MARINE MAMMAL AERIAL SURVEYS IN ECLIPSE SOUND, MILNE INLET AND POND INLET, 1 AUGUST – 17 SEPTEMBER 2015
by
for
LGL DRAFT Report FA0059-2 15 March 2016
MARINE MAMMAL AERIAL SURVEYS IN ECLIPSE SOUND, MILNE INLET AND POND INLET, 1 AUGUST – 17 SEPTEMBER 2015
by
Tannis A. Thomas, Scott Raborn, Robert E. Elliott and Valerie D. Moulton
LGL Limited, environmental research associates P.O. Box 280, King City, ON L7B 1A6 [email protected]
for
Baffinland Iron Mines Corporation 2275 Upper Middle Road East, Suite 300 Oakville, ON L6H 0C3
LGL DRAFT Report FA0059-2 15 March 2016
Suggested format for citation:
Thomas, T.A., S. Raborn, R.E. Elliott and V.D. Moulton. 2016. Marine mammal aerial surveys in Eclipse Sound, Milne Inlet and Pond Inlet, 1 August – 17 September 2015. LGL Draft Report No. FA0059-2. Prepared by LGL Limited, King City, ON for Baffinland Iron Mines Corporation, Oakville, ON. 85 p. + appendices.
Page ii
Disclaimer
This technical report is considered a draft and is subject to review and input from the Marine Environment Working Group (MEWG) established for the Mary River Project.
Page iii
This page intentionally left blank.
Table of Contents
TABLE OF CONTENTS
Summary of Approach and Key Findings ...... xii Extensive Aerial Surveys ...... xiii Photographic Surveys ...... xiii Summary ...... xiv 1 Introduction ...... 1 1.1 Background ...... 1 1.2 Objectives ...... 2 1.3 Narwhals ...... 2 1.3.1 General Distribution and Population Status ...... 2 1.3.2 Migration ...... 4 1.3.3 Summering Areas ...... 4 1.3.4 Narwhal Response to Vessels ...... 5 1.3.5 Subsistence Harvest ...... 6 1.4 Bowhead Whales ...... 7 1.4.1 General Distribution and Population Status ...... 7 1.4.2 Migration and Summering Areas ...... 8 1.5 Other Marine Mammals ...... 9 2 Description of the Study Area ...... 9 2.1 Ice ...... 10 2.2 Bathymetry ...... 10 3 Methods ...... 12 3.1 Survey Design in 2015 ...... 12 3.1.1 Extensive Survey...... 12 3.1.2 Photographic Survey ...... 13 3.2 Data Recording Procedures ...... 13 3.2.1 Extensive Survey...... 13 3.2.2 Photographic Survey ...... 16 3.3 Analysis Procedures ...... 18 3.3.1 Extensive Surveys ...... 18 3.3.1.1 Geographic Strata ...... 18 3.3.1.2 Density Estimates ...... 19 3.3.1.3 Narwhal Group Size ...... 20 3.3.1.4 Shipping Activity ...... 20 3.3.2 Photographic Surveys ...... 20 3.3.2.1 Categorizing Data Relative to Ship Transits ...... 21 3.3.2.2 Calculating Narwhal Closest Points of Approach ...... 21 3.3.2.3 Narwhal Orientation ...... 21 3.3.2.4 Density Estimates ...... 21
Page v Table of Contents
3.4 Statistical Models ...... 23 3.4.1 Extensive Surveys ...... 23 3.4.2 Photographic Surveys ...... 26 4 Results ...... 27 4.1 Survey Effort in 2015 ...... 27 4.1.1 Extensive Surveys ...... 27 4.1.2 Photographic Surveys ...... 27 4.2 Survey Conditions in 2015 ...... 29 4.2.1 Extensive Surveys ...... 29 4.2.2 Photographic Surveys ...... 29 4.3 Shipping Activity in 2015 ...... 29 4.3.1 Extensive Surveys ...... 30 4.3.1.1 Survey Period 1: 1 August 2015 ...... 30 4.3.1.2 Survey Period 2: 16−17 August 2015 ...... 32 4.3.1.3 Survey Period 3: 31 August 2015 ...... 33 4.3.1.4 Survey Period 4: 15 and 17 September 2015 ...... 34 4.3.2 Photographic Surveys ...... 34 4.3.2.1 Photo Survey 1: 18 August 2015 ...... 35 4.3.2.2 Photo Survey 2: 22 August 2015 ...... 36 4.3.2.3 Photo Survey 3: 30 August 2015 ...... 37 4.3.2.4 Photo Survey 4: 4 September 2015 ...... 38 4.4 Extensive Survey Sightings, Density Estimates, and Model Results ...... 39 4.4.1 Marine Mammal Sightings in 2015—Overview ...... 39 4.4.1.1 Pinnipeds ...... 40 4.4.2 Narwhals ...... 42 4.4.2.1 Numbers Observed and Group Size ...... 42 4.4.2.2 General Observations ...... 43 4.4.2.3 Density Estimates ...... 43 4.4.3 Generalized Non-linear Mixed Model ...... 49 4.5 Photographic Survey Results ...... 52 4.5.1 Marine Mammal Sightings—Overview ...... 52 4.5.1.1 Bowhead Whale ...... 52 4.5.1.2 Pinnipeds ...... 53 4.5.2 Narwhals ...... 54 4.5.2.1 Numbers Observed ...... 54 4.5.2.2 Group Size ...... 55 4.5.2.3 Orientation ...... 57 4.5.2.4 Density Estimates and Shipping Activity...... 60 4.5.3 Generalized Non-linear Mixed Model ...... 69 5 Discussion ...... 72
Page vi Table of Contents
5.1 Narwhal Relative Abundance and General Distribution ...... 72 5.1.1 Observed Densities and General Distribution ...... 72 5.1.2 Narwhal Movements within the Study Area ...... 73 5.1.3 Annual Variation ...... 73 5.2 Narwhals and Shipping ...... 74 5.2.1 Extensive Aerial Surveys ...... 74 5.2.2 Photographic Surveys ...... 75 5.2.2.1 Narwhal Orientation ...... 75 5.2.2.2 Narwhal Group Size ...... 75 5.2.2.3 Narwhal Distribution and Relative Abundance ...... 75 5.2.3 Summary ...... 77 6 Acknowledgements ...... 78 7 Literature Cited ...... 80 APPENDIX A: Mapping Narwhal Sightings from Photographs ...... 86 APPENDIX B: Perception Bias Calculations ...... 89 APPENDIX C: GNLMM Model Fit ...... 92 APPENDIX D: Additional Tables ...... 94
Page vii List of Figures
LIST OF FIGURES
FIGURE 1. Aerial survey study area including geographic strata used in analyses...... 10 FIGURE 2. Study area showing water depths...... 11 FIGURE 3. Aerial survey transects flown for the (A) extensive and (B) photographic surveys during early August to mid-September 2015...... 12 FIGURE 4. Twin Otter with camera port used to conduct the aerial surveys...... 15 FIGURE 5. Internal view of ventral camera port and camera installed in the Twin Otter...... 15 FIGURE 6. Geometry of oblique aerial photos (modified from Asselin and Richard 2011)...... 17 FIGURE 7. Systematic photographic image taken during a photographic survey in Tremblay Sound on 30 August 2015...... 18 FIGURE 8. Summary of shipping activity in the study area based on AIS data for 1 August 2015...... 32 FIGURE 9. Summary of shipping activity in the study area based on AIS data for 16−17 August 2015...... 33 FIGURE 10. Summary of shipping activity in the study area based on AIS data for 31 August 2015...... 34 FIGURE 11. Summary of shipping activity in the study area based on AIS data for 15&17 September 2015 ...... 35 FIGURE 12. Summary of shipping activity in the study area based on AIS data for 18 August 2015 ...... 36 FIGURE 13. Summary of shipping activity in the study area based on AIS data for 22 August 2015 ...... 37 FIGURE 14. Summary of shipping activity in the study area based on AIS data for 30 August 2015 ...... 38 FIGURE 15. Summary of shipping activity in the study area based on AIS data for 4 September 2015 ...... 39 FIGURE 16. Harp seal, ringed seal, bearded seal and walrus sightings recorded during the extensive aerial surveys (1 August-17 September 2015)...... 41 FIGURE 17. Unidentified seal sightings recorded during the extensive aerial surveys (1 August- 17 September 2015)...... 42 FIGURE 18. Total narwhal sightings with various group sizes...... 43 FIGURE 19. Narwhal sightings during Survey Period 1 (1 August 2015)...... 44 FIGURE 20. Narwhal sightings during Survey Period 2 (16−17 August 2015)...... 45 FIGURE 21. Narwhal sightings during Survey Period 3 (31 August 2015)...... 47 FIGURE 22. Narwhal sightings during Survey Period 4 (15 and 17 September 2015)...... 48 FIGURE 23. Narwhal density estimates (uncorrected) and number of vessels during the day of the aerial survey in geographic strata ...... 50 Figure 24. Output for the fixed effect interaction of the categorical variable, GeoStrat, and the continuous variable, Julian, used in the generalized nonlinear mixed model (GNLMM) ...... 51
Page viii List of Figures
FIGURE 25. Bowhead whale, bearded seal, walrus, unidentified seal, and unidentified pinniped sightings identified during the photographic surveys (18 August−4 September 2015)...... 54 FIGURE 26. Total narwhal sightings with various group sizes Before, During and After vessel passages...... 56 FIGURE 27. Headings of narwhals in the Milne Inlet survey area during 18 August, 30 August, and 4 September ...... 58 FIGURE 28. Headings of narwhals in Tremblay Sound during 22 and 30 August 2015 ...... 59 FIGURE 29. Estimated narwhal detection function based on photographic survey data...... 60 FIGURE 30. Narwhal sightings in Milne Inlet during Survey 3 (18 August 2015) ...... 61 FIGURE 31. Narwhal sightings in Tremblay Sound during Survey 4 (22 August 2015)...... 62 FIGURE 32. Narwhal sightings in Milne Inlet during Survey 5 (30 August 2015) ...... 64 FIGURE 33. Narwhal sightings in Tremblay Sound during Survey 5 (30 August 2015) ...... 65 FIGURE 34. Narwhal sightings in Milne Inlet during Survey 7, Replicates 1−4 (4 September 2015) ...... 67 FIGURE 35. Narwhal sightings in Milne Inlet during Survey 7, Replicates 5−8 (4 September 2015)...... 68 FIGURE 36. Predicted number of narwhals (individuals per photograph) versus whether the photograph was blocked from the vessel by a land barrier (variable LandBarr described in the methods) based on least square means output from the generalized nonlinear mixed model (GNLMM)...... 70 FIGURE 37. Output for the fixed effect interaction of the categorical variable, BDA (Before, During, After vessel transit), and the continuous variable, PhotoCPA, used in the generalized non-linear mixed model (GNLMM) ...... 71 FIGURE C-1. Observed (circles) and predicted (red line) number of narwhals for each sample plotted against the sorted rank of predicted values for the extensive survey model...... 92 FIGURE C-2. Moving sum of residuals plotted against the linear predictor for the number of narwhals (window size=9) for the extensive survey modelns with an absolute residual greater than the largest observed absolute residual...... 92 FIGURE C-3. Observed (circles) and predicted (red line) number of narwhals for each sample plotted against the sorted rank of predicted values for the photographic survey model ...... 93 FIGURE C-4. Moving sum of residuals plotted against the linear predictor for the number of narwhals (window size=15) for the photographic survey model...... 93
Page ix List of Tables
LIST OF TABLES
TABLE 1. Bathymetry summary (percent of each strata with various water depths) within the Study Area...... 11 2 TABLE 2. Areas (km ) of the geographic strata in the Study Area surveyed during Baffinland aerial surveys in 2007, 2008, 2013, 2014 and 2015...... 19 TABLE 3. Summary of extensive and photographic aerial survey effort in 2015...... 28 TABLE 4. Number of survey replicates flown and photos reviewed Before, During and After the transit of a vessel through the Milne Inlet and Tremblay Sound photographic study area in 2015...... 28 TABLE 5. Summary of shipping activity in the study area for vessels with AIS data during the 2015 open-water period (data from exactEarth Ltd.)...... 30 TABLE 6. Summary of Baffinland shipping activity in the study area during the 2015 open-water period (based on AIS data from exactEarth Ltd.) ...... 31 TABLE 7. Numbers of narwhals and pinnipeds recorded during extensive aerial surveys conducted from 1 August to 17 September 2015 ...... 40 TABLE 8. Observed density estimates (uncorrected) for narwhals in the study area during Survey Period 1 (1 August 2015)...... 44 TABLE 9. Observed density estimates (uncorrected) for narwhals in the study area during Survey Period 2 (16−17 August 2015)...... 46 TABLE 10. Observed density estimates (uncorrected) for narwhals in the study area during Survey Period 3 (31 August 2015)...... 47 TABLE 11. Observed density estimates (uncorrected) for narwhals in the study area during Survey Period 4 (15&17 September 2015)...... 49 TABLE 12. Covariance parameter estimates and tests of significance for GNLMM of extensive aerial survey data...... 51 TABLE 13. Output for the categorical variables used in the generalized nonlinear mixed model (GNLMM) ...... 52 TABLE 14. Numbers of narwhals, bowhead whales and pinnipeds identified during photographic surveys conducted from 18 August to 4 September 2015 ...... 53 TABLE 15. Numbers of narwhals identified during each photographic survey replicate conducted from 18 August to 4 September 2015 ...... 55 TABLE 16. Group sizes of narwhals in Milne Inlet North and Tremblay Sound, categorized by vessel activity, during the photographic aerial surveys...... 56 TABLE 17. Density estimates of narwhals in Milne Inlet during Survey 3 (18 August 2015) ..... 61 TABLE 18. Density estimates of narwhals in Milne Inlet and Tremblay Sound during Survey 4 (22 August 2015) ...... 63 TABLE 19. Density estimates of narwhals in Milne Inlet and Tremblay Sound during Survey 5 (30 August 2015) ...... 63 TABLE 20. Density estimates of narwhals in Milne Inlet during Survey 7 (4 September 2015) . 66 TABLE 21. Covariance parameter estimates and tests of significance for the GNLMM of the photographic data...... 69 TABLE 22. P-values for fixed effects included in the GNLMM of the photographic data...... 69
Page x List of Tables
TABLE 23. Weighted mean distances (m) between photographs and the closest point of approach (CPA) Before, During, and After vessel passages ...... 70 TABLE D-1. Average number of large vessels present in geographic strata per day for each year (first of August-middle of September) based on extensive survey data modelled with the GNLMM ...... 94 TABLE D-2. Narwhal orientation and group size calculated for each photographic survey relative to shipping activity...... 94
Page xi Summary
SUMMARY OF APPROACH AND KEY FINDINGS
The 2015 open-water season marked the first year of Operational shipping for the Early Revenue Phase (ERP) of the Mary River Project. Baffinland’s Marine Environmental Effects Monitoring (EEM) Plan outlined the approach for monitoring of marine mammals and emphasized the importance of documenting potential effects of shipping on narwhals during this first year of increased shipping activity. As such, the aerial survey program was modified in 2015 relative to the 2013 and 2014 when the emphasis was on baseline data collection. During August and September 2015, extensive aerial surveys of the northern shipping route (through Pond Inlet, Eclipse Sound, and Milne Inlet) and adjacent areas (Koluktoo Bay and Tremblay Sound) were conducted in conjunction with photographic surveys of Milne Inlet and Tremblay Sound (see map below). The photographic survey allowed for examination of changes in the distribution and abundance of narwhals in key summering areas relative to periods Before, During, and After ore carriers transited through northern Milne Inlet. In contrast, the extensive aerial surveys were designed to determine whether the overall spatial-temporal pattern of narwhal distribution and abundance in the study area changed and whether narwhal relative abundance within specific geographic strata changed in response to large vessels, primarily those used by Baffinland.
Aerial survey transects flown for the (A) extensive and (B) photographic surveys during early August to mid-September 2015.
Page xii Summary
Extensive Aerial Surveys Results from the statistical model analyses (Generalized Non-linear Mixed Model, GNLMM) of the 2013, 2014, and 2015 extensive aerial data indicated that there was a statistically significant difference in narwhal densities relative to the number of large vessels present in the study area. A consistent decrease in narwhal densities for a given geographic strata occurred as the number of large vessels increased from zero to greater than two vessels. However, the only statistically significant result among pairwise comparisons for this shipping level variable was during periods when there were no large vessels present versus periods when more than two vessels (i.e., three to five vessels) were present. There were about 11× more narwhals when no vessels were present compared to when more than two vessels were present. Despite this finding, and given the increase in vessel activity during 2015, the model term Year was not statistically significant, which indicates that there were no substantial differences in the overall density of narwhals in 2013, 2014 and 2015. These results suggest that while narwhals may be responding to large vessel transits by exhibiting temporary displacement and/or changes in behaviour that reduce sighting probability, large-scale decreases in their density and spatial- temporal distribution were not apparent. This is consistent with the findings of Baffinland’s shore-based study of narwhals conducted from Bruce Head (Smith et al. 2016) and the photographic survey results discussed below.
