Guest Editorial: Dumbing Down Hydrology

July 27, 2016

Why are we dumbing down hydrology for engineers? When I say “dumbing down,” what I mean is that we (the stormwater profession) are telling engineers to use pre-computer, simple methods and techniques to size stormwater facilities. I recently asked a noted, national stormwater expert this question, and his response was (to paraphrase him), “We need to keep things simple because 99% of stormwater design is being done by a general civil engineer using somewhat dated models, or simple automations of older methods, for whom the stormwater design is about 10% of what he does.” That may be true, but it is hardly an appropriate excuse. Today in the United States we spend billions of dollars on stormwater management. For that much money, isn’t it important that we use the most accurate tools and methods available? So, why in most parts of the country are we still using pre-computer, slide rule hydrology to do our calculations?

Slide Rule Hydrology
What is slide rule hydrology? Slide rule hydrology is a set of simplified engineering methods to compute runoff from rainfall. This includes methods such as the Rational Method (Q = CIA) and SCS curve numbers. These methods were developed when the only computational tool we had to use to compute runoff was the slide rule or the electro-mechanical calculator. (An aside here for anyone under the age of 50 who doesn’t know what a slide rule is: A slide rule is an ingenious analog set of sliding scales that was invented prior to electronic calculators and personal computers to make complex engineering and scientific calculations. I used a slide rule in my engineering classes all the way through graduate school [1973]. Today slide rules can be found only in museums and the desk drawers of old engineers.)

More than 100 years ago, hydrologists and engineers understood that runoff is a function of land cover, soil, precipitation, and drainage area size. They knew that soil moisture changes between storm events and even during storm events as soils alternately get wetter and drier. The more saturated the soil, the greater the runoff. Precipitation is also highly variable. No two storms are the same in their volume, intensity, or duration. But when our only tool was the slide rule to compute runoff, we had to keep hydrology simple. We only had the computational power to calculate the runoff from a single event, and even then we had to make major assumptions about the physical mechanisms that convert rainfall to runoff.

These major assumptions include assuming that a specific storm return period produces the same return period flood (e.g., a two-year storm always produces a two-year flood). They assume a representative storm of a standard shape and volume (e.g., Type 1A storm) to compute runoff. They also include the assumption of a certain average soil moisture condition at the start of the rain event. All of these assumptions are built into single-event hydrologic modeling. These assumptions are bundled into the standard runoff coefficients for the Rational Method and standardized curve numbers for SCS-based methods.

These single-event methods, by their very nature, have no mechanism to simulate back-to-back storms, variable soil moisture conditions, or the effects of long-term infiltration and evapotranspiration. In the real world, a two-year precipitation event does not always produce a two-year runoff event. Real storms do not have a standard shape, volume, or length. And the soil moisture varies from one event to the next and even during a single event, depending on the climate, season, storm, and soil type.

Today no one uses a slide rule. So why are we still using hydrologic modeling methods designed for the slide rule?

What is the alternative?

Continuous Simulation Hydrology
Continuous simulation models the entire hydrologic cycle. All of the water in the hydrologic cycle is tracked everywhere all of the time. This type of hydrologic modeling can’t be done with a slide rule. A computer is required for the needed calculations.

With the use of a computer the continuous simulation model represents all of the processes observed in the hydrologic cycle with appropriate algorithms. The model routes the rainfall through the various storage compartments of the hydrologic cycle: interception storage, shallow and deep soil moisture storage, and groundwater storage. Evaporation and transpiration return water back to the atmosphere from each of the storages at different rates as a function of vegetation and soil type. Continuous simulation models can track and route each of the three different components of runoff: surface runoff, interflow (shallow, subsurface runoff), and groundwater (or baseflow).

Continuous simulation modeling is not a new concept. This year is the 50th anniversary of the publication of the Stanford Watershed Model by Dr. Norman H. Crawford and Professor Ray K. Linsley Jr. The Stanford Watershed Model went through a number of iterations and refinements before becoming EPA’s HSPF (Hydrological Simulation Program–Fortran), which was first released in 1980 (36 years ago). Today HSPF is included in a number of continuous simulation hydrology modeling packages, including EPA BASINS, HEC-HMS, and WWHM. Some single-event hydrology models, such as SWMM, have been rewritten to do continuous simulation modeling.

Continuous simulation models provide more accurate hydrologic estimates than the pre-computer, slide rule, single-event models because fewer major assumptions are required.

No longer do we have to assume some standard soil moisture condition at the start of a storm event. Continuous simulation models track soil moisture changes, both between storm events and during storm events, and how these changes vary the rainfall-runoff relationship with time.

With the use of continuous simulation models, no longer do we have to assume a standard storm shape, intensity, volume, and duration. Continuous simulation models use long-term, measured historic precipitation data, which include big storms, little storms, extended dry periods, and back-to-back major storm events. While we know that we will not see these exact storms in the future, we can be fairly certain that we will see the same general storm patterns and climatic conditions. If we think that in the future that climatic conditions and storm patterns will change in intensity, volume, and/or duration, we have the computational tools to alter the historic record to reflect these expected changes.

