially consistent manner, which provides important scientific support for ET-based water resources management.