The impact of Land Surface Model vegetation parameterization on the terrestrial water and carbon cycles

• Main Author: Stevens, David Pierre Ann
• Publication Date: 2022
• Language: eng
• Source: Repositórios Científicos de Acesso Aberto de Portugal (RCAAP)
• Download full: http://hdl.handle.net/10451/59851
• topic: Modelos de Superfície Terrestre (MST); Parametrização da vegetação; Observação da terra; Benchmarking; Machine learning; Land surface model; Vegetation parameterization; Earth observation; Domínio/Área Científica::Ciências Naturais::Ciências da Terra e do Ambiente

The surface-atmosphere turbulent exchanges couple the water, energy, and carbon cycles in the earth system. The biosphere plays an important role in the evaporation process, and vegetation related parameters such as the leaf area index (LAI), vertical root distribution and stomatal resistance are poorly constrained due to sparse observations at the spatial-temporal scales at which land surface models (LSMs) operate. Considering the central role of the land surface vegetation in the climate system, various networks and methods of observational data and processes are used to constrain different ECMWF (European Center for Medium-Range Weather Forecasts) model configurations to better understand underlying conditions of the land surface energy, water and carbon budgets. Observed data from the FLUXNET network are used to perform offline point simulations with a strong emphasis on the representation of evaporation and its link with water stress conditions. The close relationship between LAI and the minimum canopy resistance is investigated and shows some model performance improvements potential but failed to solve other issues such as excessive evaporative drought conditions. However, the replacement of the exponential roots profile by a uniform roots distribution and associated maximum rooting depth reduced the underestimation of evaporation during water stress conditions. This result highlights the importance of root distribution in controlling soil moisture resistance in water stress conditions. The revised uniform root profile also has a positive effect on the model carbon cycle representation and brings model output closer to observations. Moreover, the positive effect of this new root scheme on model performance is intensified when coupled with the A-gs photosynthesis-conductance scheme as opposed to the traditionally used Jarvis approach, which implies that a more physiologically based model vegetation parameterization could benefit land surface models performance. The same findings also apply at the grid scale, where the replacement of the current vegetation types, cover, characteristics and LAI, by a new high resolution remote sensing derived global vegetation dataset did not significantly affect the simulated water budgets. However, the deeper and uniform root scheme allows plants to use soil water more efficiently and induce a more intense and positive response from land surface models. Even if an increase in model resolution could improve the simulations, the core of LSMs discrepancies reflect parameterization uncertainty to represent and capture smaller-scale unresolved processes and integrate them to the model grid-scale.