Tuesday 17 January 2017

The first detailed study on the social structure of Icelandic killer whales

by Sara Tavares

It was the 21st of July 2015 and we had been in the water for 2 hours observing a group of whales. We were nearing the end of our field season and had been observing this particular group of whales for the last days. The group included one adult male, 3 females/subadult males and 2 juveniles which had been travelling and feeding in close association. As the day drew to a close we returned home and wondered if we would see them again the next day. I was puzzled, since I remembered these whales from the winter season the previous year and I never saw them together. The next day we found some of the same whales but others had been replaced. The group was now composed of 2 adult males, 4 females/subadult males and 2 juveniles. Only 4 of the individuals were the same as in the first original group we had observed. To our surprise, the day after we saw the 2 missing whales from the first original group with completely different whales. What was going on here?…




Understanding the social life of killer whales requires several observations of the same individuals over time, so that patterns of association can be seen. This type of long-term data is essential to understand if whales form stable and coherent groups, and if there is a hierarchy in associations. We studied associations in the Icelandic killer whale population, a fish-eating population that seems to feed on herring, to investigate for the first time in detail what the social structure of this population might be like. Up to now, how killer whale populations are socially organised was thought to be determined solely by differences between feeding on fish and feeding on mammals.
For example, resident killer whales in the Pacific, that feed mainly on salmon, form clear family groups of individuals that associate very strongly (are always together) and their offspring stay with their mothers throughout their lifetime. These groups are called matrilines and generally all individuals within a matriline associate equally strongly. Different matrilines associate with each other for some periods of time in a hierarchical way - some associate more than others, creating a tree of associations. The mammal-eating population in the Pacific (also called transients or Bigg’s killer whales), which feeds mainly on seals, also has stable family groups but there is some dispersal of both females and males from their family groups that is not seen in the resident population. Associations between adult males are very rare and unstable in this population. The dispersal seen is needed to maintain smaller groups than the ones present in the resident population, which is optimal for feeding on seals. In our study of the Icelandic herring-eating population, curiously, we found striking differences from both mammal- and fish-eating populations in the Pacific. So what else might be shaping the social structure of killer whales worldwide?

Using photographs of 198 identified killer whales we measured their association values. We discovered that associations in the Icelandic population are not random but there are very few, very strong associations among individuals. Members of social groups were not always together and didn’t associate equally strongly, unlike what is seen in residents. On the other hand, like in residents but unlike transients, there was no dispersal of either sex from their groups. In this population there was no hierarchy of association where small groups of individuals associate preferentially with other groups – that is, not all individuals associate at similarly high levels within social groups and distinctly lower levels between social groups.

There is, in fact, large diversity in the groups formed. Some of the groups are simple with similar associations between individuals. However, the majority of groups formed have many different strengths of associations, that is, individuals associate differently with other members of the group, and some groups seem to be composed of more than one subgroup. Associations between whales can be constant, or temporary. But temporary associations were not between coherent, stable groups that associate together for some time, as in the Pacific residents. Instead, they seem to be between (sets of) individuals that associate casually but with preferences. For example, some individuals might prefer to associate in the summer but not in the winter, or in one year but not in others. So they might hang out with the same individuals but be apart for several seasons (and therefore have low association values).

One of the most interesting finds was that associations between individuals were not completely determined by their movement patterns. We know that the whales we see in Iceland have different movements; for example, some go to Scotland in the summer while others appear to stay in Icelandic waters year-round. We also know that some of these whales even have different feeding ecologies. In the Pacific the mammal-eating and the fish-eating populations seem to be socially isolated, but in Iceland we see whales with different movement and feeding habits forming groups. For example, the killer whales matched to Scotland, which feed on fish and seals, sometimes group with individuals seen in Iceland year-round, that appear to follow the herring stock.

Examples of close associations between individuals with different movement patterns, subgroups and groups: 
(a) IF-4, who travels between Iceland and Scotland, in close association with individual IS121, seen year-round in Iceland; (b) 997, another whale who travels between Iceland and Scotland, in close association with IS041, also seen year-round in Iceland.

Although our study shows that killer whales in Iceland don’t form stable, coherent family groups alike those of residents, it is possible that these do exist but are just not behaving exactly the same. So, although both residents and Icelandic killer whales feed on fish, they seem to present different association strategies.

To understand why, consider the differences between Pacific residents and Icelandic killer whales. To start with, they target different prey! Salmon and herring behave very differently and require very different feeding strategies that, in turn, shape the way each killer whale population catches its prey and, likely, the way they socially organize. For example, in Iceland the whales seem to use a group feeding strategy that requires coordination to successfully hunt herring. When herring schools are larger, the whales likely benefit from forming a larger group. We often observe whales in large aggregations in Iceland, concentrated in a relatively small area where there is a lot of herring. This creates lots of opportunities for associations with different individuals and, given the possible benefits for dealing with larger (and always changing size) herring schools, a more fluid social structure may be beneficial for these whales.

The fact that we have observed Icelandic killer whales on herring grounds revealed the fluidity in the social structure of this population, but it will require many more years of observation before we understand the full details of their social structure. Indeed, it is possible that more observations might reveal stable, coherent groups particularly at other times of the year when large aggregations are not as common. For now, we will continue using photo-identification to investigate all these questions over longer time periods and in more behavioural and environmental contexts. We are also using genetic data to learn about the relatedness between individuals that associate and have different movements. We hope to give you more exciting news of these studies soon!
Research paper:
A multilevel society of herring-eating killer whales indicates adaptation to prey characteristics
Sara B. Tavares; Filipa I.P. Samarra; Patrick J.O. Miller
Behavioral Ecology 2016;
doi: 10.1093/beheco/arw179