Is the annual maximum leaf area index an important driver of water fluxes simulated by a land surface model in temperate forests?

In land surface models, vegetation is often described using plant functional types (PFTs), a classification that aggregates plant species into a few groups based on similar characteristics. Within-PFT variability of these characteristics can introduce considerable uncertainty in the simulation of water fluxes in forests. Our objectives were to (i) compare the variability of the annual maximum leaf area index (LAI(max)) within and between PFTs and (ii) assess whether this variability leads to significant differences in simulated water fluxes at a regional scale. We classified our study region in southwestern Quebec (Canada) into three PFTs (evergreen needleleaf, deciduous broadleaf, and mixed forests) and characterized LAI(max) using remotely sensed MODIS-LAI data. We simulated water fluxes with the Canadian Land Surface Scheme (CLASS) and performed a sensitivity analysis. We found that within-PFT variability of LAI(max) was 1.7 times more important than variability between PFTs, with similar mean values for the two dominant PFTs, deciduous broadleaf forests (6.6 m(2).m(-2)) and mixed forests (6.3 m(2).m(-2)). In CLASS, varying LAI(max) within the observed range of values (4.0-7.5 m(2).m(-2)) led to changes of less than 2% in mean evapotranspiration. Overall, LAI(max) is likely not an important driver of the spatial variability of water fluxes at the regional level.