Ge et al. (2025) High-resolution GRACE-based assessment of hydrological drought patterns and recovery dynamics across the Nile Basin countries (2003–2023)

Identification

• Journal: Journal of Hydrology Regional Studies
• Year: 2025
• Date: 2025-11-15
• Authors: Ying Ge, Vagner G. Ferreira, Jia Xu, Christopher E. Ndehedehe
• DOI: 10.1016/j.ejrh.2025.102904

Research Groups

• School of Earth Sciences and Engineering, Hohai University, Nanjing, China
• School of Environment and Science, Griffith University, Nathan, QLD, Australia
• Australian Rivers Institute, Griffith University, Nathan, QLD, Australia

Short Summary

This study developed a comprehensive framework combining high-resolution GRACE terrestrial water storage (TWS) downscaling with advanced spatiotemporal pattern recognition to monitor hydrological droughts across the Nile Basin countries from 2003 to 2023. It identified seven distinct drought hotspots, quantified pixel-level drought characteristics and recovery dynamics, and revealed varied country-scale water security trends, highlighting significant improvements in some nations post-2017 while others face persistent challenges.

Objective

• To quantify pixel-level drought characteristics (magnitude, duration, and severity) across the Nile Nations using high-resolution GRACE TWS data from 2003 to 2023.
• To apply the Joint Approximate Diagonalization of Eigenmatrices (JADE) algorithm to decompose water storage deficit patterns and identify dominant modes of variability.
• To analyze country-scale drought impacts by calculating the percentage of land area affected by water storage deficits for each of the 11 Nile Basin nations, examining temporal trends and spatial patterns from 2003 to 2023.

Study Configuration

• Spatial Scale: The Nile Nations (Burundi, Democratic Republic of Congo, Egypt, Eritrea, Ethiopia, Kenya, Rwanda, South Sudan, Sudan, Tanzania, and Uganda), covering the Nile River Basin. Data resolution enhanced from approximately 400 km to 28 km (0.25° grid).
• Temporal Scale: January 2003 to December 2023 (21 years).

Methodology and Data

• Models used:
  • Gaussian Process Regression (GPR) for downscaling GRACE TWS data.
  • Joint Approximate Diagonalization of Eigenmatrices (JADE) algorithm for extracting independent spatiotemporal drought patterns.
  • Seasonal and Trend decomposition using LOESS (STL) combined with a Long Short-Term Memory (LSTM) neural network for reconstructing GRACE TWS data gaps.
  • Noah land surface model (GLDAS Version 2.1) for soil moisture content.
• Data sources:
  • Gravity Recovery and Climate Experiment (GRACE) and GRACE Follow-On (GRACE-FO) satellite missions (CSR RL06.2 mascon solutions) for terrestrial water storage (TWS).
  • Global Land Data Assimilation System (GLDAS) Version 2.1 Noah model for soil moisture storage (SMS) and annual rainfall data.

Main Results

• The downscaled GRACE TWS data (approximately 28 km resolution) revealed finer spatial structures and localized variability compared to the original 400 km resolution, particularly in hydrologically active regions like Lake Nasser, the Rift Valley, and Lake Victoria.
• An overall increase in water storage was observed across the Nile Basin from 2003 to 2023, but with regional declines, notably up to -10 mm/year in eastern Kenya and about -5 mm/year in eastern Democratic Republic of Congo.
• Seven distinct drought hotspots were identified:
  • Lake Victoria Basin (Region 1): Experienced a prolonged drought from 2003 to 2013, with a peak deficit of -211.3 km³ in October 2006 and a maximum total severity of -11,214.3 km³·month. Average recovery time was 6.5 months.
  • Lake Tanganyika (Region 2): Faced a severe drought from 2003 to 2007, with a peak deficit of -147.4 km³ in February 2006 and a maximum instantaneous severity of -3436.4 km³·month. Average recovery time was 6 to 7 months.
  • Lake Nasser and Aswan Dam (Region 3): Showed significant water deficits from 2003 to 2019, with the most severe periods in 2003–2006 and 2010–2014, peaking at -2464.3 km³·month in January 2018. Average recovery time was 8.5 months.
  • East Kenya (Region 4): Experienced a pronounced drought hotspot, with a significant event commencing in late 2019, leading to a peak deficit of -59.2 km³ in September 2021 and maximum severity of -1064.1 km³·month in October 2023. Average recovery time was 6.9 months.
  • Ethiopian Highlands (Region 5): Suffered severe water storage deficits from 2003 to 2006 and 2007 to 2012, with a peak deficit of -71.2 km³ in November 2004 and maximum severity of -1785.6 km³·month in January 2007. Average recovery time was 6 to 7 months.
  • South Sudan (Region 6): Exhibited distinct spatial patterns of water deficits, with a prolonged and severe drought event from mid-2015 to mid-2021, and a peak instantaneous severity of -1656.0 km³·month in December 2013. Average recovery time was 5.0 months.
  • Northwest Democratic Republic of Congo (Region 7): Experienced a major, prolonged drought period from 2010 to 2014, with a peak severity of -571.7 km³·month in January 2014. Mean median recovery time was 4.3 months.
• Country-scale analysis revealed dramatic improvements in water security for Rwanda, Uganda, and Burundi post-2017, with drought-affected areas dropping to 0%. Conversely, East Egypt and Kenya continue to face persistent challenges, with Egypt showing an increasing trend in affected areas (peaking at 75.7% in 2015) and Kenya exhibiting cyclical patterns.
• All regions, except East Kenya, experienced positive water storage anomalies from 2019 to 2023, indicating a recent trend towards wetter conditions across much of the study area.

Contributions

• Developed a novel framework for high-resolution (approximately 28 km) hydrological drought monitoring in data-scarce regions by integrating GRACE TWS downscaling with advanced spatiotemporal pattern recognition techniques (GPR and JADE).
• Provided the first pixel-level quantification of drought characteristics (magnitude, duration, severity, and recovery dynamics) across the entire Nile Basin countries for the period 2003–2023, offering objective and spatially explicit insights without relying on sparse ground observations.
• Identified and characterized seven distinct, physically meaningful drought hotspots and their spatiotemporal evolution, enhancing understanding of regional drought dynamics.
• Offered a comprehensive country-scale assessment of drought impacts, revealing diverse water security trends and highlighting areas of concern and improvement among the 11 Nile Basin nations.
• Addressed key limitations of GRACE data, such as coarse spatial resolution and mission gaps, through innovative downscaling and reconstruction methodologies.
• Provided a replicable methodology for drought assessment that can be applied to other transboundary river systems globally.

Funding

• National Key R&D Program of China "Joint Research, Development and Application Demonstration of Remote Sensing Monitoring Technology for Typical Natural Resources Features" (Grant No. 2023YFE0207900).
• National Natural Science Foundation of China (Grant No. W2432026) for Vagner Ferreira.
• Australian Research Council (DE230101327) for Christopher Ndehedehe.

Citation

@article{Ge2025Highresolution,
  author = {Ge, Ying and Ferreira, Vagner G. and Xu, Jia and Ndehedehe, Christopher E.},
  title = {High-resolution GRACE-based assessment of hydrological drought patterns and recovery dynamics across the Nile Basin countries (2003–2023)},
  journal = {Journal of Hydrology Regional Studies},
  year = {2025},
  doi = {10.1016/j.ejrh.2025.102904},
  url = {https://doi.org/10.1016/j.ejrh.2025.102904}
}