Estimates of Collision Risk of Harbour Porpoises and Marine Renewable Energy Devices at Sites of High Tidal-Stream Energy

Section 1: High Resolution Boat Surveys of Porpoise Occurrence in and Around Tidal Sites

1.1 Introduction

There are several well developed methods commonly used to investigate porpoise distribution and abundance. The most versatile and prevalent is the use of moving platforms (Hammond 2010). Of these the two most common platforms are 1) planes from which surfacing animals can be sighted or photographed and 2) vessels such as boats from which surfacing porpoises can be seen or submerged vocalising animals can be acoustically detected. While the former offers the opportunity to rapidly survey large areas (particularly those with complex coastlines), the latter also allows small areas to be surveyed in greater detail, potentially at lower cost and with acoustic methods incorporated. Both aerial and boat-based techniques have been used to examine porpoise occurrence over large areas off the west coast of Scotland (plane: SCANS II survey; boat: Embling et al. 2010 and Booth 2010). Because all of these surveys were synoptic in their design and implementation they did not provide definitive information on how frequently porpoises use highly tidal areas on the west coast of Scotland.

Here we focus on boat-based methods to investigate porpoise occurrence in two discrete areas of immediate interest to the tidal-stream industry. Both of these areas are likely to be developed to test and refine tidal-energy extraction device arrays. The first is the Sound of Islay (between the islands of Islay and Jura,55.840°N -6.098°W). These narrows are around 700 m wide at their most constrained and experience tidal streams to at least 2.6m.s ^-1. A development of ten Hammerfest Strøm HS1000 one megawatt tidal turbines is currently being progressed by Scottish Power Renewables. The second site is at Kyle Rhea (the narrows between Skye and the mainland 57.236°N -5.660°W). These narrows are approximately 450 m wide at their most constrained and experience tidal streams of at least to 2.3 m.s ^-1. Two companies have shown interest in this area but site evaluation and environmental consenting are at an earlier stage than the Sound of Islay scheme. Both of these areas are comparable in their size, habitats and tidal-streams and both are likely to become either the world's first or near-first demonstration arrays of full-scale tidal-stream turbines. In terms of surveying, both are directly adjacent to areas of known high porpoise abundance (Booth, 2010) and offer waters that are sufficiently sheltered to allow detailed investigations of porpoise occurrence relevant to the abundance-in-tidal-flow question.

While it may be preferable to survey these relatively small areas using a boat, the strong tidal streams themselves present a significant difficulty. When flowing, the water velocity (2.5 m.s ^-1) itself becomes a significant proportion of the boat's speed (3.6 - 5 m.s ^-1). Survey against the flow and the vessel will make comparatively little progress over the seabed; survey with the flow and the boat will cover much more ground per unit time compared with less tidal areas. These differences are likely to significantly impact the number of porpoises detected. This is because if porpoises orientate relative to the seabed, for example, piloting the boat with the flow will cover more ground and so more porpoises will be detected per unit time than otherwise. Conversely if porpoises station with the water column and not the bottom then surveying against the water flow will also overinflate the number of porpoises detected. Furthermore, with observation time per unit of sea changing then the probabilities of observing intermittently surfacing animals are also affected. Because our knowledge of harbour porpoise behaviour in tidal areas is rudimentary it is unknown how porpoises behave relative to the water column or bottom in these conditions and therefore the presence, nature and magnitude of the biases described above are unknown. Thus it became necessary for this project to modify existing line transect methods so that they were not overly biased by the water flow problem. In the rest of this section outlines the method we formulated and implemented to counter the flow issue and used it to determine and compare sighting and acoustic detection rates of porpoises in the two tidal narrows sites against adjacent waters known to have high densities of porpoises.

1.2 Methods

The Hebridean Whale and Dolphin Trust's 16 m research boat Silurian was used to survey both the Sound of Islay and Kyle Rhea. The boat was deployed as a motor boat (134 horsepower single diesel engine) for these surveys and carried a crew of 10. Visual surveys were conducted in daylight hours in sea conditions equivalent to those expected for wind speeds of Beaufort 3 or less. Two visual observers were on watch at all times during surveys. They were positioned on the foredeck at the mast with an eye height of approximately four meters. They each watched ahead of the vessel to 90° either side of port and starboard respectively. Sightings of marine mammals were called on handheld radio down to a third person below decks who ran the computer sightings database ( Logger 2000, IFAW). This programme continuously logged the boat's GPS track. It also prompted for information on weather and sea conditions every thirty minutes and every time the vessel changed course. The distance to each marine mammal sighting made by the observers was logged along with the bearing from them to the animals and the boat's apparent heading. The real heading was determined from the GPS track and often differed significantly from the boat's apparent heading especially when operating in tidal-streams. Notes were also kept on the species, number, direction of travel and behaviour of the animals sighted. Observers were on watch for a maximum of two hours at a time. The distances estimated by the observers were periodically checked using a laser range finder (Leica Rangemaster 800).

As well as the visual observations, acoustic data were collected by towing a two-element hydrophone array and depth sensor100 m behind the vessel. Porpoise echolocation clicks received by the hydrophone array and amplifiers were sent to a onboard high speed data acquisition card sampling at 500 kHz. This output was streamed to a PC running purpose written software (RainbowClick IFAW; Gillespie, 1997) for real time quality control and recorded along with time and GPS locations on a hard drive for later analysis.

