Factors affecting the distribution and density of Grey seal, (Halichoerus grypus) populations

Factors Affecting the Distribution and Density of Grey Seal, (Halichoerus grypus) Breeding Colonies.

David Blackmore 11^th May1998 Thesis submitted for the degree of BSc. (Hons.) Environmental Biology Department of Environmental & Evolutionary Biology Sea Mammal Research Unit University of St. Andrews, Fife. Scotland

Acknowledgements

I would like to thank my supervisor, Professor John Harwood of the Sea Mammal Research Unit, St. Andrews for his assistance and enthusiasm throughout the duration of the project.

Many thanks also to Callan Duck, for his not inconsiderable help and advice, and for his willingness to help at all times. Also to many others in the SMRU for their patience and help.

To Graeme for putting up with me for so long.

To the cleaner, for letting us use her room!

To my parents for believing that I'm not as stupid as I sometimes seem, for always supporting me, and for large amounts of money…….

Dave.

Abstract

Table of Contents

1: Introduction 2: Methods 2.1 Photographic sets 2.2 Mapping and recording methods 2.3 Transfer to computer 2.4 Computer analysis 3: Results 4: Discussion 5: Conclusion 6: References Appendix

Table of Figures

Table 1 Table 2 Table 3 Table 4 Figure 1 Shillay, The Monach Isles Figure 2 Shivinish, The Monach Isles Figure 3 Little Linga, The Orkney Isles Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10

1: Introduction The initial brief of this investigation was to analyze the role of terrain features on the distribution and density of Grey seals on islands where populations have reached a stable density inorder to predict the theoretical maximum population and corresponding distribution of seals on islands where no stable maximum has yet been achieved.

So that an attempt could be made to relate terrain features to seal density, testable hypotheses were formed after consultation with several people with knowledge of this field, and consultation of a previous study of terrain factors affecting Grey seal distribution on islands during the breeding season (Collins 1997).

The islands used in this study were selected for the nature of their topography, the number of seals on them, their size, and the dynamics of their population growth over the last 14 years. It was hoped that a fully computerization system could be developed that would allow the manipulation of aerial photos and data1. Unfortunately, with resources and time available, this proved unrealistic. A simpler, but more time-consuming semi-manual method, using ArcView? 3.0a GIS? was adopted and perfected with a lot of trial and error.

Rather than being a waste of time, this on going learning process was an integral part of the investigation. It provided valuable information on problems and considerations likely to be encountered developing a more wide ranging investigation, and allowing results to be as accurate and relevant as possible.

Grey seals are the largest living carnivore in Britain. Males grow to about 2.3 metres whilst females are smaller and average only 1.8 metres (SNH 1996). Like all Phocids (true seals), Grey seals are obligate land or ice breeders. They form highly synchronous breeding colonies and show high levels of site fidelity. Breeding occurs on a variety of substrates ranging from rocky shores and beaches to ice (Twiss et al, unpublished). There are clear differences in behavioural patterns, associated with this variety of habitats.

The obvious benefits of colonial breeding, mixed with higher pup mortality rates associated with male aggression has produced a breeding system which consist of a mixture of polygyny and partner fidelity (Amos et al, 1995). In a system where females make a large reproductive investment before and after their offspring are born, and males rarely contribute more than their sperm, males are free to mate with many females. If an animal has difficulty moving in its habitat, then the easiest way for a female to ensure that she will be successfully mated is if she is easily identifiable and accessible. By crowding in groups, females maximize their conspicuity, and minimize the distance a male has to cover inorder to mate with them.

A female will become pregnant for the first time when she is 3 to 5 years old. Pup gestation occurs between October and September and the pupping season lasts from late September until November (Twiss, Pomeroy & Anderson, 1994). White-coated pups will then suckle for between 16 and 21 days, staying close to the site of birth. During this time, a pup will gain an average of 30kg in weight (Boyd, 1982 Unpublished). When not suckling her pup the cow may spend time at sea, but this is dependent on the accessibility of the sea, and the ease with which she can identify her pup on returning. She will not feed for the entire suckling period, loosing on average 65kg of her own body weight (SNH, 1996). After weaning, white-coats molt, and can be identified by their grey coats. Moulters often aggregate in small groups away from adults.

