Green-Ampt Infiltration Method and Parameter Estimation

Green-Ampt Infiltration Method and Parameter Estimation

As a theory-based infiltration method, Geen-Ampt is widely used in hydrologic modeling since its parameters are easy to acquire, especially for those project sites without monitored data for model calibration. Green-Ampt method was initially developed for ponded infiltration into a homogeneous soil with a uniform initial water content and the water wetting front movement in the soil was governed by Darcy’s law (Figure 1).

Figure 1

As shown in Figure 2, generally four (4) parameters are required for Green-Ampt infiltration method: initial water content (dimensionless), saturated water content or porosity (dimensionless), wetting front soil suction head (inch), and hydraulic conductivity (in/hr). Some literatures and software list three (3) parameters by combining initial water content and saturated water content as a single parameter – initial water content deficit (saturated water content – initial water content)

Figure 2

Soil Water Content or Moisture Content is the fraction of the total soil volume is taken up by water which is a dimensionless value expressed as the ratio of V-water to V-bulk soil. The corresponding value at the commencement of a simulation is called initial water content. When the soil is saturated, the water content is called saturated water content and it can be estimated as porosity or total porosity.

The initial water content usually can not be zero in reality because of residual soil water/moisture content which is the water content of a soil where a further increase in negative pore-water pressure does not produce significant changes in water content. By definition, the residual soil water content (Unit: dimensionless, V-water/V-bulk soil) represents the remaining water content after a saturated soil is allowed to drain thoroughly for an extended period of time. Depending on antecedent moisture conditions (AMC), the initial water content is a value between residual soil water content and saturated water content, and correspondingly the initial water content deficit can range from zero (with saturated water content) to effective porosity (with residual soil water content).

Under the average antecedent moisture condition (average AMC), Dr. David Maidment in his book of Handbook of Hydrology suggested using wilting point (-1500 kPa water content) as initial water content in the western USA and using field capacity (-33kPa water content) as initial water content in the eastern USA. Correspondingly, the initial water content deficit will be calculated as (porosity – wilting point, Do NOT use effective porosity here – see Figure 3 for illustration) for Western USA or (porosity – field capacity) for Eastern USA for average AMC (Table 2). For the concepts of wilting point and field capacity, refer to this post.

Figure 3

Wetting front suction head (inch) and hydraulic conductivity (in/hr) of various soil textures were estimated by Rawls in 1982 and 1983 (Table 1, after unit conversion). It is not clear why the two papers authored by Rawls in 1982 and 1983 presented two sets of hydraulic conductivity values. For Green-Ampt infiltration method, the hydraulic conductivity values from the 1983 paper (B) are recommended as initial estimation, which are also endorsed by HEC-HMS Tutorials and Guides and EPA SWMM User’s Manual.

Table 1

In addition to the four required parameters mentioned above, when applying Green-Ampt infiltration method for HEC-RAS 2D modeling in a 2D infiltration layer, two more parameters are needed as shown in Figure 4: 1. residual moisture content, or residual soil water content; and 2. pore-size distribution index. Both of the parameters are dimensionless.

Figure 4

Per HEC-RAS 6.1 User’s Manual, the two additional Green-Ampt parameter are required to utilize the Green-Ampt with Redistribution (GAR) method. The residual moisture content, or residual soil water content and pore-size distribution index are summarized in Table 2.

Table 2

According to the study done by Tuflow, Green-Ampt infiltration method is relatively insensitive to porosity (saturated water content) and wetting front soil suction head. During model calibration, the focus should be on hydraulic conductivity and initial water content. The hydraulic conductivity impacts the runoff volume greatly: the higher the hydraulic conductivity, the more infiltration and the less runoff. When initial water content is increased, the model has a quick response to rainfall and usually ends up with more runoff. The influence of the initial water content tends to be less substantial after soil becomes saturated.

Green-Ampt infiltration method is available in InfoWorks ICM for both of its hydrology methods – the traditional subcatchment hydrology method and Rain-on-Grid on a 2D zone.

In XPSWMM, Green-Ampt infiltration method needs to be defined in Global Data… –> (R) Infiltration (Figure 5).

Figure 5

For illustration, a HEC-HMS model result is shown in Figure 6 where the loss method is Green and Ampt.

Figure 6

PS. PCSWMM help page has a soil characteristics table (Table 3) which generally matches Table 1 and Table 2 above. In Table 3, the Initial Deficit is estimated as the difference between Porosity and Wilting Point.

Table 3

6 COMMENTS

comments user
Scott Mansell

I’m very comfortable with 2D HEC-RAS modeling, but hadn’t tried the new infiltration capabilities before. I tried setting up infiltration on a 2D model for the first time using typial soil and landuse data and using the methods from the user manual, and I used Green-Ampt for the method. The plan is for this model to be used to simulate both short-term events and long term (~several month periods), so both accurate runoff prediction and recovery of the soil moisture deficit is important. I ran my first test for 4 days with rain occurring in the first 6 hours at a fairly heavy rate, and ET occurring at a low rate the whole simulation period. It works, but the ponded areas drain REALLY slowly after the rain stops, even when the Ksat is relatively high. Looking into the problem more, I plotted some timeseries of several cells for cumulative infiltration depth and the cumulative ET depth. I could see that during the rainfall the infiltration depth increased with time, and it increased faster for those areas with higher Ksat, which makes sense. However, after the rain stopped, the cumulative infiltration depth remained constant for the rest of the simulation, even when there was ponded water on the cell. Some depth loss occurred, but only in response to ET. I also noticed that ET only occurred while the cell had a non-zero ponded depth, so apparently it doesn’t evaporate from the soil and restore the moisture deficit like I had understood from the user manual? I’m very confused as to why this is happening. I guess I could switch to the Deficit-Constant method, but I think I read on here or somewhere that that method tends to underpredict runoff during the storm event because no runoff at all occurs until the moisture deficit is 0. I’d appreciate any help or advice you could give me.
Thanks,
Scott

    comments user
    Jimmy

    I would make a simple model, test run it while changing one variable at one time to debug it.

comments user
Scott Mansell

Sorry, I realized another thought. Even in the Deficit-Constant method, it looks like percolation only occurs as long as the soil is saturated. As soon as it has any non-zero moisture deficit, percolation stops and the only way for water to leave the soil is ET. I would think that is a problem, as I would expect water to percolate to some extent until field capacity is reached (although at an decreasing rate). Do you know if there is any way in HEC-RAS to simulate that? Do you know if ponded cells (or partially ponded cells) continue to percolate, or is it only based on rainfall?
Thanks,
Scott

    comments user
    Jimmy

    The deficit constant method assumes the soil (active soil layer) has to become saturated first before percolation happens.

comments user
Taehee YOON

Im a University student in South Korea.
I want to be know reference about “Table1”
Can you give me a reference??

    comments user
    Jimmy

    the reference is at the end of the table.

Leave a Reply

Your email address will not be published. Required fields are marked *