The vessel's survey track was predefined before each day's survey and was intended to survey both the area of tidal interest and less tidal waters to the north or south for comparison. The exact area surveyed was tuned so that the best surface conditions would be encountered given the expected wind conditions and so that the tidal-area itself would be examined in a variety of tidal states (slack low, flood, slack high, ebb). As described in 1.1, using conventional survey methods in tidal-streams (where the water was flowing at speeds that are a sizable proportion of the vessel speed) are likely to result in significant biases. As a result we considered methodological adaptations to reduce tidal-stream specific effects. Surveying with the tidal stream, turning and surveying against it and averaging would be one way to counter any bias but because the tidal flow was a significant fraction of boat speed the against-tide leg would amount to considerably more survey time than the down-tide leg. Furthermore at spring tides, going against the tide would be near impossible for the vessel used. Instead we developed a new method of tacking back and forth across the tidal-stream at an angle so that progress over the seabed equalled progress through the flowing water. Thus porpoises would have equal detection probabilities if they were orientating to the bottom or the water column or (more likely) some combination of the two. Tacking back and forth up-current turned out to be impossible with the rapid rates of flow experienced off Islay and Skye. Instead the boat was piloted across and with the flow at an angle of 58° off the downstream direction of the tide. This angle was arrived at after considering the expected average flow rates at these sites during the survey periods and the cruising speed of the Silurian (See Gordon et al. 2011 for further explanation of this approach).

After the survey the visual sightings were filtered for sea state, site etc. and plotted using GIS. To draw generalisations about porpoise occurrence inside and outside of the tidal stream sites of interest the survey areas were divided into zones which are marked on the maps where appropriate. Acoustic data were analysed using the software package RainbowClick (provided by IFAW; Gillespie, 1997). This package contains click classifiers for harbour porpoise clicks, allowing individual porpoise clicks to be clearly identified. These clicks were subsequently integrated into a large survey database, by matching each click with a specific GPS location along the survey trackline. The survey recorder software was set to record the GPS position every 10 seconds (although in practice this interval sometimes varied due to unexpected technical difficulties). In this manner, clicks were assigned to particular 10-second segments, allowing click rates (clicks/second) to be calculated. It also allowed for the calculation of total number of clicks per km surveyed, once transect lengths were calculated.

1.3 Results

Survey effort was performed in two sessions. In the summer of 2009 we focussed on tidal-stream waters around Sound of Islay. In 2010 we repeated similar survey efforts but this time surrounding the Kyle Rhea narrows 160 km further to the North.

1.3.1 Sound of Islay:

Seven days of survey effort were dedicated to the Sound of Islay and adjacent waters between 2 ^nd and 7 ^th July 2009. The survey vessel Silurian was mobilized to and from Oban and completed a total of 534 km of track-line effort. More than three quarters (78%) of this distance was covered by concurrent visual and acoustic observations ( Table 1). Most of the rest was either visual only(<1%) or acoustic only (18%) observations. "Off effort" work (3%) primarily involved locating overnight anchorages. Average vessel speed over the ground (recorded at 10 second intervals) was 6.4 knots, with little variation among the different survey modes ( Table 2). As expected, the greatest variability occurred during 'Off effort' periods.

Table 1. Overall summary of survey effort in and around the Sound of Islay, July 2-7, 2009, by survey type. 'Off effort' involves times when the vessel was manoeuvring to and from intended survey start / end locations from overnight anchorages.

Survey length (km), by survey type
Date	Acoustic only	Acoustic and visual	Visual only	Off effort	Total length
02/07/2009		64.0	0.5	2.0	66.4
03/07/2009		78.6	3.1	6.1	87.8
04/07/2009		89.0		1.0	90.0
05/07/2009	17.2	65.3		3.2	85.6
06/07/2009	6.2	92.0		1.5	99.8
07/07/2009	73.6	28.8		1.8	104.3
Total	97.0	417.8	3.6	15.6	534.0

Table 2. Average speed (and standard deviations) of the survey vessel during different survey types (aggregated for entire survey). The total number of records per survey type is indicated.

Speeds by survey type	# of records	Average speed (knots)	SD of speed
Acoustic only	3167	6.2	0.9
Acoustic and visual	14127	6.5	0.8
Visual only	136	5.4	1.3
Off effort	793	4.2	2.3
Total	18223	6.4	1.1

The majority of survey effort occurred within the confines of the Sound of Islay ( Figure 1). For analytical purposes, the survey area was subdivided into six broad areas (shown on Figure 1), based on a combination of coastline features and bathymetry:

1. Firth of Lorn, broadly from Mull southward to Colonsay;
2. North of Sound of Islay, including waters between Colonsay and Jura;
3. Northern Entrance to the Sound of Islay(between Islay and Jura);
4. Central Channel of the Sound of Islay;
5. Southern Entrance to the Sound of Islay;
6. South of Sound of Islay, including waters in the Sound of Jura

Outward boundaries of Areas 1 and 6 were generalised to encompass the entire survey track, but were otherwise relatively arbitrary in nature. Most survey effort was in Areas 2-5 (the areas of greatest interest) and involved both acoustic and visual survey effort. The bulk of the acoustic-only effort occurred on the voyage back to Oban when weather conditions were inappropriate for visual observations. Visual-only survey effort occurred briefly on two occasions in Areas 2 and 6, respectively.

