Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Doctorat → Pays-Bas → Simulating energy, water and CO2 fluxes at representative desert ecosystems over Central Asia

University of Twente (2015)

Simulating energy, water and CO2 fluxes at representative desert ecosystems over Central Asia

Li Longhui

Titre : Simulating energy, water and CO2 fluxes at representative desert ecosystems over Central Asia

Auteur : Li Longhui

Etablissement de soutenance : University of Twente

Grade : Doctor University of Twente 2015

Résumé
Quantifying the exchanges of energy, water and CO2 between the land surface and the atmosphere is the principal objective of Land Surface Models. Despite considerable advances in LSMs during the past few decades, predictive ability of the LSMs in dry environments remains low and unsatisfactory, and hindered our ability to understand the interactions of climate and vegetation. Therefore, this thesis is aiming at evaluating the performance of the Common Land Model (CLM) in simulating energy, water and CO2 at desert sites in Central Asia. By comparing the results from CLM against the eddy covariance measured surface flux variables, we found that, in general, the CLM is able to satisfactorily reproduce the energy fluxes at three desert sites in Central Asia but net radiation during night-time was systematically overestimated. We confirmed that more realistic representation of vertical distribution of root in soil profile and the root water uptake function has significant improvements on the simulation of water and carbon fluxes at desert sites. However, the performance of the CLM for simulating carbon flux was site-dependent and varied greatly with time scales and the simulated ecosystem respiration was poorly correlated to the observed one. Overall, the CLM was proven to simulate energy fluxes better than carbon fluxes. We concluded that there is large potential to improve the land surface model when applied in Central Asian desert ecosystems. Generating accurate input information on vegetation coverage, compositions, and improving the ability in estimating leaf area index and integrating our latest understanding on the morphological functions into the current land surface models may greatly help advancing LSMs in desert ecosystems.

Version intégrale (ITC)

Page publiée le 7 février 2018