Multicriteria evaluation of the AquaCrop crop model in a hilly rainfed Mediterranean agrosystem

Mariem Dhouib (1) , Rim Zitouna-Chebbi (2) , Laurent Prevot (1) , J. Molénat (1) , Insaf Mekki (2) , Frédéric Jacob (1)

1 UMR LISAH - Laboratoire d'étude des Interactions Sol - Agrosystème - Hydrosystème
2 INRGREF - Institut National de Recherche en Génie Rural Eaux et Forêts

Exploring crop spatial organizations within landscapes is a promising solution for agroecological transitions and climate change adaptation in Mediterranean rainfed hilly agrosystems. A prerequisite is to ensure that crop models can simulate a range of agrohydrological processes in such agrosystems. The current study deepened the evaluation of the AquaCrop model by conducting a multicriteria evaluation (canopy cover CC, dry aboveground biomass AGB, actual evapotranspiration ETa, runoff R, soil water content SWC) for a range of crop and soil combinations, and for contrasted hydroclimatic years in northeastern Tunisia. The data were collected in the Kamech catchment (OMERE Observatory) during nine measurement campaigns on predominant soils and crops. AquaCrop simulations were based on field observations and parameters from the literature. AquaCrop could simulate plant dynamics and water fluxes for contrasted hydroclimatic years, with a slight dependence on soil class and a significant dependence on crop type. Model simulations were of moderate quality for CC (R2 of 0.45, RMSE of 0.24 on average) and of acceptable quality for AGB (R2 of 0.81, RMSE of 0.85 ton ha− 1 on average). AquaCrop acceptably simulated water transfer across the soil–plant continuum (R2 of 0.62, RMSE of 0.77 mm day− 1 on average for ETa; R2 of 0.68, RMSE of 0.75 mm day− 1 on average for R; R2 of 0.86, RMSE of 27.4 mm on average for SWC). The model performances were satisfactory for most cases, with p values larger than 5 % for the Student’s t test on linear regressions of validation. Our results were similar to those reported in previous studies over flat terrain, including delayed senescence by model simulations with subsequent overestimation of CC and AGB observations. Additionally, soil cracks likely altered the AquaCrop ability to simulate runoff. Despite these limitations, our results support the application of AquaCrop to evaluate water productivity in hilly agrosystems.

Mots clés: AquaCrop model; Rainfed agrosystems; Hilly terrains; Multicriteria evaluation; Mediterranean

Domaines: Agronomie

Mariem Dhouib, Rim Zitouna-Chebbi, Laurent Prevot, J. Molénat, Insaf Mekki, et al.. Multicriteria evaluation of the AquaCrop crop model in a hilly rainfed Mediterranean agrosystem. Agricultural Water Management, 2022, 273, pp.107912. ⟨10.1016/j.agwat.2022.107912⟩. ⟨hal-03771403⟩