Southern resident growth rate ambiguity

About a month ago Orca Relief issued press releases related to the population dynamics of southern resident killer whales.  I spent a little while then marveling at how their new leader Bruce Stedman seemed to have cherry picked data with which to support a claim that the population was in precipitous decline.  This morning I took some time to revisit those plots in light of the equally disturbing statistics and graphs presented during the orca-salmon workshops conducted by Will Stelle over the last couple years (live blogs with audio and/or video here — workshop 2; workshop 3).

Basically, somebody who is statistically-talented and politically-independent needs to clarify for the conservation and stewardship community the state of the Southern Resident killer whale (SRKW) population.  The same person or entity should re-interpret the data whenever we get an updated census (from the Center for Whale Research [CWR] which — as an aside — recently put its NOAA-funded photos and census data behind a membership paywall).

The main misleading aspects of the Orca Relief plots (see below) are that they only went back to 1997 (though we have data back to 1971) and they focused on decreases in juvenile and reproductive females that were drawn from maxima in the time series, rather than being fit to the full range of data.  This is an incomplete statistical presentation, and one that is potentially misleading, just as using stacked bar graphs can cause confusion in any science presentation.  Below are two examples.

Female SRKWs of (undefined) reproductive age by pod.  Stacked bar graphs make it difficult to see and compare trends in the K and L populations; only the J and total population trends can be easily discerned.

Juvenile female SRKW population time series as presented by Orca Relief.  The rate of decline from 2001 to present is an overestimate of decline compared with one made using 1997 to present, or a best linear fit to all available data.  More importantly, why have the previous, available 22 years of data been excluded?

For comparison, here is a plot of all CWR census data (1971-2012) by pod as presented in the 2012 Puget Sound Partnerships orca vital sign:

CWR SRKW population data by pod presented by the PSP.

This shows how potentially misleading it is for an organization like Orca Relief to showcase trends using only data since 1997!  If we use 1971 as the baseline, J and K pod appear to have grown slowly, while L pod’s growth is positive but lower.  It also is apparent that the population dynamics of L pod, more than J or K, drive the sudden changes in the total population, which is growing on average over the whole time series.  As the 2012 PSP’s orca vital sign reports:

“…as of July 1st 2013, the size of the population was 82 individuals, down by four whales relative to the 2010 baseline reference of 86 whales… Although there has been no progress made since 2010, the population has been growing, albeit slowly at about 1% per year, over the longer term (1979 to 2010).”

The stories run by Q13FOX, the San Juan Islander and Island Guardian and picked up by KOMO news didn’t concern themselves with such statistical details.  However, a later piece in the San Juan Islander about reproductive (mature) male SRKWs quoted NOAA’s Eric Ward who argued that the recent decline in mature males is part of the population’s historic fluctuations, not a sign of imminent collapse as suggested by Orca Relief in the Island Guardian.  Both pointed out the good news: that juvenile male populations have been rising recently.

Mature male time series portrayed by Orca Relief (1997-2013)

Juvenile male increase graphed by Orca Relief

Eric Ward’s plot of mature male SRKW population (1971-2013)

As an aside, I was impressed that the San Juan Islander provided references to some of Eric Ward’s papers.  What wasn’t referenced though, were the many discussions of SRKW population trends in the 3 NOAA-sponsored workshops on “The Effects of Salmon Fisheries on Southern Resident Killer Whales.”  In the third workshop, it became clear that Ward presented estimates of population growth rates for the SRKWs (basically +1% annual growth rate) which were at odds with estimates made by his Canadian modeler-analog Antonio Velez-Espino (basically -1% annual growth rate).   (Velez-Espino chose to exclude pre-1987 data, stating that his choice — a compromise between extent and quality of data — was most representative of the current population ~25 years or one generation back.)

So, is the SRKW population in slow decline or slow recovery?!  How concerned should we be with the latest downturn in the overall population?  What does it mean for the recovery process if the population continues to fail to meet the recovery plan goal of maintaining 2.3% annual growth over a period of 14 years?

The final report from the orca-salmon workshops (Hilborn, Cox, Gulland, Hankin, Hobbs, Schindler, Trites, 2012) has some exemplary figures that shed a little more light on the situation, particularly the trends in the population of juvenile females (J pod increasing, L pod decreasing).  Note that the break points between Orca Relief’s and the Science Panel’s juvenile/young and breeding/reproductive age classes are different.  Nevertheless, it’s worth focusing on the  juvenile and reproductive female age groups because they have a strong influence in population models like Velez-Espino’s.  (He stated in the final workshop that the greatest increase to the modeled population growth rate comes from increasing survival and fecundity of young reproductive females.)  Below are figures for “young” (<21 year old) and “young and reproductive” (<43 year old) populations (not directly comparable to the age groups chosen by Orca Relief, but similar).

