Dynamics and controls of ecosystem scale water, carbon, and energy cycling at semiarid grassland and shrubland

Woody plants are invasive to native grasslands in arid and semiarid areas worldwide. Substantial changes in the water, carbon, and energy cycles are expected to accompany these woody invasions. At the Sevilleta National Wildlife Refuge in central New Mexico, the native grassland site is only about 2 km from the shrubland site. Because these sites are close together, differences in precipitation, incoming solar radiation, and other hydroclimatological factors are minimized. This allows for direct comparison of temporal variations in water, carbon, and energy, and the factors that control these variations between shrubland and grassland. Multi-year observations of ecosystem scale flux measurements are made using either the Bowen ratio method or the eddy covariance method, each accompanied by measurements of precipitation, soil moisture and radiation.; The research presented here falls into three parts. First, we measured temporal fluctuations of evapotranspiration (ET) and identified key sources of the observed variability as a means to understand the coupled water and energy cycles in both semiarid grassland and shrubland. We show that timescales involved with the dynamics of ET are short, on the order of three days, largely because ET is strongly correlated with surface soil moisture. Second, having identified soil moisture distribution as the key link between hydrologic and ecologic processes in semiarid grassland and shrubland, we used ecosystem scale flux measurements to demonstrate how soil moisture influences respiration and assimilation. We show that timescales associated with assimilation are on the order of months and that assimilation is correlated with deep soil moisture. Additionally we show that timescales related to respiration just after rainstorms are short, which suggests a surface soil moisture control. Third, using our newly acquired knowledge of the relationships between soil moisture and the water and carbon cycles, we improved a well known simple hydro-ecological model based on vertically averaged root zone soil moisture by dividing the root zone into two layers: a shallow surface zone and a deep zone. Although the root zone model generates a reasonable assimilation time series, ET and respiration are better modeled with the addition of a separate surface soil moisture reservoir.