Water Vapor Exchange in a Costa Rican Lower Montane Tropical Forest

Abstract:

Because of high canopy interception in tropical forests, evaporation from wet canopy surfaces makes up a sizeable portion of the total water vapor flux. The modeling complexities presented by changing canopy wetness, along with a scarcity of land-atmosphere flux exchange data from tropical forests, means evapotranspiration (ET) processes have been poorly represented in the tropics in land-surface modeling schemes. To better understand tropical forest ET, we will evaluate the influence of canopy wetness and various micrometeorological data on ET partitioning and total ET flux.

We have collected flux data from a lower montane forest in Costa Rica at a newly established AmeriFlux site, which notably has the highest mean annual precipitation of any site in the network. The site features a 39-m canopy tower, equipped with two eddy covariance systems (LI-7200, LI-COR), a CO₂/H₂O atmospheric profile system (AP200, Campbell Scientific), leaf wetness sensors (LWS, Decagon Devices), sap flow sensors, and a soil respiration chamber (LI-8100A, LI-COR) as well as an array of other micrometeorological sensors.

At the site, total ET is driven primarily by available energy, and to a lesser extent, by vapor pressure deficit. Average daily latent energy fluxes peak at values of 160, 75, and 35 W m^-2 for dry, partially wet, and wet canopy conditions respectively. Correlations between latent energy flux and all other variables are strongest for drier canopy conditions. Complex relationships between canopy wetness and tropical forest ET cause the environmental controls on these fluxes to be significantly different from those in other biomes. As a result, a new modeling paradigm is needed to more accurately model ET differences between tropical forests and other vegetation types.