Long-term streamflow relations with riparian gallery forest expansion into tallgrass prairie in the Southern Great Plains, USA

Though the effects of forests on water yields have been studied since the early 20th century, considerable uncertainty remains regarding streamflow relations in landscapes where forestation of degraded grassland is accompanied by recovery of soil hydrological function. This uncertainty arises because trees may reduce subsurface flow by transpiring water inaccessible to shallow-rooted grasses, but forestation of degraded grassland may also improve soil infiltration capacity, thereby increasing soil water recharge, a prerequisite for baseflow. To determine how riparian gallery forest expansion affected streamflow, we classified a chronosequence of historical aerial photography covering a 78km ² watershed equipped with a streamflow gauge. Here we show that in the Council Creek watershed in north-central Oklahoma, the baseflow component of streamflow increased from 1938 to 1992 during which time tree cover increased from 5% to 18% of the watershed area. During this period there were significant increasing trends in precipitation and evapotranspiration, but no significant trends in total streamflow. Stepwise regression indicated that relative to increases in precipitation, increases in forest cover were a better predictor of the observed baseflow increase. Baseflow increases over time were not distributed uniformly throughout the year, but rather occurred only from October through May, during most of which time the primarily deciduous riparian gallery forest species and the herbaceous species that they replaced would have been dormant. Due to confounding factors including elevated precipitation from 1980 to 1992, changing precipitation characteristics, and changing land-use, it was impossible to determine to what extent riparian gallery forest expansion caused the associated baseflow increases. Nonetheless, the method used in the study-relating an historical aerial photography chronosequence to existing long-term stream gauge data-proved to be an inexpensive, efficient, and effective in disproving our hypothesis that riparian gallery forest expansion would reduce baseflows. Further research is needed to better predict how long-term changes in forest cover, the position of forest cover within a watershed, and associated changes in soil and floodplain hydrology will affect streamflow in large mixed land-use watersheds.