Tuesday, April 24, 2012

Making Predictions!

I love this! A testable prediction! I wonder if the Wyoming Fish and Game Department specified what will happen if they are over or under their prediction? Oh, look! They'll use adaptive management (bottom of page 23):
The Department will use an adaptive management approach to employ harvest strategies to meet management objectives.
Hmmm, not many details on what that means however, and the peer review of their plan called them out on it. So on March 12, 2012 they released a clarification of the Adaptive Management plan. Let's see ... compare their plan to the Structured Decision Making checklist:

  1. Problem - Keep wolf numbers above the level that would trigger ESA re-listing and otherwise as low as possible. 
  2. Objectives These are present, and at least some of them are partially SMART, e.g. maintain > 10 breeding pairs and > 100 wolves in the state outside Yellowstone National Park and the Wind River Reservation. This is specific, measurable, a , relevant. Time frame? There's something in the original plan about the total Northern Rocky Mountain population meeting targets over each 3 year period, but not clear from my first read how that steps down to Wyoming. They also have objectives to maintain > 1 migrant per generation between Idaho and Wyoming subpopulations, and to minimize economic losses to the livestock industry. That last one in particular is very very fuzzy. There may be others, but there's no clear "objectives" section that spells it all out for us.
  3. Alternatives The primary alternative that is considered here is variations of quota and season length within the Wyoming Trophy Game Management Area, that is, hunting regulations. There are also variations on depredation permits and translocation is mentioned for helping with gene flow issues. The hunting regulations will be set annually, so this is an iterated decision that is appropriate for Adaptive Management. Apparently Wyoming does compensate land owners for livestock depredation, but higher or lower levels of compensation  doesn't seem to be called out as an alternative. 
  4. Consequences Consequences? There will be consequences to these actions? Nowhere do they attempt to examine how different alternatives will lead to different outcomes for the objectives, and as a result ...
  5. Tradeoffs there ain't no stinking tradeoffs to be made, we can do it all. Actually in the clarification they do point out, at length, that managing wolves to be next to the 10/100 level would reduce their flexibility to do depredation control, etc., so they won't do that. What they will do isn't clear either, but they won't be aiming to be at the minimum. 
Overall, I'd have to grade this as a D-. They're partway there, but it could be much more clearly spelled out. And someone IS predicting consequences, very precisely too, so I'd like to know how. 

P.S. I found it! 52 is the sum of the hunting quotas for the 12 designated hunting areas. So it would be more accurate to describe the prediction as a maximum, assuming that hunting closes down in an area exactly on time and no one accidentally goes over. To be fair to the story in the Ranger, only the headline suggests that 52 is a prediction, in the body of the text they say it is the number hunters are allowed to kill, not the number they will kill. Apparently all other sources of mortality will add up to 46. As far as I can tell, the total number of mortalities for 2012 is based on the assumption that a wolf population can sustain 36% annual anthropogenic mortality before declining - and LO! 0.36 (270) = 97.2, or approximately the total number of expected human caused deaths in 2012.


  1. An anonymous colleague is having trouble posting anonymously, so on their behalf:

    Fuzzy objectives are clearly that by design. I have 2 closely related hypotheses for this, in this instance. First, if an objective has a quantity associated with it, then it clearly can be defined as a success or a failure. If that objective is not meant, guess what? Costly litigation! So what would be the motivation for quantifying certain objectves? Second, the lack of a measurable objective also translates into maximizing flexibilty with management decisions, at least related to that objective. One simple event, especially if picked up by the media, can change everything in a hurry. What does a state agency do if they are meeting all objectives, but social or political pressure is intense to implement a change? The agency may then be force to ignore that objective. That could also result in a loss of management control. Despite these obvious observations, surely much thought is put into alternatives, consequences, and trade-offs even if this is not apparent in a formal management plan.

  2. Yup. All of the things you mention fall under the heading of what Sarah Michaels and I call "social indeterminism", and the high levels of social indeterminism associated with carnivore management are what make SDM/AM difficult or impossible. Rather than blythe references to Adaptive Management, I believe that in such circumstances the politics should be grappled with directly.