The differential response of transpiration and bare-soil evaporation to precipitation in a Chihuahuan Desert shrubland

In general the coupled response of bare soil evaporation and plant transpiration following precipitation pulses is not well understood in most ecosystems. To better understand how precipitation is partitioned into evaporation and transpiration following natural precipitation events we measured whole plant transpiration and evapotranspiration at a semiarid Chihuahuan desert shrubland site in southeastern Arizona. Whole plant transpiration was measured using the heat balance sap flow technique, while evapotranspiration was measured using both the eddy covariance and the Bowen ratio technique. We note that eddy covariance and bowen ratio estimates of turbulent heat fluxes and net ecosystem carbon dioxide differed considerably. To estimate ecosystem-scale transpiration, sap flow measurements were scaled up using a survey of stem density for the site. During dry periods when evaporation would have been minimal we found good agreement between estimates of evapotranspiration and sap flow estimates of transpiration further validating our scaling approach. Before the monsoon season began, transpiration was near zero as was evaporation. At the start of the mid-summer rainy season, evapotranspiration was dominated by evaporation as it took approximately ten days for the plants to respond to the precipitation input. Following this initial period, periods immediately following rain events (< 2 days) were dominated by evaporation while interstorm periods were dominated by transpiration. The surface of the coarse, well-drained soils dried quickly which rapidly reduced evaporation. Our data suggest that the partitioning of precipitation inputs in this ecosystem is very dynamic and would be missed by simply measuring whole ecosystem evapotranspiration alone.