The VIC-NL model represents surface and subsurface hydrologic processes on a spatially distributed (grid cell) basis. In typical applications, gridcells have ranged in resolution from 1/8 to 2 degrees per side. Subgrid scale variation in vegetation characteristics can be approximated by partitioning grid cell areas to different vegetation classes. For example, a single grid cell could contain 2 vegetation classes - 33% with tall coniferous trees and 36% with grassland, each represented by a different vegetation parameter set (e.g., leaf area index, surface roughness). Grid cell subsurface processes are represented using average soil characteristics for the entire cell. Energy and water balance terms are computed independently for each coverage class (vegetation and bare soil) present in the model.
Vegetation and soil characteristics associated with each gridcell are reflected in sets of vegetation and soil parameters. Parameters for vegetation types are specified in a user defined library of vegetation classes (usually derived from standard, national classification schemes), while their distribution over the gridded land surface area is specified in a vegetation parameter file. Soil characteristics (e.g., sand and clay percents, bulk density) can be represented for a user-defined number of vertical soil layers - usually two or three, divided into a thin upper layer and a secondary set of layers that extend several meters into the soil column. The sources of data and specification for each vegetation and soil parameter are described in greater detail in the links following from the VIC-NL model operations web page.
Processes governing the flux and storage of water and heat in each cell-sized system of vegetation and soil structure include evaporation from the soil layers (E), evapotranspiration (Et), canopy interception evaporation (Ec), latent heat flux (L), sensible heat flux (S), longwave radiation (RL), shortwave radiation (RS), ground heat flux (t G), infiltration (i), percolation (Q), runoff (R) and baseflow (B). A full discussion of the algorithms relating to these processes may be found in the references listed in http://www.hydro.washington.edu/Lettenmaier/Models/VIC/VIChome.html, particularly Liang, et al. (1994), Liang, et al. (1998) and Cherkauer, et al. (1998).
Features of interest include the eponymous variable infiltration curve, shown to the right of the soil and vegetation column schematic, which scales the maximum infiltration (im) by a non-linear function of fractional gridcell area to enable runoff calculations for subgrid-scale areas (such as arise from the use of multiple vegetation classes). Another feature is the specification of baseflow as a function of soil moisture in the lowest soil layer. This relationship is non-linear at high soil moisture contents, producing rapid baseflow response in wet conditions. Below a a user specified value of soil moisture, the function becomes linear reducing the responsiveness of baseflow in dry conditions. As well as for many other processes in the VIC-NL model, the parameters that define both the infiltration curve and the baseflow curves are user-specified, affording the model user a good deal of control over these processes.
At present, the model runs one gridcell at a time over a desired period (any subset of the period spanned by the model forcing data), to produce time series of runoff, baseflow, evaporation, and other physical variables for each gridcell. These time series are then routed by the VIC routing model to produce streamflow at points of interest in the watershed. In the near future, runoff calculations may be re-sequenced so that runoff is generated in "image mode" - i.e., for one timestep for all gridcells before the proceeding to the next timestep.
return to the VIC-nL Model Figure