Photographic Surveys The GNLMM of the photographic data indicated that During large vessel transits through Milne Inlet, the number of narwhals increased significantly (P-value <0.05) in areas farther from the vessel trackline relative to periods Before and After a large vessel transit. This indicates that narwhals in Milne Inlet exhibited at least temporary avoidance of large vessel transits and/or exhibited changes in behaviour closer to the ship track line that affected their detection in photographs. The data indicated that on average, narwhals were ~1.5 times farther from a large vessel track line During its transit than Before and After its transit through the Milne Inlet study area. However, based on the model findings, narwhal avoidance of large vessels seems temporary. The pattern of narwhal abundance relative to distance from the vessel track line was similar both Before and After a large vessel passage. These results should be interpreted with caution given that there were only two photographic surveys of Milne Inlet with high numbers of narwhals present and during which there were data for all periods Before, During and After a large vessel passage. Also, during these two photographic surveys (i.e., on 30 August and 4 September), there was small boat and hunting activity in the study area. Given the limited data collected in Tremblay Sound, particularly during periods Before large vessels transited through Milne Inlet, it is not possible to determine whether narwhals moved into Tremblay Sound to avoid large vessels in Milne Inlet. However, there are indications from two photographic surveys (22 and 30 August) that narwhals in Tremblay Sound possibly responded to ore carriers. Changes in narwhal orientation and locations of the narwhal herds suggested that some narwhals moved out of Tremblay Sound after ore carriers had either left the study area or anchored at Ragged Island.
Page xiii Summary
Summary Results from both the extensive and photographic surveys indicate that narwhal numbers are reduced during periods with large vessel activity, although there are indications that avoidance of vessels may have been temporary. It is uncertain how these statistically significant differences translate into biological significance for narwhals. However, there were no detectable overall changes in the spatial-temporal pattern of narwhal occurrence in their summering areas and no significant changes in their relative abundance from year-to-year. We recognize that sample size was limited, particularly for the photographic surveys, and results were potentially influenced by factors like small boat traffic and hunting in the study area. In addition, there remain important questions about how far away from a vessel narwhals respond, the duration of avoidance, as well as whether narwhals habituate to repeated ship passages.
Page xiv Marine Mammal Aerial Surveys 2015
1 INTRODUCTION This document is a draft technical report presenting results of aerial surveys for marine mammals in Eclipse Sound, Milne Inlet and Pond Inlet in Nunavut during 1 August – 17 September 2015. This is the third annual technical report for aerial surveys supported by Baffinland Iron Mines Corporation (Baffinland) as part of its Marine Environmental Effects Monitoring Plan (SEM and LGL 2016).
1.1 Background Baffinland is developing an open pit iron mine on northern Baffin Island, Nunavut Territory. The Nunavut Impact Review Board (NIRB) granted a Project Certificate (No. 005) in 2012 for Baffinland’s Mary River Project (the “Project”). The Project involved year-round shipping of ore from a port in Steensby Inlet along a southern shipping route through Foxe Basin and Hudson Strait; and shipping of construction supplies and oversized equipment to Milne Port along a northern shipping route through Eclipse Sound, Milne Inlet and Pond Inlet during the open-water season. Changes to the initial development stages of the project, termed the Early Revenue Phase (ERP), were amended and approved by NIRB in March 2014, and then modified and approved by Aboriginal Affairs and Northern Development Canada (AANDC) in April 2014. The ERP includes shipping up to 4.2 million tonnes per year (Mt/a) of ore via Milne Port during the open- water season (late July to late October), and the deferral of ore shipments from Steensby Port. During the ERP, shipment of the ore to market will occur via chartered ore carrier vessels. Shipping of ore from Milne Inlet during the open-water period began in 2015 (i.e., the first year of the ERP Operations Phase), and is expected to continue for the life of the Project (~25 years). During the peak of the Operations Phase, ~50 to 70 vessels (i.e., 100-140 one-way transits) inclusive of ore carriers, sealifts and tankers will ship to Milne Port during each open-water shipping season. During the first year of ERP Operations, Baffinland shipped ~900,000 tonnes with 13 ore carriers and this study was designed to test for the effects of this large vessel shipping on marine mammals. As a condition of the regulatory approval, Terms and Conditions were set out in the Mary River Project Certificate, many of which require monitoring the effects of shipping on marine mammals. The northern shipping route for the Mary River Project is an important summering area and migratory route for narwhals (Monodon monoceros). Narwhals are an important food and income source for the Inuit and hunters from Pond Inlet and they harvest narwhals and in the future may harvest bowhead whales, (Balaena mysticetus) near the northern shipping route during the open-water period. In support of the Environmental Impact Statement (EIS) for the Project, aerial surveys for marine mammals were conducted during the open-water seasons of 2007 and 2008; effort was focused on the peak period of cetacean abundance (late-July and August). Prior to Baffinland’s monitoring efforts, there were limited data on marine mammal distribution and abundance in Milne Inlet, Eclipse Sound, Pond Inlet and adjacent fjords during late-summer and early-fall. To help address data gaps and provide baseline data for future monitoring of marine mammals along the northern shipping route, Baffinland conducted aerial surveys for marine mammals in Eclipse
Page 1 Marine Mammal Aerial Surveys 2015
Sound, Milne Inlet, Navy Board Inlet and Pond Inlet during 31 August – 18 October 2013 and 1 August – 22 October 2014. Aerial surveys were continued in 2015 with a modified approach designed to examine potential effects of large vessel shipping on marine mammal distribution and abundance. The aerial survey results presented here are one component of Baffinland’s Marine Environmental Effects Monitoring (EEM) Plan (SEM and LGL 2016). The 2015 results are complemented by more spatially focused shore-based observations of narwhal response to shipping (conducted in 2013, 2014 and 2015; Thomas et al. 2014; Smith et al. 2015, 2016) as well as an acoustic study of ship sounds and marine mammal vocalizations conducted in 2014 and 2015 (Kim and Conrad 2015, 2016).
1.2 Objectives As in 2013 and 2014, a key objective of the aerial surveys in 2015 was to document distribution, abundance and movements of narwhals during the open-water season in Eclipse Sound, Milne Inlet and Pond Inlet. Another key objective was to assess the effects of large vessel shipping on the spatial-temporal pattern of narwhal distribution and abundance in and adjacent to Baffinland’s northern shipping route using a statistical modelling approach developed for previous technical reports (Elliott et al. 2015; Thomas et al. 2015). Finally, a third objective was to use photographic surveys of narwhals in northern Milne Inlet and Tremblay Sound to document narwhal occurrence Before, During and After a large vessel passage. This latter objective was intended to allow determination of whether there are movements of narwhals in a substantial portion of their primary summering habitat in response to repeated ship traffic. An assessment at this spatial-temporal scale is particularly important during the first operational year(s) of the ERP. By documenting narwhal relative abundance and distribution in relation to a vessel track Before, During and After a transit, it is possible to approximate the extent of potential avoidance, the area(s) where narwhals may move to avoid a ship transit, and approximately how long narwhals avoid an area around a vessel. Although the main species of concern were narwhals and to a lesser extent bowhead whales, data on all marine mammal species were collected during the aerial surveys and are presented in this report.
1.3 Narwhals 1.3.1 General Distribution and Population Status The narwhals that occur along and near the northern shipping route are considered part of the Baffin Bay population which occupies Baffin Bay and adjacent waters in winter. In summer, a large portion of the population aggregates in Canadian waters, in areas ranging from eastern Baffin Island coastal waters to the High Arctic Archipelago (Richard et al. 2010). The remainder of the population summers along West Greenland (Heide-Jørgensen et al. 2010; DFO 2012). Though many Inuit believe that at least two stocks of narwhals exist in the Arctic, narwhals in the eastern Canadian Arctic are currently assessed as a single population by the Committee on the Status of Endangered Wildlife in Canada (COSEWIC). This species is listed as Special Concern by COSEWIC (2004), and has no status under the Species at Risk Act (SARA; GC 2016).
Page 2 Marine Mammal Aerial Surveys 2015
Strong evidence from genetic, contaminant and satellite tracking data has been used to define four management stocks1 that summer entirely in Canadian waters: the Somerset Island, Admiralty Inlet, Eclipse Sound and East Baffin Island stocks (Richard et al. 2010). It is the Eclipse Sound stock that primarily occurs along the northern shipping route. However, there is interchange of narwhals amongst narwhal stocks with individuals moving to different areas within a given summer season and between years (Fallis et al. 1983; Smith et al. 1985; Koski and Davis 1994; Richard et al. 1994). Estimating the population size of narwhals is challenging given the large and remote area occupied and the often less than ideal weather conditions for aerial surveys. There is a long history of estimating portions of the Baffin Bay narwhal population in various parts of the Canadian Arctic during different seasons, but until 2013, none of the surveys covered all of the known summering areas in the same year. In August 2013, DFO undertook a broad scale aerial survey program called the High Arctic Cetacean Survey (HACS) whose objective was to produce new abundance estimates of the Baffin Bay narwhal population (and the Eastern Arctic-West Greenland bowhead whale population; DFO 2014). Surveys were conducted at this time of year to capture peak abundance of narwhals in their summer habitat. Based on these aerial survey data, narwhal abundance was estimated to be ~142,000 for the entire Baffin Bay stock. Surveys were conducted in Eclipse Sound, Milne Inlet, Pond Inlet and Navy Board Inlet on 18−19 August 2013 and an estimate for the number of narwhals that summer in the Eclipse Sound area was 10,489 (coefficient of variation (CV) 0.24; Doniol-Valcroze et al. 2015a). The previous estimate for the number of narwhals that summer in the Eclipse Sound area was 20,211; this is based on aerial survey data collected in 2002–2003 by DFO and which were subsequently reanalyzed (NAMMCO 2010). In some earlier years, narwhals that summered in the Eclipse Sound area were estimated to make up over 30% of the estimated total number of narwhals that summered in the Canadian High Arctic (~63,000 narwhals—but not all summering areas were surveyed); whereas during the 2013 survey only about 7.4% of the Canadian High Arctic narwhals were found there. In addition to the earlier surveys not covering all of the summering areas, there appears to be considerable variation in the year-to-year numbers in the Eclipse Sound area (Richard 2010). Doniol-Valcroze et al. (2015a) discussed how the stock estimates for the 2013 surveys in Admiralty Inlet (∼35,000 narwhals) and Eclipse Sound (~10,000 narwhals) differ substantially from previous survey estimates of the same stocks (18,000 for Admiralty Inlet in 2010 and 20,000 for Eclipse Sound in 2004). They noted that the combined estimate for Admiralty Inlet and Eclipse Sound (~45,000) is similar to the total of previous abundance estimates, and that documented movements of individual narwhals between Eclipse Sound and Admiralty Inlet raise the possibility of some degree of exchange between the two summering areas. They also noted that survey design and
1 DFO has provisionally identified a fifth Canadian stock called the “North Water stock”, which may be shared with Greenland. For management purposes the fifth stock has been provisionally split into two stocks—the Jones Sound and Smith Sound stocks to ensure that the quota for that stock allows a sustainable harvest. More data are needed to ascertain the status for the North Water stock because there are no harvest data from the provisional Smith Sound stock and narwhals do not enter Jones Sound, and particularly the fjords west and northwest of Jones Sound, in all years (Koski and Davis 1979; Koski 1980). For management purposes, it is conservatively assumed that there is little to no interchange amongst narwhals of the various stocks although telemetry data have confirmed that some movement occurs (DFO n.d.) and large between-year changes in population estimates for individual stocks can only be explained by movements between summering areas.
Page 3 Marine Mammal Aerial Surveys 2015
determination of how aggregation rates influence abundance estimates may also have caused the very different estimates for 2002-2003 and 2013. 1.3.2 Migration The timing of narwhal arrival in their summering areas varies considerably from year-to- year, depending on ice conditions. Baffin Bay narwhals typically enter Pond Inlet and Lancaster Sound in late-June and July after breakup of the ice edge. Those entering Lancaster Sound continue to summering areas farther west (Koski and Davis 1979; Koski 1980; Finley and Gibb 1982). Those narwhals entering Pond Inlet move through ice cracks toward Eclipse Sound (Finley and Gibb 1982). Based on Baffinland aerial surveys conducted during the open-water seasons of 2007, 2008, 2013 and 2014, narwhals were present in Pond Inlet and eastern Eclipse Sound during early August, remained in and around Milne Inlet, Koluktoo Bay and Tremblay Sound during mid- to late August, and migrated out of the study area via Pond Inlet starting in mid-September. The Eclipse Sound stock of narwhals starts migrating down the east coast of Baffin Island in late September (Koski and Davis 1994; Dietz et al. 2001). Narwhals generally arrive in their wintering areas in November (Heide-Jørgensen et al. 2003). The Baffin Bay narwhal population winters in very heavy pack ice throughout northern Davis Strait and southern Baffin Bay (McLaren and Davis 1982; Koski and Davis 1994). It is in their wintering areas that they feed heavily and primarily on Greenland halibut (Laidre and Heide-Jørgensen 2005). 1.3.3 Summering Areas Although the timing of narwhal migration into their summering areas is variable, narwhals typically exhibit site fidelity for the same fjords and bays each summer (Heide-Jørgensen et al. 2003; Laidre et al. 2004; Heide-Jørgensen et al. 2015), though some evidence suggests mixing between summering areas does occur (Dietz et al. 2001; Heide-Jørgensen et al. 2002; Watt et al. 2012). They are present in their summering area through September (Mansfield et al. 1975; Finley and Gibb 1982; Koski and Davis 1994; Elliott et al. 2015; Thomas et al. 2015). Aerial survey data collected by Baffinland during five open-water seasons have demonstrated that narwhal numbers tend to be more concentrated from mid-August to early September and that during this period narwhals are most frequently observed in Milne Inlet, Koluktoo Bay and Tremblay Sound despite considerable survey effort in Eclipse Sound, Navy Board Inlet, Pond Inlet and adjacent fjords. Aerial survey findings have also shown that there is much fine scale movement by groups of narwhals among various areas of Milne Inlet and adjacent fjords, both from day to day and over a longer time interval. There was also much variation in the numbers of narwhals observed during the shore-based study of narwhals at Bruce Head, Milne Inlet in 2013, 2014 and 2015 (Thomas et al. 2014; Smith et al. 2015, 2016). The reasons for narwhal site fidelity to summering areas like Milne Inlet are not clear. Mansfield et al. (1975) examined stomach contents of 62 narwhals captured in Koluktoo Bay (at the south end of Milne Inlet) in summer, and few were found to contain food despite extensive runs of Arctic char (Salvelinus alpinus) in the area. Similarly, Finley and Gibb (1982) examined stomach contents of 73 narwhals from Pond Inlet and Eclipse Sound, and found little evidence of feeding in the fjords in the summer. In contrast to this, foraging behaviour (i.e., fish chases and underwater blows) was observed in Milne Inlet in 2013 (Thomas et al. 2014), indicating that
Page 4 Marine Mammal Aerial Surveys 2015
some feeding does occur. Site fidelity to Milne Inlet is likely tied to its suitability for calving or rearing calves (Remnant and Thomas 1992), as evidenced by numerous observations of mother- calf pairs in the area (Mansfield et al. 1975; Kingsley et al. 1994; Marcoux et al. 2009; Thomas et al. 2014; Smith et al. 2015, 2016) and a recent observation of nursing behaviour (Smith et al. 2016). Site fidelity to Milne Inlet may also be related to its suitability as a refuge from predators; narwhals have been observed sheltering from killer whales (Orcinus orca) in the shallows of Koluktoo Bay (Campbell et al. 1988). 1.3.4 Narwhal Response to Vessels Prior to monitoring studies undertaken for Baffinland, there have been no systematic studies of narwhal response to shipping traffic during the open-water season. The shore-based study of narwhal response to shipping near Bruce Head, Milne Inlet has been conducted for three years—two before operational shipping of ore (2013, 2014; Thomas et al. 2014; Smith et al. 2015) and one during the first year of ERP Operations (2015; Smith et al. 2016). That study has collected three primary types of data: group composition, behaviour, and relative abundance and distribution (RAD). RAD data were collected on narwhal groups in a stratified study area (SSA; 82.5 km2 around Bruce Head). The numbers of narwhals in each of nine strata where nominally counted every hour and more frequently when a large vessel (primarily an ore carrier, 225 m in length, ~75,000 dead weight tonnage (DWT)) was present. A modelling approach was used to test the hypothesis that narwhal abundance in the SSA does not change in the presence of large vessel traffic. The generalized linear mixed model (GLMM) of the combined 2014 and 2015 dataset indicated that significantly lower numbers of narwhals were observed in the SSA when large vessels transited south vs. when large vessels were not present (P = 0.003, Smith et al. 2016). The lowest narwhal counts occurred when vessels transited south through the SSA, and the highest narwhal counts occurred when large vessels transited north through the SSA. Ad libetum observations made during large vessel transits northward through the SSA suggest that narwhals generally do not respond to the large vessel presence by fleeing the area. During large vessel transits on 18 and 22 August 2015, groups of narwhals were observed briefly resting while oriented toward the large vessel, before then swimming away and diving. Some narwhals were observed in relative close proximity (i.e., hundreds of metres) to the vessel, and many were observed swimming to the south. During these two large vessel transits, many narwhals were observed to remain within the SSA, though they did exhibit a change in behaviour. During the shore-based study, narwhals in the presence of large vessels were observed to display a small range of group behaviours. Travelling, milling and resting with backs exposed were the primary behaviours observed and diving and bubble rings were the secondary behaviours observed. Additional secondary behaviours that were only observed when large vessels were absent included: side swimming, back swimming, rubbing, tusking and nursing. Data on change in narwhal swim speed in response to large vessel presence were equivocal based on data collected in 2015. Narwhal group sizes were significantly larger in the presence of large vessels (P = 0.049) – though caution must be exercised because with the exception of one group of narwhals, all group composition data collected in the presence of large vessels were collected during a single vessel transit (Smith et al. 2016).
