Water is the cornerstone of mankind survival. However, water shortage/scarcity has been an outstanding issue worldwide, which significantly threatens the development of human societies. In the past two decades (1997-2017), available water resources per capita has reduced by more than 20%. To keep pace with anticipated increasing population and material needs, global water use is expected to continue growing at a rate of approximately 1%/year. At present, water shortage/scarcity impacts more than 3 billion people, and this problem will be further exacerbated due to the climate-driven changes in evaporation, precipitation, and runoff. In addition, human activities also play a significant role in water shortages, which has substantially altered the Earth system since the industrial era th...[ Read more ]
Water is the cornerstone of mankind survival. However, water shortage/scarcity has been an outstanding issue worldwide, which significantly threatens the development of human societies. In the past two decades (1997-2017), available water resources per capita has reduced by more than 20%. To keep pace with anticipated increasing population and material needs, global water use is expected to continue growing at a rate of approximately 1%/year. At present, water shortage/scarcity impacts more than 3 billion people, and this problem will be further exacerbated due to the climate-driven changes in evaporation, precipitation, and runoff. In addition, human activities also play a significant role in water shortages, which has substantially altered the Earth system since the industrial era through constructing water conservancy projects and changing land cover/land use. It has a long history that hydrological models are employed for modeling hydrological processes and water cycle. These models play an indispensable role in evaluating the influences of climate change and human activities on water resources. However, there are still some long-standing challenges, which has not been fully resolved, in water resources management. This doctoral thesis aims to address some of the challenges by three case studies, thereby providing implications for water resources management and decision making.
Irrigation consumes a huge amount of freshwater in water-limited agricultural regions. However, improving irrigation efficiency (IE) has a limited effect in reducing irrigation water consumption, which has been long recognized as the paradox of IE. Although long-recognized by researchers and decision makers, the paradox is rarely utilized to guide irrigation activities in the real word, which is largely due to the missing of rigorous evaluation of basin-scale IE. For example, most of the traditional IE indices don’t explicitly consider irrigation return flow and groundwater's direct contributions to crop evapotranspiration. Therefore, we proposed a new basin-scale IE index that factors in the two water flux accounts, and we holistically analyzed basin-scale IE of the Zhangye Basin (ZB), a typical agricultural area in China, based on integrated ecohydrological modeling. The major study findings in this case study are as follows: 1) About 13% of irrigation water in the ZB becomes return flow, producing a difference of 0.1 between the proposed basin-scale IE index and a traditional field-scale index. 2) Basin-scale IE has significant interannual variations, but its multiyear average shows stability, which may be related to the basin's characteristics (e.g., soil type, land cover, and hydrological processes). 3) Basin-scale IE reveals great spatial heterogeneity, which is attributed to the intensity of surface water-groundwater exchanges. 4) Under the warmer and slightly wetter climate change scenario, return flow in the ZB will increase by 3.2%/decade, leading to an increasing trend in basin-scale IE. Overall, rigorous evaluation of basin-scale IE is critical to making heterogeneous policies and developing adaptive management strategies under the changing climate. The integrated modeling approach developed in this case study provides a methodological foundation for overcoming misunderstandings about the IE paradox and benefits the reform of the current IE policy agenda.
The Greater Bay Area (GBA) is the largest and wealthiest region in southern China. As the current withdrawal from the Dongjiang river is perilously close to the 40% upper limit imposed to ensure a healthy and sustainable ecosystem, water scarcity becomes a crucial bottleneck to the socioeconomic development in the GBA. However, the impacts of climate change on the GBA’s water resources are still outstanding, a holistic assessment of how water resources evolve under various socioeconomic considerations is essential and pivotal for GBA’s sustainability. This case study adopted long short-term memory (LSTM), the most successful deep learning (DL) approach in hydrological modeling, to simulate the river dynamics of the three rivers (Xijiang, Beijiang, and Dongjiang river) in the GBA and projected their temporal changes under two climate change scenarios (SSP245 and SSP585). The major study findings in this case study are as follows: 1) Annual runoff of all the three rivers shows an increasing trend under approximately 90% of the model projections under both the scenarios, especially under SSP585. 2). Runoff seasonality in the Beijiang and Dongjiang river basin is projected to significantly increase under both the scenarios, while it in the Xijiang river basin remains unchanged throughout the 21st century. 3) The Beijiang river and Dongjiang river basin will be faced with new flood and drought threats triggered by climate change. 4) In the Xijiang river basin, although flood risk will increase in the future, drought risk will be alleviated, which further illustrates the strong resilience of the XRB to climate change. Overall, this case study systematically analyzes the impacts of climate change on water resources in the GBA, and provides some suggestions regarding how to reduce the vulnerability of water resources in the GBA to climate change and guarantee water security.
Excessive riverine export of nitrogen caused by human activities poses huge threats to water security in coastal areas and marine ecosystems. Given the current state of technology, long-term high-frequency monitoring of riverine nitrogen remains costly and applicable only for a small number of rivers worldwide. On the other hand, existing models, either process-based or empirical, are usually deficient in applicability beyond where they were built. A coherent view of the daily dynamics of global riverine nitrogen export is still missing. In this case study, we use LSTM-based DL approaches to model daily riverine nitrogen export in response to hydrometeorological (i.e., runoff and precipitation) and anthropogenic drivers (i.e., fertilization activities). LSTM models for the daily concentration and flux of dissolved inorganic nitrogen (DIN) were built in North River watershed, a coastal watershed in Fujian province. The major study findings in this case study are as follows: 1) The DL models performed great for both the predictions of DIN concentration and flux, with Nash-Sutcliffe efficiency coefficients (NSEs) up to 0.67 and 0.92, respectively. Under comparable data conditions, this excellent performance is unlikely to be achieved by process-based models. 2) The flux model ensemble, without retraining, performed well (mean NSE = 0.32–0.84) in seven distinct watersheds across Asia, Europe, and North America, and retraining with multi-watershed data further improved the lowest NSE from 0.32 to 0.68. 3) The multi-watershed flux model projects 0.60–12.4% increases in the nitrogen export to oceans from the studied watersheds under a 20% increase in fertilizer consumption, which rises to 6.7–20.1% with a 10% increase in runoff, indicating the synergistic effect of human activities and climate change. This case study demonstrates the great power of explainable AI in water environment modeling. It provides a potential shortcut to a consistent understanding of the global daily-resolution dynamics of riverine nitrogen export under the currently limited data conditions, which is significant for global water resources management.