-1 Simulation of stream water alkalinity concentrations using coupled models of soil air CO 2 and stream water chemistry DANIEL L. WELSCH 1, *, B. JACK COSBY 2 and GEORGE M. HORNBERGER 2 1 Department of Geography, Frostburg State University, Frostburg, MD 21532, USA; 2 Department of Environmental Sciences, The University of Virginia, Charlottesville, VA, USA; *Author for correspondence (e-mail: dwelsch@frostburg.edu; phone: +1-301-687-4891; fax: +1-301-687-4495) Received 2 February 2005; accepted in revised form 27 November 2005 Key words: Alkalinity, Carbon dioxide, CO 2 model, Simulation Abstract. The ability to predict stream alkalinity values over timescales shorter than monthly or annually is needed to understand the response of stream chemistry to acidic inputs which occur across short time scales (days). We develop and apply a coupled series of physically-based models which are able to predict daily stream alkalinity values by first calculating soil air CO 2 concen- trations. We apply the model to a 9 year record of discharge and stream chemistry from a small catchment in the Shenandoah National park of Virginia. We find that we are able to accurately predict the minimum daily stream alkalinity values for all years and we are able to accurately predict the entire annual cycle for 6 of the 9 years (Nash–Sutcliffe criterion equals 0.26). For the 3 years which we overpredict summer stream alkalinity, summer precipitation was greater than normal and much greater than the period for which the model was calibrated. Introduction Alkalinity concentrations in headwater streams vary over multiple time scales. Alkalinity varies over decadal periods in response to changes in catchment soil characteristics and atmospheric deposition of strong acid anions (Neal and Whitehead 1988). A seasonal time scale is also evident in response to seasonal fluctuations in soil temperature (Castelle and Galloway 1993; Norton et al. 2001). The shortest time scale, on the order of days or hours, results in alka- linity fluctuations in response to rapidly changing soil respiration values and changes in stream discharge from storms (Hyer et al. 1995; Wigington et al. 1996; Lawrence 2002; Welsch and Hornberger 2004). This paper seeks to answer two fundamental questions. First, how can we model stream alkalinity variations in response to environmental conditions in a humid forest catchment? Second, what about the physical, chemical, and biological processes occurring in the watershed can we learn from this mod- eling if we are successful? In order to answer these questions, we will take a coupled-model approach where the following steps are followed: (1) identify Biogeochemistry (2006) 79: 339–360 Ó Springer 2006 DOI 10.1007/s10533-005-5480-9