No longer do we have to assume that a specific storm return period produces the same return period flood. We know, just through observation, that if a two-year storm occurs when the watershed is already fully saturated then the resulting runoff will be much greater than if the storm occurs at the end of a long dry spell. This means that sometimes a two-year storm produces a 1.1-year flood (which has a 90% chance of occurring in any individual year), and sometimes it produces a five-year flood (20% chance). Rather than making an assumption about frequency or return period, with continuous simulation models we can produce a long-term continuous flow record that we then statistically analyze to independently determine the appropriate flow frequency.

These are some of the advantages of using continuous simulation:

  • We gain more accurate hydrologic results and a better understanding of the important hydrologic processes that control local, regional, and national water issues.
  • We gain the ability to reproduce historic flood events and compare our modeling results with observed flow data, where such data are available.
  • We gain the ability to evaluate how flow control facilities behave over a full range of actual hydrologic conditions, not just a single hypothetical event.
  • And we gain the ability to produce multiple-year, long-term records to statistically evaluate runoff and streamflow in terms of magnitude, frequency, and duration.

Magnitude and frequency do not tell the whole picture. Duration (the percent of time that a particular value is exceeded) is now recognized as an important flow statistic. Duration tells us how long the runoff is at or above a specific value. When it comes to stream-defining mechanisms such as erosive flows, the number of hours (or percent of time) that the flows are large enough to cause erosion is critical to the management of the riparian system. If the flow control facility maintains the existing flood frequency (in other words, the two-year flood does not increase), but the duration of erosive flows increases then the facility is not doing its job.

The common complaint against the use of continuous simulation hydrology is that it is too data intensive and too complicated for the average stormwater engineer to use. I don’t see other professions saying “Don’t include complex computer calculations in the design of our product because they will be too difficult to understand.” Structural engineers are smart enough to know how to include dynamic loads in building and bridge designs. And automotive engineers are smart enough to know how to make use of computer-controlled fuel injection in today’s cars. From what I have observed in my 40+ years in the profession is that stormwater engineers are just as smart.

What Is Needed
What the stormwater engineering profession is lacking is not intelligence; it is education. I have met very few stormwater engineers who have taken a hydrology course in college. It is not the engineers’ fault; it is the fault of our engineering colleges and universities. While many teach drainage design, few teach continuous simulation hydrology modeling. There is a difference. Drainage design typically focuses in the pre-computer single-event runoff calculation methods that I rail against above. That is hardly what is needed in today’s complex stormwater world.

Many engineers first learn about continuous simulation hydrology modeling through workshops and training seminars. While typically these are good venues for learning continuous simulation modeling, they are too brief and too infrequent. As a result, many engineers simply try to wing it. They assume that because they know how to do single-event modeling that they can use the same techniques and knowledge to do continuous simulation modeling. That is like a car mechanic assuming that because he knows how to adjust a carburetor that he can also fix a fuel injection system. Nice idea, but not realistic.

The argument that we should stay with the old methods because most stormwater design is being done by a general civil engineer, for whom the stormwater design is only about 10% of what he does, is a poor argument. If this work is outside of the general civil engineer’s area of expertise then that engineer is violating his or her professional code of ethics. This violation is specifically stated in ASCE Code of Ethics Canon 2: “Engineers shall perform services only in areas of their competence.”

The solution to this potential professional code of ethics violation is to hire someone who specializes in this continuous simulation hydrology analysis and knows how to use the latest, most accurate techniques and modeling tools. A general civil engineer certainly does this when dealing with complex structural design problems. Complex stormwater design problems are no different. (Although it should be noted that when a structural design fails it is pretty obvious, when a stormwater design fails we claim that the facility was not designed for that size of event—a convenient excuse, but one that will eventually fail; see my lawyer comment below.)

The other missing piece in what is needed to advance stormwater design and management into the 21st century is updating the stormwater engineering standards and regulations. Many, if not most, of these standards and regulations were written in the 1950s and ’60s when the focus was on flood control and the computational tools were limited to slide rules and nomographs (“a graphical calculating device in the form of a two-dimensional diagram designed to allow the approximate graphical computation of a function,” in case you have never seen one). We have come a long way since then, but our standards and regulations have not always kept up.

During my 40+ years in the profession I have met many federal, state, and local regulators and technical staff. And for five years I was one of them in the Seattle metropolitan area. I know only too well how overworked and underappreciated they are. Typically, stormwater regulations and standards are only one of their many responsibilities, and usually they do not have the technical background and education to make technical decisions that, in the end, drive the regulations and standards. They often rely upon us consultants to provide them with that technical expertise. And all too often they have received misguided and/or just bad advice all in the name of keeping things simple.

