Grizzly Twister and Other Games that Scientists Play
by David Mattson
Those of you who were born after 1970 probably don’t remember a game made by Milton Bradley called Twister that in 1966 exploded onto the national stage and into America’s living rooms and dens. Its popularity was considerably enhanced when Johnny Carson leaned scandalously over Eva Gabor’s partially clad buttocks during a live demonstration of the game on the Tonight Show. The game plays out on a large mat sporting multi-colored dots laid on the floor. The players are the figurative game pieces, contorting themselves as they place hands and feet on dots according to the dictates of a spinning dial. The winner is the one who stays upright, the loser the one who tips over in a twisted knot of limbs.
This game often comes to mind as I listen to or read quotes from Yellowstone’s grizzly bear researchers and managers, although at one level their mantra is pretty straight-forward. We’ve got ever more grizzly bears in ever more places being driven farther out onto the edges of the ecosystem by saturation of carrying capacity in the center. Moreover, carrying capacity is a static parameter creating a context within which the ever-adaptable omnivorous grizzlies are readily switching diets to fully compensate for changes in comparative availabilities of foods, drawing on a repertoire of more than 200 roughly equivalent foods.
At one level this official rhetoric can be understood simply as a brew of misrepresentation, disingenuousness, and partisan selectivity. As a literal matter of fact, carrying capacity is not static; the quality and quantity of Yellowstone’s grizzly bear foods vary by orders of magnitude; continent-wide, substantial differences in foods drive similar orders-of-magnitude differences in grizzly bear densities; nor are grizzly bears Latin taxonomists. Regarding this last point, the putative laundry list of >200 bear foods in Yellowstone was concocted by researchers based on phylogenetic differences among foods codified in Latin names. The problem is that such distinctions don’t matter to bears, whereas the many-fewer functional differences in nutritional quality and architecture of foods do.
These counter-factuals are perhaps so obvious as to beg the question of what motivates researchers and managers who spout the official rhetoric—and who should know better. Without belaboring the point, none of the explanations I’ve been able to come up with so far are very flattering.
Claims about Population Size and Trend
But my focus here is on claims of ever more bears in ever more places being driven to ecosystem extremities by progressive saturation of carrying capacity.
Perhaps the first point worth reiterating is that carrying capacity varies—potentially by a substantial amount—and has likely declined in the ecosystem core with losses of cutthroat trout, whitebark pine, and elk. All three of these foods had been high-quality dietary mainstays for Yellowstone’s grizzly bears, but are all now much diminished—cutthroat trout to the point of functional extinction as a bear food. Thus, even if true, saturation of carrying capacity has more likely occurred because of a lowering of this figurative ceiling rather than because of burgeoning bear numbers. And any changes in habitat have almost certainly been compounded by the reorientation of bears to alternative high-quality foods disproportionately concentrated on the ecosystem periphery; e.g., livestock and army cutworm moths. The data for increased consumption of moths and terrestrial meat are unambiguous, as is the sky-rocketing increase in human-bear conflicts over livestock.
But, more to the point of this piece, I find the numerous potentially contradictory statements made by agency spokespeople about population size and trend to be especially intriguing. For simplicity’s sake, its’ worth looking at a few illustrative examples from Frank van Manen, head of the Interagency Grizzly Bear Study Team (IGBST), and the man charged with overseeing grizzly bear research in the Yellowstone Ecosystem.
He has said “The population has been at a pretty stable level since the early 2000s,” in the same article as “…there’s no doubt that we are now at a point that we have the largest population size.”
He has likewise said that “grizzly numbers outside the monitoring area aren’t counted” (the “monitoring area” being the Demographic Monitoring Area [DMA] that defines the bounds within which bears are counted), while invoking bears outside the monitoring area to claim “the largest number of bears ever.”
During a recent meeting in April 2019, he was seen pointing with agitation to a graph showing that grizzly bear mortalities had mounted exponentially during recent years, while saying that “mortalities were not unusual last year, and exceeded by death rates during 2008.”
Meanwhile the press is littered with his quotes variously asserting that the distribution of Yellowstone grizzlies has expanded by 11%, 34%, 38%, and even 300%, referencing vagarious temporal benchmarks.
He and the US Fish & Wildlife Service have also repeatedly claimed that a certain method used for estimating population size and trend (the so-called Chao2 method) is “the best available science,” while in near the same breath invoking other unspecified but presumably more deficient methods to claim that Chao2 under-estimates population size by “a lot” or “by 40 to 50 percent.”
You should be confused. I’m confused. Perhaps the intended outcome is for a confused public to yield to the confusion and surrender their trust and judgement to the presumed experts—the only winners left standing in the game of Grizzly Twister.