During the summer month’s adults aggregate on sandbars, mud flats, and rocky outcrops in large haul-out groups, and appear to have a certain degree of fidelity to the same summer haul-out site from year to year (Anderson, Burton & Summers 1975). Grey seals often dive for long durations in the search for fish and other prey, during which they must conserve their oxygen supply. Grey seals have become accomplished at regulating their metabolism in order to maximize the time they are able to hunt for, often diving for as long as 20 minutes.

Unfortunately seals have gained a reputation for, “stealing” fish from the sea, hosting fish parasites (codworm) (Anderson et al., 1989, Harwood, 1984), damaging netting cages and fish rearing cages and generally threatening the livelihoods of local fishermen (Anderson, 1984, Thompson, 1984). This is a reputation that is by no means proven, since there is increasing evidence that Grey seals and fishermen often do not compete for the same fish. Seals hunt many fish species which are of no commercial value to fishermen, and often hunt in areas unpopular with fishermen (Hammond, et al., 1994, McConnell, et al., 1984). Furthermore, sometimes the presence of seals may be to the fishermen’s advantage; local lobster fishermen in Orkney have a saying, “were there are seals there will be lobsters”. Certainly seals have been shown to eat a lot of octopus, and octopus prey heavily on lobsters (Hewer, 1974).

Grey seals are primarily adapted for life at sea, and have evolved features, such as streamlining and flippers that make life on land more difficult. Since the first Grey seal studies, biologists have noted interesting observations regarding the distribution of Grey seals, but have concentrated for the most part on other aspects of their biology.

At the start of the breeding season cows arrive at the breeding colonies before bulls. Once on land they appear to congregate close to areas of water (Boyd et al 1962, Hewer, 1960). On islands where the preferred beach breeding locations appear to be available, females spend most of their time in the water. In situations where seals pup on the vegetated tops of islands, congregations of cows form around any pools of water. Later in the season, these pools are often flattened out into muddy areas (Anderson et al, 1975).

Stirling (1975) states that the Grey seal, “offers the greatest opportunity for study of the effects of different breeding habits on social behaviour”. Recent advances in the design of biological surveys, and analysis of survey data have enabled the spatial distribution of species to be mapped more accurately, but no-one has yet looked quantitatively at the relationship between Grey seals distribution and terrain type, at the scale of the seal itself.

It has been postulated that the observed affinity for water is a behavioural adaptation to living on land (Pomeroy, et al. 1994). Most mammals need to maintain a body temperature of 37 degrees celsius (SNH 1996), but the sea is much colder than this, and conducts heat from the body more efficiently than air. Adult Grey seals have evolved a thick layer of insulating blubber beneath the skin, which can comprise as much as 70% of their body weight (Ryg, et al. 1990). Sometimes seals are so well insulated from the cold that on land during the summer breeding season, they overheat. It has been suggested that by remaining close to water, seals are maintaining the ability to avoid overheating, (Heat transfer in water is approximately 25 times faster than in air)(Hart & Irvine 1959).

Net heat loss without access to water can only be achieved by avoiding direct solar radiation (McGinnis, 1975). The shade provided by cliffs and caves is also an area where groups of seals have been observed. During particularly warm days, seals have sometimes been observed fanning themselves with their flippers. Seal flippers have a large peripheral blood supply (Hart & Irvine, 1959). It is feasible that this fanning is an attempt to lose body heat by simultaneously increasing the circulation of air around the flippers and the blood supply to the flippers.

