top banner
top banner

Hydrologic Predictions - On-Going Activities
Jump to: Motivation, Premises, Approach, Activities

Improving snow modeling component in Land Surface Models

One of the most difficult tasks in hydrology is realistic prediction of the timing, amount, and patterns of snowmelt in mountainous areas. The NWSRFS Snow-17 model was developed in the 1970s ( Anderson , 1978 ). The NOAH snow model, although rooted in improved physics, still has major weaknesses (see comments in Chen and Dudhia, 2001 ). Considerable research (including our own) indicates that certain critical physical processes of snow need to be described within a snow model, including:

  • Snow in vegetated areas,
  • liquid-ice phase changes inside the snowpack,
  • variations of snow physical properties with snow compaction,
  • strong gradients of snow property variability in the top layer of a snowpack, and
  • the formation of deeply stable atmosphere over the snow cover ( Jin et al., 1999 ).

Comparison between several snow models (table 1) indicate that the SAST model considers all of the above key characteristics

Model Characteristics NWSRFS
Snow-17
(Anderson, 1978)
NOAH
Snow model
(Chen and Dudhia, 2001)
SAST
Snow-Atmo.-Soil Transfer
(Jin et al., 1999 a,b)
Use Mass-Energy Balance Yes (Lumped) Yes Yes
Application River basin Regular Grids Grids or Basin
Number of Vertical Layers 1 1 3 in variable thickness
Vegetation Influence No Yes Yes
Liquid Water-Ice in Snow Yes No Yes
Deep Stable Atmosphere Not considered Not considered Considered
Snow Physical Properties Fixed Fixed Vary with compaction

As shown, the SAST snow model, developed by our team priori to our relocation to UCI, was designed to represent most of the above- mentioned processes, while being efficient in terms of computation. The model evaluated well in coupled runs using the NCAR GCM-CCM3 and MM5 and off-line comparisons (including the international PILPS 2e experiment (when the model was applied to the high-latitude Torne-Kalix River basin shared by Sweden and Finland; see the special issue for PILPS 2e experiment in Global and Planetary Change , 2003).

Current and near-future activities will focus on

(1) implementing the SAST model into the NWSRFS and NOAH models,

(2) conducting intensive studies and validation over the CBRFC and CNRFC operational forecast areas,

(3) developing a satellite snow data assimilation scheme to improve the prediction of SWE (see below), and

(4) employing calibration procedures to improve the snowmelt runoff prediction

(please see section of hydrologic models).

 

Conceptual description of the SAST's treatment of snow-vegetation-soil distribution before (top) and after (bottom) snow.

Links
PIPLS 2E
PIPLS