Difference in distribution between the White-tailed eagle and the Steller's sea eagle on their wintering grounds.: On Hokkaido, Japan.

Difference in distribution between the

White-tailed eagle and the Steller's sea eagle on

their wintering grounds.

On Hokkaido, Japan

Bruno Eusebi

Degree project in biology, Master ofscience (2years), 2020 Examensarbete ibiologi 45 hp tillmasterexamen, 2020

Biology Education Centre, Uppsala University, and Tokyo University ofAgriculture (NODAI)

Table of Contents

Abstract ... 1

Introduction ... 1

Competition ... 1

Study system ... 2

Materials and methods ... 4

Regional scale data ... 4

Local scale data ... 5

Statistics ... 11

Regional scale data... 11

Local scale data ... 11

Results ... 11

Regional scale data ... 11

Local scale data ... 14

Discussion ... 20

Regional scale data ... 20

Local scale data ... 21

Conclusion ... 23

Acknowledgements ... 24

Citations ... 24

Image credits ... 26

Appendix ... 27

Abstract

The White-tailed eagle (H. albicilla) and the Steller's sea eagle (H. pelagicus) overwinter in the same region of Japan: North-western Hokkaido. To examine how and if these two species compete at their overwintering grounds the following questions were asked: are they evenly distributed over this region and do they compete over space and resources? For the period 2015-2019, I found that the two species were not evenly distributed over the region, and that H. pelagicus is the most common species and does occur over the whole area. At a more detailed study at Lake Abashiri performed in 2019- 2020, results showed that H. albicilla was the most common species, and that the distribution of the eagles was determined by the availability of food from anthropogenic activity. The species competed for food resource, fish, and H. pelagicus was the stronger competitor in terms of attacks against heterospecifics. They were less aggressive, but more successful when attacking compared to the White-tailed eagle.

Introduction

Competition

Competition is an interaction between individuals, populations or species that results in a negative outcome for one of the competitors (Begon et al. 2006). For competition to happen, the competitors have to share a limited resource, for example food, territory or water. Interspecific competition is the interaction between two species as opposed to intraspecific competition which occurs between individuals of the same species (Begon & Townsend 2008). There are three categories of competition:

interference, exploitative and apparent. Interference competition takes place when individuals directly interfere with each other’s foraging, survival, reproduction or settlement. Exploitative interference is when both species do not interact directly, but they utilise a common limited resource. When apparent competition occurs, the species also do not interact directly, but they do so indirectly via a predator (Holt 1977). In this case, if one prey species’ population increases, it will lead to the increase of the predator population which in turn will result in an increased predation pressure on the other prey species.

For animal species with distinct breeding and non-breeding seasons, the intensity

of competition can differ between seasons. For example, Eccard et al. (2011) showed

that the competition between two vole species, Myodes glareolus and Microtus agrestis,

differs depending on the season. During the winter (non-breeding season), there was no

direct interference, which meant no costly aggression between the species in terms of

energy constrains and thermoregulatory needs. However, when food constraints increase

exploitative interference occur. In the summer, aggressive direct interactions increased,

especially in heterospecific encounters, triggered by individual reproductive and hormone condition as well as kin protection (Eccard et al. 2011).

In birds, there are very few studies on interspecific competition on wintering grounds. However, it has been shown that resident birds do compete interspecifically with migrant birds. This competition occurs because of the temporary character of the migrant/resident interaction. It happens only in winter (when migratory birds are present) instead of all year round, which would create an unsustainable environment for both species to coexist (Powell et al. 2020). In another study it was showed that migratory movements allow birds to spread out in order to reduce the competition because on one hand they are not breeding so they do not have offspring and territory to protect and on the other hand the resources become scarce so that spreading out will make resource partitioning relaxed (Lack 1968). The principles and the reasons behind the behaviours of migrant wintering birds, when it comes to interspecific competition, are still blurry and need to be studied more extensively (Greenberg 1986, Powell et al. 2020).

Study system

Steller’s sea eagles (Haliaeetus pelagicus) and white-tailed eagles (Haliaeetus

genus, Haliaeetus, which is thought to be one of the oldest genera within birds, stemming supposedly from the Oligocene (33 million years ago) (Rasmussen et al. 1987) or from the mid-Miocene (12-16 million years ago)(Lambrecht, Kálmán 1933).

Although these birds are closely related phylogenetically, their distributions on a global scale are rather different. H. albicilla has a broad distribution, extending roughly through the whole palearctic realm. They bread as far West as South Greenland and as far East as the Russian Kamchatka, as North as Northern Scandinavia and as South as the Caspian Sea. In the winter, they can reach as South as the Gujarat state in India or the Guangdong province in China (Ferguson-Lees & Christie 2001). They are also known to winter on Hokkaido, Japan, where also, some resident couples stay throughout the year.

H. pelagicus, on the other hand, has a small and restricted distribution. It breeds

on the Northern coasts of the Okhotsk Sea and travels south during the winter to the Southern coasts of the Okhotsk Sea and on the Northern coasts of the Japan Sea (Tingay

& Katzner 2010), which comprises the island of Hokkaido in Japan.

Despite a quite large distribution both species are threatened by several

anthropogenic factors. Steller’s sea eagle, is classified as vulnerable in the IUCN red list

and its population is still declining (BirdLife International 2016a). This eagle has also been

named Treasure of Japan, which makes it a protected species. In contrast, the White-

tailed eagle is classified as least concerned, with an increasing population (BirdLife

International 2016b). Conserving these eagles is important as they play a role in the food-

symbol (for example like the bald eagle in the United States of America) and losing them would represent a big loss of both cultural and biological heritage.