Young (<21 year old) SRKW population trends and sex ratios.

Young and reproductive SRKW population trends

So, out of time again, I’m left wondering: should we be panicking as the strange year of 2013draws to a close?  What do you think?

References:

 

 Hilborn, R., Cox, S. P., Gulland, F. M. D., Hankin, D. G., Hobbs, N. T., Schindler, D. E., & Trites, A. W. (2012). The Effects of Salmon Fisheries on Southern Resident Killer Whales Final Report of the Independent Science Panel (p. 87). NMFS & DFO. Retrieved from http://www.nwr.noaa.gov/publications/protected_species/marine_mammals/cetaceans/killer_whales/esa_status/kw-chnk-final-rpt.pdf 
 

5 thoughts on “Southern resident growth rate ambiguity”

1. Scott Veirs Post author

Thanks for the baseline perspective on population trends, David. I hadn’t thought about all of the potential reference points, nor had I pondered whether the New Carissa might have affected L pod in 1999-2000.

Reading about that spill along with Fred Felleman’s excellent Crosscut piece makes me wonder whether we shouldn’t focus the bulk of our current attention (and the attention of Blackfish audiences) on preventing oil tanker traffic in SRKW critical habitat. It won’t matter if Chinook populations are re-bounding if we do to the SRKWs what we did to the Prince Williams RKWs.

Despite still being unsure of whether to panic or not, Susan and Howie have again reminded me how we might better facilitate recovery if we were more coordinated. I wish there was a council of elders (matriarchs?) that would re-assess each new population or demographic data point for us (with the best available model and a range of sensible baselines) and also initiate a ~monthly action campaign designed to affect the greatest possible reduction in extinction risks for the SRKWs. This month it might focus on oil spill risk reduction. Next month it could target a particular Chinook population that would feed SRKWs were it to recover. And so on…

Indeed, the point was made in the last orca-salmon workshop (by Mike Ford, I think) that perhaps SRKW recovery should be defined based on reducing extinction risk to a target level, rather than maintaining a growth rate, or a crossing a population threshold.

2. Scott Veirs Post author

Copying a Facebook comment by David Bain here, for the record:

