Response of leaf and whole-tree canopy conductance to wet conditions within a mature premontane tropical forest in Costa Rica

Tropical water recycling and carbon storage are dependent on canopy-atmosphere dynamics, which are substantially altered when rainfall occurs. However, models only indirectly consider leaf wetness as a driving factor for carbon and water fluxes. To better understand how leaf wetness condition affects stomatal and canopy conductance to water vapor, we tested a set of widely used models for a mature tropical forest of Costa Rica with prolonged periods of wet leaves. We relied on a year of sap flux measurements from 26 trees to estimate transpiration (Ec) and multiple micrometeorological profile measurements from a 40-m tower to be used in the models. Stomatal conductance (gs) models included those proposed by Jones (1992) (gs-J), using shaded and sunlit leaf temperatures, and Monteith and Unsworth (1990) (gs-MU), using air temperature. Canopy conductance (gc) models included those proposed by McNaughton and Jarvis (1983) (gc-MJ) and Penman-Monteith (gc-PM). Between gs and gc, gc had the largest differences within models during dry periods; while estimates were most similar during wet periods. Yet, all gc and gs estimates on wet days were at least as high as on dry days, indicative of their insensitivity to leaf wetness. Shaded leaf gs averaged 26% higher than in sunlit leaves. Additionally, the highly decoupled interface (Ω>0.90) reflected multiple environmental drivers that may influence conductance (e.g. vapor pressure deficit and leaf temperature). This was also seen through large shifts of diurnal peaks of gs and gc (up to 2 hours earlier than Ec) associated with the daily variation of air temperature and net radiation. Overall, this study led to three major insights: 1) gc and gs cannot accurately be predicted under wet conditions without accounting for leaf wetness, 2) even during dry days, low vapor pressure deficits interfere with model accuracy, and 3) intermittent rain during semi-dry and wet days cause large fluctuations in gc and gs estimates. Thus, it is advised that sub-daily scale (5- or 10-min intervals) and direct physiological measurements of conductance under wet conditions should be adopted. While methodologically challenging, improved estimates of conductance of water vapor at leaf-to-canopy scales are critical for improving the mechanistic understanding of plant water fluxes in wet environments.