Gaglo et al. (2025) Sensitivity of a Sahelian groundwater-based agroforestry system to tree density and water availability using the land surface model ORCHIDEE (r7949)

Identification

• Journal: Geoscientific model development
• Year: 2025
• Date: 2025-12-04
• Authors: Espoir Koudjo Gaglo, Émeline Chaste, Sebastiaan Luyssaert, Olivier Roupsard, Christophe Jourdan, Sidy Sow, Nicolas Vandewalle, Frédéric Dô, Daouda Ngom, Aude Valade
• DOI: 10.5194/gmd-18-9541-2025

Research Groups

• Université Cheikh Anta Diop, Dakar, Senegal
• LMI IESOL, IRD, ISRA, Bel Air, Dakar, Senegal
• Eco&Sols, University Montpellier, CIRAD, INRAE, Institut Agro, IRD, Montpellier, France
• A-LIFE, Vrije Universiteit Amsterdam, Amsterdam, Netherlands
• CIRAD, UMR Eco&Sols, Dakar, Senegal
• Université Gaston Berger, Saint-Louis, Senegal
• Earth and Life Institute, Faculty of Bioscience Engineering, Catholic University of Louvain, Louvain-la-Neuve, Belgium
• Terra Teaching and Research Center, University of Liège, Gembloux Agro-Bio Tech, Gembloux, Belgium
• IRD, UMR Eco & Sols, Univ Montpellier, IRD, CIRAD, INRAE, Montpellier, France

Short Summary

This study developed and evaluated a new configuration of the ORCHIDEE land surface model to simulate a Sahelian groundwater-based agroforestry system, revealing that increased tree density enhances carbon sequestration but reduces crop yield, and that interannual water variability differentially impacts tree and crop productivity.

Objective

• Evaluate the performance of a newly configured ORCHIDEE land surface model for simulating Faidherbia albida agroforestry ecosystems.
• Assess the sensitivity of tree and crop productivity, along with carbon and energy fluxes, to variations in tree density and water availability (rainfall and groundwater capillary fringe soil water content).

Study Configuration

• Spatial Scale: Plot scale (Niakhar agroforestry site, Senegal).
• Temporal Scale: Model spin-up and transient runs from 1860 to 2017; calibration (2018–2020) and evaluation (2021–2023) using half-hourly to daily observations; sensitivity analyses over 3- to 6-year periods.

Methodology and Data

• Models used: ORCHIDEE Land Surface Model (revision 7949), specifically modified to simulate Faidherbia albida growth, inverted phenology, deep root water uptake from groundwater, and adjusted photosynthesis/carbon allocation parameters for both trees and associated crops.
• Data sources:
  • Observation: Half-hourly eddy covariance measurements (sensible heat, latent heat, CO2), meteorological variables (air temperature, relative humidity, wind speed, atmospheric pressure, rainfall, net radiation), leaf area index (LAI) for trees and crops, soil humidity at nine depths, groundwater fluctuations, and soil water content in the capillary fringe of the groundwater table (SWCC) from the "Faidherbia-Flux" observatory in Niakhar, Senegal (2018–2023). Photosynthetic parameters (Vcmax and Jmax) were also measured.
  • Reanalysis/Gridded: CRUJRA v2.2.2 (6-hourly, 0.5° resolution) for climate forcing (1901–2020) and ERA Interim (3-hourly, 0.7° resolution) for gap-filling of meteorological data.
  • Ancillary: Site-specific soil texture and hydrological parameters from previous studies.

