Shallow groundwater often presents multiple behaviors in supporting the sustainability of arid agroecosystems. The behaviors are especially important and complicated for the arid irrigated areas with fragmented land-use types. In the view of understanding the role of shallow groundwater in such areas, the MODFLOW and MT3DMS (with the assistance of HYDRUS-1D) were used to simulate groundwater flow and salt transport in a typical observation site of the arid upper Yellow River basin (YRB). The original MODFLOW was modified with introducing a variable expression of specific yield, aiming to improve the modeling accuracy in shallow groundwater environments. The model was calibrated and validated using the observation data of groundwater depth and drainage in 2013 and 2012, respectively. Groundwater evapotranspiration and lateral water-salt exchange among different land-use types were simulated and analyzed; meanwhile, the dry drainage effect of non-cultivated natural land was quantitatively interpreted. Results indicated that the fragmented land-use/planting pattern (with different irrigation scheduling) resulted in significant lateral groundwater flow and water/salt redistribution among different land-use fields during irrigation. The groundwater evapotranspiration ranged at about 160–220 mm for different land-use types, reaching to 0.30–0.45 of plant evapotranspiration (ET_pt) during the growing seasons. This effectively supplemented the plant water use in arid climate, especially for the unirrigated cultivated land (during some stages) and the non-irrigated natural land. Meanwhile, the lateral groundwater flow drove salt migrating from the cultivated land to the natural land. The natural land accumulated 27–40% of salts introduced by irrigation, playing a significant role in accommodating salts (i.e. the dry drainage effect). Moreover, our modeling work further implied that the reliable groundwater modeling in arid and shallow groundwater areas required a more detailed model conceptualization and more accurate information of land-use distribution, irrigation scheduling, and local microrelief than the general understanding in the past.

浅层地下水通常在支持干旱农业生态系统的可持续性方面表现出多种行为。对于土地利用类型分散的干旱灌溉地区，这些行为尤为重要和复杂。为了了解浅层地下水在这些地区的作用，在干旱黄河上游典型观测点（MOD）中，使用MODFLOW和MT3DMS（在HYDRUS-1D的帮助下）模拟了地下水流和盐分的运移。 ）。原始MODFLOW进行了修改，引入了比产量的变量表达式，旨在提高浅层地下水环境中的建模精度。分别使用2013年和2012年的地下水深度和排水观测数据对模型进行了校准和验证。对不同土地利用类型之间的地下水蒸散量和侧向水盐交换进行了模拟和分析；同时，对非耕地自然干旱的排水效果进行了定量解释。结果表明，零散的土地利用/种植方式（不同的灌溉计划）导致灌溉期间不同土地利用领域之间明显的侧向地下水流和水/盐再分配。不同土地利用类型的地下水蒸散量约为160-220 mm，植物蒸散量可达0.30-0.45（结果表明，零散的土地利用/种植方式（不同的灌溉计划）导致灌溉期间不同土地利用领域之间明显的侧向地下水流和水/盐再分配。不同土地利用类型的地下水蒸散量约为160–220 mm，植物蒸散量可达0.30–0.45（结果表明，零散的土地利用/种植方式（不同的灌溉计划）导致灌溉期间不同土地利用领域之间明显的侧向地下水流和水/盐再分配。不同土地利用类型的地下水蒸散量约为160-220 mm，植物蒸散量可达0.30-0.45（ET _pt）。这有效地补充了干旱气候中的植物用水，特别是对于未灌溉的耕地（在某些阶段）和非灌溉的自然土地。同时，地下水的横向流动驱使盐从耕地迁移到自然地。自然土地积累了灌溉引入的盐的27％至40％，在容纳盐方面起着重要作用（即干排水作用）。此外，我们的建模工作还暗示，在干旱和浅水区进行可靠的地下水建模需要比过去的一般理解更详细的模型概念化和更准确的土地利用分布，灌溉调度和局部微浮雕信息。