Abidi et al. (2026) Long-term streamflow reconstruction of The Medjerda River, Tunisia, from tree rings

Identification

• Journal: Journal of Hydrology Regional Studies
• Year: 2026
• Date: 2026-02-20
• Authors: Sahar Abidi, Shulong Yu, Ramzi Touchan, Mehvish Majeed, Ali Khorchani
• DOI: 10.1016/j.ejrh.2026.103262

Research Groups

• Laboratoire Des Ressources Sylvo-Pastorales, Institut Sylvo-Pastorale de Tabarka, Universit´e de Jendouba, Tabarka, Tunisia
• Laboratory of Tree Ring Research, University of Arizona, Tucson, AZ, USA
• Department of Geography, University of Nevada, Reno, NV, USA
• University of Carthage, National Research Institute of Rural Engineering, Water, and Forestry, INRGREF, Laboratory of Management and Valorization of Forest Resources, Ariana, Tunisia

Short Summary

This study presents the first tree-ring-based reconstruction of Medjerda River discharge in Tunisia, extending the natural flow record from 1876 to 2009 CE using seven Pinus halepensis chronologies and a hydrological model. The reconstruction explains 54 % of discharge variability and provides crucial long-term context for hydroclimatic extremes, aiding water resource management in this semi-arid region.

Objective

• To reconstruct the natural-flow discharge of the Medjerda River, Tunisia, from 1876 to 2009 CE using tree-ring data and a web-based dendrohydrology platform.
• To compare recent hydrological extremes in the Medjerda River Basin with past droughts and pluvial events.
• To assess the reliability of existing tree-ring chronologies from Tunisia and Algeria in capturing the hydroclimatic signal of the Medjerda River.
• To identify limitations of the current tree-ring network for Medjerda Basin hydrology and evaluate the potential for developing longer streamflow reconstructions to inform future climate resilience strategies.

Study Configuration

• Spatial Scale: Medjerda River Basin, upstream of the Sidi Salem Big Dam, near the Bou Salem gauge, Tunisia. The drainage area is approximately 17,811 km². Tree-ring chronologies are from Tunisia and Algeria, spanning latitudes from 35.13°N to 37.25°N and longitudes from 6.47°E to 9.92°E. Climate data are at a 0.5° x 0.5° spatial resolution.
• Temporal Scale:
  • Reconstruction period: 1876–2009 CE (134 years).
  • Main calibration period: 1934–2009 CE.
  • Observed Water Balance Model (WBM) discharge: 1903–2021 CE.
  • Observed gauged discharge: 1956–2009 CE.
  • Tree-ring chronologies: Start years range from 1684 to 1934 CE, end years from 2002 to 2013 CE.

Methodology and Data

• Models used:
  • Water Balance Model (WBM): A global, modular, grid-based watershed model simulating major land surface water cycle components (evapotranspiration, soil moisture, snow storage, runoff, river discharge) at a daily time step, driven by monthly climate input.
  • TRISH (Tree-Ring Integrated System for Hydrology): A web-based dendrohydrology platform that incorporates the WBM.
  • Seascorr: A MATLAB program used to identify seasonal climate signals in tree-ring series.
  • Statistical methods: Two-stage stepwise regression, Principal Component Analysis (PCA) on the covariance matrix, cross-validation (leave-9-out), split-sample validation, Lilliefors test for normality, Durbin-Watson test for autocorrelation, Breusch-Pagan statistics for heteroscedasticity, and Mann-Kendall test for trend. Nash-Sutcliffe statistic for independent verification.
  • Software: R software program ReconAnalog, dplR for chronology standardization, COFECHA for cross-dating.
• Data sources:
  • Tree-ring data: Seven Pinus halepensis chronologies (selected from an initial network of 15) from Tunisia and Algeria, obtained from the International Tree-Ring Data Bank (ITRDB).
  • Climate data: CRU TS v4.05 dataset (Climatic Research Unit) for monthly precipitation and temperature, at a 0.5° x 0.5° spatial resolution.
  • Hydrological data:
    • Modelled natural flow discharge (WBM discharge) for the Medjerda River (1903–2021 CE).
    • Observed gauged Medjerda River flows at the Bou Salem station (1956–2009 CE).

Main Results

• The primary reconstruction (Q1), calibrated over 1934–2009 CE, explains 54 % of the observed WBM discharge variability.
• Modelled natural flow (WBM discharge) is consistently higher than gauged discharge (mean 125.3 m³/s vs 17.6 m³/s over 1956–2009 CE), but the two annual time series are highly correlated (r = 0.67).
• Pinus halepensis tree growth is predominantly influenced by precipitation accumulated over the entire water year (October-September), with strong positive correlations (e.g., r = 0.76 for a 12-month period ending in September for one chronology).
• The reconstruction extends the Medjerda River discharge record from 1876 to 2009 CE, providing 134 years of annual estimates, more than doubling the instrumental record.
• The most severe single-year drought (2002 CE, 3 % of normal discharge) and flood (2004 CE, 263 % of normal discharge) in the 134-year record occurred in recent decades.
• Prominent multi-year droughts identified include 1876–1880, the 1940s (especially 1945–1947), and 1999–2002.
• Multi-decadal variability shows high-amplitude swings (e.g., late 1870s drought to 1880s wet period) and a trend toward greater variability in recent decades.
• The reconstruction is highly synchronous (r = 0.90 for unsmoothed series, >0.74 for 30-year sliding window) with a previous October-June precipitation reconstruction for Tunisia, confirming the regional hydroclimatic relevance of the tree-ring network.

Contributions

• Presents the first tree-ring-based reconstruction of Medjerda River discharge in Tunisia, significantly extending the natural flow record from 1876 to 2009 CE.
• Demonstrates a novel approach by combining a comprehensive regional tree-ring network with a web-based hydrological modeling platform (TRISH/WBM) for streamflow reconstruction in data-scarce semi-arid regions.
• Provides critical long-term context for recent hydrological extremes, revealing that the most severe single-year drought and flood in 134 years occurred in the early 2000s.
• Highlights the necessity of considering longer time horizons and multi-decadal variability for robust water resource planning and climate adaptation strategies in the Medjerda Basin.
• Identifies limitations of the current tree-ring network and proposes future research directions, including expanding the network, particularly in Algerian headwaters, and collecting remnant wood for multi-century reconstructions.

Funding

• Agnese N. Haury Visiting Scholar & Trainee Fellowship (supported Sahar Abidi at the University of Arizona's Laboratory of Tree-Ring Research).

Citation

@article{Abidi2026Longterm,
  author = {Abidi, Sahar and Yu, Shulong and Touchan, Ramzi and Majeed, Mehvish and Khorchani, Ali},
  title = {Long-term streamflow reconstruction of The Medjerda River, Tunisia, from tree rings},
  journal = {Journal of Hydrology Regional Studies},
  year = {2026},
  doi = {10.1016/j.ejrh.2026.103262},
  url = {https://doi.org/10.1016/j.ejrh.2026.103262}
}