Use of Water and Chemical Budgets to Quantify Interbasin Groundwater Transfer in a Lowland Rainforest

Previous water quality data strongly suggested the presence of interbasin groundwater transfer (subsurface seepage beneath topographic boundaries) into lowland rainforest watersheds at our study site in Costa Rica. This hypothesis was tested by using physical hydrologic data and water quality data (major ion concentrations) together to construct water and chemical budgets for two small adjacent watersheds at the site, the Arboleda (50 ha) and Taconazo (26 ha). Water budgets were computed for four consecutive 12-month periods beginning 1 December 1998 and ending 30 November 2002. Major ion budgets were computed for the latter two of these four budget years (12/00-11/01 and 12/01-11/02). Streamflow was continuously measured at a V-notch weir on each watershed, rainfall was measured with tipping bucket gauge in a nearby clearing, the (insignificant) change in water storage each budget year was estimated from the difference in water table elevation between the start and end of the year, and ET was estimated using Li-cor pyranometer data for our budget years and an empirical relationship between annual Priestly-Taylor ET estimates and annual pyranometer-based energy inputs at our site. Chemical data were used to distinguish between two chemically different waters (high-solute "bedrock groundwater" and low-solute "local water") in streamflow and in interbasin groundwater transfer (IGT). The water budgets unambiguously confirm a large IGT into the Arboleda watershed; on average about 15,000 mm of water entered the watershed each year, 5,000 from rainfall and 10,000 from IGT. About 44% of the IGT was bedrock groundwater, and the remainder was low-solute local water, chemically identical to subsurface water derived from rainfall onto the watershed. The Taconazo had little or no IGT (the water budgets and chemical data show no IGT of bedrock groundwater and a small positive IGT of local water that is not clearly distinguishable from zero, given the uncertainties in the budgets). Chemical budgets differed even more greatly than water budgets between the two watersheds, because of the large differences (factors of 13 to 30) in major ion concentrations between bedrock groundwater and local water. Annual Na, K, Mg, Ca, Cl, and SO4 budgets in the Arboleda were dominated by the roughly 4,400 mm of bedrock groundwater IGT (this accounted for about 93% of the inputs of these six ions to the watershed). IGT of local water accounted for about 4% of major ion inputs, and atmospheric deposition less than 3%. Results show the importance of IGT to the hydrologic and geochemical status of lowland rainforest in this area, and suggest the clear need to account for IGT in small watershed studies from which conclusions are to be drawn regarding the hydrogeochemistry of a particular land use or landscape type. IGT may be an important link between distant upland and lowland areas through an interposed groundwater system, and as such should be considered in both practical and research-oriented water resource questions.