The colonization of various sites around the Scottish coast by Grey seals has been extensively studied for over 20 years (Anderson & Curry 1976, Anderson & Harwood 1985, Pomeroy et al 1994, Twiss 1994).
The Sea Mammal Research Unit (SMRU), a NERC funded body, has provided advice to the Home Office and the Scottish Office on the management of seals since 1970 (Harwood 1997). As part of its research programme, SMRU has conducted annual aerial photographic surveys of all the major Grey seal colonies in Scotland and England since the early 1960’s, involving the use of a light aircraft and (since 1985) a rocking camera. The photographs obtained have primarily been used in population monitoring, but they can also be used to identify possible trends in the movement and positioning of seals on land, common to all the colonies.

Until recently very few studies have tried to look in any detail at the relationship between the location and movements of Grey seals and the type of terrestrial habitats available to them. Collins (1997) performed a preliminary investigation into breeding Grey seals, and their choice of breeding sight. Twiss et al. performed a fine scale examination of terrain choice by breeding females on North Rona and the Isle of May. Vicinity of water has appeared as a possible predictor of seal location and numbers, and therefore formed the basis for this study.
Additionally there have been very many studies, which investigated the relationship between the location of a species and their choice of habitat.

Recent advances in the design of biological surveys and analysis of survey data have allowed the spatial distribution of species to be mapped more accurately (Lawton & May 1995). Rushton et al. (1997) modeled the distribution of red and grey squirrels using a combination of GIS and population dynamics. GIS was used to identify the location and type of habitat blocks in the study area. The Cairngorms Partnership employed a system of land-cover classification similar to that used in this study, to investigate the relationship between Black grouse distribution and differing moorland vegetation.

Much of the method and statistical analysis from these (and other studies) can be readily adapted to the study of Grey seals, and has been of some use.

“The failure to colonize is a failure to find suitable habitat, not a failure to disperse”-Lack 1968. Lack was primarily a bird scientist, and as such held the view that animals had easy access to all possible habitats. Grey seals are not so fortunate all of the time. It is true that when at sea, even juveniles have been known to travel vast distances (McConnell B.J. et al 1984), but on land their movement is restricted far more. Seals may well be able to access all suitable islands in order to make an assessment as to its suitability, but this initial judgment will be made from the areas of island easily accessible from the sea (i.e. beaches and outcrops). Females often come ashore and return to the sea several times whilst investigating a prospective breeding site (Anderson et al, 1975). During this period, any disturbance or threat is likely to drive them away.

The existence of spatial heterogeneity in biological systems has inspired many scientists to examine the consequences of modeling one or other of the aspects of this heterogeneity, but despite the large number of spatial studies, many biologists accept that spatial theory is still in its infancy (de Roos et al).
The majority of studies that have investigated searching behaviour have concentrated on the search for food. Even so, much of the theory involved is still relevant to this study, since space itself is a resource, and as such is valuable to the individual. Different strategies for resource location have been evolved by almost every animal on earth. These vary mainly in the degree to which the organism can detect environmental clues detailing the best resources. Perhaps the best-documented cases, (such as “random walk” behaviour) have been for birds (Bell, 1991).

Any study which attempts to investigate links between resource patch distribution and distribution of a species must not ignore the influence of other factors that may not be constant. Whilst a patch may be selected according to its contents, the usefulness of a habitat to a given species also depends on factors, such as temperature, humidity and solar radiation, which are additional to the patch resources. These factors are liable to change regardless of the properties of the resource patch, and may alter the availability or suitability of a particular resource, causing a change in species distribution independent of patch dynamics.

“A necessary prerequisite for estimating extinction rates…(and population growth rates in general)…is a database that adequately represents the biota and is comprehensive in its geographical coverage”, (Lawton & May 1995). This may seem like a fundamental and obvious requirement for any study attempting to investigate species distribution and density; so obvious that it is often taken for granted that such data exists. In reality, only a few long running field studies can claim to have adequate databases. The SMRU photographic sets used in this study represent one such database. At present, the massive amount of information held within these photographs has no easy way of being analyzed at the fine scale required for detailed population studies.