Page 5 Marine Mammal Aerial Surveys 2015
Narwhal response to icebreaking ships was documented in the 1980s (LGL and Greeneridge 1986; Cosens and Dueck 1986; Finley et al. 1990). The most comprehensive studies of narwhal response to icebreaking ships were undertaken during June of 1982, 1983 and 1984 in Lancaster Sound. In each study year, the M/V Arctic, an icebreaking ore carrier (20,000 DWT) was accompanied by the CCGS John A. MacDonald (1982, 1983) or the CCGS Louis S. St-Laurent (1984) in Lancaster Sound as it approached the landfast ice-edge and then moved through the fast ice enroute to the Nanisivik mine in Admiralty Inlet. Narwhals at ice edges waiting to continue their migration to summering areas responded to oncoming vessels and periodic icebreaking by 1) demonstrating a “freeze” response, typically lying motionless or swimming slowly away (as far as 37 km along the ice edge), 2) huddling in groups, and 3) ceasing sound production. The responses of narwhals observed during the three-year study were generally similar to responses to predators (i.e., killer whales) as described by Inuit hunters (Finley et al. 1990). After initially being displaced in response to relatively low levels of noise from the approaching ship, narwhals sometimes returned 1–2 days later (Finley et al. 1990) and engaged in diving and foraging behaviour. Cosens and Dueck (1986) conducted helicopter surveys of leads in the ice across the mouth of Admiralty Inlet, and examined the effects of icebreaker traffic on the relative abundance, distribution, activity and orientation of narwhals. They classified narwhals as being disturbed when ships were within 130 km of the floe edge. Cosens and Dueck (1986) found that disturbed narwhals were in larger groups, swam more slowly and with fewer changes in direction, and rested less than when vessels were absent. Narwhals were also found to change orientation when ships were ~50 km away. Cosens and Dueck (1986) did not observe the “freeze” response observed by Finley et al. (1990), and suggested that this was because there was differing ice conditions (i.e., heavy pack ice that provided cover and lack of landfast ice across Lancaster Sound which allowed movement past the area of icebreaking activity) or because narwhals were becoming habituated to ship traffic. Disturbed belugas were observed to show less directed movement, less resting, and more circling than when vessels were absent. The strong responses of narwhals to icebreaking vessels at long ranges are unique in the literature of vessel noise responses by marine mammals and seem at odds with the findings of the shore-based study of narwhal response to large vessels during the open-water season (Smith et al. 2016). Although the shore-based study is an important component of Baffinland’s marine mammal EEM program, it is limited in spatial scope. The findings of the aerial surveys presented in the present report complement the shore-based study and are essential in understanding narwhal responses to large vessel shipping on a larger spatial scale (i.e., Milne Inlet, Eclipse Sound, and Pond Inlet). 1.3.5 Subsistence Harvest Narwhals are a significant source of food and income for the residents of Pond Inlet. During the open-water period, harvesting of narwhals by Pond Inlet residents occurs in Milne Inlet, Eclipse Sound and Pond Inlet (Volume 4, Section 10 in Baffinland 2012). More specifically, narwhal harvest locations have primarily been identified in Koluktoo Bay, in Tremblay Sound, near Pond Inlet, along the west shore of Milne Inlet, and near Bruce Head,
Page 6 Marine Mammal Aerial Surveys 2015
Milne Inlet during summer and autumn, and at the Pond Inlet ice-edge during late spring or early summer. Based on the Nunavut Wildlife Harvest Study findings (1996-1999), most narwhals are harvested in August followed by July and September (Priest and Usher 2004). Hunting has frequently been observed in August during the shore-based study of narwhals at Bruce Head (Smith et al. 2016). The rocks at the base of the cliff at Bruce Head, are well used by local Inuit for hunting narwhals and less frequently seals. This method of shore-based hunting (i.e., shooting from the shoreline at the base of Bruce Head vs. from a small vessel) is practiced during the open-water season at a number of locations in Eclipse Sound (Finley and Miller 1982). In 2015, most of the hunting (i.e., shooting) activity observed from Bruce Head was conducted from the shore, though shooting was also occasionally observed from small vessels in 2015 (Smith et al. 2016). Narwhal hunting from small motorized boats has also been observed in Milne Inlet during recent Baffinland aerial surveys (this report). In Canada, the narwhal hunt is co-managed by Inuit, DFO, and Nunavut Wildlife Management Board. Recently, a Narwhal Integrated Management Plan has been approved that sets quotas based on Total Allowable Landed Catch (TALC) based on estimates of stock size. The current TALC limit for the Eclipse Sound stock is recommended as 134 (Doniol-Valcroze et al. 2015a). A robust management system for the narwhal hunt is required to meet the Convention on International Trade in Endangered Species (CITES) requirements. Narwhals are listed in Appendix II of CITES. To export narwhal products (i.e., tusks) internationally, a non- detrimental finding decision from the DFO Scientific Authority is required for issuance of a CITES permit for export.
1.4 Bowhead Whales 1.4.1 General Distribution and Population Status The Eastern Canada-West Greenland (EC-WG) population of bowhead whales occurs along the northern shipping route. This population ranges throughout the eastern and central Canadian High Arctic (Davis and Koski 1980; Finley 2001; IWC 2008). The most recent and reliable population estimate for EC-WG bowhead whales is 6446 (95% CI 3722−11,200) based on the DFO HACS conducted in August 2013 (Doniol-Valcroze et al. 2015b). This is very similar to the previous estimate of 6344 (95% CI 3119–12,906; IWC 2009) based on aerial survey data collected in 2002. Authors of both studies acknowledge that the estimates are likely negatively biased since neither study included surveys of all known summering areas. The EC- WG population of bowheads is listed as Special Concern by COSEWIC (COSEWIC 2009) and has no status under SARA (GC 2016). There is a limited subsistence hunt for bowheads in Nunavut and the Nunavik Marine Region that is co-managed by the Nunavut Wildlife Management Board, Nunavik Marine Region Wildlife Board, and DFO (Doniol-Valcroze et al. 2015b). There is also a limited hunt in West Greenland under a quota established by the International Whaling Commission (IWC 2009).
Page 7 Marine Mammal Aerial Surveys 2015
1.4.2 Migration and Summering Areas Like narwhals, bowhead whales have several primary summering areas in the Canadian High Arctic and move into these areas as the ice breaks up and access is possible. From May to early July, bowheads are found at the floe edge and in the pack ice off Lancaster Sound and Pond Inlet (Koski and Davis 1979; Reeves et al. 1983; Moore and Reeves 1993). When the pack ice melts, they quickly migrate through Lancaster Sound to the channels of the Canadian Arctic Archipelago. The migration through Lancaster Sound can extend from early–mid May to early August, with a peak in late June (Davis and Koski 1980). Some bowheads from the EC-WG population migrate northward into Prince Regent Inlet through Fury and Hecla Strait from late June to early July following the breakup of landfast ice in northern Foxe Basin. IQ holders report that bowheads without calves are usually the first to arrive in spring, often as solitary animals or in pairs; mothers with calves tend to arrive later in the season (Appendix 8A-2 in Baffinland 2012). A number of EC-WG bowheads summer in the bays and passages of the central and eastern High Arctic islands during August and September (Davis and Koski 1980; Koski and Davis 1980). Bowheads are believed to return to the same summering areas from year to year, including those along the northern shipping route, i.e., Eclipse Sound, Pond Inlet and Navy Board Inlet; as well as Isabella Bay (eastern Baffin Island); Admiralty and Prince Regent inlets; and Cumberland Sound (Lubbock 1937; Appendix 8A-2 in Baffinland 2012). Milne Inlet, Eclipse Sound and, to a lesser extent, Koluktoo Bay are used by bowhead whales, including mothers and calves, during the open-water season. According to IQ, feeding usually takes place in nearshore, sheltered, shallow waters in summer. Mating occurs in late March to late-April and single calves are born during spring migration of the following year (Nerini et al. 1984; Koski et al. 1993). The importance of Eclipse Sound and adjacent waters as a bowhead summering area is uncertain but it appears relatively low numbers of bowhead whales occur there in most years. Based on the recent HACS, the estimated number of bowhead whales that summered in the Eclipse Sound area in 2013 was 32 (Doniol-Valcroze et al. 2015b). Results of aerial surveys conducted for Baffinland indicate that there is inter-annual variation in the numbers of bowheads that occur in the Eclipse Sound area. During aerial surveys of the same area and generally the same time of year, bowheads were sighted in varying numbers in 2007 and 2008. More specifically, there were 51 sightings totalling 61 bowheads (may include more than one sighting of individual whales) in Milne Inlet, Eclipse and Tremblay sounds, and Koluktoo Bay during 7 August – 17 September 2007 and 12 sightings totalling 15 bowheads in Milne Inlet, Eclipse Sound and Koluktoo Bay during 4 – 26 August 2008 (Appendix 8A-2 in Baffinland 2012). During the 2013 aerial surveys in Eclipse Sound, Milne Inlet, Navy Board Inlet and Pond Inlet, one bowhead was sighted on the 30 September in Navy Board Inlet (Elliott et al. 2015); whereas 13 bowheads were sighted there during the 2014 aerial surveys (Thomas et al. 2015). Small numbers of bowheads (1-6) have been observed during the shore-based study at Bruce Head, Milne Inlet in each of 2013, 2014 and 2015 (Thomas et al. 2014; Smith et al. 2015, 2016). Of note, bowhead calls were detected acoustically on 13 days in August 2014 near Bruce Head, Milne Inlet (Kim and Conrad 2015). The fall migration out of the eastern Canadian Arctic Archipelago begins during late September–early October (Koski and Davis 1980; Heide-Jørgensen et al. 2006). Whales that
Page 8 Marine Mammal Aerial Surveys 2015
summer in Eclipse Sound and Milne Inlet proceed north through Navy Board Inlet and then east along north Bylot Island (Koski and Davis 1980), and then migrate along the east coast of Baffin Island to Isabella Bay where they join whales that have summered in that area (Finley 1990, 2001). The migration along Bylot Island and northeast Baffin Island is primarily along the coast, but satellite telemetry indicates that some animals move offshore (Heide-Jørgensen et al. 2006). All sightings during 1978–1979 aerial surveys were made within 1.5 km of the coast (Koski and Davis 1980) but those surveys were terminated before the end of the migration. The migration is also quite rapid, with the whales averaging speeds of 5 km/h (Koski and Davis 1980). Most EC-WG bowhead whales winter in Hudson Strait but some winter in southern Baffin Bay, e.g., in the pack ice along the coast of western Greenland (Koski et al. 2006; Heide- Jørgensen et al. 2007; Elliott et al. 2013) and Cumberland Sound. From March to May, adult bowheads are found along an ice edge southwest of Disko Island, Greenland, in flaw zones off Cumberland Sound and Frobisher Bay (Moore and Reeves 1993), and in polynyas in Baffin Bay between 60ᵒ and 70ᵒ N (Finley 2001).
1.5 Other Marine Mammals As noted earlier, the focus of this study is narwhals and to a lesser extent bowhead whales. Killer whales and beluga whales (Delphinapterus leucas) occur along the northern shipping route but in much lower numbers and their occurrence in the study area and conservation status has been reviewed in previous Baffinland technical reports (Elliott et al. 2015; Thomas et al. 2015). This study was not designed to survey pinnipeds but they were recorded during aerial surveys when possible. Ringed (Pusa hispida) and harp (Pagophilus groenlandicus) seals are considered common along the northern shipping route during the open-water season. Small numbers of bearded seals (Erignathus barbatus) and walrus (Odobenus rosmarus) may also occur there at this time. Similarly, polar bears (Ursus maritimus) were not the focus of this study but they were recorded during aerial surveys. Polar bear occurrence in the study area and their conservation status has been reviewed in previous Baffinland technical reports (Elliott et al. 2015; Thomas et al. 2015).
2 DESCRIPTION OF THE STUDY AREA The study area includes Navy Board Inlet, Eclipse Sound, Pond Inlet, Milne Inlet, Koluktoo Bay and the smaller fjords of Tremblay Sound, Eskimo Inlet, White Bay, Tay Sound, Paquet Bay and Oliver Sound (Figure 1). These marine areas encompass ~ 8354 km2. The study area is considered a large fjord system, typically bounded by cliffs and steep terrain that rise inland to mountains, glaciers and ice fields. However, some parts of Milne Inlet, Navy Board Inlet, Eclipse Sound and Pond Inlet have relatively wide coastal plains that challenge the classical definition of a fjord. The smaller embayments of Oliver Sound, Tremblay Sound, Eskimo Inlet, Tay Sound and Paquet Bay have classical fjord like qualities such as deep water, U-shaped valley cross-sections and submerged moraines or sills at the seaward end of the water bodies.
Page 9 Marine Mammal Aerial Surveys 2015
FIGURE 1. Aerial survey study area including geographic strata used in analyses.
2.1 Ice The study area is covered by landfast ice during the winter that gives way to open water frequented by ice bergs during the summer and early fall. Landfast ice has typically melted by the end of July and freeze-up begins in the fall during the first half of October. Ice cover usually begins forming within the narrow fjords and northern portion of Navy Board Inlet first likely due to the diminished wave action when compared to other parts of the study area.
2.2 Bathymetry Water depth within the study area is quite variable and reaches a maximum depth of 1100 m (see Table 1 and Figure 2). Water is relatively deep in the northern portion of Navy Board Inlet, throughout Eclipse Sound, Pond Inlet and Milne Inlet North. The southern half of Navy Board Inlet, Milne Inlet South, as well as White Bay, Eskimo Inlet and south Paquet Bay are relatively shallow water (< 200 m).
Page 10 Marine Mammal Aerial Surveys 2015
TABLE 1. Bathymetry summary (percent of each strata with various water depths) within the Study Area.
Water Milne Milne Eclipse Navy Eclipse Koluktoo Eskimo Tremblay White Tay Paquet Oliver Pond Depth Inlet Inlet Sound Board Sound Bay Inlet Sound Bay Sound Bay Sound Inlet (m) South North West Inlet East 0-50 35.4 23.3 11.7 43.7 29.5 7.6 17.4 54.9 5.7 22.7 60.2 25.0 6.4 50-100 18.4 21.2 6.3 14.7 13.7 6.6 10.2 17.8 5.0 12.1 18.1 16.8 4.5 100-200 31.5 36.5 11.4 23.1 23.3 36.6 25.4 16.0 11.4 17.4 5.4 18.3 5.3 200-300 13.7 13.3 18.3 18.5 26.1 13.1 17.6 7.4 10.3 12.3 6.8 28.3 5.7 300-400 1.0 5.6 14.3 7.3 8.6 16.8 2.7 10.4 10.7 7.0 11.4 6.0 400-500 11.3 9.6 7.8 1.1 10.9 13.4 1.8 0.2 6.9 500-600 15.0 15.5 3.1 11.7 10.3 0.7 11.7 600-700 9.2 2.4 1.9 16.9 1.2 27.6 700-800 2.6 17.5 10.2 800-900 0.2 6.6 900-1000 4.8 1000-1100 4.3
FIGURE 2. Study area showing water depths.
