The isotopic and hydrologic response of small, closed-basin lakes to climate forcing from predictive models: Application to paleoclimate studies in the upper Columbia River basin

Simulations conducted using a coupled lake-catchment, hydrologic and isotope mass-balance model indicate that small, closed-basin lakes in north-central Washington are isotopically sensitive to changes in precipitation, relative humidity, and temperature. Most notably, model simulations predicted inconsistent lake responses to precipitation changes due to differences in lake outseepage rates and surface area to volume (SA : V) ratios. Greater outseepage within model experiments resulted in increased sensitivity to changes in mean precipitation. Moreover, simulations suggest that, in lakes with appreciable outseepage, SA : V ratio changes resulting from lake-level variations control the direction of changes in lake water oxygen isotope composition (d18O). Specifically, in lakes with a SA : V ratio that increases at higher lake levels, steady state d18O values will increase in response to greater long-term average precipitation. These results suggest that closed-basin lakes with low outseepage rates will exhibit a transient isotopic response to stochastic variability in hydrologic forcing but will not strongly respond at steady state to variation in mean hydrologic conditions. Conversely, closed-basin lakes with appreciable outseepage will exhibit strong isotopic responses to both stochastic variability and variation in mean hydrologic conditions (i.e., mean precipitation, relative humidity, and temperature control of catchment hydrologic inputs to the lake). These relationships provide a mechanism for explaining inconsistencies in the isotopic responses of lakes within a given region to hydrologic forcing and demonstrate that semiquantitative models for describing the relationship between lake hydrologic and isotopic responses to climate variability are not appropriate for all closed-basin lakes. The hydrologic and chemical evolution of a lake is subject to a complex array of climate and catchment controls that can be mathematically described and related using numeric mass-balance models. Previous modeling studies have examined the influence of changes in climate (relative humidity, precipitation, temperature, solar insolation, wind speed, etc.) on the isotopic composition of lake water by simulating hydrologic and isotopic fluxes through time (Hostetler and Benson 1994; Gibson et al. 2002; Shapley et al. 2008). These climate variables, as well as catchment parameters and basin morphology, control water balance and lake residence time, and therefore define the temporal extent and magnitude of lake responses to climate dynamics. As such, mass-balance models can be used both to interpret lake sediment oxygen isotope (d18O) records by characterizing lake sensitivity to specific climate variables and to investigate underdetermined aspects of lake hydrologic systems (such as outseepage and throughflow rates) and therefore are useful to both paleoclimatologists (Rowe and Dunbar 2004; Jones et al. 2007; Rosenmeier et al. in press) and water resources scientists (Sacks 2002). In the seasonal, drought-prone climate of north-central Washington, small, closed-basin lakes (i.e., lakes with low rates of outseepage and no surficial outflow) exhibit hydrologic and isotopic instability. This instability arises primarily from the inflow of isotopically light surface runoff from snowmelt and spring precipitation and from evaporative enrichment of isotopes throughout the summer and early fall. In these lakes, the low salinity and isotopic depletion of runoff results in early spring water column stratification that is then weakened by wind action, diffusion, and evaporative enrichment in subsequent months. The persistent isotopic and chemical instability of closed-basin lakes in north-central Washington precludes the application of standard steady state analytical models and necessitates the use of numerical models to quantitatively describe lake response to climate change. In this paper, a coupled lake and catchment numeric mass-balance model is presented that simulates the hydrologic and isotopic response of two small, closedbasin lakes, Castor Lake and Scanlon Lake, to changes in the primary drought controlling climate variables (i.e., precipitation, relative humidity, and temperature). This relatively simple model is similar to the (hydrologic– isotopic balance (HIBAL) model for application to paleolake systems) model of Benson and Paillet (2002) in that it incorporates observed mixing depths and meteorological data to predict near surface and deeper lake water isotopic and hydrologic responses to climate forcing. To account for hydrologic inputs directly from the surrounding catchment (a component not represented in the HIBAL system) the model presented here incorporates catchment subroutines that describe snowpack, runoff, and soil moisture volume changes through time. Of the more sophisticated coupled lake-catchment models that are well established in the literature (e.g., the calibration free, onedimensional thermal model of Hostetler and Bartlein 1990), * Corresponding author: [email protected] Limnol. Oceanogr., 55(6), 2010, 2231–2245 E 2010, by the American Society of Limnology and Oceanography, Inc. doi:10.4319/lo.2010.55.6.2231