The ambient sea state was recorded throughout the survey to help indicate the relatively sightability of surfacing cetaceans given ambient conditions. For this surface wave conditions (rather than wind) based on the Beaufort Scale were judged and recorded. This measure therefore included wind against tide occurrences with associated choppy conditions. The weather was favourable for the majority of this survey and most "on effort" survey work (70%) occurred with sea states at or less than Beaufort 3 ( Table 3). When sea state deteriorated beyond 4 the visual surveyors were typically stood down but the acoustic surveying was maintained.

Sea state conditions varied by area. Generally, sea states were highest in waters to the north of the Sound of Islay and lowest (70-90% < SS 3) in the Central Channel (which was helpfully also the area of most interest for this study) and the Southern Entrance ( Figure 2).

Figure 1. Overview of Sound of Islay survey tracks. Numbered areas include 1) Firth of Lorn; 2) North of Sound of Islay (between Islay and Jura); 3) Northern Entrance; 4) Central Channel; 5) Southern Entrance, and 6) South of Sound of Islay.

Table 3. Lengths of survey transects undertaken at different sea states (aggregated for entire survey).

Transect length (km)	Sea state (Beaufort scale)										Total
Survey type	0.5	1	1.5	2	2.5	3	3.5	4	5	6
Acoustic only				4.6	3.8	7.8	6.1	26.8	40.7	7.2	97.0
Acoustic & visual	27.5	10.7	55.5	124.5	94.0	33.7	22.1	36.5	13.3		417.8
Visual only			0.5		3.1						3.6
Off effort			0.8		2.1	0.5	0.7				4.1
Total	27.5	10.7	56.8	129.1	103.1	41.9	28.9	63.4	53.9	7.2	522.5

Figure 2 . Distribution of total (visual and acoustic) survey effort at different sea states in different areas of the Sound of Islay (arranged from North to South on the x-axis). Higher sea states were most frequently recorded towards the northern part of the survey area.

A total of 60 sighting events (73 individuals) of five marine mammal species were recorded during the visual survey of the Sound of Islay area. The majority of cetacean sightings (89%) were of harbour porpoises (34 individuals) but two minke whales ( Balaenoptera acutorostrata) and two unidentified dolphins were also seen. Both harbour/common ( Phoca vitulina) and grey seals ( Halichoerus grypus) were also seen in the water ( Table 4). The "unidentified seals" were also likely to be one or other of these two species. No sightings of any kind occurred at sea states >4 and the majority (81%) occurred in sea states of 2 or less ( Table 4).

The spatial distribution of sightings varied from species to species. Harbour porpoises were mainly sighted towards the southern end of the Sound of Islay and beyond, with a single sighting in the Central Channel (involving three animals) and two sightings (involving a single individual and a pair) north of the Sound of Islay (Figure 3A). In contrast, seal sightings were concentrated within the Central Channel area (Figure 3C). There were too few sightings of minke whales or dolphins to infer distribution but it is interesting to note that both were observed in the Central Channel section of the Sound of Islay as well as elsewhere (Figure 3C).

Visual sightings data were further investigated for harbour porpoises to determine whether the distribution of sightings deviated significantly from what might be expected if porpoises were distributed randomly across the area. To do this the relationship between sightings and sea state were investigated in combination with the Kyle Rhea data (see below).

Table 4. Summary of marine mammal sightings, sighting rates (#per km surveyed), and standard deviation ( SD) of sighting rates at different sea states during the July 2009 Sound of Islay survey (visual transects only, all areas combined). NB: average sighting rates and SDs were calculated by averaging sighting rates across all transects at each sea state.

Sighting rates per sea state during VISUAL SURVEY		SEASTATE
		0	0.5	1	1.5	2	2.5	3	3.5	4	4.5	5	Total
Transect length (km)		-	27.5	10.7	56.0	124.5	97.1	33.7	22.1	36.5	-	13.3	421.4 km
Harbour porpoise	# of animals sighted		11		5	15	3						34 porpoises
	#/km		0.19		0.07	0.05	0.05						0.04/km
	SD (sightings rate)		0.32		0.42	0.24	0.30						0.25
Minke whale	# of animals sighted					2							2 minkes [1]
	#/km					0.01							0.004/km
	SD (sightings rate)					0.10							0.06
Unidentified dolphin	# of animals sighted				1			1					2 dolphins
	#/km				0.006			0.05					0.005/km
	SD (sightings rate)				0.043			0.26					0.08
Harbour seal	# of animals sighted		1		6	4	2	4					17 seals
	#/km		0.05		0.19	0.04	0.03	0.17					0.06/km
	SD (sightings rate)		0.18		0.59	0.24	0.14	0.92					0.39
Grey seal	# of animals sighted		1	1	3	3							8 seals
	#/km		0.06	0.09	0.07	0.03							0.03/km
	SD (sightings rate)		0.19	0.32	0.28	0.20							0.17
Unidentified seal	# of animals sighted		2	1	3	2	1	1					10 seals
	#/km		0.11	0.07	0.07	0.04	0.01	0.01					0.03/km
	SD (sightings rate)		0.24	0.25	0.31	0.31	0.09	0.06					0.23

Figure 3. Visual sightings of A) harbour porpoises (group size of each sighting indicated), B) other cetaceans and C) seals, observed during the July 2009 survey of the Sound of Islay area. Most porpoise sightings occurred towards the southern part of the Sound of Islay, whereas most in-water seal observations took place in the central channel of the Sound.

