Stream Gaugers vs. Beavers: the Beavers are Winning!

Canada was built on the back of beavers but — much to the chagrin of stream gaugers ever since — beavers managed to outlive the fashion of top hats. After the end of the fur trade, beavers started to re-populate the north with impunity.

Anyone unlucky enough to have the responsibility of stream gauging on a beaver infested stream knows how difficult it is to produce an accurate discharge hydrograph when beavers are around.

No matter how carefully you locate your gauge initially, it will only be a matter of time before a beaver comes along and decides the velocity in that reach is a bit too fast and he will build a dam to slow the water down. Once that happens stage becomes very insensitive to change in discharge. You can move your gauge, but the beaver will eventually build a dam downstream of your new site. You can blow up the dam, but the beaver will re-build it. You can develop a new rating curve or apply a shift correction, but every time you do that the beaver will modify the dam to invalidate your effort.

The real challenge for stream hydrographers is figuring out whether the stage-discharge relation is positive or negative at any given time. Rising stage may mean increasing discharge or it may mean an industrious rodent is busy making improvements to his dam, which reduces discharge. Falling stage may mean decreasing discharge or it may mean a breach in the dam, which increases discharge.

A Swedish biologist – Lars Wilsson – discovered that beavers respond to the sound of flowing water and are hard-wired to build dams to stop the flow. The more swiftly the water flows, the faster the beaver works. Silent leaks are not repaired, but water overflowing the dam is quickly repaired. This means that beaver dams wind up being more like sieves than weirs.

This dam(n) behavior is being increasingly celebrated all over the world.

The bring back the beaver campaign is seen as a bio-engineering solution for drought in California. Beavers are credited with saving salmon. Beavers have been reintroduced to Scotland. Beavers are also proposed as being the solution for water pollution with the potential to remove 5 to 45% of nitrogen loading in southern New England.

This interest in restoring beavers to the wild has been around for a while. This “fur for the future” video from the 1950’s shows beavers being trapped and then air-dropped into the mountains of Idaho. The airborne beaver corps could be revived to wreak havoc on stream gauges everywhere!

Things are not looking too good for stream hydrographers.

We need to become better at monitoring discharge in beaver infested streams. Index velocity might help, but good luck developing a stage-area relation for a stream that is persistently over-bank with water flowing amongst the dead/dying trees of what was the riparian zone. Good luck maintaining an index-mean velocity relation when the only part of the channel with enough water to deploy an ADVM has almost zero velocity. Good luck convincing your client to pay for the needed hardware.

In the meantime we’re left with shifting a rating curve to gaugings obtained by frequent field visits.

As long as we are stuck doing that, we should endeavor to do the best job possible. McCullough et al. 2010 investigated the hydraulic characteristics of beaver dams and found that they can be modeled using the weir equation Q=CL(H-e)3/2 where C is a coefficient, L is the length of the crest of the dam, H is stage, and e is an offset value.

The coefficient can be thought of as a composite roughness coefficient for water flowing through/over/under/around the dam. Using this as the basis for managing rating curves in beaver-controlled streams constrains the solution for fitting the series of deviations from the base curve. Basically, you can assume that L and e (at least as long as the dam stays the same length and height) are constant so the only thing that is changing from gauging to gauging is the coefficient C. This is different than the more typical method of applying a single point shift correction, which is equivalent to adjusting the offset (e). I sure wish I knew this back in the day when I was still doing field work!

I am interested in hearing from anyone who is struggling with beavers at their gauging stations. Does this reflect your experience? If not, why not? How do you deal with beavers? Please share your experiences below, for the benefit of others.

McCullough, M.C., D.E. Eisenhauer, M.G. Dosskey, and D.M. Admiraal. 2006. Hydraulic characteristics of beaver dams in a Midwestern U.S. agricultural watershed. World Environment and Water Resource Congress. pp. 1-10 doi: 1061/40856(200)182

Wilsson, L. 1971. Observations and experiments on the ethology of the European beaver (Castor fiber L.): A study in the development of phylogenetically adapted behavior in a highly specialized mammal. Viltrevy. 8(3) pp.115-205.


Free Whitepaper: 5 Best Practices for Building Better Stage-Discharge Rating Curves

A reliable rating curve is one that is credible, defensible, and minimizes re-work. This paper outlines 5 modern best practices used by highly effective hydrographers. Read whitepaper here.

8 responses to “Stream Gaugers vs. Beavers: the Beavers are Winning!”

  1. It is an interesting concept.
    I will try to see if I can make it work in practice.
    Has anyone out there had any luck using conductivity as the primary data source and using a conductivity-discharge relationship when beaver dams are a problem?