Both species have fish as one of their main food resource. One factor for the threat to these species is therefore depleting food resources, caused by overfishing, which in turn can lead to lead poisoning (Saito 2009). In Japan, the two species mainly feed in the rivers during salmon (Onchorhynchus sp.) migration season, since food is very abundant there because of the salmon run. When this period is over, they shift to the lakes or the seas where it is easier to find fish (Shiraki 2001). If fish is not available they rely on secondary sources of food such as mainly big mammal carrions on land or smaller seabirds on the shores. Sika deer carrions can be found effortlessly in the Japanese forests because hunters are allowed by tradition and by law to remove the desired meat from the carcass and leave the rest where the animal has been shot. On top of that, the authorities have a program in place to reduce the deer population from 200000 individuals to 30000 individuals. As a result of those two events, the availability of deer left by hunters or deer simply wounded and dead afterwards has dramatically increased. Since the deer still contain the rifle bullets and shotgun slugs, the ammunition fragments in those carrions are ingested in heavy quantities by the raptors. Data show levels of lead in the liver of deceased eagles that are particularly high, ranging from 2 to 89 ppm while the normal range is 0 to 0.2ppm (Saito 2009). This indicates that the deer carrion can be highly poisonous. Consequences of this poisoning include severe weight loss (especially pectoral muscle and visceral fat), atrophy of the liver and distended gallbladder which directly results in death or, if not death, at least weakening of the individual, which results to sub-lethal effects such as being hit by cars.

Another threat to the eagles are wind turbines collisions (Ueta et al. 2010). Even though White-tailed eagles are more affected by these than Steller’s sea eagles due to their overall lower flight altitude and higher flight frequency, both species suffer from the installation of wind farms around their wintering grounds. Even if the migrating flight altitudes can be quite high, the wintering flight altitudes, between the food sources and their roosting grounds are low enough for both species to be threatened by surrounding wind farms.

With regard to the distribution of the two species, it is interesting to note that there

is an overlap between the widely distributed species (H. albicilla) and the more endemic

species (H. pelagicus) at the wintering grounds. Hence, these two species are competing

for the same resources, which are, as stated before, decreasing. Analysing their

distributions, especially on the locations where they overlap might give us an insight on

how each species is affected by interspecific competition and if there has been changes

in these reactions, over time, as the pressure in terms of competition for food increases

(Helander et al. 2016).

The main aim of this project is to study the difference in distribution between Steller’s sea eagles (Haliaeetus pelagicus) and white-tailed eagles (Haliaeetus albicilla) on their wintering grounds, on East Hokkaido Island in Japan. To be more specific, I will try to answer the following questions: 1) is there a difference in distribution between the species on a regional scale (i.e. on the whole eastern region of Hokkaido), and 2) is there a difference in distribution between the two species on a local scale (i.e. on and around Lake Abashiri, on Hokkaido)? If there are differences, I will pinpoint exactly what these differences are, and identify what possible factors that could explain them.

Materials and methods

The project was divided into two main parts. The first one was analysis of data from a regional scale. This data was collected by members and collaborators of Tokyo University of Agriculture on Hokkaido, over a period of 5 years and from different locations around the whole eastern region of the island of Hokkaido. The second part was the data that I collected during my stay 2019-2020. This data includes the individual counts around Lake Abashiri (local scale): direct observation of feeding behaviour and video taping of feeding behaviour.

Regional scale data

For this data collection, members and collaborators at the Tokyo University of Agriculture have counted individuals on their wintering grounds between 2015-2019.

Approximately 200 counting places were visited per year. Each place is identified with a combination of two numbers, where the first number corresponds to a subzone and the second number identifies the exact location within the zone. Upon analyses, the different sites were merged into 16 larger zones, according to their geographical proximity. On top of that, each location was also put into one of 4 categories; if the counting place was located along the coast, on a cliff or on a beach, it was classified as C – Sea Coast; if the counting place was located on a river bank or on a bridge crossing a river, it was classified as R – River; if the counting place was located around a lake or on a lake, it was classified as L – Lake; and if the counting place was located on land, away from big water sources, for example in waste disposal areas or in forests, it was classified as IR – Inland.

During monitoring of the area, the participants were standing at the counting point and surveyed the area with binoculars, focusing mainly on the tree lines and on the ice.

With help of a clicker, they counted the number of individuals and classified them at the

same time in different categories: adult H. pelagicus, young H. pelagicus, adult H. albicilla,

young H. albicilla and unknown eagle.

Local scale data

For the local scale data, I collected data of several types focusing on behaviour especially with regard to food competition at one lake site: Lake Abashiri (Figure 1).

For the first and main part of my project I identified nine spots for where to stop and gather data around Lake Abashiri. At these points I estimated number of eagles, and there were four zone categories: zones where commercial fisheries and professional fishermen were active (massive ice fishing with nets, on the lake), zones where recreational fishermen were active (small-scale ice fishing with fishing poles, on the lake), zones with a view on the tributary and distributary rivers of the lake (no fishing, on the river) and zones were the eagles were undisturbed by any human activity (no fishing, on the lake).

Figure 1. Satellite image of the lake with the different zones pictured. The 9 circled areas represent the area I could observe while accurately identifying the eagles.

The zones where commercial fisheries are present are:

• Zone 1: Yobito harbour, which is the primary location for fishermen to depart from and come back to. Fishing holes are set up with nets, in the ice, surrounding the port. Unwanted fish (e.g. stone flounder) are discarded next to the ice fishing holes, when the nets are brought up. Yobito harbour is also the place where they bring the fish back and sort it, discarding any unwanted fish that has slipped through the first sorting.

• Zone 3: a zone where ice fishing holes with nets are set up for commercial fisheries

but there is no harbour to come back to and sort the fish.

• Zone 5: a zone where ice fishing holes with nets are set up for commercial fisheries but there is no harbour to come back to and sort the fish.

The zones where recreational fisheries are present are:

• Zone 2: Connectrip zone, a company that organises, rents material and rents ice holes for ice-fishing to tourists and locals. Hence a lot of recreational fishermen and families come there but they do not discard any fish here.