Nonetheless, there is an imperative to penetrate the veil of obfuscations and contradictions to help foster a truly informed rather than faith-based public conversation about how best to conserve our iconic Yellowstone grizzly bears. So, in service of that end, what follows is my best shot at penetrating the tangled official rhetoric, and elucidating what’s probably been happening with size, trend, and mortality of Yellowstone’s grizzly bears…
Imposition of a Political Boundary
A main source of confusion arises from the fact that the IGBST estimates population size and sex-specific mortality rates only for the Demographic Monitoring Area (DMA), and not for any areas outside. Estimates of population size are derived, at root, from estimated numbers of reproductive females within the DMA, which are then multiplied by various more-or-less static factors to account for other sex and age classes of bears.
Leaving intrinsic problems with methods aside for the moment, up until recently the DMA encompassed all of the sightings of reproductive females and nearly all of the documented grizzly bear mortalities. But during recent years an increasing fraction of both mortalities and sightings of reproductive females have occurred outside the DMA, leaving a presumably increasing but unknown fraction of the total population unaccounted for in estimates of population size and mortality rates that are limited to the DMA.
Importantly, estimated numbers of reproductive females have been more-or-less static within the DMA since roughly 2002, hence a population at a “pretty static level.” I’ve shown trends in numbers of reproductive females inside the DMA derived from two different methods in figure 1 below. Chao2 represents the presumed “best available science,” whereas MR refers to the less esteemed Mark-Resight method—the apparent (although unacknowledged) basis for claims that there may be 40-50% more bears in “the population” than indicated by the Chao2 method.
Outside the DMA, there are scant grounds for determining how many bears might be unaccounted for, although agency researchers and managers seem to be using aerial extent of the population's distribution outside the DMA as a surrogate--despite the fact that bear densities are certainly lower on the periphery and comprised of a greater proportion of males. Regardless, these unaccounted-for bears seem to leave agency spokespeople ample latitude for claims that we have “record numbers” of bears.
As a bottom line, we don’t, in fact, know total size of the population, or what the trend has been, simply because the bounds within which numbers of live bears are tracked were arbitrarily truncated at the DMA through a process that had more to do with partisan politics than bear biology. What we do know is that numbers of reproductive females within the DMA have been static for the last 15 plus years; that the DMA accounts for most—a super-majority—of bears in the ecosystem; but with the proviso that total population distribution is increasingly extending beyond the DMA.
A Compounding Problem
An additional and on-going source of confusion arises from the fact that researchers and managers have changed the fundamental basis for estimates of trend since 2002, but with each successive change leading to an artificial inflation of population estimates over what had come before. More specifically, trend was tracked on the basis of estimated MINIMUM numbers of bears up through 2006, accounting for the totality of the distribution. Beginning in 2007, trend was tracked on the basis of TOTAL population size, still accounting for all bears in all areas.
But with this change came the chronic problem of managers claiming a substantial increase in the population, but based almost solely on comparing the apples of a minimum estimate with the oranges of a total estimate. This conflation has been egregious and obvious enough to, again, beg the question of motives.
Nonetheless, a further change was introduced in 2012 when the IGBST started using a larger multiplier to account for numbers of males in the population. So, here again, we had a seeming instantaneous increase in estimated numbers of bears, but based largely on a change of method and the addition of a putative burgeoning of demographically extraneous males.
Moreover, beginning a few years later, estimates neglected what turned into an increasing (but small) fraction of the population ranging outside the DMA, which introduced yet more confounding variation in any estimates of trend. I show all of this in figure 2 above, along with a rough estimate of trend spanning all of these changes in methods and bias.
The upshot is that we have essentially no basis for estimating trend of the entire population from estimates of either minimum or total population size for the period 2002-present. Likewise, unless the known and probable biases are corrected, there is essentially no basis for estimating so-called mortality rates for the total population, especially in ways that are comparable from one era to another.
What’s Been Happening with Mortality?
Frank van Manen and Wyoming’s wildlife managers offered the public a stereo dose of rhetoric during summer and fall of 2018, insisting that numbers of grizzly bear mortalities and conflicts were “not unusual.” This claim was essentially reiterated by Dr. van Manen in April of 2019 when he told grizzly bear managers assembled for their spring meeting that death rates last year were normal, not alarming, and less that what had been previously observed during 2008. On the face of it, such claims must have been profoundly confusing for any consumers of mass media given that rote numbers would suggest that 2018 was indeed a record-breaking year.
So what was going on?
The data are pretty unambiguous. Figure 3 below shows numbers of documented grizzly bear deaths by year, with trend indicated by a burgundy line. I’ve shown deaths for the entire population with pink dots, and deaths recorded inside the DMA with gray dots. By either reckoning, 2015-2018 were alarming years. Considering the entire population, 2018 was indisputably record-breaking, with mortalities during 2015-2017 at levels higher than any before. Even limiting the scope to deaths recorded within the DMA, the trend is dramatically upward.
The most charitable way to characterize statements by Yellowstone’s grizzly bear researchers and managers on the topic of grizzly bear mortalities is…disingenuous.
