The role of scattered trees in soil water dynamics of pastures and agricultural lands in the Central American Tropics

Sammanfattning: Trees affect the water cycle, and thus the amount of groundwater and surface water, through their effects upon the local microclimate and edaphic properties; in particular through rainfall interception, evapotranspiration, and infiltration. Empirical data about the effects of scattered trees upon hydrological processes have been largely lacking, especially for tropical regions. Although benefits of scattered trees for poverty alleviation and for ecosystem services like carbon sequestration, soil enrichment, biodiversity conservation, and air and water quality have been acknowledged, a balance is needed between conservation and restoration of scattered trees in agricultural landscapes. The overall aim of this thesis is to contribute to the understanding of the role of low-density stands of trees upon soil water dynamics. The general hypothesis was that trees improve the soil water dynamics of degraded soils significantly by increasing soil water contents, infiltrability, and preferential water flows, but that those benefits are not as pronounced in locations that have soils whose inherently good properties are maintained through appropriate management. I also tested whether trees and grasses in pastoral landscapes draw water from different sources, and whether their proportions of use from each source change seasonally. Furthermore, as an evaluation of an alternative tool for research, I explored a new conceptual model relating the effect of vegetation cover on the spatial variability of line conditioned excess (lc-excess) of water stable isotopes and soil water content (SWC). To make that evaluation, a combination of measurements were taken in two contrasting study locations: an agroforestry coffee farm in Central Costa Rica (assumed to be in good physical condition), and a pasture landscape with scattered trees in Copan, Honduras (assumed to be degraded). Measurements included soil infiltrability, SWC, preferential flow, and water stable isotopes. Results showed that trees induced higher infiltrability and preferential flow in the agroecosystem with soil degradation (the pasture), but did not affect infiltrability in the coffee agroforestry system. In the latter, soil moisture was lower under trees than underneath coffee due to the trees’ greater transpiration. During the dry season, preferential flow was greater under coffee shrubs than under neighboring trees. The relationship between soil moisture and spatial variation of lc-excess depended on vegetation type and season. Therefore, the conceptual model gave insights into the dominating processes affected by the vegetation during dry and wet seasons. In the pasture with scattered trees, I found a vertical partitioning of soil water between trees and grasses: (1) in the dry season, trees use groundwater preferentially, while grasses used subsurface water without reaching groundwater; and (2) in the wet season, both trees and grasses use soil surface water. In conclusion, my research showed that when soils are prone to degradation, trees may provide benefits in the form of infiltrability and preferential flow that are commonly neglected in hydrological models used currently for policy decisions. However, such benefits are absent when soils are inherently stable. I also found that when water becomes limiting in the surface soils, trees and grasses partition their water use between the subsoil and groundwater, thereby reducing their competition.