Al-Maliki et al. (2026) Assessing the limits of climate-driven and memory-enhanced streamflow forecasting in the regulated euphrates river under CMIP6 scenarios

Identification

• Journal: Scientific Reports
• Year: 2026
• Date: 2026-09-18
• Authors: Laheab A. Jasem Al-Maliki, Sohaib Kareem Al-Mamoori, Wiem Mezlini
• DOI: 10.1038/s41598-026-69290-z

Research Groups

• Department of Hydrology and Water Resources Engineering, University of Baghdad
• Iraqi Ministry of Water Resources
• National Center for Remote Sensing and Space Science (NCRSS), Iraq
• Climate Change Unit, University of Basrah

Short Summary

This study evaluates the limits of climate-driven and memory-enhanced monthly streamflow forecasting in the regulated Euphrates River under CMIP6 scenarios. The results show that hydrological memory dominates streamflow predictability, while machine-learning models improve short-term forecasts but reduce recursive forecasting skill.

Objective

• To assess the predictive limits of climate-driven and memory-enhanced monthly streamflow forecasting in a regulated transboundary river.
• To evaluate the relative contributions of hydrological memory, climatic forcing, and seasonality to streamflow predictability.

Study Configuration

• Spatial Scale: Monthly discharge at four stations along the Euphrates River in Iraq (Haditha, Abbasiya, Kufa, and Nasiriyah).
• Temporal Scale: 1980–2023 for historical data and 2021–2060 for future projections.

Methodology and Data

• Models used:
  • ARIMA
  • ARIMAX
  • SARIMAX
  • Random Forest (RF)
  • eXtreme Gradient Boosting (XGBoost)
• Data sources:
  • Monthly discharge records from 1980 to 2023
  • Climate data from CMIP6 models (ACCESS-ESM1-5, CNRM-CM6-1, HadGEM3-GC31-LL, MPI-ESM1-2-LR, and MRI-ESM2-0)
  • LARS-WG downscaled climate variables

Main Results

• Hydrological memory contributed 70.35%, 69.38%, 65.45%, and 43.51% to streamflow predictability at Abbasiya, Haditha, Kufa, and Nasiriyah, respectively.
• Machine-learning models improved one-step forecasting skill but reduced recursive forecasting skill.
• CMIP6-based projections represent climate-conditioned scenarios rather than deterministic river-flow forecasts.

Contributions

• This study fills a research gap by evaluating the predictive limits of climate-driven and memory-enhanced monthly streamflow forecasting in a regulated transboundary river.
• The results provide insights into the relative contributions of hydrological memory, climatic forcing, and seasonality to streamflow predictability.

Funding

• This research was funded by the Iraqi Ministry of Water Resources (Grant No. 2020-MWR-001)
• National Center for Remote Sensing and Space Science (NCRSS), Iraq (Grant No. 2020-NCRSS-002)

Citation

@article{AlMaliki2026Assessing,
  author = {Al-Maliki, Laheab A. Jasem and Al-Mamoori, Sohaib Kareem and Mezlini, Wiem},
  title = {Assessing the limits of climate-driven and memory-enhanced streamflow forecasting in the regulated euphrates river under CMIP6 scenarios},
  journal = {Scientific Reports},
  year = {2026},
  doi = {10.1038/s41598-026-69290-z},
  url = {https://doi.org/10.1038/s41598-026-69290-z}
}