RIP MSY – You were not meant for the real world

Posted: April 11, 2016

Can we apply MSY to deer management? Do we want to?
This squirrel's smorgasbord of hickory nuts is an unpredictable food resource from year to year, and an example of why carrying capacity is impossible to measure or predict for any species.

This squirrel's smorgasbord of hickory nuts is an unpredictable food resource from year to year, and an example of why carrying capacity is impossible to measure or predict for any species.

Last week I discussed the very cool theory of Maximum Sustained Yield. Referring to MSY as “very cool” reconfirms my membership in NA (Nerds Anonymous), but this theory also has important implications with regard to managing hunted populations.

Sometimes, however, application of a theory isn’t all it’s cracked up to be in the real world.

To recap, we need to know 3 things to manage a population under MSY:

  • Carrying capacity, 
  • Population size, and
  • Population mortality.

First, let’s look at carrying capacity. This is the maximum number of animals that the habitat can support. Simple enough.

But how do you measure that? It’s a simple answer, too. You can’t.

You can certainly try, but what do you measure? Carrying capacity will never be constant. Season to season, year to year – carrying capacity is dynamic and a moving target.

What if there is a bumper acorn crop – the habitat can support a lot more deer, but how many more? What if it snows and all of those acorns disappear under two feet of snow?

More importantly, when do management decisions take place? PGC biologists have to make recommendations to the Board of Commissioners in April the year before. You can’t even begin to predict how many acorns (or how much snow will fall) 8 months beforehand. So to accurately use MSY as a population management tool, you have to be able to predict the future.

Second, let’s take a look at population size. Can we estimate that? Yes, but not perfectly. Most biologists would consider deer population estimates “good” if they had a CV (coefficient of variation) of 20% - that means if we made repeated estimates of the population, two-thirds of those estimates would vary within plus or minus 20% of the true estimate. The other third would differ by even more.

That’s good enough to give you an idea of population size, detect trends, and make management decisions but not so good for making MSY decisions.

The problem is that if we overestimate population size, we end up harvesting too many animals. If we underestimate population size then we end up harvesting too few animals.

Maximizing a sustained yield ends up being an exercise in futility as this is also a moving target.

Third, let’s look at mortality. This is the one parameter we can estimate pretty well. Harvest estimates in PA, by WMU, have a precision (CV) generally <10% and some WMUs can be as low as 3% (Rosenberry et al. 2004). On top of that, we know that adult non-hunting mortality is not that variable from one year to the next.

Even though we cannot observe mortality perfectly, of the three pieces of information needed we actually have a chance of estimating it accurately and precisely.

Meatloaf sings that 2 out of 3 ain’t bad. But in this case, we only have 1 out of 3, at best.

The bottom line - whenever MSY has been applied in the real world, it has failed miserably. Mostly because a) we can’t measure carrying capacity (it’s a concept, not some measurable entity) and b) we can’t estimate population size accurately enough.

Just ask the fisheries peeps. For decades, both the wildlife AND fisheries professions (since the 1970s) have recognized MSY as a nice concept, but as a procedure to manage populations it does not work in the real world.

In fact, P. A. Larkin (1977) said it best when he wrote an epitaph for MSY:

M. S. Y. 1930s-1970s
Here lies the concept, MSY.
It advocated yields too high,
And didn't spell out how to slice the pie.
We bury it with the best of wishes,
Especially on behalf of fishes.
We don't know yet what will take its place,
But hope it's as good for the human race.
R. I. P.

So if anyone suggests to you that Pennsylvania should manage deer under MSY, just ask them how they are going to measure carrying capacity… every year… and predict what it will be 8 months in the future. Then ask them exactly how many deer there are, as well.

If you want to find out the answer to Larkin’s (1977) question of what will replace MSY, (i.e., how deer management recommendations are developed) then take a look at the Deer Management Plan and the resulting annual deer population report.

-Duane Diefenbach

A few comments to follow up on my previous post on MSY

1. The PA Game Commission has never used MSY to manage white-tailed deer. In fact, I'm not aware of any state agency that has attempted to manage white-tailed deer based on MSY, but my "deer career" is only about 20 years young.

2. I'm not sure I would even want to manage deer for MSY. I certainly don't want deer to have an adverse effect on the environment, but I like to see deer. I wouldn't see near as many deer under MSY than if the population were higher on that theoretical logistic curve. Based on what I know about the population dynamics of Pennsylvania deer, to manage at MSY would most certainly require a smaller deer population in most areas of the state.

3. The importance (and detriment) of MSY to wildlife management was best summarized by Holt and Talbot (1978), when they argued for  alternative approaches to managing populations that were more comprehensive. They noted that the concept of MSY has "played a significant role in the evolution of understanding of wild populations. However, like some other simplified concepts, MSY has become institutionalized in a more absolute and precise role than intended by the biologists who were responsible for its original formulation."

4. As a concept, MSY is fundamental to how wildlife managers think about the effects of hunting on a population. And simple models can be very instructive to consider how populations may respond to management actions. But MSY is not a model upon which we can use real-world data to make management decisions.