Ideally, a fully computerized method involving georeferenced maps underlying digitized aerial photographs would allow the exact position, age, sex, etc. of every seal to be accurately plotted, together with information about terrain types and environmental conditions. The benefits of this system would mainly be due to standardization of a counting and classification procedure, the elimination of much human error, the speed with which a comprehensive database could be constructed, and the accuracy of spatial data obtained. Initially it was intended to setup such a system, where photographs would be montaged together, seal positions marked, and habitat types identified, all using computers.

Unfortunately, with the resources and time available this proved to be impossible. The underlying problem was how to load the pictures into the computer in a way that maintained picture resolution, but was not excessively memory greedy. Presently the maximum resolution of scanners (~2000 dpi) and the size of the memory files created (many gigabytes) makes the use of scanned images impractical (Duck C, Pers. Com.). An alternative method involving video images was also rejected1.

After this failure, a workable second method had to be devised inorder to obtain results, in the remaining time available. It was decided that a system, which involved manually inputting data into a GIS (Geographical Information System) software package, would provide results. ArcView 3.0a GIS by ESRI was selected. Major problems getting fully aquainted with this complex software, further delayed the study, with frequent loss of data files, and other unexplained system crashes.

This study therefore points the way for those who which to develop a system for the comprehensive study of many aspects of Grey seal ecology and biology and the terrain on which they breed. It attempts to define a basic workable rationale and method for the fine-scale study of Grey seal distribution and terrain type. It also attempts to find a correlation between the spatially distributed pattern of Grey seal pups and six categories of terrain types. The criterion for selection of these categories is discussed later.

2: Method
2.1 The Islands
In this study, SMRU aerial photos where analyzed in an attempt to identify correlations between fine scale topographical features and the distribution of seals. Existing SMRU Grey seal pup census data was examined inorder to select suitable islands. The clarity of photographs, although taken from approx. 1,200 feet with a pre-focus camera, was sufficient to allow identification and location of seals and topographical features. Islands whose populations seem to have reached a maximum upper limit were selected. These islands should allow the identification of terrain features which seals seem to prefer. Islands were the animal population is still increasing, contain vacant areas that seals have simply not needed to disperse into, and cannot be used to measure correlations. Increasing populations can be studied if correlations are found in stable populations, and estimates of the maximum seal population, and the area of island covered by the colony can be derived.

Initially four stable islands, (Little-Linga (59? 9’30”N, 2? 41’W), Shillay (57? 31’30”N, 7? 40’45”W), Stockay (NW), and Gasker (NW)) and two increasing populations, (Shivinish (57? 31’N, 7? 38’10”W) and Copinsay (58? 54’N, 2? 41’W)) were selected. For the purposes of this project the islands of the Outer Hebrides and the Orkney isles have been generalized into two topographic types; high, hilly, islands with gullies leading to the sea, and low lying, flat, sandy islands. All islands selected for these studies are of the low-lying type. Stockay, Gasker and Copinsay, although originally considered, were later eliminated due to technical problems with software and incorrect information supplied by the Ordinance Survey.

All three remaining islands (Shillay, Shivinish and Little-Linga) are topographically very similar. Low, grassy middles, with small pools of water and mud patches, are bordered by alternating rocky outcrops and sandy beaches. This topography allows seals access sites to most the island. On all three islands, seals are found throughout the grassy tops, but particularly around areas of water.

2.2 Photographic sets
Photographs were chosen from the annual sets of SMRU photos taken of each island. The data used in this study were collected from aerial photographs taken during the 1996 breeding season. Sets of photos that corresponded best with the known height of the pupping season (Harwood, J. unpublished) were selected in order to maximize the number of pups visible on the island. Each Island had one set of photos selected for it: (#20)Shillay 26/10/96; (#20)Shivinish 26/10/96; (#24)Little-Linga 5/11/96.
When the photographs are taken, there is always a certain amount of overlap between adjacent images. These areas of overlap would result in an exaggerated count of seals, if not taken into account. The SMRU mark identical points that occur in both photos by punching small holes in the photographs. Straight lines joining these points are drawn onto the films, to denote the boundary between differing areas on each photo.