• Zone 9: a zone where locals come for recreational fishing, although less crowded than zone 2, because people have to bring their material, dig their holes and set up equipment by themselves.

The zones where rivers are present:

• Zone 4: the distributary part of the Abashiri River, which links Abashiri Lake to the Okhotsk Sea. It is surrounded by trees and there is no fishing activity taking place there.

• Zone 8: the tributary part of the Abashiri River which links Abashiri Lake to its source, Mount Ahoro in the Akan volcanic complex, near Tsubetsu. It is surrounded by trees and there is no fishing activity taking place there.

The zones without human activity:

• Zone 6: no fishing activity or fishing holes present, just plain frozen lake water surrounded by a tree line.

• Zone 7: no fishing activity or fishing holes present, just plain frozen lake water surrounded by a tree line and a small hamlet.

At least once every three days, in the early morning I spent at least 20 minutes of observation of the zones at each of these points. Preliminary observations suggested that three days corresponds to good collection interval. One day intervals yield too much data and are not profitable, as population movement and changes are usually not noticeable overnight, but rather over several days, on top of being expensive and time consuming.

One-week intervals might be too much as one can miss changes in occurrence and

activity of the eagles. Two-times-a-week intervals also give a bit of flexibility to account

for unfavourable weather conditions (heavy fog, heavy snow: anything that leads to bad

visibility). The material used for these censuses were a pair of binoculars, a field scope

and a clicker. While counting the number of individuals in the different zone, each bird

was recorded in a specific category according to species and age, determined by the

phenotypical characteristics of both species:

• Adult Steller’s sea eagles (H. pelagicus: Image 1) – Easily recognisable with their bright white shoulders, thighs and tail, contrasting with their dark blueish plumage on the rest of the body. Another criterion is the bright colour of their legs as well as the bright colour of their large bill. A last criterion, useable only in flight, though, is the wedge-shape of the tail.

Image 1. Adult/Mature H. pelagicus

• Adult White-tailed eagles (H. albicilla: Image 2) – Smaller than H. pelagicus.

Recognisable by their bright white tail, contrasting with the brown plumage over

the rest of the body (no other colour on the shoulders or the thighs). The brown

tone can differ between individuals, but also for the same individual throughout its

lifetime (older White-tailed eagles can even have a whitish head which makes them

confusable for vagrant American Bald eagles, H. leucocephalus). The bill and the

legs are yellow but the colours appear washed out compared to H. pelagicus. The

bill is also smaller than the one of the Steller’s sea eagle. In flight, their tail is

rounder and flatter than H. pelagicus.

Image 2. Adult/Mature H. albicilla

• Young Steller’s sea eagles (H. pelagicus: Image 3) – Brown plumage all over the body, no different colour on the tail, shoulders or legs. The big bill is usually pale instead of bright yellow; same for the feet. They can express a white tail when they are at the stage between first year younglings and adulthood, which does not really help identifying them. They get their adult plumage at around four to five years of age. The shape of the tail in flight remains a solid criterion for identification.

Image 3. Young/Immature H. pelagicus

• Young White-tailed eagles (H. albicilla: Image 4) – Brown plumage all over the body, no different colour on the tail. Bill smaller than young Steller’s sea eagles, very pale. Same colouration for the feet. Do not express a white-tail before adulthood. Overall smaller than young Steller’s sea eagles. The shape of the tail in flight remains a solid criterion for identification.

Image 4. Young/Immature H. albicilla

• Unknown: All the eagles that cannot be identified with certainty fall into this category. As the adults are easily distinguishable from each other, mainly young individuals are represented here.

For the second part of the project, I focused on the behaviour of both species with

regard to food. This could give an insight in interspecific and intraspecific interactions as

a result of competition for the limited food resources. To do so, I gathered data in Yobito

harbour. Here the fishermen discard unwanted fish, which the eagles feed on when all

the humans have left the area. The discarded fish species are mainly stone flounder

(Kareius bicoloratus), white-spotted char (Salvelinus leucomaensis), saffron cod

(Eleginus gracilis) and Japanese smelt (Hypomesus nipponensis). I also gathered data

in Furen Lake where a hotel owner disposes fish on the ice to attract eagles for

photographers to take pictures of the birds while they are feeding. Two methods were

used in this part.

The first one was a direct observation with a technique called focal sampling. With help of binoculars and/or a field scope, I focused on one individual for 5 minutes or until it flew out of sight. During this time, I observed all the interactions between the selected individual and any other eagle. After the 5 minutes had passed, I chose another individual, usually from another category (ad/y H. pelagicus, ad/y H. albicilla). I used the following rotation for these obsevations:

• 1

individual: adult H. pelagicus

• 2

individual: young H. pelagicus

• 3

individual: adult H. albicilla

• 4

individual: young H. albicilla

• 5

individual: adult H. pelagicus

• Etc.

The second method was indirect observation with a camera. With this, a large angle camera was set up pointing towards the pile of fish. Once it is activated, I went into a hide in order to not disturb the birds. After the eagles had left due to lack of food, I went back and deactivates the camera. The footage of these recordings was then analysed as for the direct observations described above.

The interactions and behaviour where classified as follows: while tracking an individual, I first noted how they were feeding: grabbing a fish without landing and eating it somewhere else (usually in a tree) or by landing, grabbing a fish and taking off to eat it somewhere else, or by landing and eating the fish right where it stands. Secondly, I noted if the eagle was attacked or not by another eagle. If an aggressive behaviour was recorded, I classified it as ground attack (at least one eagle is grounded) or aerial attack (both eagles are flying). For each attack, I also classified it as success if the attacked individual drops its fish and as a failure if it kept the fish. The method for the classification can be represented as follows:

Attack? No Yes

Ground Success Failure Aerial Success

Failure

Statistics Regional scale data

I used a linear regression to check if the number of individuals in the different regions increased or decreased between the years or if they remained stable from a year to another. Furthermore, I used a repeated measures ANOVA adjusted with the Bonferroni correction (when adequate) to check if on one hand there is a difference in distribution between both species across the regions and a similar ANOVA to check for differences between age groups. I used pairwise t-tests to compare between groups. I used Pearson’s chi-squared tests with Bonferroni correction to determine if there are differences in the distribution between the species according to their environment.