An example of this “keep it simple” approach is the idea of containing onsite the runoff from small storms up to some percentile (85th or 95th). This is based on the idea that for predevelopment conditions there is no stormwater runoff from these smaller storms. That assumption depends on a number of factors—the most important being the soil type and antecedent soil conditions (infiltration) and vegetation (evapotranspiration). The goal is then to provide for a developed land-use condition sufficient storage onsite to prevent runoff from these same smaller storms. It is an admirable goal, but how does it really work?

With land-use development comes impervious surfaces and the loss of both infiltration and evapotranspiration opportunities. Now the engineer is tasked with providing sufficient storage to compensate for these losses. The question is: Once you provide this storage, where does the water go? If you leave it sitting in a pond or a vault with little chance to infiltrate into the soil or evaporate into the atmosphere, what happens when the next storm arrives? If you discharge it into the stormwater system, what rate of discharge is acceptable, and who decides? And if you discharge it into the stormwater system, then you are not really mimicking the predevelopment runoff conditions.

You discover these problems only when you are forced to account for all of the water all of the time, and that is what continuous simulation hydrology modeling requires you to do.

So what is the solution when it comes to setting stormwater standards and regulations? Just as there is a saying at crosswalks, “Look before crossing,” there should be an equivalent saying in stormwater: “Model before regulating.” Spend the time and money to find out what the real-world impacts are going to be before setting a standard that may sound good on paper but makes little or no sense in the real world. Think of it as regulatory insurance. You spend a little up front to save you from a world of pain later on.

Some may argue that there is no harm in using these “tried and true” old methods used by our grandfathers. They argue that these methods worked then and they work now. But have they ever really worked?

In the Puget Sound region, King County (the Seattle metropolitan area) found that stormwater control facilities designed using SCS curve number methods consistently failed in urban and suburban communities. The SCS-based facilities were undersized and discharged too-large flows that caused downstream erosion and property damage (Booth and Jackson 1997). King County’s solution was to change to a standard (flow duration) and method (continuous simulation hydrology) that provided the actual protection specified by the stormwater regulations (Jackson et al. 2001).

This standard (flow duration) and method (continuous simulation hydrology) were incorporated into the Washington State Department of Ecology’s regulations for the 19 counties of western Washington (Ecology 2005) and the development of the Western Washington Hydrology Model (WWHM). (In the spirit of full disclosure I should note that I, as an advocate for continuous simulation hydrology and a practitioner of it for over 40 years, strongly encouraged the Department of Ecology to convert from a single-event standard to a continuous simulation standard for its stormwater regulations and later assisted in the development of WWHM.)

And that is the final problem with dumbing down hydrology for engineers. What if, by doing so, you are not getting the actual protection specified by the stormwater regulations? Can you hide behind the law and claim that failure was an Act of God? Maybe. But, knowing lawyers, it is my prediction that it is only a matter of time before some young, savvy environmental attorneys realize the potential liability caused by state and local municipalities not fully implementing the newest and best technology and standards to meet Clean Water Act NPDES permit requirements and file class action lawsuits on the behalf of flood-damaged homeowners. Win or lose, it is going to be a costly and unnecessary battle.

Conclusions
It is time we stop dumbing down hydrology for engineers. Years ago we traded in our slide rules for computers. It is time to trade our slide rule–based, single-event hydrology methods for computer-based continuous simulation hydrology models and join the 21st century.

References
Booth, D. B., and C. R. Jackson. 1997. Urbanization of Aquatic Systems, Degradation Thresholds, Stormwater Detention, and the Limits of Mitigation. University of Washington. Published in the Journal of American Water Resources Association, Volume 33, Issue 5, pp. 1077–1090.

Ecology. 2005. Stormwater Management Manual for Western Washington. Washington State Department of Ecology. Olympia, WA.

Jackson, C. R., S. J. Burges, X. Liang, K. M. Leytham, K. R. Whiting, D. M. Hartley, C. W. Crawford, B. N. Johnson, and R. R. Horner. 2001. “Development and Application of Simplified Continuous Hydrologic Modeling for Drainage Design and Analysis.” King County Department of Natural Resources, Water and Land Resources Division. Seattle, WA. Published by the American Geophysical Union in Land Use and Watersheds: Human Influence on Hydrology and Geomorphology in Urban and Forest Areas, Water Science and Application, Volume 2, pp. 39–58.

Doug Beyerlein will give a presentation on this topic Wednesday, August 24, 2016, at StormCon in Indianapolis, IN. See www.StormCon.com for details.

About the Author

Doug Beyerlein

Doug Beyerlein, co-founder of Clear Creek Solutions, has led the engineering community in the development of new, more accurate tools to analyze the hydrologic impacts of human activities on rivers, streams, and fish habitat. Doug is licensed as a Registered Professional Engineer in Washington and California and is certified as a Professional Hydrologist by the American Institute of Hydrology. He has over 40 years of experience in stormwater modeling.

Photo 45833406 © Kenishirotie | Dreamstime.com
Photo 60969407 © Hin255 | Dreamstime.com
Photo 15149780 © Jaromír Chalabala | Dreamstime.com
Photo 51983635 © Joe Sohm | Dreamstime.com