The Problem of Death Rates
But, to be fair, statements about death rates introduce a level of complexity and nuance that could conceivably accommodate some of the claims made by van Manen during spring of 2019 given that these so-called rates include estimates of both dead bears and live bears—so, a ratio or rate. Again, leaving problems intrinsic to methods aside, it is possible to estimate rates in a way that not only allows for comparisons among the last 15 or so years, but also accounts for total population size--without being bound to truncations imposed by the DMA.
We know numbers of documented mortalities for the entire population inside and out of the DMA and, from that, can come up with a rough estimate total mortality that accounts for undocumented deaths. The tricky bit is coming up with a comparable estimate of total population size that accounts for not only what I would consider to be the unjustified inflation that occurred around 2012, but also for the fraction of the population outside the DMA.
I was able to arrive at an estimate of total population size that included bears both inside and out of the DMA using some plausible adjustments to account for the bears outside.[1] I’ve shown the resulting estimates of total population size for 2007-2018 by gray dots and a green tend line in figure 4 below. Notably, these estimates are not limited to the DMA. The red dots and connecting pink line show estimated annual mortality rates for the entire population based on a simple ratio of total estimated deaths to total estimated live bears.
So what about death rates for the entire rather than truncated population? Death rates for 2015-2018 were likely higher than any recorded during recent decades, including during 2008, with 2018 clearly breaking all records dating back to at least 2007. Interestingly, the down-turn in total population size evident after 2014 is entirely plausible given that it coincides with the dramatically elevated death rates during 2015-2018 in comparison to 2013-2014.
Parenthetically, I’ve shown in shades of yellow in all of the figures above the period following when we lost most whitebark pine in the Yellowstone Ecosystem—a catastrophe that unfolded mostly between 2000 and 2009. The dramatic escalation of grizzly bear deaths in the aftermath is stark, along with the problematic change in methods that obscured the probable substantial demographic changes that ensued. This timing for change in methods was probably not a conspiracy, but unfortunate nonetheless.
Taken altogether, these results render absurd the claims about deaths and death rates made by Dr. van Manen and Wyoming’s wildlife managers. Likewise, their claims about population size are disingenuous at best, especially if applied to the totality rather than the fraction partitioned off by the DMA for political purposes.
A Concluding Thought
For good or bad, demography has ended up being central to the on-going political debate about the future for Yellowstone’s grizzly bears. Perhaps as a direct consequence, the relevant science has been badly politicized, with implications not only for grizzly bears and the people who care about them, but also for the integrity of relations between science and society. The game of Grizzly Twister is not an innocent one.
Regardless of whether I interpreted the data here correctly or not, it is incumbent on IGBST researchers to make the science and its limitations as transparent and understandable as possible to a lay audience. Yet I know this will not happen, if for no other reason than the corporate mindset of Dr. van Manen’ current employer, the US Geological Survey (USGS). This agency treats production of science as tantamount to a for-profit business, with the profits buried in book-keeping and largely devoted to off-setting declining budgets as well as a bloated administrative bureaucracy.
The upshot, though, is that people such as Dr. van Manen see state wildlife managers and the US Fish & Wildlife Service literally as “customers,” and, because of that, “the product” ends up being a scientific endeavor designed to serve the customers’ needs. In this case, the customers clearly expect timely delivery of a scientific justification for removing Endangered Species Act (ESA) protections from Yellowstone’s grizzly bears.
In addition to pervasive disinformation, evidence of corrosive influences is not too hard to find. If government researchers were simply “objective scientists,” they would have remained focused on the totality of the population, not an expedient fraction parsed out according to the political boundaries of the Demographic Monitoring Area. More than that, these researchers would not have presumed to offer interpretations of the ESA, or actively supported trophy hunting and trucking bears among ecosystems to address genetic concerns—as Dr. van Manen has done. And the list goes on.
In the case of grizzly bears, our taxpayer-supported scientists have not been acting as honest brokers of scientific information. This cankerous problem is compounded by the uncritical embrace of their platitudes and disinformation by managers, undiscerning journalists, and a vulnerable public. Until this root problem is corrected, we will continue to be at the mercy of decision-making driven by corrupted science.
[1] My approach to this challenge was relatively straight-forward. I multiplied the DMA-specific estimate of total population by a factor proportional to the fraction of the distribution outside the DMA each year as well as by a factor proportional to the fraction of unduplicated females with COY annually documented outside the DMA, and then took the average. The inflation based on distribution was probably a high bound, and the inflation based on females with COY probably a low bound, which leaves the average as being likely. I moreover took an average of the male-specific multipliers assuming a 0.64:1 and 1:1 sex ratio for application to the entire 2007-2018 period given that the major presumed increase in male survival after 2012 is prima facie implausible given what was, in fact, a substantial comparative increase in numbers of male deaths—the topic of perhaps another essay.