Local scale data

On the local scale part, I performed 9 unpaired t-tests to assess if there is a difference in both species’ individuals’ numbers within each site (1 unpaired t-test per site). I carried out two one-way ANOVAs with the Tuckey post-hoc tests (one for each species), to evaluate the difference between each site in the number of individuals of both species. To judge the feeding methods, I used chi squared tests comparing my estimate to the expected values of 1/3,1/3,1/3, which would mean that they pick a feeding method at random, with no pattern or preference. If the test does not show a significant difference, it means that the actual values are close to this 3*1/3 ratio. To assess the differences in attack success rate between both species, when attacking as well as when being attacked, I used McNemar’s Chi-square test with continuity correction. Finally, I used chi squared tests to rate the proportions of attacks, comparing my values to the expected values of 1/4,1/4,1/4,1/4, which would mean the attacks are random, with no pattern or preference. If the test does not show a significant difference, it means that the actual values are close to this 4*1/4 ratio.

Results

Regional scale data

Over all the research area, 8834 individuals were counted, with the biggest portion being adult Steller’s sea eagles, Haliaeetus pelagicus (around 3500 individuals), then adult white-tailed eagles, Haliaeetus albicilla (around 2750 individuals), young Steller’s sea eagles (around 1250 individuals) and finally young white-tailed eagles (around 1200 individuals) (Figure 2-4). The rest of the counted individuals were scored unknown species/age

A linear regression showed that the number of individuals did not increase or

decrease significantly between the years (p = 0,385; Fig 2). A repeated measures ANOVA

showed that there is a difference in the distribution of both species (2 groups, H. albicilla

and H. pelagicus), across the years (p = 0,012; Fig 2), with H. pelagicus being more

common. A second repeated measures ANOVA showed that when splitting the groups (4 groups: adult H. albicilla, young H. albicilla, adult H. pelagicus, young H. pelagicus) this difference was still present (p << 0,001)

With help of pairwise t-tests, it is clear that the difference stems from all the groups (adj. p range from 0,032 to p << 0,001) except between the young individuals of both species (adj. p = 1,00). Figures 3 and 4 show example of these differences in distribution between both species, over the whole eastern region of Hokkaido, where adult Steller’s sea eagle is the dominant group.

Figure 2. Number of eagles in function of the environment type and the year. // IR=inland, L=lake, R=river, S=sea

The statistical analysis, using Pearson’s chi-square tests and Bonferroni

corrections showed that there are significant differences between the number of

individuals of both species in any of the environments. (P-values: IR < 0,001, L<<0,001,

R<<0,001, S=0,001, Fig 2).

Figure 3. Number of eagles in function of the regions (see in the appendix for the numerical references)

Out of the 16 established regions, there are two that have an extrodinary high abundance of eagles over the past five years with just over 2000 individuals: the Notsuke/Furen/Nemuro region and the Hamanaka/Teshikaga/Shibecha/Kushiro/Tsurui region (Figure 3; region 5 and 6). Both these regions are dominated by H. pelagicus (up to twice the number of white-tailed eagles). The second most abundant region is Kitami/Monbetsu/Abashiri with around 1000 individuals of which around 625 are white- tailed eagles. The next four regions are again dominated by Steller’s sea eagles (Soya, Shari, Rausu/Shibetsu, Tokachi) with between 500 and 800 individuals. The other 9 regions, show an abundance bellow 300 individuals, and are dominated by H. albicilla (Furano/Asahikawa, Haboro, Ishikari, Chitose, Utonai, Hidaka, Otaru, Yakumo/Hakudate, Iwate/Miyagi/Fukui/Izumo).

When the data is broken down by year, the same trend is observed with the exception of the Rausu/Shibetsu region in 2015 being dominated by white-tailed eagles and having an overall bigger eagle population than the Kitami/Monbetsu/Abashiri region.

In 2016, there were once again more eagles in Rausu/Shibetsu, but this year Steller’s

sea eagles were dominant, like the other years.

Figure 4. Number of eagles in the different points of interest. N/A represents the eagles that have been counted somewhere other than those counting points of interest.

Points of interest were established as places where eagles are known to be easily observed in significant numbers. Out of the 23 established points of interest, two are standing out as having more individuals: Lake Furen with more than 1500 individuals, and dominated by H. pelagicus and Rausu-Shiretoko which seems to have a fairly equal repartition (perhaps leaning toward H. pelagicus dominance) (Fig. 4). The next 6 points have over 250 individuals and are dominated by Steller’s sea eagles, with the exception of Lake Abashiri, which is dominated by white-tailed eagles. The rest of the interest points are all bellow 250 individuals. The trends are overall pretty similar when the data is decomposed by year, with few interesting changes like, for example, the jump to 250 individuals in Utoro-Shiretoko, in 2019, which differs from the usual number of about 50.

Local scale data

Over the time of the data collection at Lake Abashiri, I counted 1641 individuals,

with the biggest proportion being adult White-tailed eagles, Haliaeetus albicilla (757

individuals), then adult Steller’s sea eagles, Haliaeetus pelagicus (569 individuals), young

White-tailed eagles (174 individuals), young Steller’s sea eagles (118 individuals) and the rest being classified as “unknown eagles” (23 individuals) (Figure 5 and 6).

Unpaired t-tests showed that there is no significant difference between the number of individuals of both species in the different sites except in site 2 (p-value = 0,041), that seems to be dominated by White-tailed eagles.