“When discussing Southern Residents back in the 90’s, Paul Wade pointed out that a population (of any species) around 100, whether increasing or decreasing, is in trouble. A major issue in how badly the population is perceived as doing is the shifting baseline problem. Orca Relief picked a baseline early in the decline, when it was near its maximum for the last 50 years. Instead, one could pick the end of the capture era and talk about how the population has increased. One good go back to the start of the capture era and talk about decline. I suspect if we could go back 100 years, we’d talk about a substantial increase over the last 100 years. If we went back 200 years, we’d talk about how the population has been decimated. So, the real questions are: what time frame reflects the status quo; based on the age-sex structure of the population, what is its reproductive value; based on time in the genetic bottleneck, how much genetic variation is left? Finally, can we expect things to change for better or worse? Salmon populations have always exhibited natural fluctuation. Major fluctuations are tied to the Pacific Decadal Oscillation. The decline in the 90’s, corresponded to an unfavorable phase of the PDO. We saw signs of recovery when the PDO became favorable. However, that recovery has proven temporary, as the population has dropped again. We’ve been in the favorable phase long enough that things are likely to get worse in the near future, although it’s impossible to predict whether it will flip back in a couple years or a couple decades. An increase in vessel traffic also occurred during the 90’s decline. Whale oriented traffic leveled off after the 9/11 attack, but is likely to increase as the economy recovers. Shipping traffic is also likely to increase as the economy recovers. Fish oriented traffic is likely to continue to track fish abundance. As with fish abundance, disturbance is likely to get worse in the near future. As for toxic chemicals, DDT and PCBs were banned about 40 years ago, and it’s likely that the peak of this problem is occurring about now, meaning things should start getting better there. However, PBDEs may mask that recovery, as their levels are rising and are likely to continue to do so for many years. We also need to worry about catastrophic events. For example, did the New Charissa kill L Pod whales, explaining a peak in mortality at that time? Did the Cryptococcus outbreak that killed many harbor porpoises also kill a few Southern Residents? These issues all have timelines associated with them, and their past influences what we should select as the baseline. How we will manage them in the future influences how much panic we should exhibit. Orca Relief’s contribution is suggesting we look at age-sex structure. Fisher defined the concept of reproductive value. In essence, males cannot give birth, so they do not contribute (mathematically) to the reproductive value of the population. Similarly, post-reproductive females have no (mathematical) reproductive value. For the rest of the females, their reproductive value depends on age. A female about to give birth for the first time has maximal reproductive value. Younger females have less, because they may die before they start reproducing. Older reproductive females also have lower reproductive value, because there is a limit on how many more calves they can have before becoming post-reproductive or dying. An impartial analysis would need to go through the females on a case-by-case basis and calculate the population’s reproductive value. Another of Fisher’s contributions was showing that the population should invest equally in male and female offspring. Subsequent work by Hamilton and Trivers explored conditions under which deviation from 50:50 would be adaptive. Trivers found maternal condition could be important. That is, if we’ve seen a change in sex ratio at birth, that’s likely to be a symptom of a rapid change in female condition on a population scale. As a practical matter, if female condition results in a male-biased sex ratio at birth, the potential for population growth goes down. I suggested in my dissertation that females are the more expensive sex, so females in poor condition are more likely to have sons, but that’s an issue that’s worth further review rather than immediate acceptance as fact. As for genetic variability, it’s likely Southern Residents started out with relatively low genetic variability. They’d be a peripheral population founded by a small number of individuals drawn from a larger population. I.e., they’re a subspecies within the resident species, and the variable part of the species seems to be centered just south of the Bering Sea. When you get out to BC’s Northern Resident Community and the Southern Resident subspecies, there’s always been limited variability. Southern Residents have probably been in a further genetic bottleneck for about 6 generations, due to destruction of salmon spawning habitat in the California gold rush, overfishing made possible by the invention of salmon canning, and salmon habitat loss due to dam building (all dating back to the mid-1800’s). Pragmatically, paternity tests suggest only two males are fathers of a high proportion of juvenile southern residents, meaning inbreeding depression is probably not far away. As for the future, some things like vessel traffic, fishery management policies, and toxins are subject to human control. Others, like disease, ocean acidification, and climate change are bigger problems than we know how to control–we can only control the degree to which we choose to mitigate them. So, should we panic? The article Scott cited might have had a more concerned tone if it used current information. It used the 2012 population count, so missed the drop to 82 in the 2013 count. It’s missed the two additional losses since then, which dropped the population to 80. We’ll need to re-evaluate after the 2014 count. It may go up from 80 if calves are born over the winter, or drop if we have more deaths than births. I’d suggest we at least get a sense of urgency. Instead of the policy of wait and see on the Snake River dams that Bush and Obama have been offering to the courts, we need to start the process of tearing down those dams immediately to increase salmon habitat. Instead of spending 17 years removing culverts that block fish passage, we should probably try to replace all of them within about 5 years. We should eliminate salmon farms immediately. We should scale back oil shipping rather than consider expanding it. We should prepare to intervene in disease outbreaks. We should look into ways to reduce disturbance. We should make Hood Canal critical habitat, and figure out how to get Southern Residents using it again. We should not experiment with tidal and wave energy in critical habitat. We should get serious about cleaning up toxic waste sites, eliminate toxins at their source, and set stormwater toxin levels based on the assumption that whales (and many people) eat lots of fish. We should start taking action as soon as we identify a problem, instead of waiting until we know exactly how bad a problem is, or hoping we can get recovery by solving a different problem instead. That’s a long to-do list. Some will say that it’s too many things to focus on, and activists will be spread too thin to make progress on any of them (panic!), or give up on southern residents altogether and go after Sea World instead (displacement). Others will see lots of room for improvement and get to work (hope!).”

3. Lisa Cunningham

Thank you, Howard for explaining what the Center for Whale Research has done and why. I recently became a member of the CWR and the Orca Network in order to support efforts to document the SRKW population.

4. Howard Garrett

Just as an aside to the aside that the Center for Whale Research recently put its NOAA-funded photos and census data behind a membership paywall, the context for the Center’s attempt to raise money through memberships by offering a membership benefit was a major cut in that NOAA funding, ironically. Ideally people would support the Center without the need for a paywall.

1. Scott Veirs Post author

Thanks, Howie, for the insight. Beam Reach and I became CWR members today, so I support the idea that if the research is supported by private donations, access to results could be restricted to the donors. It’s the topic of another post, really, but in pursuit of more open science, I’m trying to understand NOAA’s evolving access policies. In my ideal world, research funded publicly should be made promptly available to the public — both as results and as raw data. Sadly, the historic CWR census data that were funded (at least in part?) by NOAA are very difficult to find on the Internet.

Comments are closed.