Jim Dumont is the Engineering Applications Authority for the Water Balance Model program,. He is a recognized specialist in hydrologic modelling. For many years, he has been teaching modelling seminars as part of the professional development program provided by the Association of Professional Engineers and Geoscientists of BC. Jim Dumont evolved the Water Balance Methodology to address the relationship between rainfall volume control and resulting flow rates in streams; and developed the Stream Health Methodology as the technical foundation for the ‘Beyond the Guidebook’ initiative in 2007. This methodology correlated stream erosion as a measure of stream health.
The Partnership for Water Sustainability receives questions on a continuing basis from Water Balance Model users. Some are experienced water resource practitioners. Many are not. Regardless, Jim Dumont will take the time necessary to patiently explain basic principles.
I’m currently using the online Water Balance Model to develop a pre-to-post development water balance for a site in Ontario. I’m struggling to find agreement between the infiltration values from the online model and the ‘Thornthwaite and Mather’ (T&M) method of estimating infiltration.
The infiltration value using the online model exceeds 82% of the precipitation, whereas the T&M method gives a value of around 13% infiltration. The soil modeled in both cases is silty clay.
Obviously the two methods are very different in estimating infiltration, but I was not expecting such a large difference in values. Would you please provide me with guidance on how I could improve the model that I am working on?
“The original work by ‘Thornthwaite and Mather’ related to the estimation of Potential Evapotranspiration. This is total Potential but not the Actual amount. It is the an estimate of the amount that would occur if there were sufficient water available to supply both the evaporation and the infiltration plus plant uptake of infiltrated water for transpiration.”
“Keep in mind that there is evaporation from the ground after infiltration and there is some losses to deeper penetration of water beyond the depth of the plant roots. As such this methodology has little to do with estimation of actual infiltration rates.
“I cannot help but think that you are trying to compare two different processes and the results of calculations done by different methods, even though some of the words describing them may be the same. I am also aware that there are software packages that use the work of ‘Thornthwaite and Mather’ to estimate a ‘water balance’; however, I have not examined them in detail except to note that they generally utilize monthly climate data.”
“The Water Balance Model utilizes hourly climate data and the calculation engine has been thoroughly tested over the past thirty years. We have confidence in the results of the calculation engine, provided user input data is reasonable. In my experience, one of the least understood components of the hydrologic calculations is the soil and its condition. The second misunderstanding that occurs is the use of infiltration rates used in the WBM. These refer to the water being discharged from the bottom of volume control facilities, not the surface infiltration rate used in the hydrologic calculations.”
About the Thornthwaite and Mather Method:
Chapter 3 in the Province of Ontario’s 2003 Stormwater Management Planning and Design Manual provides this description of the ‘Thornthwaite and Mather’ method:
3.2.2 Water Balance Methods: In cases in which the available data cannot support more sophisticated approaches, water balance methods are more appropriate for predicting the changes to the hydrologic cycle that may result from urban development. They can be used to determine amounts of water that should be infiltrated to compensate for reductions caused by large paved areas or changes to vegetation. The water balance method developed by Thornthwaite and Mather (1957) determines the potential and actual amounts of evapotranspiration and water surplus (or excess of precipitation over evapotranspiration). Infiltration factors are used to determine the fraction of water surplus that infiltrates into the ground and the fraction that runs off to nearby streams. Thornthwaite and Mather’s method requires monthly or daily precipitation, monthly or daily temperature, latitude of the site, vegetation type, soil type, and a series of tables. The tables define a heat index, potential evapotranspiration, water holding capacity, and soil moisture retention. Snowfall, and alternating wet and dry cycles are included. Soil water holding capacity is dependent upon the soil type, soil structure and the type of vegetation growing on it. The Thornthwaite and Mather water balance method assumes mature vegetation and does not account for growing seasons where evapotranspiration would be less for immature vegetation.