Page 11 Marine Mammal Aerial Surveys 2015
3 METHODS
3.1 Survey Design in 2015 The aerial survey study design was modified in 2015 to include a photographic survey (Figure 3B) of narwhals in northern Milne Inlet and Tremblay Sound that focused on times Before, During and After vessel passages. The extensive survey (Figure 3A), similar to those conducted during baseline data collection in 2013 and 2014 (and 2007 to 2008) was also conducted. This “two-pronged approach” was undertaken to allow for detection of large-scale as well as finer-scale changes in narwhal distribution and abundance.
FIGURE 3. Aerial survey transects flown for the (A) extensive and (B) photographic surveys during early August to mid-September 2015. 3.1.1 Extensive Survey From 1 August to 17 September, an extensive grid was surveyed during four periods, i.e., every two weeks (early August, mid-August, late August, mid-September). The extensive aerial surveys were designed to have repeat coverage of Eclipse Sound, Pond Inlet and Milne Inlet on a bi-weekly schedule. Each extensive survey consisted of 21 linear transects in Eclipse Sound, Milne Inlet and Pond Inlet; 23 “zigzag” transects in the southern part of Milne Inlet and Koluktoo Bay; and a non-linear transect flown in Tremblay Sound. Linear transects varied in length from 9–53 km, 30–40 km, and 12–25 km within Eclipse Sound, Milne Inlet and Pond Inlet, respectively. A non-linear transect was flown in Tremblay Sound, where the survey areas were too narrow to allow linear transects. “Zig−Zag” transects were flown in the southern part of Milne Inlet and Koluktoo Bay. “Zig-Zag” transects were flown at times that had minimal effect on the shore-based study of narwhals at Bruce Head (Smith et al. 2016). The “Tremblay Transect” and “Zig−Zag Transects” were 54 km and 116 km in total length, respectively.
Page 12 Marine Mammal Aerial Surveys 2015
The extensive survey grid was surveyed once during each survey period. The transect locations, orientation, and spacing in Eclipse Sound, Pond Inlet, Milne Inlet (northern) and Tremblay Sound follow those flown during previous aerial surveys for Baffinland (i.e., 2007, 2008, 2013, and 2014). In 2015, transect location and orientation in Koluktoo Bay and southern Milne Inlet were the same as the 2014 design (Thomas et al. 2015). Navy Board Inlet was not surveyed in 2015. 3.1.2 Photographic Survey Photographic surveys of northern Milne Inlet and Tremblay Sound (see Figure 3B) were conducted periodically over ~30 days during the peak period when narwhals occur in these key summering areas and when the majority of Baffinland vessel transits were expected during 2015. From 17 August to 17 September 2015, four photographic aerial surveys were flown. Photographic surveys were “centered” around Baffinland vessel (ore carriers) transits in Milne Inlet to the extent possible— i.e., aerial surveys were conducted Before, During and After a vessel passage. Each photographic survey consisted of four linear transects in Milne Inlet and/or one non-linear transect in Tremblay Sound, flown repeatedly to obtain multiple replicates during periods Before, During and After a large vessel passage. Linear transects for the photographic survey varied in length from 30–40 km within Milne Inlet and the Tremblay Sound transect extended 54 km in length. The photographic study area encompassed the four transects in Milne Inlet North with an area of 690.48 km2 (slightly larger than the Milne Inlet North stratum in Figure 1 for the extensive surveys) and the transect in Tremblay Sound, with an area of 155.9 km2. On 30 August, six transects (oriented north-south and 23-35 km in length) in Eclipse Sound West were photographically surveyed. This survey was conducted in an attempt to locate narwhals after photographic surveys of Milne Inlet North revealed low numbers of narwhals present in the area. The Eclipse Sound photographic transects were only flown once. The timing of the photographic surveys was dependant on information obtained on estimated times of large vessel passage and on useable daylight hours. Two sources of information were accessed to determine when a ship would be transiting through Milne Inlet: a web-based Automatic Information System (AIS) and information provided by Baffinland personnel on the expected departure and arrival times for chartered vessels transiting to/from Milne Port. To the extent possible, aerial survey departure times were chosen to optimize the “Before, During, and After” survey coverage of the vessel passage through Milne Inlet.
3.2 Data Recording Procedures 3.2.1 Extensive Survey Strip-transect survey methodology (Caughley 1977; Eberhardt 1978) was used for the extensive aerial surveys. Two primary observers occupied seats on opposite sides of the aircraft. Two secondary observers were located behind the primary observers. Observers were able to rest during transit and refuelling stops. Observers focused their effort on an ~ 1,000 m strip on each side of the aircraft while on transect. These strips extended from 135.7 m to 1,137.9 m from the aircraft. By focusing on a
Page 13 Marine Mammal Aerial Surveys 2015
defined strip, better coverage is obtained in the area near the aircraft. Although this reduces the number of overall sightings, it reduces the number of animals that are missed in the strip near the aircraft and increases the reliability of density estimates. Observers also scanned for and recorded sightings beyond the defined strip at a reduced level of effort from that described above when no animals were seen close to the aircraft. When possible (see Section 3.3.1.2 Density Estimates for further details), sightings were recorded with the associated clinometer angle to estimate the distance to the sighting from the aircraft. Surveys were conducted at an altitude of ~305 m (1,000 ft) above sea level and a ground speed of ~222 km/h (120 kts). Survey transects often started and ended at the shoreline. The survey aircraft was a de Havilland DHC-6 Series 300 Twin Otter (Figure 4) with extended-range fuel tanks (maximum range of 4.9 h at nominal speed of 220 km/h). The aircraft was equipped with four bubble windows; two were installed in the emergency exit windows of the second passenger row for use by the primary observers and two were installed in the rear cargo doors for use by the secondary observers. An aft ventral camera port was installed in the aircraft with a custom manufactured frame that was designed specifically for the belly camera port built into almost all twin otters that the aircraft provider operates (Figure 4). Two digital single lens reflex (DSLR) cameras (Nikon D810 with a 24–70 mm lens locked at 35 mm, with the focus locked at 1,000 ft, 36 mega-pixel images) were mounted on the frame. Both cameras had dual media card slots, one secure digital 512 GB (SD) and one compact flash 512 GB (CF) card slot. The cameras were oriented widthwise (long side perpendicular to the trackline) and angled obliquely (pointed 27 degrees out from the trackline): one to the right side and the other to the left side of the trackline (Figure 5). The cameras were connected to an AC power supply and interval shutter releases (Nikon) that allowed photos to be taken of the sea surface and marine mammals every 2 seconds. The 2-second interval provided an overlap of 65.0% and 40.5% along the distal and proximal edges of consecutive photos, respectively. The camera produced images with pixel sizes of 3.9 cm and 11.6 cm on the ground at the proximal and distal edges of the photographs, respectively. Images were saved on secure digital and compact flash cards installed in each camera’s dual card slots. Flight environmental parameters were recorded at the start and end of each transect by the pilots. A coding sheet was provided to them before the flights to record their observations. The variables recorded were cloud cover (in tenths), height of cloud ceiling (ft), flight altitude (ft), visibility (n.mi.), wind speed (knots), wind direction (ºT), crabbing (±º) and air temperature (°C). Wind data were estimated by the pilots based on the ground speed from the aircraft’s global positioning system (GPS) and the air speed from the aircraft instruments, and air temperature was acquired from a thermometer mounted externally on the aircraft. At the end of each 2-min time period (~7.4 km) along a transect (and whenever noticeable changes occurred), the two primary observers each dictated onto a digital voice recorder the time, ice cover (intervals of 100%), ice type, Beaufort wind force (in reference to wave description; hereafter referred to as sea state), sun glare (none, moderate, severe) and overall sightability conditions (subjectively classified as excellent, good, moderately impaired, severely impaired or impossible). A timer, initialized at the start of a flight, provided an audible signal at 2-min intervals.
Page 14 Marine Mammal Aerial Surveys 2015
FIGURE 4. Twin Otter with camera port used to conduct the aerial surveys.
FIGURE 5. Internal view of ventral camera port and camera installed in the Twin Otter.
For each cetacean sighting, the observer dictated onto a digital voice recorder the time, species, number of animals in the group, age class, behaviour, direction of movement (relative to aircraft heading) and angle of declination to the sighting location using a Suunto clinometer. Observers also recorded sightings of polar bears, seals, walruses and seabirds (when possible). Marine mammal sightings were recorded to species only when there was a positive identification otherwise sightings were recorded to the highest taxonomic level possible (e.g., seal species). Narwhals that were in a group were often touching or within half a body length of each other, oriented in the same direction and swimming at the same speed.
Page 15 Marine Mammal Aerial Surveys 2015
A notebook computer was used onboard the aircraft to store flight data. The computer system was connected to a primary Garmin GPS receiver that had an external antenna placed on the dashboard of the aircraft. The latitude, longitude and flight altitude from this receiver was transmitted to the computer through a USB cable approximately every second. All sighting locations were interpolated from date-time tagged GPS tracklog fixes based on the time recorded by the observer for each sighting. When large narwhal herds were encountered the observer would often leave the voice recorder on, indicating the time it was turned on and turned off; the sighting time would be estimated based on the tape counter minutes and seconds compared to the start and end of the continuous recording period. A second handheld GPS receiver placed on the aircraft dashboard was operated as a backup to the primary GPS receiver. 3.2.2 Photographic Survey To minimize potential responses by narwhals to overflights of the Twin Otter, photographic surveys were conducted at a higher altitude than the extensive surveys. To determine a suitable altitude to fly the photographic surveys that would minimize narwhal disturbance and still be able to discern calves and the presence of narwhals with tusks, test flights were flown in Tremblay Sound on the 17 August. Test flights were undertaken at various altitudes ranging from 1500 ft to 4500 ft in 500 ft increments. After reviewing the images of narwhals at the various altitudes, an altitude of 2500 ft (~762 m) above sea level was selected for the photographic surveys. The photographic survey was conducted at a ground speed of ~280 km/h (150 kts). One or two primary observer(s) occupied seats on opposite sides of the aircraft. Periodically one observer looked through the camera port in the belly of the aircraft. Observers focused their effort on a ~ 1,000 m strip on each side of the aircraft or directly downwards if positioned at the camera port while on transect. Although observers recorded environmental and sightings data similar to that collected during the extensive surveys, the data collected by observers was used primarily for determining the start and end of the transect for the photographic image data. At a later date, comparisons between the photographic data and the observer data could be used to develop correction factors for observer bias during aerial surveys. During photographic surveys, observers were also tasked with camera set-up and ensuring photographs were taken and images were properly stored. The same DSLR cameras and setup were used in the photographic surveys as were used in the extensive surveys. The cameras captured images at 3−second intervals. The sensor on the Nikon D810 is 24.0 mm by 35.9 mm and the focal length of the lens used was 35.0 mm. The viewing angle of the cameras (α) was 27°. The swath width for the two camera system was calculated using the methods described in Grendzdörffer et al. (2008), (eq. 1) (Figure 6): A = Altitude × tan[radians(α + β)] (1) B = {Altitude × cos[radians(β)]/FocalLength × cos[radians (α - β)]}× sensor width C = {Altitude × cos[radians(β)]/FocalLength × cos[radians (α + β)]}× sensor width Where: α = angle of camera β = half the field of view of the lens
Page 16 Marine Mammal Aerial Surveys 2015
FIGURE 6. Geometry of oblique aerial photos (modified from Asselin and Richard 2011). There was a 3 m overlap of images on either side of the centerline. Photographic images covered an area extending ∼1046 m from the centre line of the survey track (length of A in Figure 6 minus 3 m for overlap on centerline). The distal edge of the photographs covered a length of ∼792 m (length of C in Figure 6) and the proximal edge of the photographs covered a length of ∼466 m (length of B in Figure 6). Each photograph covered an area on the surface of the water of ∼0.657 km2. During the photographic surveys, 46.9% coverage of Milne Inlet North area and 54.8% coverage of Tremblay Sound was photographed during a typical survey. Digital images taken every 6 seconds (i.e., every second photograph) provided enough overlap between images to provide 100% coverage of the area under the aircraft. The 6-second interval provided an overlap of 59% and 30% along the distal and proximal edges of consecutive photos, respectively. The camera produced images with pixel sizes of 9.8 cm and 28.0 cm on the ground at the proximal and distal edges of the photographs, respectively. The overall number of narwhal sightings, their position relative to a large vessel, and their orientation were reviewed for a subset of each of the photographic surveys (see Section 4.5 later for details). Images were viewed on 2560×1600 pixel resolution monitors connected to computers; see Figure 7 for a sample photo. The screen resolution of the monitor is only about one-ninth of the resolution of the DSLR imagery, but this resolution is as good as or better than is obtained by observers looking out of bubble windows, particularly toward the edges of the outer viewing area. Photo analysts extracted the following data from the photographs: species, number in group, orientation (recorded as a compass point and converted to degrees True), x and y coordinates in photograph (measured in pixels using Adobe Photoshop CC version 14.1.1 and used to calculate distance from vessel), and sighting confidence (subjectively classified as 100% confident, high confidence, moderate confidence, or low confidence in the species identification). A ‘group’ was defined as narwhals within one body length of each other. If the same sighting was seen on two separate images (i.e., in the overlap zone or on the trackline on both the right and left image) it was recorded in the database only the first time it was observed. Narwhal sightings were mapped by calculating the geographic coordinates for each sighting. The calculation process involved several steps that integrate the camera image size, camera lens distortion parameters, camera frame versus ground geometry, and GPS data collected during the survey flights (see Appendix A for details).
Page 17 Marine Mammal Aerial Surveys 2015
FIGURE 7. Systematic photographic image taken during a photographic survey in Tremblay Sound on 30 August 2015. The photograph contains 27 sightings of narwhals, with one sighting zoomed in on a group of 8 narwhals all travelling in a southwesterly direction.
3.3 Analysis Procedures This section presents analysis procedures of the extensive and photographic survey data with the exception of the statistical model methods, which are presented in Section 3.4. 3.3.1 Extensive Surveys All aerial survey data were reviewed for accuracy and consistency after they were entered into an Excel database. Survey data were excluded from analyses when sightability was recorded as “Severely Impaired” or “Impossible” due to low fog or cloud, severe glare, and/or sea state (Beaufort scale) 5 or greater. This resulted in the exclusion of 4% of data. 3.3.1.1 Geographic Strata The study area was divided into 13 geographic strata or zones that correspond to identified water bodies (see Figure 1 earlier; Table 2). Eclipse Sound was divided into two strata (East and West), based upon differences in survey effort in 2013. Milne Inlet was divided into two strata—‘Milne Inlet North’ and ‘Milne Inlet South’ in an attempt to account for the very high numbers of narwhal sightings along the north and south shoreline of the Bruce Head peninsula and along the eastern shoreline of Milne Inlet from the Baffinland port site to the north end of Stephens Island. The intent of subdividing the Milne Inlet stratum was to produce more reliable
Page 18 Marine Mammal Aerial Surveys 2015
2 TABLE 2. Areas (km ) of the geographic strata in the Study Area surveyed during Baffinland aerial surveys in 2007, 2008, 2013, 2014 and 2015.
Geographic Strata Years Surveyed Total Area (km2)
Eclipse Sound East 2007, 2008, 2013, 2014, 2015 1956.7 Eclipse Sound West 2007, 2008, 2013, 2014, 2015 835.5 Koluktoo Bay 2007, 2008, 2013, 2014, 2015 75.8 Milne Inlet North 2007, 2008, 2013, 2014, 2015 657.7 Milne Inlet South 2007, 2008, 2013, 2014, 2015 180.7 Pond Inlet 2007, 2008, 2013, 2014, 2015 1432.2 Tremblay Sound 2007, 2008, 2013, 2014, 2015 155.9 Navy Board Inlet 2007, 2008, 2013, 2014 2103.6 Eskimo Inlet 2007, 2008, 2014 67.5 Oliver Sound 2007, 2008, 2014 232.3 Paquet Bay 2007, 2008, 2014 221.2 Tay Sound 2007, 2008, 2014 312.3 White Bay 2007, 2008, 2014 122.6
Total 8354.0
estimates of narwhal densities when survey effort was not able to systematically account for the clumped nature of narwhal sightings. In 2015, aerial survey effort included Eclipse Sound (east, west), Pond Inlet, Milne Inlet (north, south), Koluktoo Bay, and Tremblay Sound. Survey coverage in these geographic strata varied from year-to-year (Table 2). Of note, the 2007, 2008, 2013 and 2014 aerial surveys were conducted using the same aircraft type, survey altitude, and survey speed as was used in 2015. 3.3.1.2 Density Estimates Narwhal density estimates were calculated for each survey and geographic strata in 2015. Observed (uncorrected) density estimates were calculated based on strip-transect survey methodology because clinometer angles were recorded for only about half of narwhal sightings2 in 2015 due to the clumped distribution of the animals3. A strip width of 400 m was selected based on previous aerial surveys of narwhals conducted by DFO (Richard et al. 2010). This strip width was offset from the centreline of the survey aircraft by 135.7 m (i.e., equivalent to a clinometer depression angle of 66 degrees at 300 m altitude). Observed densities presented in this report provide indexes suitable to detect relative change and patterns in narwhal distribution and abundance.