Table 5. Summary of harbour porpoise sightings, sighting rates (# seen per km surveyed), and standard deviation ( SD) of sighting rates at different sea states in different areas during the July 2009 Sound of Islay survey (visual transects only).

Harbour porpoise sighting rates per area, per sea state during VISUAL SURVEY		SEASTATE
Area		0	0.5	1	1.5	2	2.5	3	3.5	4	4.5	5	Total
North of Sound of Islay	# of porpoises						3						3
	Survey length (km)					2.3	30.0	14.1	6.6	8.1		61.2	2.3
	Avg. sighting rate (#/km)						0.20						0.09
	SD (sighting rate)						0.58						0.40
Northern Entrance	# of porpoises												0
	Survey length (km)					27.6	24.6	3.5		6.7		8.1	70.4
	Avg. sighting rate (#/km)												-
	SD (sighting rate)												-
Central Channel	# of porpoises				3								3
	Survey length (km)		9.1	7.8	37.8	67.0	14.6	12.1	10.5	13.5		5.1	177.7
	Avg. sighting rate (#/km)				0.07								0.01
	SD (sighting rate)				0.44								0.19
Southern Entrance	# of porpoises		2			6							8
	Survey length (km)		6.7	2.9	3.7	10.4	13.7		4.9	6.3		0.1	48.7
	Avg. sighting rate (#/km)		0.09			0.28							0.08
	SD (sighting rate)		0.17			0.43							0.25
South of Sound of Islay	# of porpoises		9		2	9							20
	Survey length (km)		11.8		14.4	17.1	14.2	3.9		2.0			63.5
	Avg. sighting rate (#/km)		0.76		0.12	0.53							0.24
	SD (sighting rate)		0.04		0.27	0.50							0.38
Overall	# of porpoises		11		5	15	3						34
	Survey length (km)		27.5	10.7	56.0	124.5	97.1	33.7	22.1	36.5		13.3	421.4
	Avg. sighting rate (#/km)		0.17		0.07	0.05	0.05						0.04
	SD (sighting rate)		0.32		0.41	0.26	0.28						0.25

In addition to the visual surveys, hydrophones were towed behind the boat to pick-up the echolocation signals of porpoises in the vicinity of the survey vessel. These data are similar to sightings in that the presence of porpoises can be deduced. However it is important to note that because porpoises do not echolocate all of the time an absence of detections does not confirm absence. That said acoustic detection is less prone to bias from adverse sea-states so provides parallel perspectives on porpoise distribution. Slightly more acoustic effort was collected than visual because of the sea states during the survey (acoustic: 514, visual: 421 km effort)

As with visual sightings, porpoise clicks were not particularly frequently recorded across the whole study area. Overall, a total of 169 clicks were detected in 45 events, of which a substantial number (n=19) involved the detection of only a single click. As these events could have been artefacts caused by high ambient noise or other non-porpoise sources, they were discarded from further analysis ( Table 6).

Table 6. Summary of harbour porpoise click event detections, detection rates (# events per km surveyed), and standard deviation ( SD) of detection rates in different areas during the July 2009 Sound of Islay survey (acoustic transects only). Events involving only a single click have been excluded from the bottom half of this table.

Harbour porpoise click events detected during ACOUSTIC survey effort	Area
	Firth of Lorn	North of Sound of Islay	Northern Entrance	Central Channel	Southern Entrance	South of Sound of Islay	Total
Survey length (km)	42.1	77.6	76.4	190.6	48.7	79.4	514.8
# of click events detected	21	1		7	2	14	45
Average detection rate (click events/km)	0.25	0.002		0.03	0.03	0.11	0.03
SD of detection rate (click events/km)	0.34	0.011		0.38	0.11	0.29	0.31
# of click events detected (>1click per event)	11			4	1	10	26
Average detection rate (click events/km, >1 click per event)	0.13			0.02	0.01	0.09	0.02
SD of detection rate (click events/km, >1 click per event)	0.18			0.24	0.07	0.24	0.20

Remaining click events were used to calculate click rates (clicks per second) for each point along the transect where such an event was detected ( Figure 6). For each click event, the number of clicks within each time period from one GPS position reading to the next (typically 10 seconds) was calculated to generate an average click rate (clicks/second). All such click rates were aggregated according to surveyed area and plotted in 1 click/second bins ( Figure 5). Sea states were not found to have an obvious impact on acoustic detection rates. Generally speaking, click rates were low in all areas, rarely exceeding 1 click/second.

Figure 4. Frequency distribution of harbour porpoise click rates at events involving >1click, as detected during the acoustic survey of the Sound of Islay area, July 2009. Note that detection rates are quite low overall.