Main Results

• Model Evaluation: The model reproduced the Faidherbia albida reverse phenology (dry season growth) with a tree LAI RMSE of 0.15 and crop LAI RMSE of 0.26, though interannual variability was underestimated for both. Daily tree GPP peak (4.75 gC m−2 d−1) matched observed (4.74 gC m−2 d−1), but annual tree GPP was underestimated by 2 % to 22 %. Daily crop GPP peak (13.41 gC m−2 d−1) was lower than observed (15.76 gC m−2 d−1), with annual crop GPP underestimated by 9 % to 19 %. Ecosystem GPP was underestimated by 6 % to 18 %. Energy fluxes (LE, H, Rn) showed reasonable fit, with RMSEs of 2.12 MJ m−2 d−1, 2.40 MJ m−2 d−1, and 2.31 MJ m−2 d−1, respectively, but tended to overestimate low values and dry season net radiation.
• Sensitivity to Tree Density:
  • Annual ecosystem GPP significantly increased with tree density, from 9.91 ± 0.64 tC ha−1 yr−1 at 0 trees ha−1 to 14.97 ± 1.32 tC ha−1 yr−1 at 26 trees ha−1.
  • Above-ground harvested crop biomass decreased with increasing tree density, from 3.33 ± 0.26 tC ha−1 yr−1 at 0 trees ha−1 to 2.08 ± 0.34 tC ha−1 yr−1 at 26 trees ha−1, highlighting a trade-off between carbon sequestration and crop yield.
  • Dry season latent heat flux (LE) increased with tree density (0.12 ± 0.07 GJ m−2 yr−1 to 0.85 ± 0.12 GJ m−2 yr−1), while rainy season LE decreased. Sensible heat flux (H) showed inverse trends.
• Sensitivity to Interannual Water Variability:
  • Tree GPP showed limited responsiveness to interannual SWCC variability, with a slight shift in growing period resulting in an overall decrease of less than 10 gC m−2 yr−1 (approx. 2%) during dry seasons.
  • Crop GPP was highly sensitive to rainfall timing and distribution, with early-season rainfall surpluses buffering the effects of late-season deficits, leading to GPP increases of 11 % to 14 % in some years despite overall deficits.
  • Ecosystem GPP anomalies ranged from an 8 % decrease to a 14 % increase depending on the combined variability of rainfall and SWCC. Energy flux anomalies were amplified when both rainfall and SWCC varied.

Contributions

• Developed and evaluated a novel configuration of the ORCHIDEE land surface model to explicitly represent the phreatophytic behavior and reverse phenology of Faidherbia albida in Sahelian agroforestry systems, including groundwater dependence.
• Quantified the trade-offs between carbon sequestration and crop yield in Faidherbia albida parklands across varying tree densities.
• Demonstrated the differential sensitivity of tree and crop carbon and energy fluxes to the timing and interannual variability of rainfall and groundwater availability in a semi-arid agroforestry context.
• Highlighted the critical need for improved representation of deep-rooted phreatophytic species, groundwater dynamics, and belowground processes in land surface models for accurate simulations in semi-arid ecosystems.

Funding

• DM-TropAFS project (ANR, French National Research Agency) under "Investissements d’avenir" programme, reference ANR-10-LABX-001-01 Labex Agro, coordinated by Agropolis Fondation.
• High-Performance Computing (HPC) resources by Grand Équipement National de Calcul Intensif – Très Grand Centre de Calcul (GENCi–TGCC) under grant 2024-06328.
• German Academic Exchange Service (DAAD) for a scholarship (In Region Scholarship Programme – CERAAS Senegal/57504744).
• CIRAD for Action Incitative (AI) Fellowship.
• EU Eranet LEAP Agri-215 RAMSES II (EU) grant no. 215.
• EU DeSIRA CASSECS (grant no. FOOD/2019/410-169).
• EU H2020 SUSTAIN-SAHEL (grant No. 861974).
• EU HORIZON EUROPE GALILEO (grant agreement no. 101181623).
• ANR and France 2030 program PEPR FairCarboN RIFT project (ANR-22-PEXF-0004).

Citation

@article{Gaglo2025Sensitivity,
  author = {Gaglo, Espoir Koudjo and Chaste, Émeline and Luyssaert, Sebastiaan and Roupsard, Olivier and Jourdan, Christophe and Sow, Sidy and Vandewalle, Nicolas and Dô, Frédéric and Ngom, Daouda and Valade, Aude},
  title = {Sensitivity of a Sahelian groundwater-based agroforestry system to tree density and water availability using the land surface model ORCHIDEE (r7949)},
  journal = {Geoscientific model development},
  year = {2025},
  doi = {10.5194/gmd-18-9541-2025},
  url = {https://doi.org/10.5194/gmd-18-9541-2025}
}