Ecohydrology of a semi-arid forest: effects of precipitation pattern and canopy structure on forest productivity
Abstract
In dry forest ecosystems water is often the main limitation to carbon sequestration and water availability to the trees is further limited by losses to soil evaporation (E). While it is important to assess E, it is also a difficult variable to directly measure, and can be expected to be highly heterogeneous in open-canopies. We investigated the links between the water and carbon cycles in a semi-arid forest in Southern-Israel (40-year-old P. halepensis forest, with LAI of 1.5 and mean annual precipitation of 285 mm). During a four-year research period we measured precipitation (P, event rain gauges), soil water content (TDRs), evapotranspiration (ET, eddy covariance), tree transpiration (T, sap flux), soil evaporation (E, soil chambers) and intercepted precipitation (calculated). These were combined with on-going measurements of net ecosystem CO2 exchange (NEE), radiation budget and meteorological measurements. On average for the study period, ET accounted for 94% of P, varying between 100% when P<250 mm and 85% when P>300 mm (with indications for losses to subsurface flow and soil moisture storage in wetter years). Both T and E were major fluxes in this dry forest: 45% and 36% of ET, respectively. Two main processes influenced the partitioning of ET between E and T: precipitation patterns and canopy structure. The pulsed-storm pattern, characteristic of semi-arid climates, was sufficient to maintain the topsoil layer wet during the whole wet season (November to March), producing high E during high radiation days (up to 0.70 mm d-1 in the wetting and drying seasons). Only infrequent and relatively intensive storms resulted in infiltration to the depth of maximum rooting density (~25cm) and below, increasing water availability to the trees. As a result, two years with similar P (231 and 224 mm) had different ecosystem water use efficiency: 1.06 (gCO2/kgH2O) in a year with few large storms, but only 0.74 in a year with many small storms. The spatial variability in E was large, with SE of ±47% between 14 measurement sites. This variability was correlated with PAR, which was up to 92% higher in sun-exposed compared to shaded sites. In this open canopy system, E fluxes measured in sun-exposed areas were on average double those in shaded areas. There is clearly a tradeoff between increasing water availability to trees by increasing canopy cover (i.e. shade fraction) resulting with reduced E losses on the one hand, and increasing losses to canopy interception and increased demands for T, on the other hand. According to our analysis, optimization between losses to E and maximizing T/P is obtain in our research site at a canopy cover of 65% (400 trees ha-1). Our results also indicated that the expected intensification of the hydrological budget may be beneficial for this system. These results have implications for assessing climate change predictions for the regions (reduced precipitation), and for forest management (optimize tree density).
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2009
- Bibcode:
- 2009AGUFM.B51A0296R
- Keywords:
-
- 0426 BIOGEOSCIENCES / Biosphere/atmosphere interactions;
- 0439 BIOGEOSCIENCES / Ecosystems;
- structure and dynamics;
- 1813 HYDROLOGY / Eco-hydrology;
- 1836 HYDROLOGY / Hydrological cycles and budgets