Numerous catchment studies all over the world indicate that the presence of forests normally decreases runoff (Fig. 3.2, from Nemec and Rodier,1979). At the same time runoff peaks are reduced and dry-season runoff may be increased. The causal background for the differences is to be found in higher evapotranspiration from forests because of their large transpiring leaf surface (which also intercepts much rain, rapidly evaporating after a shower) and longer transpiration period (particularly evident when evergreen forests are compared with arable land). Forests also have more permeable soils than arable land, and many types of grasslands, also often have deeper roots using a larger soil volume for their water supply.

   The actual precipitation in a forest may be higher than in the open, as the rough surface of the forest canopy may capture fog and rain droplets which have a low sedimentation rate and therefore are measured very incompletely in conventional rain gauges. The effect is particularly evident in some mountain forests where 'fog-drip' may form a considerable part of the total precipitation, Also, forest edges may get more precipitation than both open fields and the interior of the forest. This more or less 'occult' precipitation can seldom change the general picture, viz. that afforestation decreases and deforestation increases runoff. Yet it should be kept in mind that precipitation measured in standard gauges does not reflect the heterogeneity in actual amount of water supplied to the ground, and that this heterogeneity is much larger in a forest (stem-flow, drip from branch terminal shoots in some tree species, etc.), and that the deviations from standard meteorological data are considerable in forest edges and generally in mosaic landscapes with both tree groves and grassy patches.

   The high permeability of forest soils for water is connected with the high heterogeneity of both vegetation and soil. Vertical channels formed by large roots may persist long after the death of the root and the tree. Both forests and grasslands may also have channels formed by burrowing animals (earth-worms, moles, and others), but grasslands are more subjected to soil compaction, e.g. by trampling by large herbivores. There is thus less surface runoff from a forest than from open fields. The transport of water through the unsaturated zone, and further with the groundwater gradients, occurs both in forest soils and outside the forest, but this process is much slower than surface runoff. The low proportion of surface runoff and high proportion of ground-water runoff in forests is the main reason why erosion is much lower in forested areas than in areas with high proportions of pastures and, in particular, arable land. A further reason why erosion is less from permanently vegetated areas is the protective effect of roots, both living and recently dead ones (Fig. 3.3, from Swanson et al. , 1982). The living vegetation and the forest litter layer also offer mechanical protection of the soil, when heavy raindrops first hit the vegetation and litter and thereby lose much of their kinetic energy.

   As will be discussed later, changes in river transport of material are among the process in the past which we can measure and relate to human impacts on the ecosystems.