水文学(农业)
土壤盐分
地下水
构造盆地
地质学
环境科学
冲积扇
干旱
盐度
湿地
土壤水分
蒸散量
地表水
冲积层
旱地盐分
地下水位
含水层
水平衡
流域
横断面
冲积平原
浸出模型
地下水补给
底土
地下水流
作者
Ziyue Wang,Chaoyao Zan,Y. X. Zhao,Bo Xu,Rui Long,Xiaoyong Wang,Jun Zhang,Tianming Huang
出处
期刊:Water
[Multidisciplinary Digital Publishing Institute]
日期:2025-12-05
卷期号:17 (24): 3462-3462
摘要
Identifying the dominant mechanisms of water and soil salinization in arid and semi-arid endorheic basins is fundamental for our understanding of basin-scale water–salt balance and supports water resources management. In many inland basins, mineral dissolution, evaporation, and transpiration govern salinization, but disentangling these processes remains difficult. Using the Barkol Basin in northwestern China as a representative endorheic system, we sampled waters and soils along a transect from the mountain front through alluvial fan springs and rivers to the terminal lake. We integrated δ18O–δ2H with hydrochemical analyses, employing deuterium excess (d-excess) to partition salinity sources and quantify contributions. The results showed that mineral dissolution predominated, contributing 65.8–81.8% of groundwater salinity in alluvial fan settings and ~99.7% in the terminal lake, whereas direct evapoconcentration was minor (springs and rivers ≤ 4%; lake ≤ 0.2%). Water chemistry types evolved from Ca-HCO3 in mountainous runoff, to Ca·Na-HCO3·SO4 in groundwater and groundwater-fed rivers, and finally to Na-SO4·Cl in the terminal lake. The soil profiles showed that groundwater flow and vadose-zone water–salt transport control spatial patterns: surface salinity rises from basin margins (<1 mg/g) to the lakeshore and is extremely high near the lake (23.85–244.77 mg/g). In spring discharge belts and downstream wetlands, the sustained evapotranspiration of groundwater-supported soil moisture drives surface salt accumulation, making lakeshores and wetlands into terminal sinks. The d-excess-based method can robustly separate the salinization processes despite its initial isotopic variability.
科研通智能强力驱动
Strongly Powered by AbleSci AI