As with sightings, the distribution of porpoise click event detections in the areas immediately surrounding the Sound of Islay varied from one area to another ( Figure 6). Most click events were recorded in the Firth of Lorne during the return voyage to Oban at the end of the survey, in an area where porpoises have regularly been reported in the past (Booth 2010). There were no click events (involving >1 click) detected anywhere around the northern section of the Sound of Islay, with a small number of events detected in the Central Channel itself and more in areas 5 and 6 further south. When the two independent methods of visual and acoustic detections were superimposed, the distribution of both broadly correspond (Figure 7 & 8). Because the acoustic methods are less prone to interference from sea conditions, the coincidence of results from these two independent techniques suggest that the finding of a predominantly southern distribution pattern derived from the visual sightings was not an artefact of just the sea conditions at the time of the survey.

Figure 5. Acoustic detections of harbour porpoise clicks recorded during the July 2009 survey of the Sound of Islay area. Asterisks denote events where only single clicks were recorded (which might have been artefacts caused by background noise).

Figure 6 . Detailed view of distribution of harbour porpoise sightings and click detections within the Sound of Islay area. Visual and acoustic observations were closely correlated in several instances. The concentration of observations towards the southern end of the Sound of Islay is notable.

1.3.2 Kyle Rhea

We performed nine days of survey effort in and around the Kyle Rhea area on-board the Silurian between May 13 and May 21 2010.The survey ran out of Tobermory on Mull with a primary focus on the Kyle Rhea but also included significant effort in the Sound of Sleat, some in Loch Alsh and a small amount in Loch Duich. A total of 776 km of track line was surveyed during this period, with concurrent visual and acoustic observations collected during the vast majority (78%) of the time spent "on effort" surveying which itself accounted for 84% of the distance covered ( Table 7).

Table 7. Overall summary of survey effort in the Kyle Rhea, May 13-21 2010, by survey type. 'Off effort' involved times when the vessel was manoeuvring to and from intended survey start / end locations from overnight anchorages or during the vertical array trials (see Section 2).

Survey length (km), by survey type
Date	Acoustic only	Acoustic and visual	Visual only	Off effort	Total length
13/05/2010	69.3			30.3	99.6
14/05/2010	14.4	77.6	2.6	10.6	105.3
16/05/2010	19.2	62.4		10.6	92.2
17/05/2010	0.8	43.2		27.0	71.0
18/05/2010	22.8	19.7		15.1	57.6
19/05/2010		115.2	5.4	6.1	126.7
20/05/2010	10.4	79.5		17.9	107.8
21/05/2010		112.0		3.6	115.6
Total	136.9	509.8	8.0	121.2	775.9

Average vessel speed, recorded at 10-second intervals, was 5.0 knots, with little variation among different survey types and significantly lower speeds during "Off effort" transects ( Table 8). As in the Sound of Islay survey, the greatest variability occurred during 'Off effort' periods which involved a range of vessel movements unrelated to the actual survey.

Table 8. Average speeds, and standard deviations ( SD), of the survey vessel Silurian during different survey types (aggregated for entire survey). The total number of records ( GPS locations at 10-second intervals) per survey type is indicated.

Speeds by survey type	# of records	Average speed	SD of speed
Acoustic only	4702	6.2	1.3
Acoustic and visual	21082	5.8	2.1
Visual only	282	5.7	1.1
Off effort	9224	2.5	2.6
Total	35290	5.0	2.6

For analytical purposes, the survey area was subdivided into four broad areas (see Figures 8 and 9 for boundaries and place names) based on a combination of coastline features and bathymetry:

1. Loch Alsh/Loch Duich (north of the Central Channel of Kyle Rhea);
2. Central Channel of Kyle Rhea (from the northern entrance south to Bernera Bay beyond the southern entrance);
3. Sound of Sleat North (the area immediately south of the Central Channel of Kyle Rhea, down to the Sandaig Islands);
4. Sound of Sleat - Mull Approaches (the remaining southern portion of the Sound of Sleat, as well as waters between Ardnamurchan and the Small Isles and the Sound of Mull that were traversed on the way to/from Kyle Rhea)

Outward boundaries of Areas 1 and 4 were generalised to encompass the entire survey track, but were otherwise arbitrary. Most survey effort in Areas 1, 2 and 3 (the areas of greatest interest) involved both acoustic and visual survey effort. Due to weather a considerable amount of acoustic-only effort was collected on the voyage out from Tobermory to Kyle Rhea on the first survey day. Due to technical difficulties visual-only survey effort occurred briefly on three occasions in Areas 1 and 4 (2 transects). Area 4 was by far the largest area and the absolute majority of survey effort occurred within its boundaries (Figures 7 and 8).

As with the Sound of Islay, the ambient sea state was recorded throughout the survey to help indicate the relatively sightability of surfacing cetaceans given local surface conditions. The weather was generally more favourable with 92% of surveying being conducted in waters of sea state 3 or less ( Table 9).

Table 9. Lengths of survey transects undertaken at different sea states (aggregated for entire survey).