2 In 2015, 52.8% of narwhal sightings did not have a clinometer reading. In 2007, 2008, 2013, and 2014, 92.9%, 95.0%, 89.7%, and 57.5% of narwhal sightings did not have a clinometer reading.
3 There were long periods with few or no sightings, and when animals were encountered, sightings came too quickly to allow clinometer angles to be taken for most of them.
Page 19 Marine Mammal Aerial Surveys 2015
3.3.1.3 Narwhal Group Size Shore-based studies of narwhals have reported maximum narwhal group sizes of 21–45 (Marcoux 2009; Thomas et al. 2014; Smith et al. 2016). Previous aerial surveys (2007, 2008, 2013, and 2014) reported sightings of narwhals with group sizes far greater than these maximum values. Some observers recorded aggregate group sizes for narwhals when large numbers of animals were encountered—i.e., smaller group sizes were lumped together resulting in group sizes as high as 200. These aggregated group sizes were typically recorded when groups were so numerous in a short time period that aggregates were the only way to estimate the total number of animals present. For those sightings with group size > 21, the records were split up into several smaller “dummy” records based upon the overall proportion of various group sizes recorded for narwhal groups ≤ 21 individuals. For the 2015 aerial surveys, all group sizes were < 21. 3.3.1.4 Shipping Activity Vessel Automatic Identification System (AIS) data were downloaded daily from a commercial data provider (exactEarth Ltd., Cambridge, ON) and summarized for the study area to provide an indication of shipping activity in 2013, 2014 and 2015. In addition, vessels that were contracted by Baffinland provided trackline data that was integrated into the AIS dataset. AIS data are broadcast by vessels equipped with appropriate transponders, which are mandatory for all vessels > 300 gross tonnage (GT) engaged in international voyages, cargo ships > 500 GT, and all passenger ships irrespective of size (IMO 2014). Many smaller vessels that frequent commercial shipping lanes also use AIS transponders to inform other vessels of their presence. Some vessels (e.g. sailboats, Navy, Coast Guard) turn their AIS transponders off periodically in order to evade detection by AIS receivers or to save power. Vessels are identified using some combination of: vessel name, call sign, Maritime Mobile Service Identity (MMSI) number, and International Maritime Organization (IMO) number. Maps were generated for each extensive survey period and photographic survey to provide an indication of the frequency and patterns of shipping. Within each geographic stratum, the number of vessels (unique MMSI) with AIS data on the day before and the day of an aerial survey were tabulated for the 2013, 2014 and 2015 datasets. Only vessels in transit (>0.5 knots) were included in the dataset. 3.3.2 Photographic Surveys All image data were reviewed for accuracy and consistency after they were entered into an Excel database. Observations were geo-referenced using the GPS track data. To estimate the flight altitude for each image, an 11 second rolling average of the GPS altitude output was used. Aircraft headings were also extracted from the GPS track data and matched to the photographic database. Approximately 28% or 1223 sightings identified by photo reviewers were not considered suitable for data analysis4.
4 Sighting coded with a low sighting confidence, meaning the reviewer had a low confidence in the species identification. Sightings coded with a low confidence were generally located at the distal edge of the photo (farthest from the trackline) where image quality dropped, the sighting was below the water surface and difficult to identify, or sighting conditions were poor (i.e. glare or waves with whitecaps on photograph).
Page 20 Marine Mammal Aerial Surveys 2015
3.3.2.1 Categorizing Data Relative to Ship Transits Photographic survey replicates and each photograph were categorized as occurring either Before, During or After a ship transit depending on ship position relative to the study area. For each photograph and corresponding sightings, shipping activity was coded based on the position of the vessel at the time of the sighting as either Before, During or After the vessel was present in the study area. All analyses except density calculations were analyzed at the photograph level. In situations where a vessel entered or exited the study area part way through a survey replicate, the replicate and corresponding sightings were coded as During. Density estimates were calculated at the replicate level. If multiple ships transited the study area on a given day, the ship closest to the study area at the time of the photographic survey was used to categorize the data. During the photographic surveys, there was never more than one ship transiting in the study area. 3.3.2.2 Calculating Narwhal Closest Points of Approach The closest point of approach (CPA) of each narwhal sighting to a ship trackline was calculated using a search routine in ArcGIS. More specifically, the distance (m) between each narwhal sighting and each AIS fix of a vessel transiting through the study area were calculated. The closest AIS fix was determined and the CPA (m) of the sightings to the trackline was calculated. In some cases, there was a land barrier (e.g., Stephens Island) between narwhal sightings and the ship trackline. ArcGIS was used to identify these narwhal sightings by overlaying land barriers with a line connecting the narwhal with its CPA. In addition, the distance from the center of each photograph was calculated relative to the ship trackline. 3.3.2.3 Narwhal Orientation Narwhal sightings with identified orientation (degrees True in 15° degree increments) were plotted relative to ship activity (Before, During and After) and ship travel direction (northbound vs. southbound), and data were summarized. Sightings with unknown orientation were excluded from analyses (431 of 3103 sightings were excluded). To test for uniformity in distribution of narwhal orientation relative to shipping activity, Rayleigh uniformity tests were conducted. Photo reviewers did not attempt to code whether narwhals were moving or stationary. A review of observer data collected during photographic surveys indicated that 10% (24 of 222 sightings) of narwhal sightings were considered resting/milling and 90% were considered travelling. As such, the orientation data were used as a proxy for the direction of narwhal movements or travel. 3.3.2.4 Density Estimates Density estimates of narwhals were calculated using distance sampling methods (Buckland et al. 2001) and the associated DISTANCE software (Version 6.2, release 1; Thomas et al. 2010). Separate density estimates for each survey replicate were calculated for the Milne Inlet North and Tremblay Sound strata. The analyses were conducted using the Conventional Distance Sampling (CDS) module to allow the use of correction factors (“multipliers” within DISTANCE; see below). Effort and sighting data recorded by photo reviewers were reformatted to allow for import into a DISTANCE database.
Page 21 Marine Mammal Aerial Surveys 2015
The DISTANCE sampling estimates were calculated with the following parameters specified within the DISTANCE software: • line transect survey; • single observer; • perpendicular distance to sighting; • clusters of object (with associated group size); • estimates at the stratum level except for the detection function which is estimated with the pooled stratum data; and • automatic model selection from either half-normal or hazard rate. Estimated Distance to Sightings.—The distance sampling theory and DISTANCE software use the perpendicular distance from the camera to each sighting in order to estimate abundance and density. As noted earlier, Adobe Photoshop (version 14.1.1) was used to calculate the distance of the sightings from the aerial trackline by converting the number of pixels to the distance in metres from the trackline. This provides a precise detection function. The perpendicular distances to sightings from the aerial trackline were then calculated for each sighting. Correction Factors.—To modify numbers of sightings and density estimates to account for sightings unavailable or missed, two “multipliers” were applied to the dataset: • corrections for individuals that were not seen because they were “hidden” (availability bias); and • corrections for individuals that were missed by the photo reviewer (perception bias). The availability bias is the percentage of time (expressed as a ratio) that a species spends at the surface. Availability bias estimates and the associated standard errors (SE) were obtained from the published literature. The perception bias is the ratio of the number of sightings made by the primary observer to the number of unique sightings made by both the primary and secondary observers on the trackline. To calculate perception bias for the photographic data, a sample of the images was set up in a similar fashion as the double independent observer data are for aerial surveys. One photo reviewer analysed a sample of photographs and recorded their sightings, and a second reviewer analysed the same sample of photographs and recorded their sightings. The first reviewer and the second reviewer were treated like the primary and secondary observers in double independent observer data for aerial surveys. Perception bias estimates, standard error, and degrees of freedom were calculated from double reviewer (first and second reviewers) data. Computational details are based on Magnusson (1978), and Davis et al. (1982) (see Appendix B). In addition to the availability and perception bias correction factors, the DISTANCE sampling analysis was used to generate a correction for detectability for the distance of marine mammal sightings from the survey trackline. A detection probability versus distance from the survey trackline plot was generated for narwhals.
Page 22 Marine Mammal Aerial Surveys 2015
3.4 Statistical Models 3.4.1 Extensive Surveys An understanding of narwhal spatial-seasonal distribution along and adjacent to the northern shipping route was required in order to assess the effects of shipping on narwhals. Moreover, annual variability in densities and distribution had to be understood before potential effects of shipping could be quantified in any particular localized area. The analysis presented here is the first time shipping has been included as a variable in the statistical model. Many statistical challenges were inherent with these data owing to the behaviour and life history of narwhals and the logistical limitations in the approaches available for measuring their abundance and distribution. The final analytical approach (i.e., generalized nonlinear mixed model, GNLMM) was derived based on a series of questions concerning (1) which statistical distribution to assume for the response variable, (2) how to account for correlation among samples adjacent in time and space, (3) how the random effects were defined, and (4) how fixed effects of the model were specified. Convergence for varying combinations of these decisions was time consuming, and prevented the use of multi-model inference such as the information theoretic approach suggested by Burnham and Anderson (2003). Many specifications failed to converge during the numerical search for parameter solutions or produced non-estimable confidence intervals because of unbalanced/missing data and the high percentage of zero values. Other specifications were over- parameterized resulting in unrealistic confidence intervals or simply had poor model fits. 3.4.1.1 Response Variable for Extensive Survey Data The data generated by the extensive aerial surveys afforded several responses (dependent variables) that could be used to index narwhal distribution. Observers recorded each narwhal sighting and these sightings were later pooled into a 2-min sampling interval (i.e., the number of sightings, hereafter referred to as the dependent response variable Sightings). For each narwhal sighting, group size was recorded which yielded the total number of individuals (dependent variable Number). These two dependent variables represented discrete data (i.e., values = 0, 1, 2, etc…) with the majority of observations being zeroes. This artifact created certain problems during statistical modelling, some of which could be addressed by reducing the data to the binary response Presence/Absence (dependent variable P/A). More specifically, if narwhals were seen during a 2-min interval then the value = 1; if not, the value = 0. With this binary response variable approach, the model produced the probability of observing a narwhal as opposed to total Sightings or Number. Each of these three response variables (Sightings, Number, and P/A) had potential advantages and disadvantages over the others with respect to their susceptibility to the assumptions required to create appropriate statistical models. However, analyses of all three response variables yielded similar conclusions based on data through 2013 (Elliott et al. 2015); therefore, only Number was modelled after adding the 2015 data to the existing database (i.e., 2013 and 2014 data). A generalized mixed modelling approach was used to model the Number of narwhals. This approach involved three basic steps:
Page 23 Marine Mammal Aerial Surveys 2015
1. Constructing a model with parameters of interest to predict the counts for all observations; 2. Multiplying the predicted counts from Step (1) by the sampling effort (called an effort offset) to obtain the predicted (expected) number of individuals comparable to the observed counts; 3. Computing the likelihood of the observed counts given the expected counts assuming some discrete distribution. The negative binomial distribution was selected, which accounted for overdispersion and used the log link function to portray the predicted response Number:
loge (λi ⋅ωi ) = µ + xi β + zib (1)
where, λi = predicted Number for the ith observation, ωi = Effort offset (km2 in this analysis), μ = overall mean, xi = the vector of fixed effects, β their corresponding vector of coefficients, and zi and b represent the random effects5 and coefficients. This 2-parameter discrete distribution is commonly used for this type of data to handle overdispersion and the high frequency of zero values typically seen in datasets for marine mammals (e.g., Boveng et al. 2003; Ver Hoef and Boveng 2007) and fisheries (e.g., Terceiro 2003; Minami et al. 2007; Arab et al. 2008; Shono 2008; Dunn 2009). 3.4.1.2 Independent Variables and Model Specification for Extensive Survey Data The aerial surveys produced consecutive 2-min samples along defined Transects. For each sample there were two observations (one each from the left and right observers). In order to reduce the number of zero values and promote independence among samples, the Number and km2 surveyed (i.e., Effort) from both left and right samples were pooled within each 2-min sample. Thus, the experimental unit (EU) for this analysis was defined by each Year–Julian– Transect–2-min sample (Julian refers to Julian date) combination. Continuous variables such as Water Depth and Distance from Shore were averaged within each 2-min sample, as well as across left and right subsamples. Because considerable variability in these metrics occurred within the defined EUs the usefulness of these covariates for explaining variability in the responses was limited. Moreover, the categorical variable Transect, which partially defined the random block effect accounted for much of the variability due to Water Depth and Distance from Shore. As such, these variables were not considered further in the modelling. To test for the effects of large vessels on the Number of narwhals, large vessels were categorized as the number of vessels (NumVess—0, 1, 2, >2) present in a given geographic stratum (GeoStrat) on the day of each survey. This information (derived from AIS data) was not available in 2007 and 2008. Therefore, the current analysis restricted the variable Year to 2013,
5 A random effect is a term in a generalized mixed model that accounts for variability in the response across levels of the effect. Variability across levels is assumed to be greater than variability within levels. Differences among levels are typically not the target of investigation, but the random effect is added to the model to remove noise around fixed effects that are of interest. Levels of a random effect represent a sample from a larger pool of possible levels; in this case, an infinite number of possible 2-min samples were available for sampling on any given day.
Page 24 Marine Mammal Aerial Surveys 2015
2014, and 2015. Ignoring subscripts and parameters, the response variable Number (output as Number/km2) was related to the independent variables as follows: Number = Year + GeoStrat + Julian + GeoStrat×Julian + NumVess Year×Julian×GeoStrat×Transect (2) where, GeoStrat was categorical with levels restricted to Eclipse Sound West, Eclipse Sound East, Koluktoo Bay/Milne Inlet (combined) and Tremblay Sound. The seasonal pattern in Number was never linear or consistent across GeoStrat. As such, the linear assumption was relaxed by specifying Julian as a continuous variable fit using a natural cubic spline and further allowed to vary across GeoStrat (i.e., the GeoStrat×Julian interaction term was included in the model). The first five terms were considered fixed effects (i.e., the xi part of Equation 1). Year×Julian×GeoStrat×Transect was considered a random block effect (zi) to partially account for correlation among observations within each Transect. This specification formed a generalized non-linear mixed model (GNLMM)6 whose maximum likelihood parameters were estimated using the Laplace approximation in the GLIMMIX Procedure of the statistical software SAS 9.4 (SAS Institute, Inc. 2012). All pairwise tests between levels within each categorical effect were Tukey adjusted for multiple comparisons. Note that the SAS code used for the model is available upon request. 3.4.1.3 Testing Differences among Factor Levels and Gauging Effect Size When all combinations of levels from multiple categorical variables receive equal sampling effort (i.e., a completely balanced dataset) and there are no continuous variables then tests among factor levels are straightforward. However, when sampling effort is unbalanced across level combinations, and worse when some combinations are missing, then comparisons of observed means are statistically invalid. The same is true for balanced datasets with continuous variables if not all level combination experience the exact same values for the continuous variables. For these reasons, the LSMEANS option was used during GLIMMIX Procedure in SAS. This option provides means that would have been observed had the dataset been balanced and the continuous variables been held constant at their means (or at whatever values were specified). Thus, a more “pure effect” of any one variable can be assessed by comparing the marginal means output by the LSMEANS option. However, for any non-Gaussian distribution (such as the negative binomial) combined with a random effect, the marginal distribution is highly skewed, and the estimated mean translates into the median of the marginal distribution. This explains in part why values may be far less than simple observed arithmetic means for any particular factor level; for right-skewed distributions, the median is always less than the mean. In other words, these "means" may not be typical of observed arithmetic means, but were
6 Switching to a GNLMM in lieu of the Generalized Linear Mixed Model (GLMM) used in the 2013 report relaxed the assumption that Number changed linearly with Julian date. Nonlinear patterns could have also been fit with polynomial regression, but the natural cubic spline approach used in the current analysis provided a more accurate model fit, while maintaining stable parameter estimates. The use of splines is a statistical approach that can fit complex patterns known (or at least strongly suspected) to be real and not due to random variation. They essentially consist of fitting polynomial functions to segments of a continuous variable separated by and constrained to meet at designated knots.
