Dear Dr. Yavitt, Thank you for your question. Hatchery salmon do have higher energy content at release than wild salmon of the same year, although they lose body condition after release as they struggle to adapt to feeding in the wild (Sturdevant, M. V. et al. Environ Biol of Fish 94, 101–116 (2011). Although humpback whales are known to feed on juvenile wild salmon, more common prey sources in this area are krill and forage fish like herring, capelin and sand lance. Energy content is certainly something I will be looking at (via bomb calorimetry and literature values) but other factors like prey density and escape ability are also important factors for humpback whale efficiency because they determine how many individuals a whale captures per feeding lunge (crucial for a large predator and tiny prey). I suspect the annual predictability of this prey source, its shallow distribution and predator-naive fish all aid feeding efficiency. Still, humpback whales do not aggregate in large numbers at hatcheries like they would at seasonal herring spawning events, for example. In Alaska, hatchery salmon are differentiated from farmed salmon that spend their entire lives in captivity. This distinction is important because salmon farming is illegal in Alaska.

Dear Dr. Bhatttacharya, Thank you for your question. It might be possible to do that, but the Marine Mammal Protection Act prohibits the feeding of marine mammals. The National Marine Fisheries Service defines feeding as “offering, giving, or attempting to give food or non-food items to marine mammals in the wild” (NMFS “MMPA Regulations (50 CFR 216).”http://www.nmfs.noaa.gov/pr/laws/mmpa/policies...., p9) That is why I am comparing hatchery prey sources to prey sources that are available naturally in the environment. In the long run, what you propose could cause some problems such as bringing more whales in to the area around the hatcheries leading to the need for releasing an ever-increasing amount of distraction food. In addition to protecting hatchery salmon, our goal with this project is to discourage humpback whale reliance on human-produced sources of prey, which could disappear if hatcheries cease being profitable. I believe that reducing the potential foraging efficiency for whales at hatcheries is the best way to protect whales and fisheries in the long-term.

Dear Dr. McGarvey,

Thank you for your comments which, as you will see, don’t lend themselves to brief responses! So one quick clarification: Hatchery salmon are not farmed salmon. Farmed salmon are raised to adulthood in captivity whereas these fish are released. Its an important distinction in Alaska where salmon farming is illegal.

As for your first question, humpbacks are feeding on juvenile salmon whereas I believe killer whales would only target larger salmon. In the two years of hatchery observation data I have processed I have not seen any of the hatcheries comment on killer whales in the area. However there are many other predators on these salmon fry including gulls, eagles, seabirds, seals, sea lions, minks, river otters, porpoises and other fish all of which the hatchery staff record during their twice-daily observation periods. I am excited to analyze this data further to see if a few key predators, such as whales or sea lions, change the populations of other predators. While I haven’t documented any overtly negative or positive interactions among these predators, it is certainly possible that killer whales could at some time in the future stake out hatcheries waiting for seals or sea lions to come to feed. Sea lions and seals could reduce the predation by predatory fish.

I had heard about the sea lions on salmon ladders in the Colombia River. Unlike in that system, our salmon are not endangered and hatcheries do not serve a conservation function. In terms of competition between the predators drawn together by hatchery releases, this doesn’t seem qualitatively different than competition that would emerge at any shared prey source, anthropogenic or wild. That is to say, none of these species has a competitive advantage resulting form the fact that the prey source is introduced by humans or in the example you gave, artificial aggregation at salmon ladders.

As for the second question: It helps to know that foraging opportunities for humpback whales are far from homogenous. Humpback whales are generalist predators feeding on a wide variety of prey types (krill, small schooling fish and adult herring). So it isn’t a simple one-to-one comparison. True the hatchery fish are more energetically dense and shallower than some prey options and this offers advantages. I would add to this list of advantages that hatcheries provide a relatively predictable food source in space and time (reducing “search time”) and “predator naive” prey. However there are disadvantages to feeding on hatchery salmon as well. The biomass released from hatcheries is relatively small compared to some wild prey patches I have observed (piles of herring 12 stories high!). While they are shallow, they form a rather diffuse layer unlike humpback whales’ more common prey. Prey density is probably the most important factor energetically for humpback whales because as enormous filter feeders targeting tiny prey, they have to capture multiple individuals to compensate for the energy required to overcome drag and make a single open-mouthed lunge through the water. First quantifying and then weighting the relative importance of all these prey patch characteristics (density, depth, avoidance, energy content, persistence, predictability) at a diversity of foraging opportunities (krill, herring and hatchery salmon) is far from trivial! I am glad you think it is a cool project.