When looking at the difference between the sites for each species, a one-way ANOVA show that there are significant differences (p < 0,001 for Steller’s sea eagles and p < 0,001 for White-tailed eagles). When applying the Tuckey’s test for post-hoc analysis, it appears that for White-tailed eagles, the difference stems from site 1 being different from all the other sites (adjusted p-values were <<0.001) because the other sites do not show significant differences between them (adj. p >> 0,05). On the side of the Steller’s sea eagles, it is similar with site 1 being different from all the others (adjusted p-values were >> 0.001) but other sites show significant dissimilarities among them as well, like between sites 3 and 2 (adj. p=0,005), sites 5 and 2 (adj. p=0,001), sites 4 and 3 (adj.

p=0,005), sites 6 and 3 (adj. p=0,036), sites 7 and 3 (adj. p=0,049), sites 8 and 3 (adj.

p=0,006), sites 5 and 4 (adj. p=0,001), sites 6 and 5 (adj. p=0,010), sites 7 and 5 (adj.

p=0,015), sites 8 and 5 (adj. p=0,001) and sites 9 and 5 (adj. p=0,019).

Figure 5. Total abundance of each category of eagle on the different sites around Lake Abashiri

Figure 6. Abundance of the different species of eagles through time, around Lake Abashiri. The yellow line shows the average total abundance of eagles before and during opened fisheries season while the light blue line shows the average total eagle abundance only during the opened fisheries season.

The data I collected for feeding behaviour and competition by indirect and direct observation is shown in Table 1-4. I observed the behaviour of 570 individuals (100 young

resulting in a total of 204 interactions.

Table 1. Proportion of each species group’s feeding method. Classified in three categories: Fly-by+tree is when the individual grabs fish without touching the ground and eats it somewhere else (usually on a tree); Ground+tree is when the individual lands, grabs fish after some time and takes off to eat it somewhere else (usually on a tree); Ground is when the individual lands and eats the fish where it is.

Feeding method

AD H. pelagicus

36,76 23,78 39,46

Y H. pelagicus

32 35 33

AD H. albicilla

37,84 25,95 36,21

Y H. albicilla

28 37 35

0 20 40 60 80 100 120 140 160 180

12/31/19 1/10/20 1/20/20 1/30/20 2/9/20 2/19/20 2/29/20 3/10/20

Eagles count around lake Abashiri

SSE WTE Total Average Average (post-fisheries)

• Feeding method:

There was no significant difference between the observed values and the expected values for any of the species: p = 0,12 for H. pelagicus and p = 0,43 for H. albicilla (Table 1). However, when decomposing the species into their age groups, and applying the same tests, there is a significant difference in the feeding method for the adult Steller’s sea eagle (p = 0,02), but no differences for adult White-tailed eagles (p = 0,099), young Steller’s sea eagles (p=0,93) or young White-tailed eagles (p=0,51).

Table 2. Attacking individual. Represents the individual that is trying to steal food from another one. Ground attack is when one individual tries to steal fish from another one while it is not in the air, with the two outcomes being success (the attacked individual leaves the fish for the attacking individual) or failure (the attacking individual leaves without the other one’s fish). Air attack is when one individual tries to steal fish from another one while they are flying, with the two outcomes being success (the attacked individual drops the fish) or failure (the attacked individual still has the fish). Success rate is the number of successes divided by the number of attempts.

Attacking Ground attack Air attack

Success rate

(%) Attempts

Success rate

(%) Attempts

AD H.

pelagicus 63,34 30 0 3

Y H. pelagicus 42,86 28 0 5

AD H. albicilla 45,65 46 7,69 13

Y H. albicilla 38,46 65 0 14

• Attacking:

When it comes to aggression, adult Steller’s sea eagles are overall the individuals that attack the least as they are usually in the middle of the group, as close as possible to the food, fending off some offense rather than attacking (Table 2). However, when they do attack (in 33 recorded cases), they have the most success rate with a food steal in 58%

of the cases. When young H. pelagicus attacked, in 33 instances, they had an 36%

success rate. Young White-tailed eagles are the ones that attack the most, with 79 occurrences, but they also have the worst success rate with only 31%. This could be explained by the fact that they are smaller so they cannot stay in the middle of the group, on the food and fight but they are also not very experienced, which reduces their capacity to lead successful attacks. Adults H. albicilla have the only recorded successful air attack and are overall successful in 37% of the cases (for 59 attacks).

When looking at the statistics, the McNemar’s Chi-squared test reveals that there’s a

significant difference in the success rate of both species (p=0,035) with H. pelagicus being

more successful than its White-tailed counterpart.

Table 3. Individuals being attacked. Represents the individual that is being observed. No attack means that the individual is not being bothered by another one during the observation time. Ground attack is when one individual tries to steal fish from the observed one while it is not in the air, with the two outcomes being success (the attacked individual leaves the fish for the attacking individual) or failure (the attacking individual leaves without the other one’s fish). Air attack is when one individual tries to steal fish from the observed one while they are flying, with the two outcomes being success (the attacked individual drops the fish) or failure (the attacked individual still has the fish).

Success rate is the number of successes divided by the number of attempts.

Being attacked No

attack Ground attack Air attack

Success rate % Attempts Success rate % Attemps

AD H. pelagicus 146 16,67 24 0 15

Y H. pelagicus 52 61,90 42 0 6

AD H. albicilla 136 46,34 41 0 8

Y H. albicilla 32 45,16 62 16,67 6

• Being attacked:

Looking at the attacked instead of the attacker, adults are overall not attacked as much as young ones (Table 3). The adults are left alone, in 79% and 73% of the cases, for respectively H. pelagicus and H. albicilla, whereas young individuals are left alone in only 52% and 32% of the cases respectively. When adult Steller’s sea eagles are being attacked, it mostly ends in failure of the attacker, as success only occurred in 10% of the cases. On top of that, air attacks never succeeded against adults H. pelagicus. On the other hand, the biggest attack success rate was against young Steller’s sea eagle with a 54% of success. Again, this high rate comes solely from ground attacks, as none of the six air attacks were successful against them. Against the adult White-tailed eagles, the success rate is around 39%, coming also from ground attacks only, while the young ones face an attack success rate of 43% against them. They also constitute the only category to have dropped their food after being attacked in the air, which counts as a success for air attack.

