Integrating topography, hydrology and rock structure in weathering rate models of spring watersheds

Abstract

Weathering rate models designed for watersheds combine chemical data of discharging waters with morphologic and hydrologic parameters of the catchments. At the spring watershed scale, evaluation of morphologic parameters is subjective due to difficulties in conceiving the catchment geometry. Besides, when springs emerge from crystalline massifs, rock structure must be accounted in formulas describing the area of minerals exposed to the percolating fluids, for a realistic evaluation of the rates. These particular features are not included in the available approaches and for that reason a new model was developed, coined THROW model. This is a lumped approach that integrates (T)opography, (H)ydrology, (RO)ck structure and (W)eathering in a single algorithm. The study area comprises several stream watersheds and spring sites of the Vouga River basin (northern Portugal), shaped on granites. Firstly, the THROW model couples a terrain modeling analysis with hydrologic models based on discharge rates, to determine hydraulic conductivities (K), effective porosities (ne) and annual recharges (Vr) at the stream watershed scale. Subsequently, these parameters are used in a water balance model to estimate concomitant groundwater travel times (t). The mean K [(4.7 ± 3.2) × 10−7 m s−1] and ne [(2.0 ± 1.3) × 10−2] values are adopted as proxies for the spring watersheds and a firm regression equation is defined between time and stream watershed area (A). Secondly, two more runs of terrain modeling analysis are executed to extrapolate morphologic parameters for the spring watersheds. The first run hinges on scaling properties of the drainage networks, known as Horton laws, and is used to scale watershed areas across stream orders (i). The scaling function is described by another regression equation. The second run evaluates the order of a spring watershed, defined as equivalent order (ieq) and equated to the mean order of the surrounding stream watersheds. Having calculated the ieq, spring watershed areas and travel times were downscaled using the regression equations (A < 10 km2 and t = 1.4–2.8 year). Standing on the physical and hydrologic parameters of the spring watersheds, the THROW model finally calculates plagioclase weathering rates in the vicinity of the spring sites. The SiB model (Pacheco and Van der Weijden, 1996) was used before to estimate the contribution of plagioclase dissolution to the chemical composition of these springs (Van der Weijden and Pacheco, 2006). The chemical data were now coupled with K, ne and t in a rate equation to estimate chemical weathering rates of plagioclase in the basin. In the THROW model, the rate equation describes the exposed surface area as a function of fracture spacings, openings and porosities (Pacheco and Alencoão, 2006). The calculated rates (WPl = (2.5 ± 1.2) × 10−14 mol m−2 s−1) are consistent with previous reports and with results of experimental kinetic models. The SiB results predict formation of halloysite and gibbsite along the flow path, which were indeed close to equilibrium with the dissolved Al and Si activities. ⺠Develop a weathering model that incorporates rock structure in the rate equation. ⺠Conceive a weathering model especially designed for fracture artesian springs. ⺠Create a model that integrates topography, hydrology, rock structure and weathering