Month: February 2024

Abstract

In recent years, there has been an increased interest in understanding and predicting the weather using weather station data with Spatial-Temporal Graph Neural Networks (STGNN). However, it has large prediction errors as a result of the inherent non-linearities and the influence of dynamic spatio-temporal auto-correlation. Using a continuously-varying graph topology chronologically, while embedding domain knowledge to enforce validity, can effectively resolve the issue, but the implementation of such concept constitutes an interdisciplinary challenge for researchers. A Dynamic Graph Former (DGFormer) model is proposed to address this challenge. It combines a topology learner through a deep generative layer with domain knowledge enhancement inserted into the STGNN structure, where the derived physics-guided method allows for an efficient integration with the earth system. For capture of the optimal topology, we merge a node-embedding-based similarity metric learning and the superposition principle as physical assistants into the dynamic graph module. We evaluate our model with a real-world weather dataset on short-term (12 hours) and medium-range (360 hours) prediction tasks. DGFormer achieves outstanding performance with obvious improvements by up to 34.84% at short-term prediction and by up to 23.25% at medium-range prediction compared with the state-of-the-art methods. We also conducted detailed analyses for cities in three regions and visualized the dynamic graph, revealing the characteristics, advantages, and graph visualization of our model.

Abstract

In recent years, there has been an increased interest in understanding and predicting the weather using weather station data with Spatial-Temporal Graph Neural Networks (STGNN). However, it has large prediction errors as a result of the inherent non-linearities and the influence of dynamic spatio-temporal auto-correlation. Using a continuously-varying graph topology chronologically, while embedding domain knowledge to enforce validity, can effectively resolve the issue, but the implementation of such concept constitutes an interdisciplinary challenge for researchers. A Dynamic Graph Former (DGFormer) model is proposed to address this challenge. It combines a topology learner through a deep generative layer with domain knowledge enhancement inserted into the STGNN structure, where the derived physics-guided method allows for an efficient integration with the earth system. For capture of the optimal topology, we merge a node-embedding-based similarity metric learning and the superposition principle as physical assistants into the dynamic graph module. We evaluate our model with a real-world weather dataset on short-term (12 hours) and medium-range (360 hours) prediction tasks. DGFormer achieves outstanding performance with obvious improvements by up to 34.84% at short-term prediction and by up to 23.25% at medium-range prediction compared with the state-of-the-art methods. We also conducted detailed analyses for cities in three regions and visualized the dynamic graph, revealing the characteristics, advantages, and graph visualization of our model.

Abstract

In regions where gauging is lacking or limited, the development of a methodology for water balance assessment is a key challenge. Water balance assessment is a significant task in managing the sustainable use of water resources. This study aims to improve the accuracy of water balance components including actual evapotranspiration and groundwater recharge rate in the Hashtgerd study area, Iran. Groundwater extraction volumes are determined based on two-period data collection; while precipitation, evapotranspiration, and groundwater storage change are calculated from annual datasets. To address the data-scarce problem, as an additional measure of actual evapotranspiration, satellite measurements are used to improve recharge rate accuracy. To understand the impact of satellite data uncertainties on water resources studies, each estimated water balance component is compared to observation data, and the uncertainty of these components is quantified using statistical methods, variability, and standard error. The results show that the Hashtgerd aquifer receives 199 MCM from the main drains which is a major part of the water inflow. The extraction of wells is one of the most significant outflow components of the aquifer (i.e., 284 MCM); while water loss by evaporation is estimated to be 207 MCM of the outflow. The finding shows that satellite-based evapotranspiration can reduce recharge uncertainty which can improve the resolution of groundwater balance. This study supports that the aquifer is under severe environmental pressure. One of those concerns is that groundwater levels decreased by 0.92 m per year between 2000 and 2019.

Abstract

In regions where gauging is lacking or limited, the development of a methodology for water balance assessment is a key challenge. Water balance assessment is a significant task in managing the sustainable use of water resources. This study aims to improve the accuracy of water balance components including actual evapotranspiration and groundwater recharge rate in the Hashtgerd study area, Iran. Groundwater extraction volumes are determined based on two-period data collection; while precipitation, evapotranspiration, and groundwater storage change are calculated from annual datasets. To address the data-scarce problem, as an additional measure of actual evapotranspiration, satellite measurements are used to improve recharge rate accuracy. To understand the impact of satellite data uncertainties on water resources studies, each estimated water balance component is compared to observation data, and the uncertainty of these components is quantified using statistical methods, variability, and standard error. The results show that the Hashtgerd aquifer receives 199 MCM from the main drains which is a major part of the water inflow. The extraction of wells is one of the most significant outflow components of the aquifer (i.e., 284 MCM); while water loss by evaporation is estimated to be 207 MCM of the outflow. The finding shows that satellite-based evapotranspiration can reduce recharge uncertainty which can improve the resolution of groundwater balance. This study supports that the aquifer is under severe environmental pressure. One of those concerns is that groundwater levels decreased by 0.92 m per year between 2000 and 2019.

Abstract

The study investigated the impact of climate and land cover change on water quality. The novel contribution of the study was to investigate the individual and combined impacts of climate and land cover change on water quality with high spatial and temporal resolution in a basin in Turkey. The global circulation model MPI-ESM-MR was dynamically downscaled to 10-km resolution under the RCP8.5 emission scenario. The Soil and Water Assessment Tool (SWAT) was used to model stream flow and nitrate loads. The land cover model outputs that were produced by the Land Change Modeler (LCM) were used for these simulation studies. Results revealed that decreasing precipitation intensity driven by climate change could significantly reduce nitrate transport to surface waters. In the 2075–2100 period, nitrate-nitrogen (NO₃-N) loads transported to surface water decreased by more than 75%. Furthermore, the transition predominantly from forestry to pastoral farming systems increased loads by about 6%. The study results indicated that fine-resolution land use and climate data lead to better model performance. Environmental managers can also benefit greatly from the LCM-based forecast of land use changes and the SWAT model’s attribution of changes in water quality to land use changes.

Graphical abstract