However, the statistical analysis does not support the described trend as the McNemar’s Chi-squared test does not highlight any significant difference between both species (p=0,272).

Table 4. Overview of the attacks. This table shows between which categories the interactions occurred, as a percentage.

attacked by à AD H. pelagicus Y H. pelagicus AD H. albicilla Y H. albicilla

AD H. pelagicus 15,38 20,51 30,77 33,34

Y H. pelagicus 14,58 20,84 25,00 39,58

AD H. albicilla 8,16 4,08 36,73 51,02

Y H. albicilla 23,53 19,12 25,00 32,35

• Attacks:

When merging those two last observations (from Tables 2 and 3), into Table 4, one can notice that adult Steller’s sea eagles are being attacked mostly by young eagles, probably due to their inexperience and inability to recognise that these attacks are a waste of energy and will lead in most case to failure. Young Steller’s sea eagles are also mostly attacked by young eagles, which makes more sense for young H. pelagicus but is tricky for young H. albicilla, as they face a size disadvantage. Youngs and adults H. albicilla are mainly attacked by other H. albicilla individuals which makes sense as they have more chance being successful against an opponent of their size rather than a bigger one.

There was no significant difference between the observed values and the expected values in Table 4 except for when the adult H. albicilla was attacked: p < 0,001. It seems that there is a bias towards the attackers being other H. albicilla individuals (especially young ones).

Discussion

Regional scale data

This study showed that the number of eagles overall does not change from one year to another over the five years studied. However, there is a difference in distribution among regions between the two species and also between the different age groups, as shown by the two repeated measure ANOVAs.

There seems to be a trend when looking at the graphs such that the dominance will usually go towards Steller’s sea eagles with high abundance of eagles, whereas when the overall number of eagles is low, the dominance will tend be dominated by white-tailed eagles. However, there were no statistical support for this pattern, which actually show that there is not that much difference after all (see local scale data results). It is also important to note that the data for the regional scale are collected at optimal times for each environment. The counting of eagles in the lake takes place when fisheries are opened, the counting in the rivers takes place when the salmon is plentiful during the salmon run and the counting in the sea takes place aboard an icebreaker, when drift ice is present for the eagles to land on and fish.

When the data of all the years are combined, one can notice that the eagles have

the highest abundance at the sea areas (approx. 3300 individuals), followed by the lakes

(approx. 2800 individuals), on the rivers (approx. 2400 individuals) and finally inland

(approx. 200 individuals). If the data is broken down by year, it shows that usually, the

sea areas are the preferred habitat but it also depends on the year. For example, in 2015

and 2019 abundance was highest at the lakes, whereas in 2017 and 2018 it was highest

at the rivers. In all the cases, the number of eagles inland is very small compared to the

tracked overwintering Steller’s sea eagles follow the same pattern as with my data: a majority at sea, then in the rivers or the lakes depending on the salmon run and a minority inlands (Shiraki 2001, Ueta et al. 2003).

There is a dominance of Steller’s sea eagles, over the past 5 years in the sea areas and on the lakes. It seems to be an equitable distribution on the rivers and inland. At the lakes they seem to always dominate independent of year. For the rivers, on the other hand, only 2017 was widely dominated by Steller’s sea eagles, whereas the other years were slightly dominated by white-tailed eagles. On the sea, it seems like both species are usually equally distributed except for 2019 where the number of Steller’s sea eagles is 150% higher compared to the white-tailed eagles. The differences concerning the inland areas are negligible.

It is also important to note that the trends observed in general are similar to the ones observed when decomposing the data year by year, which suggests that the individuals behave similarly from one year to another. With this idea pushed to the extreme, one could think that the individuals have some kinds of migratory habits, repeating the same migration pattern and using the same regions/points of interests from one year to another. Such “site fidelity” behaviours have already been acknowledged by several studies in the past, showing that individuals do come back to the same spots year after year (Ralph & Mewaldt 1976, Ketterson & Nolan 1990).

Local scale data

When looking at the overall number of individuals around Lake Abashiri during the period of census, it is clear that it follows the trend previously established by the regional scale data over the past 5 years: it was dominated by White-tailed eagles.

Another clear distinction is the role played by ice fisheries in the presence of eagles around the Lake, regardless of the species. Firstly, the sudden rise of the number of individuals counted after the fisheries have opened is a good indicator of abundance.

Before ice fishing season started (11

January), the average count around the lake was

22 individuals whereas after this date one counting session around the lake averages 103

individuals. Secondly, the distribution of the eagles around the lake is not random. It is

clear that most of them are found around commercial fishing zones where food is

abundant. This influence of the fisheries is also shown by Shiraki (2001) where she

compared the mean number of Steller’s sea eagles per square kilometre at commercial

fisheries sites and non-fisheries sites. She found a significant difference in the distribution

during the winter (Shiraki 2001). In addition to the commercial fishing, recreational fishing

also affects foraging of the eagles since they can acquire some leftover fish after the

humans are gone. The less disturbed zones have considerably less individuals but still,

some individuals are present as it offers a quiet resting place away from anthropogenic

disturbance and competition. The rivers at this time are not attractive at all for the eagles

as the salmons, which are the main food resource in the rivers in November/December,

have already migrated back to the ocean. This has also been noted by Ueta et al (2003) who states that “the change in habitat use by eagles during this time is a response to the declining availability of salmon as rivers become covered by ice and snow” (Ueta et al.

2003).

