In recent years, there has been a growing interest in the study and monitoring of Total Water Storage Anomalies (TWSAs). TWSAs provide crucial insights into the impact of natural climate variability and human activities on the water cycle. However, the current methodologies employed to generate high-resolution TWSAs have certain limitations. This article explores an innovative study that proposes improvements in these methodologies, specifically in modelling glaciers, human intervention modelling, and deep learning models. The study also discusses the potential applications of these enhanced TWSAs in environmental monitoring.
Towards an Enhanced Modelling Approach
The study makes use of mascon solutions from NASA's Jet Propulsion Laboratory (JPL), WaterGAP v.2.2d, and hydrological information from the Global Land Data Assimilation System (GLDAS). It introduces a new deep learning model designed to assimilate satellite observations and hydrological simulations to generate high-resolution TWSAs. The model's distinctive feature is its ability to balance accurate values from the Gravity Recovery and Climate Experiment (GRACE) observations and high-resolution structures from the Water Global Hydrological Model (WGHM) simulations.
A New Era of Satellite Data Assimilation
The proposed model employs a self-supervised data assimilation approach with a new loss function. This approach provides global high-resolution TWSAs with a spatial resolution of 0.5. The model combines hydrological simulations and measurements from GRACE and the follow-on GRACE FO satellite missions. It has proven its efficiency by closing the water balance equation in small basins while preserving large-scale accuracy, as inherited from the GRACE FO measurements.
Natural Hazards Monitoring and Beyond
These advancements contribute significantly to monitoring natural hazards locally and offer potential for better understanding the impacts of natural and anthropogenic activities on the water cycle. The highly resolved TWSAs find applications in environmental monitoring indices such as the Flood Potential Index (FPI) and Drought Severity Index (DSI).
Case Study: Ground Water Storage in Semi-Arid Regions
As an example of the practical application of these technologies, a study conducted in a semi-arid region in Iran used the GRACE dataset and the GLDAS hydrological model to analyze declining trends in Terrestrial Water Storage (TWS) and Ground Water Storage (GWS). The study found evidence of indiscriminate groundwater extraction and a decrease in average per capita water per person. This case study highlights the potential of remote sensing as an effective tool for estimating GWS.
A Step Forward
These advancements in the generation of high-resolution TWSAs represent an important step forward in the field. The use of self-supervised learning to downscale satellite gravimetry data from the GRACE satellites contributes greatly to our understanding of the impact of natural climate variability and human activities at basin scales. The study's objective is to solve the inadequacies of existing deep learning algorithms, providing global generalizability and relieving assumptions between different domains.
In conclusion, these advancements in generating high-resolution TWSAs, underscored by the application of deep learning models, mark a significant stride in water monitoring and management. They offer promising prospects for further research and practical applications in understanding the impacts of natural and anthropogenic activities on the water cycle, thus contributing to the sustainability of our planet's water resources.