2.3 Mapping and recording methods Prominent topographic features including the waterline, vegetation boundaries, beaches and rock outcrops were first traced from the individual photographs in each set, to create a map of the island (FIG). The area in each photo, which does not overlap with any other photo is marked on these photographs with red pen by the SMRU. This boundary was also traced onto 5mm graph tracing paper to form a set of overlays to the traced map (FIG). Individual photos were examined using a microfiche reader (22? magnification) adapted to take a 5mm gridded glass base-plate (FIG). In this way, the position and number of pups and topography type could be recorded on the 5mm graph tracing paper, and related to the traced map.
Topography type was categorized into six basic components, which were each given a code number, i.e.:

**Table1**
Code	Description
1	Water
2	Rock
3	Mud
4	Sand
5	Vegetation
6	Seaweed

The prominent two or three vegetation types in each square were recorded as a combination number, i.e.:

**Table 2**
Code	Description
12	Water & Rock
45	Sand & Vegetation
456	Vegetation, Sand & Seaweed
23	Rock & Mud
234	Rock & Mud & Vegetation
Relevant coding	Any prominent combination

Having recorded the required information from the photographs on to paper, the next stage was to transfer this data into a software package that would allow its manipulation.

2.4 About ArcView 3.0a
It is important to explain the theory behind GIS, before describing the method used in this study. A GIS is not just a digital map! Geographical information systems consists of computer software and hardware for entering, storing, retrieving, transforming, measuring, combining, sub-setting, and displaying spatial data that have been digitized and registered to a common coordinate system. (See Appendix 1 for a more detailed description of GIS and ArcView 3.0a GIS).

2.5 Inputting data into GIS Software
Geo-referenced, digitized Ordinance Survey maps were imported into ArcView 3.0a GIS as Tiff files. The tracing maps created earlier were photocopied onto acetates. By blu-tacking these to the computer screen, the shape of the photographs could be accurately drawn onto the maps, by creating a polygon shapefile (ESRI, 1996)(see Appendix 1). The graph tracing paper templates were also photocopied to acetates, and each 5mm cell was drawn as a polygon in the same manner.

Having replicated the 5mm cells exactly, as a polygon shapefile on top of the Tiff image map, a blank table of, “attributes”(ESRI, 1996)(Fig. 1) was created for it.
“Fields” for each set of data were created and named appropriately to create a table ready for the input of polygon attributes i.e.

Coded data recorded earlier, on the tracing graph-paper templates was inputted onto the rows of this table corresponding to the correct 5mm polygon. It was intended that the area of each 5mm polygon would be calculated in meters. Arcview uses information from the georeferenced Tiff map to do this. Unfortunately the Ordinance Survey failed to provide the relevant (*.twf) files so areas were actually calculated in unknown values. These had to be manually translated to meters and added to the attributes file.

Using the calculations function, the density of pups per meter squared was calculated by dividing the “Pup Numbers” filed by the, “Polygon Area” field. This data was also added to a new field in the table of attributes.

The polygons in the previous shapefile were not regular or all in the same alignment, due to unavoidable problems inserting photographs into the microfiche reader.
Using the Pup density and Terrain type fields, a grid theme (raster type) was imposed as another theme, over the shapefile and tiff image, creating a uniform matrix.
Using the Spatial Analyst extension, (ESRI, 1996) contours of both terrain type and Pup density were created from this matrix and stored as individual themes.

Examination of these contours identified obvious, “hotspots” of high Pup density and allowed an initial appraisal of relationships between Pup density and terrain types. Use of the Spatial analyst extension also allowed distances from grid cells containing specific attributes, to be calculated and recorded as an additional cell attribute (see Appendix 2 for procedure). Unfortunately, a bug in the software was discovered3, and at first this was not possible. To correct the problem, a section of Script was written as a patch (see Appendix 3). Once fixed, the Spatial analyst was used to create additional themes, whose cell attributes were distance from cells containing other, specified attributes, such as the six terrain classifications, or combinations of them.