In zones where the overall abundance of eagles is low (for example zone 2), the Steller’s sea eagles are almost absent compared to the White-tailed. This could be due to the fact that H. pelagicus is strictly migratory on Hokkaido. There are no breeding couples or residents on the island. In this regard, the individuals travel in groups during their migration to the island, but also between their roosting place and their feeding places. Because of this, it ends up a high number of H. pelagicus in crowded places and almost none in other places. On the other hand, some residents and breeding couples of

explain why White-tailed eagles are more often found in places where food is quite limited and other individuals are not present, as they could have a nest nearby or they could simply be used to coming often to this particular place throughout the whole year. Another explanation could be that because of their bigger size, Steller’s sea eagles can stay in the middle of the ice and feed on the fish while White-tailed eagles have to wait for food to become available. They may do so by staying on trees which makes them less noticeable and easily missed by the observer (especially because their plumage is well camouflaged with the trees while the Steller’s sea eagle’s plumage stands out).

The observation of the feeding method resulted in several interesting results. Adult Steller’s sea eagles seem to prefer eating on the spot rather than taking their food to eat elsewhere after landing or grabbing food on the fly. They are the biggest and the heaviest individuals out of the four categories, which would make sense as to why they are able to stand on the ground while eating and perhaps fending off other individuals’ intrusions. For younger H. pelagicus individuals, it does not seem like any method is preferred over another, although the number of individuals who grab food after landing and take-off to eat it elsewhere is slightly higher than the other methods of feeding. Adult White-tailed eagles usually grab fish on the fly, without landing and eat it someplace else, away from the rest of the group. If they land, they usually eat there, on the spot, although they sometimes fly out after grabbing food but to a lesser extent. The young H. albicilla prefer overall to land at some point before eating and they rarely grab their food while flying.

From these observations, one can make statements of different strategies when it comes to food. Theory suggests two main patterns when it comes to resources distribution. One states that the “resources are equally available and uniformly utilized across body sizes”

and the other states that “resources are differentially available to organisms of different body sizes” (Ernest 2005). My results seem to favour the second pattern. Adult H.

pelagicus, with their bulky body are less prone to be attacked and therefore can afford to

stay around the food source without being bothered too much (as shown in Table 1). Adult

and young H. albicilla, being smaller than the Steller’s sea eagles have a harder time

standing their grounds and fending off attacks, which could explain the thieving behaviour, as seen in Table 4. They prefer attacking each other instead of the bigger bird and when they do, they steal what they can and after they succeed, they fly away to avoid any repercussion or to avoid the food being stolen back. This supports a publication by Leyequién et al. (2007) who found that there is a negative relationship between body mass and the strength of the interspecific competition. They found evidence that the distribution of resources can be size-related and therefore can dictate the behaviour of two interacting species between one another (Leyequién et al. 2007).

Overall, it seems like the attackers are rather young eagles and White-tailed eagles. The reasons for the pattern could be several. H. albicilla could be a more aggressive species in general, young eagles could be unexperienced and would fail to recognise a battle that is already lost, or adults H. pelagicus being big and sturdy would more often assume a more defensive position on top of the food rather than having to attack as much. Although a study on two sister species of rock crabs showed the opposite in terms of aggression, there were some common points on the outcomes. The bigger crabs, usually initiated the attacks and the food contests while the smaller ones suffered from the attacks. However, both my study and theirs suggest that whenever there is a contest, the size of the bigger individual has an impact on the outcome of the competition (Matheson & Gagnon 2012). Furthermore, in a study on body size, interspecific competition and use of foraging sites in Paridae (the tits bird family), showed similar results to mine. It shows that the individuals of the bigger species (Parus montanus) influence social dominance in a way that relegates individuals of the smaller species (Parus ater) to poorer foraging sites and overall poorer food quality (Alatalo & Moreno 1987). This is a good analogy to my study; P. montanus has access to the inner parts of the trees (less energy needed to exploit, less predatory risk, more food) and relegates P.

ater to the canopy (more energy needed to exploit, less food, more predatory risk);

analogically H. pelagicus has access to the fish while staying on the fish pile (less needed energy because no need to attack, more food) while H. albicilla stays on the side, needs to attack more and is not always successful (more energy needed, less food). In other words, body size matters!

Conclusion

In conclusion, my results suggest that on regional scale, there is a difference in

distribution between the species because one is a strictly migratory species and the other

one a half migratory/half resident species. That difference is evident by the fact that a lot

of Steller’s sea eagles are present when the overall abundance of eagles is high but they

are almost absent when the overall abundance of eagles is low. On a local scale, my

results suggest that the distribution follows roughly the same rule. However, the overall

distribution of eagles on Lake Abashiri is dictated by one anthropogenic factor: the

fisheries. When it comes to food competition on Lake Abashiri, it is clear that the Steller’s

sea eagle is the dominant species, due to its larger size, while H. albicilla is more aggressive in order to steal food from the larger individuals. Immaturity also plays a role in these interactions as young individuals have not yet learned who they are able to challenge for food and who they will not get food from.

Acknowledgements

I would like to thank Saiko Shiraki for letting me conduct this project under her supervision, for her valuable technical and practical advices and for her help in settling myself in a foreign country.

In this regard, I would also like to thank the Tokyo University of Agriculture (NODAI) for providing me with a room and the material necessary to proceed with the data collection.

I am also immensely grateful to Frank Johansson, my supervisor at Uppsala University, for his help in writing this paper, his patience with me as well as the motivation he gave me to thrive for the best thesis output possible.

I would like to thank Uppsala University for giving me the opportunity to conduct this research.

Finally, I would like to thank my opponents, Èlia Padrós and Niki Chondrelli for their valuable input.

On a side note, I would also like to thank the students from the biology bachelor degree (So Nomura, Seina Tani, Hilomi Ikeda, Kai Iefuji, Kazuyoshi Tachibana, Ryosuke Murai,…) for their warm welcome as well as their help and advices in collecting data and living in Japan overall.