When I read the engineer’s report that summarized the results of the scenario comparison, I was surprised to find that there really was not a significant difference in the VOLUME summary between the pre-logged and post-logged conditions. So I decided to try playing with some of the Water Balance Model input variables. What I found left me confused.
- When I varied the soil type, infiltration was the highest and discharge was the lowest with Clay and silty clay.
- The next thing I did was vary the soil depth. The deeper the soil, the higher the infiltration. But with all other parameters being the same, and 300 mm soil depth, when I then removed the forest and changed it to grass (“building lot”), the infiltration was higher, discharge was only slightly higher, and the total losses were lower, than when the site was forested.
Because of the trees’ evapotranspiration capabilities, we were expecting that discharge and losses would be significantly increased following deforestation. I am wondering if I am doing something wrong, or interpreting something wrong (i.e. the terms “discharge”, “losses”). Any insight you may have would be extremely helpful!
“The volume summary in the Water Balance Model report does not describe surface runoff alone. Rather, it describes water entering the stream from the project site, or contributing shallow groundwater to the system feeding the stream.”
“This description ignores completely the complex and variable soil moisture reservoir that is always operating just below the surface. The typical engineering approach is extremely simplistic which is why it is really only effective for sizing culverts or storm sewers.”
About Varying of Soil Type:
“While sand can see the greatest volumes being absorbed into the soil, that water cannot be retained as the sand’s water holding capacity is the smallest of any soil. In fact what happens is the water passes through the sand and discharges to the stream in a short period of time.”
“Only a soil with some fines of silt and clay can effectively retain water within its reservoir for use by a plant. This is why sandy soils need far more water to maintain a vegetative cover or to produce a crop.”
About Varying of Soil Depth:
“The soil depth is a factor that can make a difference, but only if it is not over-estimated. The hydrologically active zone may be 300 mm deep, but most forested areas do not have this depth of soil. For example, I am working on a couple of watersheds that behave as though they have only about 100 mm of soil. This may be a result of either bedrock, or an underlying marine clay.”
“It is important to provide a reasonable estimate of depth, and to not overstate it. It is also not reasonable to expect that increasing the depth to values of 300 mm or more will show a change in the hydrology of a project area. I see many people input soil depths which are overly large into the model.”
“The best way to judge the best value to use is to go to the site and look at an excavation or road cut in a ditch. The colour variation near the surface as compared to the underlying soil horizons will be quite obvious, and will correspond to the appropriate depth for the model describing that project.”
About Forest versus Grassland Surface Condition:
“The thought that a forest will absorb a great deal more water than a healthy grassland is not correct. Both will have quite similar values. The forest, however, will intercept slightly more rainfall because of its greater surface area.”
“Where forests make a significant impact on interception is in colder climates where they intercept snow and allow it to sublimate from the branches without ever touching the ground or being absorbed and transpired.”
“Trees in an urban setting behave in a different manner due to the enhanced tree canopy which is generally thicker and larger than for trees in close proximity in a forest.”
“If you wish to visualize the difference between a forest and grasslands imagine yourself as an ant on the ground looking upward. You might also go into your yard and spray your long grass for a few minutes. Then examine the ground below the grass to see if it is wet. Next, see how far the water has penetrated into the soil. This will replicate a very typical rainfall day with very little in the way of actual penetration of the water into the soil. Even an hour of sprinkling your lawn, equivalent of 25 mm of rainfall, will only have moisture wetting the upper 75 mm of the soil.”
“The discussion of interflow seems to be generally new in the presentation but welcome nonetheless. Is the software addressing movement in all directions? Jim mentioned the “sideways” component which is typically less than the vertical migration component in unsaturated soils. Does the program address saturated conditions that are common in fine grain soils during the winter?”
“The discussion suggests that the upper 600mm is the limit of impact that soils have on infiltration or storage capacity. I would assert that it depends on the nature of the material and any boundary layers at depth. I have done hundreds of test pits in fine grained soils with saturated zones well below 600mm in native soils. I have also done dozens of holes through turf and topsoil to consolidated clays showing little to no penetration as might be expected leading to the valid conclusion that 300mm of topsoil can have a huge impact on rainfall storage.”