    • Hi Kim,
      Take a look at:
      Could electrical conductivity replace water level in rating curves for alpine streams?
      Steven V. Weijs RaphaëL. Mutzner Marc B. Parlange Water Resources Research 01/2013; 49(1):343-351. DOI:10.1029/2012WR012181


      Using discharge-electrical conductivity relationship in a mediterranean catchment: The t. ciciriello in the cilento, vallo diano and alburni european geopark (southern italy)
      D. Guida A. Cuomo

      However I suspect that a beaver dammed stream would have too much storage to yield much of a relation between conductivity and discharge. I would gladly defer to evidence that contradicts this pessimistic opinion.


  2. Because the stage-discharge relation can change daily, if the beaver are active, developing a shift is futile. Estimates are done instead. Interesting concept to use the weir equation. But if the beaver are active, then L and e are changing, and those efforts are futile as well because the system is unstable. Best to eradicate the critters. Have an on-site game-type camera take pictures daily (or more frequently) can help to produce more certain discharge data too. Maybe The Onion should do a piece on fur to get it back in vogue. Beavers are certainly a battle for stream gagers.

    • Hi Laura,
      Game cameras are a great idea. Knowing exactly when beavers are actively modifying the dam can greatly improve the interpretation of variations in stage. Developing estimates for those periods of time that the critters are active is even more of an art than estimating winter streamflow. In the winter, you can at least base your estimate on a recession curve during sustained sub-freezing temperatures. There are a lot more hydrological variables to consider in the summer. If game cameras can help to identify exactly when estimation is required as a last resort, then the cost of the cameras will pay for themselves in no time. But maybe the best idea is that we should all start wearing Top Hats.

  3. Hi Laura/Stu,

    Beaver back-watering at gauging stations has definitely been a big problem for us as well. I do like Laura’s idea of installing wildlife cameras, however in my experience, beavers (in general) operate at night, so unless you have a camera with night vision capabilities, you may not see much.

    We handle beaver back-water using shifts, similar to what Water Survey has started doing. We start by using a manual measurement shift value. Then between manual measurement shifts, and during obvious beaver activity (which we validate with data from nearby climate stations (to negate precip. events) and nearby hydrometric stations with similar drainage characteristics), we calculate daily shifts based on stage change. This can be very tedious, and result in the calculation of many shifts, but we’ve found that spending the time calculating and applying these shifts has produced pretty accurate discharge data. Obviously a lot of calculations and assumptions are included in this process, but when dealing with back-water conditions, derived discharge is estimated anyways. Let me know if you have any experience using this method, or if you have any thoughts or questions.

    I will definitely check out the McCullough et. al. paper in the meantime. Sounds like an interesting read at the very least.



  4. Top hats may be a little iffy but I’m told baked beaver is quite tasty.

  5. I am interested and pleased to see that there has been some considerable effort on increasing accuracy/consistency in stream flow monitoring by some, with respect to beaver modified systems. I am assuming that those people commenting above are scientists, perhaps professional biologists or similar? I am concerned to see the repeat use of terms ‘beaver infestation/infested’ by some commenters, suggesting that there is an invasion or infestation of a nasty/non-native species. Unless you are talking about areas beavers were introduced to where they never evolved (eg. Chile, Argentina), I am surprised to see (what I assume are) scientists suggesting eradicating a native species…as a means of making your job of measuring flows easier, as if somehow that should be the goal of effective gauging stations. It seems to me that for much of the last 100 years in North America, with the extensive trapping of beavers, it has been an unrealistically ‘easy’ job to monitor flows, since flows have not been restricted by beavers as much as they would naturally have been, and getting ‘accurate’ flow monitoring was not the same as ‘natural’/’expected’ flows, due to the absence of beavers. Beavers are a keystone species, for many reasons, including being integral to watershed health, development of our valleys since glaciation (where applicable), carbon storage, water quality/quantity control (improving water quality and reducing peak flows) and fish habitat improvement. For those suggesting beaver eradication, instead of better sampling protocols, may I urge you to focus on better understanding of ecology and things altering flows and then figure out the math and sampling approach to get the best scientific info, rather than eliminating the problem (beavers) to make getting those numbers easier.

    • Well said Norine,
      I think the comments need to be taken in context: I appreciate squirrels unless they are in my attic; I appreciate skunks unless they are in my basement; I appreciate termites unless they are in my walls; I appreciate beavers unless they are in my control reach. The post does explicitly point to the many benefits that are realized by reintroducing beavers to the wild. I also provided a link to the most comprehensive, but extremely obscure, study of beaver ecology and ethology ever written as a resource for learning to live with beavers. Ultimately, there are many more beavers than there are stream gauges so stream gauging will never be a serious threat to any beaver population. However beavers are a serious threat to stream gauges. Bad data leads to bad water resources management. Bad water resources management leads to bad environmental outcomes. The best outcomes will be if we, as a community, can figure out how to produce good data in bad circumstances, which is the whole point of the conversation. The most obvious answer is a always a good starting point for any discussion, however getting rid of beavers is completely unrealistic and so it is not the end of the story.

Join the conversation