Page 25 Marine Mammal Aerial Surveys 2015
appropriate to test for statistical significance and gauge proportional differences among effect levels. Moreover, use of the median instead of the mean is probably more meaningful with respect to biological inference. Conveniently, statistical models using the log-link function test for proportional and not absolute differences among factor levels. Therefore, effect size is more appropriately assessed by comparing the ratio of marginal means between two factor levels as opposed to their difference. 3.4.1.4 Model Diagnostics Aside from comparing predicted versus observed values, residual plots are often used to assess model fit. Standard residual plots can be difficult to interpret during generalized linear modelling of discrete datasets containing lots of zeroes. Therefore, goodness-of-fit tests were conducted according to the recommendations of Lin et al (2002), whereby the observed moving sum of residuals across the range of the predicted responses were compared to randomized realizations. Lin et al. (2002) introduced the idea of cumulative residuals to assess model fit, but recommended moving sums for when the cumulative is dominated by small values on the x-axis over a wide range of x (x=either the predicted value or covariate being assessed). That is, residuals are made to be cumulative within a sliding window of size equal to the range for the lower half of x. A poor fit is indicated by a low P-value (estimated from the Kolmogorov-type supremum test) and visual deviation of the observed fit from the randomized realizations. The supremum test is essentially a Monte Carlo test whereby random residuals around the predicted value are generated for a number of iterations (1000 iterations were used here) by parametrically bootstrapping from an assumed Gaussian distribution, whose parameters are estimated from the observed residuals. The percentage of these iterations whose maximum absolute residual is ≥ the observed maximum represents the P-value. Low P-values show areas along the x continuum that deviated more than expected from chance alone and indicate a misspecification of some sort. Residual plots against the predicted value test the adequacy of the link function and overall model specification. 3.4.2 Photographic Surveys Statistical modelling methods used for the photographic survey data were similar to those used for the extensive surveys. Exceptions are noted below. 3.4.2.1 Response Variable for Photographic Survey Data For the photographic data, the response was defined as the Number of narwhals observed within each photograph, which represented the EU. Each photograph was considered equal with respect to the area covered, so no Effort offset was required as with the 2-min EUs for the GNLMM of the extensive aerial survey data. Photographic surveys were purposive in nature, which limited the number of surveys to four large vessel events. During one of these four large vessel events (Survey 4 on 22 August 2015), no narwhals were detected in Milne Inlet; see Table 18 later). As such, Survey 4 was excluded from the analysis. Furthermore, photographic data used in the model were restricted to Milne Inlet, and all narwhal observations with low sighting confidence were excluded.
Page 26 Marine Mammal Aerial Surveys 2015
3.4.2.2 Independent Variables and Model Specification for Photographic Survey Data Large vessel activity was categorized by designating replicate surveys as occurring Before, During or After a vessel transit—this variable was termed BDA. Whether a land barrier occurred between the photograph and vessel in question served as another categorical variable (LandBarr). Finally, the distance between each photograph’s center and the CPA (PhotoCPA) to the large vessel trackline was modelled as a continuous variable. The response variable Number (output as Number/photograph) was related to the independent variables as follows:
Number = LandBarr + BDA + PhotoCPA + BDA×PhotoCPA + Julian×Replicate (3)
where, Julian×Replicate defined the random block effect. All other statistical methods were consistent with those used for the extensive aerial survey data.
4 RESULTS
4.1 Survey Effort in 2015 4.1.1 Extensive Surveys The extensive survey grid was flown four times during 1 August to 17 September 2015 (Table 3). During each of the four bi-weekly surveys, all planned transects were completed within one to two days and survey effort totalled 3273 km. Approximately 4% of this survey effort was not considered suitable for data analysis7. 4.1.2 Photographic Surveys From 18 August to 4 September 2015, photographic surveys were conducted on four days (18, 22, 30 August and 4 September). A total of 4545 km of transects (excluding 480 km of transects flown on the 17 August for the photographic test flight) were surveyed with most survey effort occurring in Milne Inlet North (Table 3). In Milne Inlet North, 4, 15 and 4 replicates were flown Before, During and After a vessel transit, respectively (Table 4). Of these 23 replicates, 14 (1 Before, 10 During and 3 After a vessel transit) were selected for photo review resulting in the analysis of 11,103 photos (Table 4). Photo review efforts were focused on the two surveys (survey 5 and 7) that had the majority of the narwhal sightings. In Tremblay Sound, two, two and three replicates were flown Before, During and After a vessel transit (Table 4). Of these seven replicates, five (1 Before, 2 During and 2 After a vessel transit) were selected for photo review resulting in the analysis of 999 photographs (Table 4).
7 In some areas, the sighting conditions were rated as “severely impaired” or “impossible” due to heavy fog or cloud.
Page 27 Marine Mammal Aerial Surveys 2015
TABLE 3. Summary of extensive and photographic aerial survey effort in 2015.
Transects Completed Milne Inlet Total Date in Flight Eclipse Sound, Milne Inlet Tremblay South, Transect 2015 No. Survey Survey Type Pond Inlet North Sound Koluktoo Bay Distance (km) 1-Aug 1 1 Extensive 15, 19-26 16-18 Y Y 632 1-Aug 2 1 Extensive 27-35 178 16-Aug 3 2 Extensive 15, 19-26 16-18 Y Y 632 16-Aug 4 2 Extensive 27-31 87 17-Aug 5 2 Extensive 31-35 102 17-Aug 6 - Test Photographic Y 480 18-Aug 7 3 Photographic 3 Replicates 1 Replicate 580 18-Aug 8 3 Photographic 3 Replicates 532 22-Aug 9 4 Photographic 4 Replicates 709 22-Aug 10 4 Photographic 2 Replicates 2 Replicates 451 30-Aug 11 5 Photographic 1 Replicate 1 Replicate 2 Replicates 452 30-Aug 12 5 Photographic 2 Replicates 2 Replicates 403 31-Aug 13 6 Extensive 19-35 17,18 562 31-Aug 14 6 Extensive 15 16 Y Y 248 4-Sep 15 7 Photographic 4 Replicates 709 4-Sep 16 7 Photographic 4 Replicates 709 13-Sep 17 - Aborted 0 15-Sep 18 8 Extensive 23-35 324 17-Sep 19 8 Extensive 15,19-23 16-18 Y Y 508 Total 8298
TABLE 4. Number of survey replicates flown and photos reviewed Before, During and After the transit of a vessel through the Milne Inlet and Tremblay Sound photographic study area in 2015.
Date in Before Vessel During Vessel After Vessel No. Images 2015 Survey Survey Type Area Flown Reviewed Flown Reviewed Flown Reviewed Reviewed Milne Inlet North 3 1 3 1 0 - 1521 18-Aug 3 Photographic Tremblay Sound 1 0 0 - 0 - 0
Milne Inlet North 1 0 3 0 2 1 784 22-Aug 4 Photographic Tremblay Sound 0 - 1 1 1 1 407
Milne Inlet North 0 - 1 1 2 2 2412 30-Aug 5 Photographic Tremblay Sound 1 1 1 1 2 1 592 Eclipse Sound 1 0 0 - 0 - 0
Milne Inlet North 0 - 8 8 0 - 6386 4-Sep 7 Photographic Tremblay Sound 0 - 0 - 0 - - Note: Start and end flight times for each replicate with reviewed data are labelled on Figures 30-35 in Section 4.5.2.4.
Page 28 Marine Mammal Aerial Surveys 2015
On 30 August, six transects (oriented north-south and 23-35 km in length) in Eclipse Sound West were photographically surveyed. As discussed previously, this survey was conducted in an attempt to locate narwhals after photographic surveys of Milne Inlet North indicated low numbers of narwhals present in the area. The six transects were flown before the transit of a vessel and were not included for photo review because the area was flown only once and no narwhals were seen by observers.
4.2 Survey Conditions in 2015 4.2.1 Extensive Surveys The study area had ice cover ranging from 0 to 100% during the early August survey (Survey 1, 1 August). Most of the ice during this survey was concentrated in Eclipse Sound West and Milne Inlet (transects 15-20 and zig-zag transects). By the mid-August survey (Survey 2, 16 and 17 August), the study area was ice free except for the occasional iceberg. The sea state (Beaufort scale) conditions during extensive surveys varied between values of 0 and 5. Sea state values varied between 0 and 3 for 85% of the survey effort. 4.2.2 Photographic Surveys The study area was ice free except for occasional icebergs during the four photographic surveys. During the four photographic surveys, sea states (Beaufort scale) ranged between 0 and 4 with most effort occurring during periods with values between 1 and 3. Sightability was typically recorded as moderately impaired, good or excellent. During periods with higher sea states (levels = Beaufort scale >3), more time was required for photo analysis to identify narwhals. Sighting confidence was also affected by sea state because white caps would obscure parts of the animals. Sea state did not appear to affect the quality of the photos as much as lighting conditions. Two surveys were flown during low light conditions and although the observers were able to see the narwhals during the survey, the light levels were not sufficient for the photographs and the images were underexposed. Photographs taken after 2100 h on the 18 August and before 1000 h on the 22 August were underexposed and not considered for photo analysis.
4.3 Shipping Activity in 2015 Based on AIS data, shipping activity involving a variety of vessel types occurred within the study area during the 2015 open-water period (Table 5). Along Baffinland’s northern shipping route (2015 BIM Shipping Route in Table 5), shipping activity was most prevalent in August and September and involved vessels associated with tourism, the Canadian government (coast guard, navy), community resupply and industry (vessels chartered by Baffinland; see Table 6). Also included in Table 5 are the vessels that occurred along Other Routes including Navy Board Inlet. Shipping along these other routes occurred primarily during August and September. Table 5 does not include the small boats associated with hunting and other small vessels that do not provide AIS data.
Page 29 Marine Mammal Aerial Surveys 2015
TABLE 5. Summary of shipping activity in the study area for vessels with AIS data during the 2015 open- water period (data from exactEarth Ltd.).
Type of Vessel Date Range When Present Ore Coast Cruise Fuel All Start Date End Date Month Carrier Cargo Guard Ship Fishing Tanker Navy Sailboat Tug Yacht Vessel BIM Shipping Route August 10 3 9 1 2 1 5 2 33 1-Aug-15 31-Aug-15 September 8 4 1 3 2 1 2 21 1-Sep-15 30-Sep-15 October 2 1 1 1 2 7 1-Oct-15 28-Oct-15 November 1 1 8-Nov-15 9-Nov-15 Other Routes August 2 2 9 2 1 2 8 26 1-Aug-15 31-Aug-15 September 5 2 3 2 2 2 3 1 20 1-Sep-15 30-Sep-15 October 1 2 3 1-Oct-15 14-Oct-15 November 1 1 8-Nov-15 9-Nov-15 Note: “BIM Shipping Route” includes vessels that occurred in Pond Inlet, Eclipse Sound, Milne Inlet, Tremblay Sound and/or Koluktoo Bay. “Other Routes” includes vessels that occurred in Navy Board Inlet and in an area around Bylot Island bounded by 74.5°N, extending to 85°W and 73°W.
In 2015, 13 ore carriers were involved in transporting ore from Milne Port to market (Table 6). At least one ore carrier was present within the study area from 1 August to 12 October and quite often 2-3 ore carriers were present either at Milne Port and/or at Ragged Island—one of Baffinland’s identified anchorage sites. There were three shipments of cargo and one fuel tanker trip in 2015. Relative to 2015, Baffinland shipping activity in the study area was much lower in 2013 and 2014 (Table 6). 4.3.1 Extensive Surveys 4.3.1.1 Survey Period 1: 1 August 2015 Based on AIS data, only Baffinland chartered vessels were present within the study area on 1 August. The M/V Federal Tiber, the first Baffinland vessel to transport ore during the ERP, (identified as BC1 in Figure 8) transited across Eclipse Sound enroute to the Ragged Island anchorage site. In addition, two tug boats (M/V Svitzer Nerthus and M/V Svitzer Njal; TU1 and TU2 in Figure 8) chartered by Baffinland were moving south through Milne Inlet enroute to Milne Port during the aerial survey. These two tug boats would remain within the Milne Port area until after Survey Period 4.
Page 30 Marine Mammal Aerial Surveys 2015
TABLE 6. Summary of Baffinland shipping activity in the study area during the 2015 open-water period (based on AIS data from exactEarth Ltd.). Also provided are corresponding shipping data for 2013 and 2014.
Day/Month in BIM Ve sse l Type Study Area Ore Carrier Cargo Fuel Tanker 2015 1−8 Aug M/V Federal Tiber 3−13 Aug M/V Nordic Odin 6−19 Aug M/V Golden Ice 7−22 Aug M/V Nordic Odyssey 10−30 Aug M/V Nordic Orion 22−30 Aug M/V Avataq 23−31 Aug M/V Sedna Desgagnes 13 Aug−4 Sep M/V Golden Brilliant 18 Aug−9 Sep M/V Golden Saguenay 19 Aug−11 Sep M/V Nordic Olympic 27 Aug−15 Sep M/V Nordic Oshima 31 Aug−17 Sep M/V Golden Opportunity 3−20 Sep M/V Golden Ruby 1 Sep−1 Oct M/V Sarah Desgagnes 27 Sep−4 Oct M/V Sedna Desgagnes 3 Sep−4 Oct M/V Nordic Odin 6 Sep−12 Oct M/V Federal Tiber 2014 22−31 Jul M/V Rosaire A. Desgagnés 11−15 Aug M/V Maria Desgagnés 11−18 Aug M/V Claude A. Desgagnés 31 Aug−4 Sept M/V Happy Delta 7−11 Sept M/V Claude A. Desgagnés 7−13 Sept M/V Maria Desgagnés 9−16 Sept M/V Happy Dover 28 Sept−3 Oct M/V Claude A. Desgagnés 2013 6−14 Aug M/V Jana Desgagnés 7−11 Aug M/V Claude A. Desgagnés 12−19 Aug M/V Qamutik 21−25 Aug M/V Avataq 30 Aug−4 Sept M/V Claude A. Desgagnés 7−15 Sept M/V Qamutik 15−24 Sept M/V Avataq 26 Sept−1 Oct M/V Claude A. Desgagnés − 30 Sept 7 Oct M/V Jana Desgagnés Note: M/V Nordic and Golden ore carriers are 225 m in length and the Federal Tiber is 190 m in length.
Page 31 Marine Mammal Aerial Surveys 2015
FIGURE 8. Summary of shipping activity in the study area based on AIS data for 1 August 2015. Ship tracks during the aerial survey are highlighted in bold red. [BC=bulk carrier; TU=tugboat] 4.3.1.2 Survey Period 2: 16−17 August 2015 Shipping activity was much higher during Survey Period 2 (16−17 August 2015) compared to the previous survey period. During the aerial survey, two ore carriers (M/V Nordic Odyssey, M/V Golden Ice) were anchored at Milne Port and two ore carriers (M/V Nordic Orion, M/V Golden Brilliant) were anchored at Ragged Island (Table 6 and see BC1−4 in Figure 9). One of the tug boats (M/V Svitzer Nerthus; TU1 in Figure 9) was working close to Milne Port while the other tug boat (M/V Svitzer Njal; TU2 in Figure 9) was docked during the aerial survey. A project ore carrier (M/V Golden Saguenay; BC5 in Figure 9) was located in Pond Inlet inbound to Milne Inlet. In addition, several non-project vessels were transiting in Pond Inlet and Eclipse Sound. More specifically: a cruise ship (M/V Hanseatic) departed from Pond Inlet heading to Navy Board Inlet; the cruise ship (M/V Akademik S. Vavilov) departed from Pond Inlet for the southern shore of Bylot Island; and a cruise ship (M/V Akademik Ioffe) was inbound to Pond Inlet from the east, arriving at an anchorage off the community. Finally, the fuel tanker M/V Travestern was anchored at Pond Inlet.
Page 32 Marine Mammal Aerial Surveys 2015
FIGURE 9. Summary of shipping activity in the study area based on AIS data for 16−17 August 2015. Ship tracks during the aerial survey are highlighted in bold red. [BC=bulk carrier; CS=cruise ship; FU=fuel tanker; SB=sailboat; TU=tugboat]
4.3.1.3 Survey Period 3: 31 August 2015 Survey Period 3 (31 August 2015) had the highest level of shipping activity relative to the other three survey periods. An ore carrier (M/V Nordic Olympic) transited south from the Ragged Island anchorage to Milne Port. The ore carrier (M/V Nordic Oshima) remained anchored at Ragged Island (Figure 10). Two other Baffinland vessels transited through Pond Inlet and Eclipse Sound during the aerial survey, a cargo vessel (M/V Sedna Desgagnes) heading eastward, and an ore carrier (M/V Golden Opportunity) was inbound heading westward to the Ragged Island anchorage. During the aerial survey, two ore carriers (M/V Golden Brilliant, M/V Golden Saguenay) were anchored at Milne Port, the two tug boats (M/V Svitzer Nerthus, M/V Svitzer Njal) were active in Assomption Harbour. In addition, during the aerial survey a cruise ship (M/V Hanseatic) entered Pond Inlet and travelled westward through Eclipse Sound. This vessel heading north through Navy Board Inlet after the aerial survey was completed.