“I agree that all soils are different and site specific conditions need to be considered. Having said that there are some generalizations that can be made regarding the flow of water through soils. If I know the processes under which the soil formed I can tell you how the water moves. I have provided a second set of graphics that shows to very different soil types.”
“The figure on the left shows a very typical Podzol that is found in BC and that has formed under a forest cover. The red stains are iron oxides that are leached out to the humus and upper portions of the soil. This indicates a primary chemical process that includes oxidation and exposure to oxygen in the form of air from the surface. Because the rust is transported by water it is very easy to see the depth to which the water typically penetrates the soil profile. Since the underlying soils started out the same as the upper soils following the last glaciation the differences have taken some 10,000 years to develop. I would say that the predominant soil in BC are the Podzols and that the typical depth of water penetration is not more than 600 mm. If you were to dig down through the reddish soil horizon you would find that the soil is dry and that the water table is some depth below the level where the colour changes.”
“The figure on the right shows a Gleysol which is formed under conditions that include long periods of saturation. The colors indicate considerable periods of limited oxygen and the chemical processes are primarily reduction. In this case the water would have moved up from below as well as down from the surface to saturate the soil for extended periods. There are small pockets of Gleysolic soils found throughout BC and they are often readily identified because the surface vegetation tends to be water tolerant plants. If you were to dig down through these soil layers you may find that occasionally it may be dry but that the water table would be quite near the surface.
“The lesson in this and the study of soil formation is that water does not move continuously downward as a result of saturated conditions. Generally the hydrologically active soil layers are limited to the upper 600 mm or so with a majority of situations the groundwater levels are deeper. In these cases the water will move sideways until one of two things occurs: 1) it reaches a stream, or 2) it finds a discontinuity in the soil that would allow it to move downward. As always, there are exceptions to these rules and that indicates a need to understand the soils that are present on a site so as to be able to describe how the water moves through the soil.”
“I hope this information can start an in-depth discussion and consideration of soil conditions particularly how the soils are hydrologically active.”
“Jim, I have a quick question regarding the Waterbalance Modelling program and I was wondering if you could help me? I’m trying to model a grass swale next to a road, however, my results seem almost ‘too good to be true’.”
“I added an infiltration swale without underdrain within the “Design Source Controls – Surface Enhancements” tab. And then added the grass swale area and impervious paving under ‘Apply Source Controls’ tab.”
” Is this the correct route to go if I just want to model a grass swale adjacent to a road segment?”
“Your results may be too good to be true.”
“The swale without storage acts just like any other grassy area in that it only captures the rainfall that directly lands upon it. If you apply a grass swale to a paved area then what the model will do is replace a portion of the area of the paved area with the area of the grass swale. An effective way of reducing imperviousness but I suspect that was not your intended method of reducing runoff.”
“If you wish to direct surface runoff to a Source Control then you must apply a Source Control with Storage. There is again the little quirk in the model for which we are contemplating a revision but for this version you still need to keep in mind. The surface area of the swale will replace an equal area of impervious pavement. This means the impervious area needs to be increased to allow the model to reduce it when it applies the surface area of the source control.”
“This should allow you to model the system and obtain reasonable results.”
The manager of development engineering services and deputy approving officer for a Vancouver Island municipality posed these two questions to Jim Dumont:
- Does the WBM require accurate detailed information to be entered to achieve meaningful results, information that the City does not have at this time?
- Is there an extensive track record of the effectiveness of the WBM?
The context for these questions was an exchange of perspectives that revolved around whether and when it is appropriate for a municipality to specifically require use of the WBM.
“Typical information needed as part of any development design process includes site specific data regarding area, slope, runoff coefficients or imperviousness, and a soil description.”
