Understanding plant water consumption is crucial for artificial afforestation under drought environments and water stress in desert regions. However, the water consumption characteristics of desert species responding to the irrigation regimes are often neglected. By conducting a field test in the Taklimakan Desert Highway shelterbelt, this study examines the sap flow traits of two typical woody halophyte species (Calligonum mongolicum and Haloxylon ammodendron) and how they react to weather conditions and watering practices. Under the same irrigation treatment, the stem flux of C. mongolicum on sunny days was 1.5–5.3 times that on dusty days, while the stem flux of H. ammodendron on sunny days was 3.5–5.5 times that on dusty days. Both species demonstrated some sap flow during the night, representing 14.3%–24.9% and 7.3%–10.4% of the total sap flow for C. mongolicum and H. ammodendron, respectively. H. ammodendron maintained a higher stem flow during daytime and was more drought resistant than C. mongolicum. The daily sap flow patterns of these two species varied, showing both ‘single’ and ‘double peak’ curves depending on the watering conditions. A delay was also observed between the sap flow of these two species and the environmental factors. The factors influencing plant sap flow were found to be in the order of solar radiation, temperature, relative humidity, and saturated water vapor pressure difference. A BP-neural network proved highly effective for accurately simulating the sap flow of these two species. This research provides insights into how two common desert tree species adapt their water use in response to drought conditions, which is vital for artificial forest creation in desert areas.

了解植物耗水量对于干旱环境和沙漠地区缺水条件下的人工造林至关重要。然而，沙漠物种对灌溉制度的耗水特征往往被忽视。通过在塔克拉玛干沙漠公路防护林进行现场试验，本研究研究了两种典型木本盐生植物（沙葱和梭梭）的液流特征以及它们对天气条件和浇水方法的反应。相同灌溉处理下，蒙古沙晴天茎通量是沙尘天的1.5~5.3倍，梭梭晴天茎通量是沙尘天的3.5~5.5倍。两种物种在夜间都表现出一定的液流，分别占蒙古梭梭和梭梭总液流的 14.3%–24.9% 和 7.3%–10.4% 。梭梭白天保持较高的茎流量，比蒙古梭梭更耐旱。这两个物种的日常液流模式各不相同，根据浇水条件呈现“单峰”和“双峰”曲线。还观察到这两个物种的液流与环境因素之间存在延迟。影响植物液流的因素依次为太阳辐射、温度、相对湿度和饱和水蒸气压差。事实证明，BP 神经网络对于准确模拟这两个物种的液流非常有效。这项研究深入了解了两种常见的沙漠树种如何调整水分利用以应对干旱条件，这对于沙漠地区的人工林创建至关重要。