Page 33 Marine Mammal Aerial Surveys 2015
FIGURE 10. Summary of shipping activity in the study area based on AIS data for 31 August 2015. Ship tracks during the aerial survey are highlighted in bold red. [BC=bulk carrier; CA=cargo ship; CS=cruise ship; TU=tugboat] 4.3.1.4 Survey Period 4: 15 and 17 September 2015 On 15 September, a single ore carrier (M/V Federal Tiber) transited to Milne Port from the Ragged Island anchorage, and was anchored at the port site during the second day of the survey (Figure 11). Three ore carriers (M/V Golden Opportunity, M/V Golden Ruby and M/V Nordic Odin) were anchored at Milne Port along with the tug boats (M/V Svitzer Nerthus and M/V Svitzer Njal). During the aerial survey, three Canadian government vessels were active in Eclipse Sound and Pond Inlet: two Royal Canadian Navy (RCN) ships (HMCS Shawinigan and HMCS Moncton) and a Canadian Coast Guard vessel (CCGS Pierre Radisson). 4.3.2 Photographic Surveys Four photographic surveys were conducted on four days when project vessels were moving through Milne Inlet, either outbound from Milne Port, moving from the Ragged Island anchorage site to Milne Port or moving from Eclipse Sound to the Ragged Island anchorage site. The photo surveys were typically conducted over Milne Inlet North and Tremblay Sound. Survey coverage included Eclipse Sound West on 30 August.
Page 34 Marine Mammal Aerial Surveys 2015
FIGURE 11. Summary of shipping activity in the study area based on AIS data for 15 and 17 September 2015. Ship tracks during the aerial survey are highlighted in bold red. [BC=bulk carrier; CG=Coast Guard vessel; NA=navy vessel; TU=tugboat]
4.3.2.1 Photo Survey 1: 18 August 2015 On 18 August, an ore carrier (M/V Golden Ice; BC5 in Figure 12) departed from Milne Port and transited northward through Milne Inlet during the photographic survey. The ore carrier (M/V Nordic Odyssey; BC6 in Figure 12) was anchored at Milne Port and the two tug boats (M/V Svitzer Nerthus, M/V Svitzer Njal; TU1 and TU2 in Figure 12) moved within Assomption Harbour. Two ore carriers (M/V Golden Brilliant, M/V Nordic Orion; BC3 and BC4 in Figure 12) were located at the Ragged Island anchorage site during the survey. A sailboat (SB3 in Figure 12) was active just south of Ragged Island during the photo survey. There was also vessel activity in Eclipse Sound, Pond Inlet, and Navy Board Inlet during Photo Survey 1 (M/V Akademik Ioffe, M/V Golden Saguenay; sailboat Aventura, and M/V Nordic Olympic).
Page 35 Marine Mammal Aerial Surveys 2015
FIGURE 12. Summary of shipping activity in the study area based on AIS data for 18 August 2015. Ship tracks during the aerial survey are highlighted in bold red. [BC=bulk carrier; CS=cruise ship; SB=sailboat; TU=tugboat]
4.3.2.2 Photo Survey 2: 22 August 2015 During Photo Survey 2, an ore carrier (M/V Nordic Odyssey) transited northward through Milne Inlet into western Eclipse Sound (Figure 13). Two ore carriers (M/V Golden Brilliant, M/V Golden Saguenay) were anchored at Ragged Island and one ore carrier (M/V Nordic Orion) and two tug boats (M/V Svitzer Nerthus, M/V Svitzer Njal) were anchored off Milne Port. On 22 August, an ore carrier (M/V Nordic Olympic) was circling in central Eclipse Sound waiting for an anchorage location to become available at Ragged Island. A cargo ship (M/V Avataq) was anchored offshore of Pond Inlet during the survey.
Page 36 Marine Mammal Aerial Surveys 2015
FIGURE 13. Summary of shipping activity in the study area based on AIS data for 22 August 2015. Ship tracks during the aerial survey are highlighted in bold red. [BC=bulk carrier; CA=cargo ship; TU=tugboat]
4.3.2.3 Photo Survey 3: 30 August 2015 There was substantial shipping activity in Milne Inlet, Eclipse Sound and Pond Inlet on 30 August (Figure 14). During Photo Survey 3, two project tug boats, a cargo ship (M/V Sedna Desgagnes), and an ore carrier (M/V Golden Brilliant) occurred in the Milne Port area. During the survey, an ore carrier (M/V Nordic Oshima) entered Milne Inlet from the north and anchored at Ragged Island. Prior to the survey, an ore carrier (M/V Golden Saguenay) had left the anchorage at Ragged Island and transited south to Milne Port, and by the time the survey had begun the ship was south of Bruce Head. During the survey, a project bulk carrier (M/V Nordic Olympic) remained at the Ragged Island anchorage site. Other vessels (three cruise ships: M/V Akademik Ioffe, M/V Le Soleal, M/V Sea Explorer I; ore carrier: M/V Nordic Orion) were active outside of Milne Inlet (Figure 14).
Page 37 Marine Mammal Aerial Surveys 2015
FIGURE 14. Summary of shipping activity in the study area based on AIS data for 30 August 2015. Ship tracks during the aerial survey are highlighted in bold red. [BC=bulk carrier; CA=cargo ship; CS=cruise ship; TU=tugboat]
4.3.2.4 Photo Survey 4: 4 September 2015 During Photo Survey 4 (4 September 2015), the ore carrier (M/V Golden Brilliant) transited north through Milne Inlet into Eclipse Sound, and shortly afterwards a second ore carrier (M/V Nordic Oshima) departed the anchorage at Ragged Island and headed south toward Milne Port (Figure 15). After the M/V Nordic Oshima left Ragged Island, a third ore carrier (M/V Golden Ruby) left Eclipse Sound and entered Milne Inlet enroute to the Ragged Island anchorage site. During the survey, two ore carriers (M/V Nordic Odin, M/V Golden Opportunity) remained at the Ragged Island anchorage. Both project tug boats were located in the Milne Port area. At Milne Port, two ships remained at anchor during the survey: a fuel tanker (M/V Sarah Desgagnes) and an ore carrier (M/V Nordic Olympic). In addition to the project ships, one cargo ship (M/V Anna Desgagnes) departed Pond Inlet and headed west across Eclipse Sound into southern Navy Board Inlet (Figure 15).
Page 38 Marine Mammal Aerial Surveys 2015
FIGURE 15. Summary of shipping activity in the study area based on AIS data for 4 September 2015. Ship tracks during the aerial survey are highlighted in bold red. [BC=bulk carrier; CA=cargo ship; FU=fuel tanker; TU=tugboat]
4.4 Extensive Survey Sightings, Density Estimates, and Model Results 4.4.1 Marine Mammal Sightings in 2015—Overview Five species of marine mammals were recorded during extensive aerial surveys: narwhal, harp seal, ringed seal, bearded seal and walrus (Table 7). Narwhals were the most abundant marine mammal recorded and the results for this species are presented in detail in Section 4.4.2. Marine mammals other than narwhals are discussed briefly below. Bowhead whales and polar bears were not observed during the extensive aerial surveys in 2015.
Page 39 Marine Mammal Aerial Surveys 2015
TABLE 7. Numbers of narwhals and pinnipeds recorded during extensive aerial surveys conducted from 1 August to 17 September 2015. Includes all sightings including those recorded during connect legs. Survey Survey Survey Survey Period 1 Period 2 Period 3 Period 4 Total No. Survey 1 Survey 2 Survey 6 Survey 8 Species (1 Aug) (16-17 Aug) (31 Aug) (15-17 Sep) Narwhal No. of Sightings 106 159 396 191 852 No. of Individuals 375 336 894 319 1924 Harp Seal No. of Sightings 2 1 3 7 13 No. of Individuals 15 1 43 89 148 Ringed Seal No. of Sightings 0 0 4 0 4 No. of Individuals 0 0 4 0 4 Bearded Seal No. of Sightings 0 0 1 2 3 No. of Individuals 0 0 1 2 3 Walrus No. of Sightings 0 0 1 0 1 No. of Individuals 0 0 1 0 1 Unidentified Seal No. of Sightings 1 2 11 47 61 No. of Individuals 1 2 12 51 66
4.4.1.1 Pinnipeds The aerial surveys were not designed to survey for seals. However, observers recorded seal sightings whenever possible. Observations of seals were more frequently recorded during periods of low sea state and when group sizes were larger. Of the four pinniped species observed, harp seals were the most frequently recorded followed by ringed seal, bearded seal and walrus (Table 7). Harp seals were observed throughout the survey period and were seen in herds ranging from 1 to 30 with an average group size of 11 (Table 7). Most harp seals (77% or 10 of 13 sightings) were observed in Eclipse Sound East and Pond Inlet. The remaining three sightings occurred in Eclipse Sound West and Milne Inlet North, near Ragged Island (Figure 16).
Page 40 Marine Mammal Aerial Surveys 2015
FIGURE 16. Harp seal, ringed seal, bearded seal and walrus sightings recorded during the extensive aerial surveys (1 August-17 September 2015).
There were four sightings of ringed seals recorded during Survey Period 3; all were seen as individual animals and in Eclipse Sound East (Table 7). Three bearded seals were observed in Pond Inlet and Milne Inlet North (Table 7; Figure 16). A single walrus was observed on 31 August in Eclipse Sound West (Figure 16). Unidentified seals were observed throughout the study area and were seen as either individuals or in groups of two animals (Figure 17).
Page 41 Marine Mammal Aerial Surveys 2015
FIGURE 17. Unidentified seal sightings recorded during the extensive aerial surveys (1 August-17 September 2015).
4.4.2 Narwhals 4.4.2.1 Numbers Observed and Group Size Overall, there were 852 sightings of narwhals totalling 1924 individuals observed during extensive surveys in 2015 (Table 7). In 2015, the number of narwhals observed (both sightings and individuals) was highest during late August (Survey Period 3; Table 7). Narwhals were typically observed as individuals or groups of two or three (84.0% of sightings; Figure 18). Group size ranged from 1 to 20 and the average group size was 2.26 individuals.
Page 42 Marine Mammal Aerial Surveys 2015
450 408 400 350 300 250 210 200 150 98 100 Total Narwhal Sightings Narwhal Total 54 50 32 25 3 8 4 3 5 2 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Group Size
FIGURE 18. Total narwhal sightings with various group sizes.
4.4.2.2 General Observations During the first three survey periods (1 August, 16−17 August and 31 August) narwhal sightings occurred almost exclusively in Milne Inlet, Tremblay Sound, and Koluktoo Bay (Figures 19−21). By the last survey period (15 and 17 September), narwhal sightings were predominantly recorded in Eclipse Sound West with the exception of one sighting in each of Tremblay Sound and Milne Inlet South (Figure 22).
4.4.2.3 Density Estimates Narwhal density estimates were calculated for each survey period and geographic stratum in 2015. Results are presented below organized by survey period. For the purposes of this section of the report, observed (uncorrected) density estimates are provided. Survey Period 1: 1 August.— Narwhals were present in the southwestern portion of the study area during Survey Period 1 (Figure 19). Observed narwhal densities were higher in Koluktoo Bay and Milne Inlet South (1.90 and 2.21 individuals/km2, respectively) versus Tremblay Sound and Milne Inlet North (0.91 and 0.60 individuals/km2, respectively; Table 8). Densities in Eclipse Sound West and Pond Inlet were low—0.04 individuals/km2 and 0.02 individuals/km2, respectively (Table 8).
Page 43 Marine Mammal Aerial Surveys 2015
FIGURE 19. Narwhal sightings during Survey Period 1 (1 August 2015).
TABLE 8. Observed density estimates (uncorrected) for narwhals in the study area during Survey Period 1 (1 August 2015).
Survey Period 1 (1 Aug) 2 No. of No. of Density (per km ) Geographic Strata Sightings Individ. Sightings Individ. Koluktoo Bay 27 64 0.80 1.90 Tremblay Sound 10 32 0.28 0.91 Milne Inlet South 42 177 0.52 2.21 Milne Inlet North 15 56 0.16 0.60 Eclipse Sound West 2 3 0.03 0.04 Eclipse Sound East 0 0 0.00 0.00
Pond Inlet 1 2 0.01 0.02
Page 44 Marine Mammal Aerial Surveys 2015
Survey Period 2: 16−17 August.—During this survey, narwhals were observed exclusively in Tremblay Sound, Milne Inlet and Koluktoo Bay (Figure 20). The observed densities of narwhals were highest in Tremblay Sound (2.61 individuals/km2) followed by Milne Inlet North (1.23 individual/km2), Milne Inlet South (0.87 individual/km2), and Koluktoo Bay (0.20 individual/km2; Table 9).
FIGURE 20. Narwhal sightings during Survey Period 2 (16−17 August 2015).
Page 45 Marine Mammal Aerial Surveys 2015
TABLE 9. Observed density estimates (uncorrected) for narwhals in the study area during Survey Period 2 (16−17 August 2015).
Survey Period 2 (16-17 Aug) 2 No. of No. of Density (per km ) Geographic Strata Sightings Individ. Sightings Individ. Koluktoo Bay 1 4 0.05 0.20 Tremblay Sound 42 86 1.28 2.61 Milne Inlet South 23 57 0.35 0.87 Milne Inlet North 64 121 0.65 1.23 Eclipse Sound West 0 0 0.00 0.00 Eclipse Sound East 0 0 0.00 0.00
Pond Inlet 0 0 0.00 0.00
Survey Period 3: 31 August.—Once again narwhals were only observed in the southwestern portion of the study area and not in Eclipse Sound and Pond Inlet (Figure 21). Observed narwhal density (18.26 individuals/km2) was quite high in Tremblay Sound (Table 10). Densities were also relatively high in Milne Inlet South (4.93 individual/km2) with lower densities observed in Koluktoo Bay (0.89 individual/km2) and Milne Inlet North (0.14 individual/km2; Table 10).
Page 46 Marine Mammal Aerial Surveys 2015
FIGURE 21. Narwhal sightings during Survey Period 3 (31 August 2015).
TABLE 10. Observed density estimates (uncorrected) for narwhals in the study area during Survey Period 3 (31 August 2015).
Survey Period 3 (31 Aug) 2 No. of No. of Density (per km ) Geographic Strata Sightings Individ. Sightings Individ. Koluktoo Bay 9 18 0.45 0.89 Tremblay Sound 209 558 6.84 18.26 Milne Inlet South 148 278 2.62 4.93 Milne Inlet North 13 13 0.14 0.14 Eclipse Sound West 0 0 0.00 0.00 Eclipse Sound East 0 0 0.00 0.00 Pond Inlet 0 0 0.00 0.00
Page 47 Marine Mammal Aerial Surveys 2015
Survey Period 4: 15−17 September.—In mid-September 2015, most narwhal sightings (3.88 individuals/km2) were observed in Eclipse Sound West (Figure 22). No narwhals were observed in Milne Inlet North or Koluktoo Bay, and only one sighting was observed in each of Tremblay Sound and Mine Inlet South.
FIGURE 22. Narwhal sightings during Survey Period 4 (15 and 17 September 2015).
Page 48 Marine Mammal Aerial Surveys 2015
TABLE 11. Observed density estimates (uncorrected) for narwhals in the study area during Survey Period 4 (15&17 September 2015).
