Apparently, some biologists think that outputs from complex computer models will convince native Arctic residents that invasive mark-recapture work has no long-term effect on the health and well-being of polar bears.
A new paper by Karyn Rode and colleagues (which includes Bromaghin and others (e.g. Amstrup) from the population estimate paper), summarized in a USGS press release issued on Monday and published online Tuesday, utilized comprehensive data collected during mark-recapture work carried out in spring (mid-March to mid-May) from 1982 to 2013 in the Southern Beaufort Sea (discussed previously here).
The authors used this data to construct complex computer models which they claim provide “empirical evidence” that their capture of polar bears had “no adverse long-term effects on feeding behavior, body condition, and reproduction.” See examples above and below from the paper.
Here is another example of the work, as described by the authors:
“We used generalised additive and generalised additive mixed models (GAMs and GAMMs, Wood 2006) to examine effects of capture history and collaring on bear body mass and body condition. All modelling was conducted in R, version 3.0.1 (R Development Core Team 2013), using the mgcv. Package (Wood 2006). Model selection was conducted by comparing AIC values between a set of predetermined models (Burnham and Anderson 2002).
Candidate models were formed from various combinations of variables that allowed evaluation of the hypotheses of interest and a fixed combination of covariates for each model set to account for suspected effects that were not of explicit interest (Table 3).”
Oddly, the press release promoting the paper (copied below) and the paper’s abstract (also copied below) mention models only once each. Since the paper is available to subscribers only, the press release and abstract with only cursory mention of the extensive modeling involved in the analyses are all that most people will see — few readers will realize this is yet another piece of complex modeling work passed off as rigorous observational science.
A CBC news story about the study published yesterday made no mention of models at all — its likely the writer didn’t know models were used to “predict” whether tranquilizers and radio collars had any lasting effects on polar bear condition, reproduction or survival. In other words, the authors did not observe a lack of long term effects, they predicted no effects (or statistically insignificant ones) using computer models.
Inuit objections to invasive research are a major impediment to polar bear biologists’ preferred method of research – mark-recapture work — in fact, mark-recapture studies are no longer permitted in Nunavut (although still allowed in Manitoba and Alaska). Clearly presented, straight-forward observational research showing no long-term effects of tranquilizers and collars could have helped dispel those objections. However, this new paper is hardly straight-forward and the promotional material is anything but transparent.
I suspect it won’t be long before these biologists will be told just how far off the mark they have been with their assumption that Nunavut Inuit will find complex model outputs to be convincing “empirical evidence” for lack of long-term harm to polar bears from invasive research methods.
I’m not sure I’m convinced myself.
Rode, K. D., Pagano, A.M., Bromaghin, J.F., Atwood, T.C., Durner, G.M. and Simac K.S. 2014.
in press. Effects of capturing and collaring on polar bears: Findings from long-term research on the southern Beaufort population. Wildlife Research 41(4):311-322. [paywalled]
http://www.publish.csiro.au/nid/144/paper/WR13225.htm Supplementary Info: 10.1071/WR13225_AC Pdf here.
Context.The potential for research methods to affect wildlife is an increasing concern among both scientists and the public. This topic has a particular urgency for polar bears because additional research is needed to monitor and understand population responses to rapid loss of sea ice habitat.
Aims. This study used data collected from polar bears sampled in the Alaska portion of the southern Beaufort Sea to investigate the potential for capture to adversely affect behaviour and vital rates. We evaluated the extent to which capture, collaring and handling may influence activity and movement days to weeks post-capture, and body mass, body condition, reproduction and survival over 6 months or more.
Methods. We compared post-capture activity and movement rates, and relationships between prior capture history and body mass, body condition and reproductive success. We also summarised data on capture-related mortality.
Key results. Individual-based estimates of activity and movement rates reached near-normal levels within 2–3 days and fully normal levels within 5 days post-capture. Models of activity and movement rates among all bears had poor fit, but suggested potential for prolonged, lower-level rate reductions. Repeated captures was not related to negative effects on body condition, reproduction or cub growth or survival. Capture-related mortality was substantially reduced after 1986, when immobilisation drugs were changed, with only 3 mortalities in 2517 captures from 1987–2013.
Conclusions. Polar bears in the southern Beaufort Sea exhibited the greatest reductions in activity and movement rates 3.5 days post-capture. These shorter-term, post-capture effects do not appear to have translated into any long-term effects on body condition, reproduction, or cub survival. Additionally, collaring had no effect on polar bear recovery rates, body condition, reproduction or cub survival.
Implications. This study provides empirical evidence that current capture-based research methods do not have long-term implications, and are not contributing to observed changes in body condition, reproduction or survival in the southern Beaufort Sea. Continued refinement of capture protocols, such as the use of low-impact dart rifles and reversible drug combinations, might improve polar bear response to capture and abate short-term reductions in activity and movement post-capture.
USGS press release 12/15/2014. “New Scientific Study Supports that Capture-based Research is Safe for Polar Bears.”
ANCHORAGE, Alaska — A polar bear capture and release-based research program had no adverse long-term effects on feeding behavior, body condition, and reproduction, according to a new study by the U.S. Geological Survey.
The study used over 40 years of capture-based data collected by USGS from polar bears in the Alaska portion of the southern Beaufort Sea. Scientists looked for short and long-term effects of capture and release and deployment of various types of satellite transmitters.
“We dug deeply into one of the most comprehensive capture-based data sets for polar bears in the world looking for any signs that our research activities might be negatively affecting polar bears,” said Karyn Rode, lead author of the study and scientist with the USGS Polar Bear Research Program.
The study found that, following capture, transmitter-tagged bears returned to near-normal rates of movement and activity within 2-3 days, and that the presence of tags had no effect on a bear’s subsequent physical condition, reproductive success, or ability to successfully raise cubs.
“Importantly, we found no indication that neck collars, the primary means for obtaining critical information on polar bear movement patterns and habitat use, adversely affected polar bear health or reproduction,” said Rode.
The study also found that repeated capture of 3 or more times was not related to effects on health and reproduction.
“We care about the animals we study and want to be certain that our research efforts are not contributing to any negative effects,” said Rode. “I expected we might find some sign that certain aspects of our studies, such as repeated capture, would negatively affect bears, and I was pleased that we could not find any negative implications.”
Efforts to conserve polar bears will require a greater understanding of how populations are responding to the loss of sea ice habitat. Capture-based methods are required to assess individual bear health and to deploy transmitters that provide information on bear movement patterns and habitat use. These methods have been used for decades in many parts of the polar bear’s range. New less invasive techniques have been developed to identify individuals via hair and biopsy samples, but these techniques do not provide
complete information on bear health, movements or habitat use. Capture is likely to continue to be an important technique for monitoring polar bears. This study is reassurance that capture, handling, and tagging can be used as research and monitoring techniques with no long-term effects on polar bear populations.
The paper “Effects of capturing and collaring on polar bears: findings from long-term research on the southern Beaufort Sea population” was published today in the journal Wildlife Research.