Edge effects and forest water use: A field study in a mixed deciduous woodland

A field study was carried out in a mixed deciduous forest in order to measure the spatial variability of evapotranspiration in relation to distance from the nearest forest edge. Throughfall was collected in storage gauges in a transect across the edge. Transpiration was measured at the tree scale by means of the sap flux technique. Thermal dissipation probes were inserted into the hydro-active sapwood of 12–16 sample trees at a time covering four species. The sample trees were located close to a north- and a south-facing forest edge and between 3 and 69 m away from the nearest edge. The probes were moved to new trees about once a month and in total 71 trees were sampled. Sap flux densities were compared with potential evaporation and scaled up to the stand through multiplication with sapwood area per unit ground area. No significant edge effect on interception evaporation could be detected but there was a large influence on stand transpiration which increased towards the edge. In ash (Fraxinus excelsior L.), this increase resulted mainly from enhanced sap flux density (by 33–82%, depending on the size class) in trees located at the edge, whereas in oak (Quercus robur L.) the sap flux density was similar in edge and inner trees and an effect was only found at the stand scale in the way that the total basal area, per unit ground area, was larger near the forest edge than in the forest interior. Hawthorn (Crataegus monogyna L.) and field maple (Acer campestre L.), which occurred mainly in the understorey, were only weakly affected by the proximity to an edge. At the stand scale the total seasonal transpiration varied between 354 mm in the forest interior (>45 m away from the edge) and 565 mm at the forest edge (<15 m away from the edge), whilst the potential evaporation over the same period was 571 mm. This corresponds to Priestley–Taylor coefficients of 0.78 in the interior and 1.25 at the edge, whilst intermediate numbers were found for the area between the edge and inner zones. Using these results to calculate the average water loss per unit ground area of hypothetical woodlands of various sizes, it is shown that the edge effect dominates the water use of small forests and becomes negligible only for woodlands larger than 100 ha.