Citations

Alatalo RV, Moreno J. 1987. Body Size, Interspecific Interactions, and Use of Foraging Sites in Tits (Paridae). Ecology 68: 1773–1777.

BirdLife International. 2016a. Haliaeetus pelagicus. The IUCN Red List of Threatened Species 2016: e.T22695147A93492859.

BirdLife International. 2016b. Haliaeetus albicilla. The IUCN Red List of Threatened Species 2016: e.T22695137A93491570.

Eccard JA, Fey K, Caspers BA, Ylönen H. 2011. Breeding state and season affect interspecific interaction types: indirect resource competition and direct interference. Oecologia 167:

623–633.

Ernest SKM. 2005. BODY SIZE, ENERGY USE, AND COMMUNITY STRUCTURE OF SMALL MAMMALS. Ecology 86: 1407–1413.

Greenberg R. 1986. Competition in Migrant Birds in the Nonbreeding Season. In: Johnston RF (ed.). Current Ornithology, pp. 281–307. Springer US, Boston, MA.

Helander, Björn, Herrmann, Christof, Hellström, Peter. 2016. HELCOM core indicator report on white-tailed eagle productivity from January 2016.

Holt RD. 1977. Predation, apparent competition, and the structure of prey communities.

Theoretical Population Biology 12: 197–229.

James Ferguson-Lees, David A. Christie. 2001. White-tailed fish-eagle. Raptors of the World, pp. 402–408.

Ketterson, Nolan. 1990. Site Attachment and Site Fidelity in Migratory Birds: Experimental Evidence from the Field and Analogies from Neurobiology. 13.

Lack D. 1968. Bird Migration and Natural Selection. Oikos 19: 1.

Lambrecht, Kálmán. 1933. Handbuch der Palaeornithologie.

Leyequién E, de Boer WF, Cleef A. 2007. Influence of body size on coexistence of bird species.

Ecological Research 22: 735–741.

Matheson K, Gagnon P. 2012. Effects of temperature, body size, and chela loss on competition for a limited food resource between indigenous rock crab (Cancer irroratus Say) and recently introduced green crab (Carcinus maenas L.). Journal of Experimental Marine Biology and Ecology 428: 49–56.

Michael Begon, Colin R. Townsend. 2008. Essentials of Ecology.

Mutsuyuki Ueta, Yoshihiro Fukuda, Reiko Takada. 2010. Difference in flight behavior between White-tailed and Steller’s Sea Eagle in Hokkaido. Bird Research 6: 43–52.

Powell LL, Ames EM, Wright JR, Matthiopoulos J, Marra PP. 2020. Interspecific competition between resident and wintering birds: experimental evidence and consequences of coexistence. Ecology, doi 10.1002/ecy.3208.

Ralph J, Mewaldt R. 1976. Homing Success in Wintering Sparrows. 15.

Rasmussen DT, Olson SL, Simons EL. 1987. Fossil Birds from the Oligocene Jebel Qatrani Formation Fayum Province, Egypt. Smithsonian Contributions to Paleobiology 1–20.

Saito K. 2009. Lead Poisoning of Steller’s Sea Eagle (Haliaeetus pelagicus) and White-tailed Eagle (Haliaeetus albicilla) Caused by the Ingestion of Lead Bullets and Slugs, in

Hokkaido, Japan. Ingestion of Lead from Spent Ammunition: Implications for Wildlife and Humans, doi 10.4080/ilsa.2009.0304.

Shiraki S. 2001. Foraging habitats of Steller’s sea-eagles during the wintering season in Hokkaido, Japan. J Raptor Res 35: 91–97.

Shiraki S. 1994. Characteristics of White-Tailed Sea Eagle Nest Sites in Hokkaido, Japan. The Condor 96: 1003–1008.

Tingay RE, Katzner TE (eds). 2010. The eagle watchers: observing and conserving raptors around the world. Comstock Pub. Associates, Ithaca.

Ueta M, McGrady MJ, Nakagawa H, Sato F, Masterov VB. 2003. Seasonal change in habitat use in Steller’s sea eagles. Oryx, doi 10.1017/S003060530300019X.

Image credits

Image 1. Julie Edgley. An adult Steller's sea eagle (Haliaeetus pelagicus) makes a shallow dive to grab a fish. 6 February 2019. Downloaded at

https://commons.wikimedia.org/wiki/File:Adult_Steller%27s_sea_eagle_fishing.jpg

Image 2. Yathin S Krishnappa. White-tailed eagle in Svolvaer, Norway. 18 August 2012.

Downloaded at

https://commons.wikimedia.org/wiki/File:Haliaeetus_albicilla_(Svolvær,_2012).jpg

Image 3. Chrumps. Haliaeetus pelagicus. 6 June 2009. Downloaded at https://commons.wikimedia.org/wiki/File:Haliaeetus-pelagicus.jpg

Image 4. Christoph Müller. Juvenile White-tailed eagle (Haliaeetus albicilla), Raftsund, Lofoten/Norway. Downloaded at

https://commons.wikimedia.org/wiki/File:White_tailed_eagle_raftsund_juvenile.jpg

Appendix

Numerical reference for the regions in Fig2:

1= Soya

2= Kitami, Monbetsu, Abashiri 3= Shari

4= Rausu, Shibetsu

5= Notsuke, Furen, Nemuro

6= Hamanaka, Teshikaga, Shibecha, Kushiro, Tsurui 7= Tokachi

8= Furano, Asahikawa 9= Haboro

10= Ishikari

11= Ishikari & Chitose 12= Utonai

13= Hidaka 14= Otaru

15= Yakumo, Hakudate

16= Iwate, Miyagi, Fukui, Izumo

Satellite picture of Hokkaido Island with the different survey areas highlighted.