Transect length (km)		Sea state
Survey type	0	0.5	1	1.5	2	2.5	3	3.5	4	Total
Acoustic only	0.1	10.4	0.4		20.5	45.4	12.9		46.7	136.4
Acoustic & visual	76.2	120.5	62.0	71.5	76.3	69.1	26.7	4.8	1.5	508.7
Visual only				5.4						5.4
Off effort	12.4	8.5	4.4	28.7	27.6	11.1	9.8		6.0	108.3
Total	88.6	139.4	66.8	105.6	124.4	125.6	49.4	4.8	54.2	758.9

Sea state conditions varied by area both because of the boat's position as weather systems passed over but also due to wind against tide circumstances. This was particularly apparent in the Sound of Sleat North area where the ebbing tide from the Kyle Rhea met the Sound of Sleat and its long south-westerly fetch ( Fig.9). During periods of particularly strong south westerly winds we took the opportunity to survey Loch Alsh and Loch Duich instead. No survey effort was undertaken in sea states>4.

Figure 7. Overview of survey track in and around Kyle Rhea area. Numbered areas include 1Loch Alsh/Loch Duich; 2) Central Channel of Kyle Rhea; 3) Sound of Sleat North; 4) Sound of Sleat - Mull Approaches. Dark grey boundaries delineate different Areas discussed in the text.

Figure 8. A higher resolution overview of survey tracks in and around the Kyle Rhea area. Numbered Areas include 1Loch Alsh/Loch Duich; 2) Central Channel of Kyle Rhea; 3) Sound of Sleat North; 4) Sound of Sleat - Mull Approaches. Dark grey boundaries delineate different Areas discussed in the text.

Figure 9. Distribution of total (visual and acoustic) survey effort at different sea states in areas of the Kyle Rhea area (arranged from North to South on the x-axis). Higher sea states were most frequently recorded within the central parts of the survey area where wind against tide conditions were frequent.

A total of 338 individuals were recorded during the visual survey of the Kyle Rhea area, involving four different species, including harbour porpoise, harbour/common seal, grey seal and a single swimming otter ( Lutra lutra, Table 10). Some seal sightings could not be conclusively identified to harbour or grey and were recorded as Unknown Seal. Overall, harbour porpoises and harbour seals were the most frequently encountered species. A further 8 sightings (involving a total of 14 harbour porpoises) were made opportunistically during acoustics-only transects or during off-effort periods, but were not considered during subsequent analyses. The vast majority (90%) of sightings occurred in sea states of 2 or less.

Table 10. Summary of marine mammal sightings, sighting rates (# seen per km surveyed), and standard deviation ( SD) of sighting rates at different sea states during the May 2010 Kyle Rhea survey (visual transects only, all areas combined). NB: average sighting rates and SDs were calculated by averaging sighting rates across all transect segments with a particular sea state, however short.

Sighting rates per sea state during VISUAL SURVEY		SEASTATE
		0	0.5	1	1.5	2	2.5	3	3.5	4	Not recorded	Total
Transect length (km)		76.2	120.5	62.0	77.0	76.3	69.1	26.7	4.8	1.5	3.6	517.8 km
Harbour porpoise	# of animals sighted	33	30	16	4	3	6	-	-	-	-	92 porpoises
	Avg. sighting rate (#/km)	0.36	0.10	0.09	0.04	0.03	0.06	-	-	-	-	0.07/km
	SD (sightings rate)	0.66	0.27	0.28	0.23	0.17	0.26	-	-	-	-	0.27
Common seal	# of animals sighted	6	44	26	43	28	11	1	-	1	-	160 seals
	Avg. sighting rate (#/km)	0.56	0.75	0.66	0.84	0.50	0.37	0.07	-	1.03	-	0.57/km
	SD (sightings rate)	1.54	3.05	1.73	2.34	1.39	1.53	0.37	-	1.78	-	2.00
Grey seal	# of animals sighted	1	8	8	16	7	3	-	-	1	-	44 seals
	Avg. sighting rate (#/km)	0.006	0.13	0.24	0.36	0.10	0.08	-	-	1.03	-	0.16/km
	SD (sightings rate)	0.025	0.48	0.61	1.09	0.59	0.46	-	-	1.78	-	0.68
Unidentified seal	# of animals sighted	6	10	3	6	5	10	1	-	-	-	41 seals
	Avg. sighting rate (#/km)	0.51	0.08	0.08	0.12	0.09	0.21	0.13	-	-	-	0.13/km
	SD (sightings rate)	1.26	0.38	0.34	0.53	0.49	1.00	0.71	-	-	-	0.64
Otter	# of animals sighted	-	-	-	-	-	1	-	-	-	-	1 otter
	Avg. sighting rate (#/km)	-	-	-	-	-	0.03	-	-	-	-	0.005/km
	SD (sightings rate)	-	-	-	-	-	0.25	-	-	-	-	0.100

There were no absolute patterns of geographical distribution between species across the survey area (Figures 10 and 11). Most porpoise sightings occurred within the Sound of Sleat - Mull Approaches area. Around the Kyle Rhea itself, porpoises were seen predominantly in the Sound of Sleat North and though they clearly use the Central Channel of Kyle Rhea only one sighting occurred despite much effort and more than half of this in near-ideal sea conditions (sea state ≤1.5, Figures 10 and 12). Seal sightings occurred throughout the survey but were very abundant in the Central Channel of Kyle Rhea ( Figure 11). The otter was seen swimming in open water in the Central Channel of Kyle Rhea ( Figure 11).

