Variations in Soil Evaporation and its Contribution to the Water Balance of a Semi-arid Forest

Variability of components and processes is present in all forests but is enhanced at dry ones due to the contrasting differences between shaded and exposed areas and the large amplitudes of diurnal, seasonal and inter-annual weather conditions. In such ecosystems, soil evaporation (E) is expected to have large variability and account for a significant part of the total evapotranspiration flux (ET), therefore important for the understanding of processes such as water use, respiration and ecosystem productivity. We report on a four years study (2003-2007) in the Yatir semi-arid pine forest in Southern Israel (40 years old P. halepensis; LAI=1.5; mean precipitation 280 mm/yr). E was directly measured using a modified soil respiration chamber (LI-COR) on 14 permanent soil collars in sites covering the forest spatial variability, carried out on a weekly basis during the research period. Field-averaged E was compared to eddy flux measurements of ET, heat-pulse measurements of tree transpiration (T) and environmental measurements (precipitation, P; soil water content, SWC; radiation, measured as PAR). Our results showed large spatial variability in E (daily SD between sites ~47%), with fluxes measured at exposed areas doubled that of shaded areas, on average. Variability was found to be correlated with PAR (up to 92% higher in exposed compared to shaded sites) and SWC (which was higher in exposed areas during the wetting seasons but higher in shaded areas during the drying seasons). The proportion of forest floor shaded fraction was shown to be a function of tree height, canopy width and tree density. Differential seasonal patterns in E and T were observed: E peaked twice - during early and late winter (up to 0.80 mm/day) and T peaked in spring (up to 1.20 mm/d; coinciding with maximum SWC in the root zone). Low E (0.10 mm/day) was measured during mid winter (max daily temp below 15°) and low E and T in summer (SWC below 10%). The E/ET ratio varied seasonally and on an annual basis E accounted for 44% of ET (102±8 mm). The results indicated that E was a significant part of the hydrological budget in this forest. Simulations based on quantitative relations between E and canopy structure developed here could provide a simple predictive/management tool to optimize tree water use and efficiency.