Boussaadi A. (2022).
Ensemble modelling of the carbon sequestration potential of agricultural soils: a case-study with the model FarmSim. Mémoire (Master 2 MIDAS) : CIHEAM-IAMM, Montpellier (France). 90 p. Master 2 Thesis. Economics. Programme: Mediterranean farming system design for a sustainable food-system [MIDAS]. Co-accreditation University of Montpellier, Institut Agro Montpellier, CIHEAM-IAMM.
Titre : | Ensemble modelling of the carbon sequestration potential of agricultural soils: a case-study with the model FarmSim |
Auteurs : | A. Boussaadi |
Type de document : | Thèse, Mémoire, Master |
Année de publication : | 2022 |
Format : | 90 p. |
Note générale : | Master 2 Thesis. Economics. Programme: Mediterranean farming system design for a sustainable food-system [MIDAS]. Co-accreditation University of Montpellier, Institut Agro Montpellier, CIHEAM-IAMM. |
Langues : | Anglais |
Langues du résumé : | Anglais ; Français |
Catégories : |
Catégories principales 06 - AGRICULTURE. FORÊTS. PÊCHES ; 6.6 - Technique Agricole (sols, engrais, mécanisation)Thésaurus IAMM SEQUESTRATION DU CARBONE ; MODELE DE SIMULATION ; UTILISATION DES TERRES ; ITALIE ; PUGLIA |
Résumé : | Soil organic carbon (SOC) stocks were estimated over a 10-year period (19942004) using the farm-scale biogeochemical model FarmSim, which couples the CERES-EGC arable crop model and the PaSim grassland model. FarmSim was run with spatially explicit input data for 48 pixels of agricultural soils located in a pilot area in southern Italy. The main objective of this work was to support an ensemble modelling approach by assessing whether the outputs given by FarmSim in the year 2004 (end of the simulation period) similar to the observations of the same year by using only generic land-use inputs, beyond soil properties (including SOC in 1994) and daily weather data. The simulations showed some underestimation of SOC by FarmSim, with a CRM (coefficient of residual mass) of 0.39 and RRMSE (relative root mean square error) of ~79%. The lower relative mean absolute error (RMAE) equal to ~53% reflects the inaccuracies associated with one or a few data points (amplified by squared differences). However, the results are likely to be improved as more detailed datasets are provided for model setup and calibration. |
Nature du diplôme : | Mémoire (Master 2 MIDAS) |
Université de soutenance : | CIHEAM-IAMM |
Ville de l'université de soutenance : | Montpellier (France) |
Cote : | Réservé lecteur CIHEAM |
Directeur de Thèse : | Belhouchette H. |
Membres du Jury : | Kleftodimos G.; Belhouchette H.; Bellocchi G.; Martin R. |