For harbour porpoise, sightings data were stratified spatially to determine whether the distribution of sightings varied significantly from what might be expected if porpoises were distributed randomly across the area ( Table 11). Sighting rates were highest in the Sound of Sleat - Mull Approaches area, and lowest in the Central Channel of Kyle Rhea. Sighting rates in Loch Alsh and Loch Duich were somewhat lower than anticipated given historic observations of porpoises in this area. In the immediate area around the Central Channel of Kyle Rhea, highest sighting rates were recorded in the Sound of Sleat North.

Figure 10. A general overview of marine mammal sightings and porpoise click detections in and around the Kyle Rhea area. Porpoises were encountered more frequently in the Sound of Sleat (particularly Sound of Sleat North), whereas most seal sightings occurred within the Central Channel of Kyle Rhea.

Figure 11. A more detailed look at marine mammal sightings and porpoise click detections in the immediate Kyle Rhea area. Numbered Areas are described in the text. Within this area, porpoises were encountered more frequently in the Sound of Sleat (particularly Sound of Sleat North, immediately south of the Central Channel), whereas most seal sightings occurred within the Central Channel of Kyle Rhea. Click events detected within the Central Channel itself typically involved only small numbers of clicks.

Table 11. Summary of harbour porpoise sightings, sighting rates (# seen per km surveyed), and standard deviation ( SD) of sighting rates at different sea states in different areas during the May 2010 Kyle Rhea survey (visual transects only).

Harbour porpoise sighting rates per area, per sea state during VISUAL SURVEY		SEASTATE
Area		0	0.5	1	1.5	2	2.5	3	3.5	4	Not recorded	Total
Loch Alsh/Loch Duich	# of porpoises					1	2					3
	Survey length (km)		15.8	7.2	15.1	25.6	8.2				3.6	75.4
	Avg. sighting rate (#/km)					0.04	0.12					0.03
	SD (sighting rate)					0.21	0.38					0.19
Central Channel of Kyle Rhea	# of porpoises						1					1
	Survey length (km)	2.5	19.5	8.0	18.0	20.1	14.1	5.5		0.3		88.0
	Avg. sighting rate (#/km)						0.05					0.01
	SD (sighting rate)						0.25					0.10
Sound of Sleat North	# of porpoises		2	2	3		3					10
	Survey length (km)		7.2	16.1	16.4	8.0	28.9	10.9	4.2	1.2		92.9
	Avg. sighting rate (#/km)		0.26	0.10	0.21		0.07					0.09
	SD (sighting rate)		0.50	0.36	0.55		0.25					0.32
Sound of Sleat - Mull Approaches	# of porpoises	33	28	14	1	2						78
	Survey length (km)	73.7	78.0	30.8	27.5	22.6	17.9	10.2	0.6			261.4
	Avg. sighting rate (#/km)	0.53	0.28	0.26	0.01	0.19						0.22
	SD (sighting rate)	0.75	0.37	0.39	0.05	0.41						0.44
Overall	# of porpoises	33	30	16	4	3	6					92
	Survey length (km)	76.2	120.5	62.0	77.0	76.3	69.1	26.7	4.8	1.5	3.6	517.8
	Avg. sighting rate (#/km)	0.36	0.10	0.09	0.04	0.03	0.06					0.07
	SD (sighting rate)	0.66	0.27	0.28	0.23	0.17	0.26					0.27

As with the Sound of Islay survey porpoise-like clicks were recorded throughout the Kyle Rhea acoustic survey. A total of 5,301 clicks were detected in 765 events across all areas. The distribution of these porpoise click events in the areas immediately surrounding the Kyle Rhea area varied considerably from one area to another (Figures 10 and 11). There were particularly frequent click detections in the northern part of the Sound of Sleat, immediately south of Kyle Rhea, as well as locally within the southern Sound of Sleat and Loch Alsh/Loch Duich. Within the Central Channel of Kyle Rhea itself most click events consisted of only small numbers of clicks per event with the majority probably being an artefact of background noise. However there was one recognisable echolocation train. This result is consistent with the visual observations and suggests that porpoises do occur in the narrows but at low densities. Porpoise clicks were frequently detected in the area immediately to the south ( Sound of Sleat North) as well as elsewhere within the greater Sound of Sleat area. Detection rates were lower than expected in the Loch Alsh/Loch Duich area given that this was a site historically targeted for its high incidence of porpoises in order to develop porpoise detection methodologies (Goodson et al., 1997).

Undoubtedly some of the single click events detected during these surveys were artefacts of ambient noise. So if all 287 single click events are excluded a less sensitive but cleaner picture of porpoise acoustic presence is revealed (lower half of Table 12).

Table 12 . Summary of porpoise-like click detections, detection rates (# events per km surveyed), and standard deviation ( SD) of detection rates in different areas during the May 2010 Kyle Rhea survey. Events involving only a single click have been excluded from the bottom half of this table.