“The WBM uses the information gathered as part of a normal design process. It does not require the user to have more site information than required by any other system of design.”
“The WBM provides a substantial amount of information in the form of climate data and municipal zoning where individual municipalities have provided the zoning information. In this manner the WBM actually provides information not available in any commercial software package.”
The Calculation Engine:
“The QUALHYMO calculation engine within the WBM has been used and has been enhanced since its first application in the early 1980’s. It certainly has an extensive track record of successful use spanning some thirty (30) years across Canada and in other parts of the world.”
“We recognize that other analysis tools are available to municipalities and that they can be used in the design of drainage and rainwater management systems for developments and municipal works. However, we believe that the WBM has significant advantages over other available tools and is simple to use.”
Land is regularly being cleared by land owners before a development permit is sought from the municipality. This practice is creating significant rainwater runoff and impacting negatively on stream values. The debate I find myself in relates to determining the baseline condition.
Once land clearing has taken place, it has been suggested that a natural state condition cannot be determined. It has been further suggested that Water Balance Model analyses should be based on the ‘current condition’ of the site – that is, ‘pre-development’, and not ‘pre-deforestation’.
How can the Water Balance Model be applied to establish performance targets relative to an appropriate baseline condition that is acceptable to the municipality?
“To specifically answer the question, you can start with the forested condition even though some or all of the trees have been removed,” replied Jim Dumont. “This will provide you with an opportunity to compare what was lost to what might have been, and in so doing allow you to create a vision of the future watershed.”
“For the purposes of a Water Balance Model simulation, the starting scenario can be the watershed in any state, whether that is forested, existing urban, future planning, or just about any condition that you may wish to assess. Yet another key message is that the existing watershed condition should not be seen as a limiting condition; rather, it is just one of many potential conditions.”
“This is where the WBM shines as it is not constrained by starting or ending points. It compares whatever you can envision.”
Water Balance Model Creates Understanding:
To illustrate the last point, Jim Dumont described two potential applications that demonstrate the nature of scenario comparison options:
- Example #1: “You could compare a forested watershed to shopping centres, housing, parks, roads, or any mix of the foregoing. The scenarios could include previous engineering standards that would see all the rainwater piped to the stream; OR you can add in any number of rainfall capture techniques for runoff mitigation.”
- Example #2: “You could also compare existing urban mixes to potential future redevelopment changes. The choices for analysis could include: maintain current engineering approaches; apply mixed mitigation works; establish more forested parks or green areas, or whatever. Your imagination is your only limitation.”
“As you can see from these two examples, the Water Balance Model can be a tool to allow you to create an understanding of the past and compare it to many possible futures. This will allow you to see how the watershed can be altered, for good or bad. Then you can create a vision of where you would like to go and how the watershed can meet your vision,” concluded Jim Dumont.
The site in question is a 5-hectare site situated within a 60-hectare drainage catchment.
The pre-existing condition was forest, and the post-development condition is cleared (deforested) land. For the purposes of comparing before and after conditions, the engineer selected “grass – building lot” as the deforested surface condition.
We were quite surprised to see a very minimal difference in discharge, losses and infiltration between the forested and deforested conditions. Then we noticed that the analysis had been based on a 60-hectare catchment area. What is the significance of selecting 60 versus 5 hectares.
“If the site has 5 ha then why is the modeller using 60 ha? This is a very good way of masking the changes happening on a small site by averaging the lack of change across a much larger area.”
“In a very temperate coastal area we do not anticipate large snow pack development during winters. It is the large snowpacks that accumulate in clear-cut forests that alter the hydrology of the interior portions of the province.”
“On the coast there will be some increases in stream discharge resulting from clear-cuts, but not nearly as great as in areas that accumulate a great deal of snow.”
“If your site is up and away from the climate gauge used in the Water Balance Model, then there will be some differences. Even the 1 degree per 1,000 foot elevation loss as you go upslope will make a big difference in snow accumulation.”