Tree transpiration well simulated by the Canadian Land Surface Scheme (CLASS) but not during drought

Abstract

Transpiration, a key component of the hydrological cycle, contributes greatly to the climate system by transferring large amount of water from soils to the atmosphere. Its correct representation within Land Surface Schemes in climate models is crucial to provide accurate and reliable climate projections. In this study, transpiration simulated by the Canadian Land Surface Scheme (CLASS) was compared to long-term observations of sap flow measurements in two boreal forest sites of eastern Canada dominated by balsam fir and black spruce. In general, CLASS adequately models daily transpiration during the growing season for most of the years at both sites. During the tree rehydration period (preceding the growing season), modeled transpiration was greatly underestimated because of overestimating the duration of the snowpack, the latter restricting transpiration. Moreover, CLASS did not capture the impact of extreme events on tree physiology and maintained high transpiration rates during a heat stress and a drought. During both observed and simulated drought events, transpiration modeled using CLASS was overestimated, due to insensitivity to substantial decreases in soil water content; modeled transpiration being strictly controlled by atmospheric variables (vapour pressure deficit and radiations). Thus, we also proposed and implemented a new equation that was able to increase the sensitivity of CLASS to decreasing soil water content. However, this equation needs to be further tested on different sites and tree species.