Harbour porpoise click events detected during ACOUSTIC survey effort	Area
	Loch Alsh/ Loch Duich	Central Channel of Kyle Rhea	Sound of Sleat North	Sound of Sleat - Mull Approaches	Total
Survey length (km)	88.8	91.0	117.0	349.9	646.7
# of click events detected	51	112	244	358	765
Average detection rate (click events/km)	0.26	5.40	6.64	1.07	3.96
SD of detection rate (click events/km)	0.90	60.09	42.40	1.48	43.54
# of click events detected (>1click per event)	35	30	179	234	478
Average detection rate (click events/km, >1 click per event)	0.17	0.20	5.41	0.73	1.44
SD of detection rate (click events/km, >1 click per event)	0.70	0.81	37.61	1.32	17.76

Once single clicks had been excluded, the remaining click events were used to calculate click rates along transects. For each click event, the number of clicks within each time period from one GPS position reading to the next (typically 10 seconds) was calculated to generate the click rate (clicks/second). All such click rates were aggregated according to the surveyed area and plotted in 1 click/second bins ( Figure 12). Sea states were found to have no apparent impact on acoustic detection rates. Generally speaking, click events were less common (~1 event every 5 km) in more northerly areas (Loch Alsh / Loch Duich and Central Channel of Kyle Rhea) and more common in the southern Sound of Sleat (~3.6 every 5 km) and much more common in the northern Sound of Sleat (~27 every 5 km).

Figure 12. Frequency distribution of harbour porpoise click rates at events involving >1click, as detected during the acoustic survey of the Kyle Rhea area, May 2010. Note that detection rates are quite low in more northerly areas, and increase further to the south in the Sound of Sleat areas.

1.3.3 Comparing results between surveys

The two surveys were carried out in distant but geographically similar areas on western Scotland. Both sites are of interest to the tidal-stream energy sector and at the heart of each is a long, relatively smooth tidal narrows with open, deeper water at either end. The surveys were designed to survey the narrows for porpoise occurrence and the waters either side for comparison. To optimise our opportunities for favourable weather, both surveys were conducted in summer (Sound of Islay: early-July, Kyle Rhea: mid-May) of consecutive years and both covered similar distances of visual survey effort (421: 581 km). The weather in both sites was mixed but overall relatively good and slightly better on the Kyle Rhea July survey. Furthermore a slightly longer survey trip in the latter allowed us to avoid periods of adverse weather (including blizzards) by engaging in other activities (see Section 2).

As is well established from other harbour porpoise visual surveys, the chance of seeing porpoises is highly influenced by sea conditions, primarily the smoothness of the surface ("sea state"). The magnitude of this effect varies by survey platform and because the same boat was used in both surveys the results from the two sites are directly comparable. Indeed when the sighting rate and sea state from the two surveys were compared the expected relationship was apparent and similar in both ( Figure 13). A total of 126 porpoises were seen during these surveys, but no sightings occurred in sea states >2.5, and over half (74 animals) were seen in sea state <1. The strength of this relationship means that simply summing sightings (regardless of weather) is misleading. Accordingly, sightings from the visual surveys were displayed in Tables 5 and 10 stratified by sea conditions. There were no obvious changes in porpoise group sizes detected at higher sea states (all sightings on these surveys involved 1 to 3 animals).

Figure 13. Average sighting rate (porpoises/km surveyed) at different sea states, for the 2009 Sound of Islay survey, the 2010 Kyle Rhea survey, and both surveys combined. Only Kyle Rhea data were available at sea state 0. There is an obvious decline in sightings beyond sea state 0.5.

The decline in sighting rates with increasing sea state implies that porpoises went undetected (by the visual observers) in rougher conditions. In order to estimate how many porpoises might have been missed, correction factors were calculated for each sea state from the above graph, by taking the ratio of the mean sightings rate at a particular sea state to the mean of sightings rate at sea state 0 (c _n), using the equation:

c _n = sr _n/sr ₀

where sr _n is the mean sightings rate at sea state =n, and sr ₀ is the mean sightings rate at sea state =0 (Evans & Hammond 2004). Porpoise sighting events ( i.e. reports of at least 1 porpoise sighted) were multiplied by these correction factors to estimate how many such events might have been missed. If data from both surveys are combined, a total of 209 additional sighting events would have been detected if sea states had been 0 throughout.

When accounting for sea state, overall harbour porpoise sighting rates (#/km) were very similar between the two survey areas (0.04 km ^-1 in Sound of Islay, vs. 0.07 km ^-1 in Kyle Rhea). When looking at smaller spatial scales, the highest sighting rates (0.22 km ^-1 and 0.24 km ^-1) occurred in the Sound of Sleat - Mull Approaches and South of the Sound of Islay, respectively (see Tables 5, 10). Both of these areas contain large expanses of deeper open water known to be favoured by porpoises at larger spatial scales (Booth 2010). While sightings did occur in both tidal-stream channels, they were much lower (1 - 2 orders of magnitude) than in the open waters to the south.

Because there is far less influence of sea conditions on the detectability of porpoises during the acoustic surveys the results from this method were less weather dependent. Instead acoustics methods have their own biases (ambient background noise which affects detectability is analogous to sea state and echolocation rate affects cue production and is comparable to surfacing rate) but the two methods together offer independent perspectives of spatial porpoise occurrence. Overall the two methods produced very similar results from both sites, namely that porpoises did occur in the tidal narrows but at much lower densities than in adjacent more open waters (in both cases to the south).