Frozen Soil:Frozen Soil

From Gsshawiki
Revision as of 19:04, 25 September 2018 by Mfollum (talk | contribs) (modCFGI)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Substained low air temperatures can result in the ground freezing with a dramatic lowering of vertical soil hydraulic conductivity and infiltration, and increased runoff. Freezing of the soil can have a large effect on outflow from a watershed. When coupled with the effects of snow and snow melt, frozen soil can have an extremely large effect on runoff volume and timing. Simulating the effects of frozen soil on hydrology is an active research area at ERDC. This section will discuss the ability to simulate the effects of frozen soil in the current release version 7.0, and previous releases.


In GSSHA versions up through 6.12 the Continuous Frozen Ground Index (CFGI) model (Molanau and Bissell, 1983) is active during any long term simlation. The CFGI model computes a running total of negative degree days using this equation:


where: T is the air temperature (C), snow_depth is the depth of snow in the cell (cm), A is a factor that accounts for degradation of the factor over time (0.99875 for hourly calculations), and K is the snow thermal insulation factor (default of 0.5). The air temperature (T) is divided by 24 in GSSHA because the calculations are peformed hourly and the model is formulated for daily computations.

When the total exceeds a threshold, the ground is considered frozen.

In GSSHA, the CFGI is computed in every cell using the air temperature from the hydrometeorlogical input data, Section 9.3. When the CFGI model exceeds the threshold and the ground is considered frozen,infiltration will cease in that cell until the CFGI value falls below the threshold. The theshold is 83.0. This value represents conditions in the Northwest United States (Molanau and Bissell, 1983). The applicability of the threshold in other regions is unknown.

In v6.1 and beyond the user may also specify that the soil is frozen anywhere there is snow on the ground by including the SNOW_NO_INFILTRATE card in the project file. CFGI will still be computed and there will be no infiltration in any cell with a CFGI above the threshold value, or with a snow pack.

In the current release version 7.0 (previously Beta V6.2) the use of the CFGI model is made optional and the ability to specify the threshold value for frozen soil and the snow thermal insulation factor, K, are included. In v7.0 the default is no frozen soil computations. To use the CFGI model, the CFGI card is included in the project file. If just this card is included, then the CFGI model will function largely as before with the default values. If the user wishes to change the threshold value for frozen soil, then the card CFGI_INDEX is included in the project file, along with the numeric value for the threshold. Values higher than 83.0 will delay the freezing of soil; values lower than 83.0 will hasten the freezing of the soil. The snow thermal factor (K) is specified with the CFGI_K card. Values higher than 0.5 increase the effect of snow; lower values decrease the effect. A project file with default values would include these cards:

CFGI_K 0.5

The effect of the SNOW_NO_INFILTRATE card remains the same in v7.0.

GSSHA v7.0 has greatly expanded for the computation of snow accumulation and melt and the CFGI index model reflects these changes. The most important is that in v7.0 air temperature is computed in each cell based on elevation. Also, based on limited experience with the model in areas with deep snow packs, we've modified the degradation factor, A, to account for the insulating effects of snow on this factor. We noticed that in a deep snow pack the current formulation prevents cold air temps from lowering the CFGI value. However, the degradation factor, A, causes the CFGI to slowly degrade over time, resulting in early thaw. To account for the effect of the snow pack on the degradation factor, A, we add the following quantity to A:


so that as the snow pack increases A tends toward 1.0. With no snow pack, there is no adjustment.

GSSHA has been coupled to the soil thermal regime model GIPL to allow the temperature profile in the soil, and deep into the underlying bedrock, to be computed. In this version of GSSHA, we are able to better estimate the location of frozen soils and the location of permafrost. This is a research version of the code that will be released after sufficient testing. More information on these efforts will be forth coming.


The CFGI method was modified by Follum et al. (2018) to account better account for topography and landcover on the available energy to thaw the ground, as well as inclusion of ground cover (i.e. leaf litter) as an additional insulator of the ground. The modified CFGI (modCFGI) was shown to better represent frozen ground temporally and spatially when tested at Sleepers River Experimental Watershed in Vermont.

modCFGI is not currently available in the release version of GSSHA, but will in the next release.

The following cards can be used within the modCFGI model:

CFGI_RTI => Uses the Trad proxy temperature instead of air temperature
CFGI_INDEX 10.0 => Same as in the CFGI model, but is lower if you account for ground cover as an insulator
CFGI_K_LOW 0.08 => Lower value of CFGI_K, based on Molnau and Bissel (1983)
CFGI_K_HIGH 0.5 => Upper value of CFGI_K, based on Molnau and Bissel (1983)
CFGI_A 0.97 => You can now specify the CFGI_A value
OUTPUT_FG_FILE "Filename.fgd" => You can now output a gridded file of frost depth

The modCFGI is similar to CFGI, but calculates the following Frost Index: F = F*CFGI_A - Ta * exp(-0.4(CFGI_K*snow-depth + CFGI_GC*depth_groundcover))

Trad is used in place of Ta when the CFGI_RTI card is present. CFGI_GC and depth_groundcover are defined in the mapping table (more documentation to follow).


  • Follum M.L., Niemann J.D., Parno J.T., Downer C.W. 2018. A simple temperature-based method to estimate heterogeneous frozen ground within a distributed watershed model. Hydrology and Earth System Sciences. 22(5):2669-88.
  • Molnau M. and V. C. Bissell. 1983. A continuous frozen ground index for flood forecasting. Proceedings of the Western Snow Conference. 743-83. pp 109-119.

GSSHA User's Manual

20 Frozen Soil