Great answer, Ellen. Thanks for clarifying a few points, and for engaging the meat of my questions. (I deliberately tried to challenge you.) You’re lucky to be working in the SE AK region. It’s been some years since I’ve spent time up there. Gotta’ find a way to get my 15 month old twins up to Sitka. . .

I feel fortunate every day to be living and working up here. Sitka is a rather kid-friendly town. My sister brought my 3 year old niece up here last summer and she had a blast. Tide pooling, berry picking, the raptor center and sea otters were all big excitement.

Dear Dr. Anderson,

I am so glad you enjoyed the presentation. Thanks for the opportunity to explain my plan for analysis a little further.

First: why use a model?
The difficulty with evaluating foraging efficiency for humpback whales is that they are generalist predators and feed on so many different kinds of prey (krill, small schooling fish and adult herring) with many different prey patch characteristics (density, depth, avoidance, energy content, size, persistence, predictability), all of which should in theory affect a whale’s foraging efficiency. I will collect data on all of these characteristics for each patch and calculate the whales’ foraging efficiency at each of these patches based on the prey characteristics and our behavior data from the tags that will help me calculate the energetic output. That way, I can compare foraging efficiency across observed prey patches. But what about unobserved prey patches? For this, I need to build a model. I can use the efficiencies calculated from known prey patches to predict the energetic efficiency at hypothetical prey patches. Then I can ask questions like: what is more efficient for a whale, a prey patch that is dense but deep or one that is diffuse and shallow?

So how does it work?
Maximum likelihood estimation is a way of taking observed data and inferring a relationship between the predictor variables (prey patch characteristics) and the response variable (foraging efficiency). This relationship is expressed in the form of parameter values for each predictor. The parameter values are the result of the computer finding values that result in the smallest amount of statistical error between what the model predicts and what was actually observed. It is basically making a best-fit line in n-dimensional space (n is the number of predictor variables that you have plus the response variable).

Thank you- you have an excellent way of demystifing abstract concepts!

That is such a nice thing to hear. I have been working on that particular skill as I hope to be a professor someday.

Dear Volker,

Obviously, this work has only been possible because of coordination with the hatcheries. The collaboration began when they approached Jan Straley, one of my committee members and my supervisor at the time. Since this behavior had not been quantitatively documented, I put together some observation sheets and distributed them to hatcheries along the eastern coast of Baranof Island. The hatchery staff have been collecting data during releases for 4 years. I compiled the first year of this data and presented it to them as a scientific report. Hatchery staff have been very helpful in sharing their knowledge of how their fish behave. Bart Watson, the General Manager of a private nonprofit hatchery organization called “Armstrong-Keta”suggested that salmon are perhaps more likely to scatter when threatened than other prey like herring that ball up. Armstrong-Keta’s hatchery manager, Ben Contag, offered the observation that gull and fish predation is probably picking off the smaller, weaker, or sicker salmon that would not have survived to spawn anyway. (Return rates vary from 0.05%-10% depending on year and species). This got me thinking that humpback whale predation might be more detrimental to return rates than other predators per salmon consumed because whales are filter feeding prey somewhat indiscriminately rather than picking off sick individuals. Watson also serves as a co-investigator on a funding proposal through Alaska Sea Grant. The hatcheries have also been generous in supplying housing and seat fares on float planes for me when I come down for observations.

As for how our results have been received. At hatcheries so far I have been able to photographically identify one animal feeding at the hatchery over several years. The staff at hatcheries have begun taking pictures of their whales to see if they can determine how many individuals are feeding near hatcheries. Most of our tagging and prey mapping has occurred outside of hatcheries because of the opportunity to collaborate with other researchers doing similar work on wild prey in the area. However, my collaborators and I did attempt to tag a whale for a few days at Hidden Falls this past April. Unfortunately due to weather and the whale’s unpredictable behavior, we were not successful. (The video makes it look easy, but it can be quite challenging.) Our Alaska Sea Grant proposal, if funded, will allow me focus my efforts at hatcheries next field season.

Yeah! Thanks Amell!

Stephanie, you are not the only one! Imagine how the hatchery staff members feel!

Thanks Jan!

Thanks, my fellow grad students were enthusiastic actors! (free pizza helped)

When I first saw that animation software I was totally amazed. We get so used to seeing just the tiniest sliver of a whales’ back before it disappears into the unknown. This new technology is really drawing back the curtain.

Thank you!

Thanks, Dad :)

Mario kart, whale edition. Sounds like a money maker to me.

Thank you

Thanks, Troy.

Thanks Kristine!

Thanks for taking the time, Ben!

Thanks! Tell Levi he has full creative rights to make a rap about it.

Thanks Leah!

Thanks for taking time time to check it out.