Langur: A snowpack simulation model

The Langur model simulates the daily accumulation and ablation of the snowpack, throughout every part of a heterogeneous landscape. It was developed for use in studies of wildlife ecology and management in Yellowstone National Park. See applications of the Langur model.

Development and testing of the original version is described in two papers:

Watson, F.G.R., Anderson, T., Newman, W., Alexander, S., & Garrott, R.A. (2006) Optimal sampling schemes for estimating mean snow water equivalents in stratified heterogeneous landscapes. J. Hydrol. 328: 432-452.

Watson, F.G.R., Newman, W., Coughlan, J.C., & Garrott, R.A. (2006) Testing a distributed snowpack simulation model against diverse observations. J. Hydrol. 328: 453-466.

The current version (2006) incorporates many improvements over the original version.

The model is implemented as original code within the Tarsier Environmental Modeling Framework:

Watson, F.G.R. & Rahman, J.M., 2003. Tarsier: A Practical Software Framework for Model Development, Testing and Deployment.
Environmental Modelling and Software. 19:245-260.


Much newer:
There's a few videos utilizing Langur output on the EcoViz home page. Scroll down to the one that says "Snow model", and also the "Virtual Interpretive Trail".

Rather old:

To see a video of the Langur model and stills from the video, click the image at right.


  • Original code: based on previous models (UEB, RSM, Macaque)
  • Spatial scale: typically 30 x 30 m resolution over 100 x 100 km landscapes
  • Temporal scale: daily simulation over many decades
  • Spatial inputs: model is responsive to raster maps of :
    • elevation
    • mean annual precipitation
    • forest cover
    • geothermal heat flux
  • Temporal inputs: requires only :
    • daily precipitation at up to 5 stations
    • daily maxamimum & minimum temperature at up to 5 stations
  • Does not require or use daily snowpack data as input
    • This reduces the circularity (in assessing the accuracy of a model)
      involved with testing a model against (snowpack) data that are
      somewhat non-independent of the input (snowpack) data
  • Outputs:
    • snow water equivalent (SWE)
    • snowpack denisty
    • snowpack temperature
    • snowpack depth
    • snow intercepted by vegetation
    • runoff
  • Internal structure:
    • 4-layers: 2 snow, 1 soil , 1 vegetation
  • Validation data:
    • >3000 snowpack cores taken in Yellowstone National Park 2001-2006
    • >100 temperature profiles
    • ~100 hardness profiles
  • Platform: PC / Windows

Here is early screen grab image of the model in operation (circa. mid-2001)
29 Nov 2006 Webmaster