Use of Small-scale Parameterizations to Improve Stream Flow Simulations using a Regional-scale Hydrologic Model in a Sub-arctic, Boreal Forest Environment
In the sub-Arctic, the presence or absence of permafrost dominates the response of many of the hydrologic processes including stream flow, soil moisture dynamics, and water storage processes. In areas underlain by permafrost, ice-rich conditions at the permafrost table inhibit surface water percolation to the deep subsurface soils, resulting in an increased runoff generation during precipitation events, decreased baseflow between precipitation events, and relatively wetter soils compared to permafrost-free areas. Permafrost also has a strong influence on ecosystem composition and function. Permafrost-free soils are dominated by deciduous vegetation, that appear to play a major role in taking up and transpiring liquid precipitation to back to the atmosphere. Permafrost soils are dominated by coniferous vegetation that have a more minor role in transpiring water to the atmosphere. As a result, the sub-arctic environment is extremely challenging to simulate, as these two very different environments--one dominated by lateral water fluxes and relatively low transpiration rates and the other dominated by vertical fluxes and relatively high transpiration rates--transition from one system to the other over extremely short spatial distances (on the order of meters). A major goal of this study is to improve the performance of regional-scale hydrologic models through improved parameterizations derived from field observation and fine- and small-scale modeling efforts. The Variable Infiltration Capacity (VIC) model, a regional-scale, hydrologic model, was applied to two catchments in the Caribou-Poker Creeks Research Watershed (CPCRW), which is located in the discontinuous permafrost zone and in the boreal forest. The first catchment (LoP) is nearly permafrost-free (<2%) and is dominated by a relatively high transpiring deciduous vegetation. The second catchment (HiP), is underlain with approximately 50% permafrost and is dominated by the relatively low-transpiring coniferous vegetation. Baseline simulations were made using the VIC calibration process and the large-scale parameterizations based upon products from the World Soil Classification product for soil properties and the Scenarios Network for Alaska and Arctic Planning product for vegetation composition., resulting in R2-values of 0.47 and 0.55 for the LoP and HiP catchments, respectively. Simulations using small-scale parameterizations derived permafrost distribution (for soil properties and vegetation composition), aspect (soil properties and vegetation composition). Simulations were also made using Results indicate that small-scale parameterizations (all combinations) result in an improved R2-vale over baseline simulations, with the largest difference using the permafrost map for both soil properties and vegetation ~0.71 and 0.65 for the LoP and HiP catchments.