Survey Period 4 (15&17 Sept) 2 No. of No. of Density (per km ) Geographic Strata Sightings Individ. Sightings Individ. Koluktoo Bay 0 0 0.00 0.00 Tremblay Sound 1 1 0.03 0.03 Milne Inlet South 1 2 0.02 0.04 Milne Inlet North 0 0 0.00 0.00 Eclipse Sound West 188 314 2.32 3.88 Eclipse Sound East 0 0 0.00 0.00
Pond Inlet 0 0 0.00 0.00
Comparison of Observed Densities in 2013, 2014, and 2015.—Figure 23 summarizes observed narwhal densities in each geographic stratum by survey period for each of 2013, 2014 and 2015; corresponding shipping activity (based on AIS data) are also presented. In all years, narwhal densities were highest during Survey Period 3 (late August/early September; Figure 23A3) compared to earlier and later in the season. By mid-September in each year, narwhals had mostly moved out of Milne Inlet, Koluktoo Bay and Tremblay Sound with the exception of 2013 when narwhals occurred in relatively high densities in Koluktoo Bay (Figure 23A4). Later in the open-water season, narwhal densities generally decreased as narwhals migrated eastward out of
their primary summering areas (Figure 23A5, A6). As detailed in Section 4.3 and shown in Figure23B, shipping activity was more frequent in 2015 relative to 2013 and 2014. Given the inherent variability in narwhal occurrence in the study area, it is not appropriate to analyze observed densities relative to shipping activity without accounting for this variation. However, one observation stands out in Figure 23—the relatively high density of narwhals in Tremblay Sound and low density of narwhals in Milne Inlet South in late August 2015 versus the same survey period in 2013 and 2014 (Figure 23A3). This may be related to increased vessel traffic in Milne Inlet in late August 2015 versus other years (Figure 23B3). 4.4.3 Generalized Non-linear Mixed Model After pooling data within each 2-min sample, the final data matrix used for the model represented a total of 1598 EUs (86% were zeroes—no narwhals present in the EU). For the final model specification, the plot of observed versus predicted values indicated random fluctuation around an unbiased prediction with reasonable prediction intervals (Appendix C, Figure C-1). The moving sum residual plot further indicated the model specification was appropriate (Appendix C, Figure C-2). Independence across subjects (subject = Year×Julian×GeoStrat×Transect) was tested with a likelihood ratio test and indicated that the random block effect was successful in partitioning residual variance (P < 0.0001; Table 12).
Page 49 Marine Mammal Aerial Surveys 2015
FIGURE 23. Narwhal density estimates (uncorrected) and number of vessels during the day of the aerial survey in geographic strata during (A1−B1) Survey Period 1, (A2−B2) Survey Period 2, (A3− B3) Survey Period 3, (A4−B4) Survey Period 4, (A5−B5) Survey Period 5, and (A6−B6) Survey Period 6 for the three years that AIS data were available. Only vessels that were moving were included in the analysis. [KB=Koluktoo Bay, TS=Tremblay Sound; MIS=Milne Inlet South; MIN=Milne Inlet North; ESW=Eclipse Sound West; ESE=Eclipse Sound East; NBI=Navy Board Inlet; PI=Pond Inlet]
Page 50 Marine Mammal Aerial Surveys 2015
TABLE 12. Covariance parameter estimates and tests of significance for GNLMM of extensive aerial survey data.
Likelihood ratio Parameter Covariance structure Subject Estimate SE test (p-value) Intercept Variance components Year×Julian×Zone×Transect 4.54 1.07 <0.0001 Scale for NegBin 11.19 2.13
The model interaction term GeoStrat×Julian was significant (Type III P < 0.0001), which indicated seasonal change in narwhal distribution across geographical strata within the study area (Figure 24). This pattern was consistent with that reported in 2013 (Elliott et al. 2015) and 2014 (Thomas et al. 2015). Number of narwhals was highest in Eclipse Sound East and West during the earliest and latest parts of the season as narwhals entered and exited the study area through these strata. Narwhal occurrence in Eclipse Sound West was a little more stable throughout the season as animals used this stratum for transitioning among all other strata. Based on GNLMM results, peak numbers for both Milne Inlet/Koluktoo Bay and Tremblay Sound occurred from mid to late August.
Figure 24. Output for the fixed effect interaction of the categorical variable, GeoStrat, and the continuous variable, Julian, used in the generalized nonlinear mixed model (GNLMM). Type III P-value <0.0001 for the GeoStrat×Julian interaction. Dashed lines represent 90% confidence limits.
Page 51 Marine Mammal Aerial Surveys 2015
The number of narwhals significantly decreased as the number of vessels (see Table D-1 in Appendix D for a summary of vessel numbers) present in the geographic strata increased (Type III P = 0.024; Table 13). While this trend was consistently observed as the number of vessels increased from zero to >2, only the pairwise comparison between vessel number level = 0 and level = >2 was statistically significant at α = 0.05 (Tukey adjusted P = 0.024). For this comparison, there were about 11 times more narwhals when no vessels were present compared to when more than two vessels were present. Despite this finding, and given the increase in vessel activity during 2015 the term Year in the GNLMM was not statistically significant (Type III P = 0.513). This suggests that there was no substantial or consistent change in the overall number of narwhals across 2013, 2014 and 2015.
TABLE 13. Output for the categorical variables used in the generalized nonlinear mixed model (GNLMM). Left of the vertical partition—overall Type III P-values are given for the effects, as well as marginal means (narwhal densities) and confidence intervals for each of their levels. Bars overlaying the marginal means reflect magnitude. Right of the partition—pairwise comparisons among levels within each effect. For each comparison, the ratio of the greater mean to the lesser mean is provided above the diagonal line; below the diagonal are the corresponding pairwise P-values (Tukey adjusted for multiple comparisons).
90% Effect (Type III p-value) Marginal mean LCL UCL Levels for each effect Number of vessels (0.0244) 0 1 2 >2 0 0.051 0.016 0.160 1.8 3.8 11.2 1 0.029 0.007 0.120 0.907 2.1 6.3 2 0.013 0.001 0.189 0.762 0.948 2.9 >2 0.005 0.001 0.020 0.024 0.102 0.865 Year (0.5135) 2013 2014 2015 2013 0.021 0.003 0.135 1.9 1.0
2014 0.011 0.003 0.038 0.728 2.0 2015 0.022 0.007 0.073 0.999 0.542
4.5 Photographic Survey Results 4.5.1 Marine Mammal Sightings—Overview Four species of marine mammals were identified during the review of the photographic data: narwhal, bowhead whale, bearded seal and walrus (Table 14). Narwhals were the focus of the photographic surveys and are further analysed in Section 4.5.2, Narwhals. Marine mammal sightings other than narwhals are discussed briefly below. 4.5.1.1 Bowhead Whale One bowhead whale sighting of two individuals was detected during photo review (Table 14). The bowheads occurred in Tremblay Sound on 22 August 2015 (Figure 25). The photo reviewer was unable to estimate size or age class of the bowheads.
Page 52 Marine Mammal Aerial Surveys 2015
TABLE 14. Numbers of narwhals, bowhead whales and pinnipeds identified during photographic surveys conducted from 18 August to 4 September 2015. Excludes all sightings coded as low confidence (sighting confidence low). Survey 3 Survey 4 Survey 5 Survey 7 Milne Milne Tremblay Milne Tremblay Milne Total No. Species Inlet Inlet Sound Inlet Sound Inlet Narwhal No. of Sightings 25 0 631 349 1299 799 3103 No. of Individuals 43 0 954 588 2568 1583 5736
Bowhead No. of Sightings 0 0 1 0 0 0 1 No. of Individuals 0 0 2 0 0 0 2 Bearded Seal No. of Sightings 0 0 0 1 0 1 2 No. of Individuals 0 0 0 1 0 1 2 Walrus No. of Sightings 0 0 0 0 0 1 1 No. of Individuals 0 0 0 0 0 1 1 Unidentified Seal No. of Sightings 0 0 0 7 0 5 12 No. of Individuals 0 0 0 8 0 5 13 Unidentified Pinniped No. of Sightings 0 0 0 1 0 0 1 No. of Individuals 0 0 0 1 0 0 1
4.5.1.2 Pinnipeds The photographic surveys were not designed for pinnipeds. However, reviewers identified seal sightings whenever possible. Observations of seals were more frequently recorded during periods of low sea state. Bearded seal and walrus were the only pinnipeds identified to species probably due to their larger size (Table 14).
There were two sightings of bearded seals in Milne Inlet (30 August and 4 September)— both of individual seals (Table 14; Figure 25). There was one sighting of a walrus (one individual) on 4 September in Milne Inlet (Figure 25). Unidentified seals and pinnipeds were observed in Milne Inlet and were seen as either individual or pairs of animals (Table 14; Figure 25).
Page 53 Marine Mammal Aerial Surveys 2015
FIGURE 25. Bowhead whale, bearded seal, walrus, unidentified seal, and unidentified pinniped sightings identified during the photographic surveys (18 August−4 September 2015).
4.5.2 Narwhals 4.5.2.1 Numbers Observed Overall, there were 3103 sightings of narwhals totalling 5736 individuals identified across all photographic surveys in 2015 (Table 14). Tremblay Sound had the three highest counts of narwhal sightings and individuals, with the highest count of 823 sightings (1663 individuals) recorded during Replicate 2 on the 30 August (Table 15).
Page 54 Marine Mammal Aerial Surveys 2015
TABLE 15. Numbers of narwhals identified during each photographic survey replicate conducted from 18 August to 4 September 2015. Excludes all sightings coded as low confidence. Milne Inlet Tremblay Sound Shipping No. of No. of No. of No. of Vessel Name Activity Survey/Replicate Sightings Indiv. Sightings Indiv. Survey 3 (18 Aug) Replicate 2 8 8 − − Before M/V Golden Ice Replicate 4 13 19 − − Before M/V Golden Ice Replicate 4 4 16 − − During M/V Golden Ice
Survey 4 (22 Aug) Replicate 1 − − 273 417 During M/V Nordic Odyssey Replicate 2 − − 358 537 After M/V Nordic Odyssey Replicate 6 0 0 − − After M/V Nordic Odyssey
Survey 5 (30 Aug) Replicate 1 10 12 27 67 During/Before M/V Nordic Oshima Replicate 2 64 79 823 1663 After/During M/V Nordic Oshima Replicate 3 275 497 − − After M/V Nordic Oshima Replicate 4 − − 449 838 After M/V Nordic Oshima
Survey 7 (4 Sept) Replicate 1 98 148 − − Before M/V Golden Brilliant Replicate 1 8 16 − − During M/V Golden Brilliant Replicate 2 179 298 − − During M/V Golden Brilliant Replicate 3 131 199 − − During M/V Golden Brilliant
Replicate 4 263 469 − − During M/V Golden Brilliant Replicate 5 35 136 − − During M/V Nordic Oshima Replicate 6 59 219 − − During M/V Nordic Oshima Replicate 7 0 0 During M/V Nordic Oshima Replicate 7 21 81 − − After M/V Nordic Oshima Replicate 8 0 0 After M/V Nordic Oshima
Replicate 8 5 17 − − During M/V Golden Ruby Total 1173 2214 1930 3522
4.5.2.2 Group Size Narwhals were typically observed as individuals or groups of two or three (91% of sightings; Figure 26). Group size ranged from 1 to 22 and the average group size was 1.8 individuals. In Milne Inlet, narwhals observed Before a vessel entered the study area, group size ranged from 1 to 7, with an average group size of 1.5 individuals (n = 119). During passage of a vessel through the study area, narwhal group size ranged from 1 to 14, with an average group size of 2.0 individuals (n = 694). For narwhals observed After a vessel left the study area, group size ranged from 1 to 15, with an average group size of 1.8 individuals (n = 360). The differences between the proportions of observed group size in Milne Inlet North was not significant during periods Before, During and After vessel passages (χ2 = 7.2, df = 4, P = 0.126; Table 16).
Page 55 Marine Mammal Aerial Surveys 2015
1800 1600 1400 After 1200 During 1000 Bef ore 800 600 400 Total Narwhal Sightings Narwhal Total 200 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Group Size
FIGURE 26. Total narwhal sightings with various group sizes Before, During and After vessel passages.
TABLE 16. Group sizes of narwhals in Milne Inlet North and Tremblay Sound, categorized by vessel activity, during the photographic aerial surveys.
Vessel Activity Area / Before During After Group Size No. % No. % No. % Milne Inlet North
1 83 69.7 362 52.2 223 61.9
2 24 20.2 177 25.5 71 19.7
3 9 7.6 77 11.1 25 6.9
>3 3 2.5 78 11.2 41 11.4
Tremblay Sound
1 9 33.3 618 56.4 432 53.5
2 10 37.0 258 23.5 262 32.5
3 2 7.4 108 9.9 74 9.2
>3 6 22.2 112 10.2 39 4.8
In Tremblay Sound, group size ranged from 1 to 8, with an average group size of 2.5 individuals (n = 27) for narwhals recorded Before a vessel entered the study area. During passage of a vessel through the study area, narwhal group size ranged from 1 to 22, with an average group size of 1.9 individuals (n = 1096). For narwhals observed After a vessel left the study area, group size ranged from 1 to 8, with an average group size of 1.7 individuals (n = 807). The differences between the proportion of observed group size was not significant at times
Page 56 Marine Mammal Aerial Surveys 2015
During and After large vessel passages (χ2 = 3.5, df = 3, P = 0.321; Table 16). Sightings made Before a vessel passage are excluded from the analysis due to the small sample size. 4.5.2.3 Orientation Narwhal orientation (heading) was coded in the database to the nearest 15°. Orientation data were grouped into three categories based on survey timing relative to vessel passage through the study area (i.e. Before, During and After). If narwhals were responding to a large vessel passing through the study area by swimming away, their orientations would depend on whether the vessel was travelling north or south through the study area. As such, the data were also grouped into northbound and southbound vessels. Because vessels did not transit through Tremblay Sound, sightings in Tremblay Sound and Milne Inlet were examined separately. As a reference, only six narwhal sightings were identified east of the vessel trackline in Milne Inlet. It should be noted that the following analyses do not account for the influences of small boat traffic (notably narwhal hunting) on narwhal behaviour. As noted previously, during the photographic surveys, there was never more than one ship transiting in the study area. Milne Inlet.—In Milne Inlet, the vector mean (i.e., average) heading of 103 narwhal sightings (from two surveys and three replicates) observed Before a northbound vessel (M/V Golden Ice and M/V Golden Brilliant) entered the study area was 295°T (angular standard deviation (SD) of 73°T); narwhal headings were not randomly distributed (Z = 20.20, P < 0.001; Figure 27A). A northwest orientation indicates that on average narwhals were orientated away from the vessel trackline prior to vessel arrival in the study area. See also Table D-2 which presents data for each photographic survey. The average heading of 511 narwhal sightings (from two surveys and five replicates) observed During a northbound vessel (M/V Golden Ice and M/V Golden Brilliant) passage through the study area was 229°T (angular SD 64°T); narwhal headings were not randomly distributed (Z = 146.08, P < 0.001; Figure 27B). On average, narwhals were oriented in a direction away from the vessel trackline. The average heading of 108 narwhal sightings (from two surveys and five replicates) observed During a southbound vessel (M/V Nordic Oshima (30 August and 4 September), and
M/V Golden Ruby) passage in the study area was 142°T (angular SD 64°T); narwhal headings were not randomly distributed (Z = 146.08, P < 0.001; Figure 27C). A southeast orientation indicates that on average, narwhals were orientated away from the vessel as it approached, or toward if the vessel had already passed the narwhal sighting. See also Table D-2 which presents data for each photographic survey. The average heading of 330 narwhal sightings (from two surveys and three replicates) observed After a southbound vessel (M/V Nordic Oshima (30 August and 4 September), and M/V Golden Ruby) passed through the study area or anchored at Ragged Island was 31°T with an (angular SD 38°T); narwhal headings were not randomly distributed (Z = 209.79, P < 0.001; Figure 27D). A northeast heading suggests that on average narwhals were orientated away from the vessel.
Page 57 Marine Mammal Aerial Surveys 2015
FIGURE 27. Headings of narwhals in the Milne Inlet survey area during 18 August, 30 August, and 4 September, for periods A. Before the passage of northbound vessels (n = 103), B. During the passage of
northbound vessels (n = 511), C. During the passage of southbound vessels (n = 108), and D. After the passage of southbound vessels (n = 330). Figures are based on sightings from photographic surveys.
Tremblay Sound.—In Tremblay Sound, the average heading of 26 narwhal sightings (from one survey and one replicate) Before a southbound vessel (M/V Nordic Oshima) moved into the Milne Inlet study area was 179°T (angular SD 95°T); narwhal headings were randomly distributed (Z = 1.59, P = 0.204; Figure 28A). The average heading of 761 narwhal sightings (from one survey and one replicate) observed During a southbound vessel (M/V Nordic Oshima) passage in the Milne Inlet study area was 212°T (angular SD 92°T); narwhal headings were not randomly distributed (Z = 58.33, P < 0.001; Figure 28B). A southwestern heading suggests that, on average, narwhals were headed into Tremblay Sound during a vessel passage in northern Milne Inlet.
Page 58 Marine Mammal Aerial Surveys 2015
During a northbound vessel (M/V Nordic Odyssey) passage through the Milne Inlet study area, the average heading of narwhals (205 sightings during one survey and one replicate) was 189°T (angular SD 85°T); headings were not randomly distributed (Z = 22.38, P < 0.001; Figure 28C). A southerly heading suggests that, on average, narwhals were headed into Tremblay Sound during a vessel passage through northern Milne Inlet The average heading of 628 narwhal sightings (from two surveys and two replicates) observed After a southbound vessel (M/V Nordic Odyssey and M/V Nordic Oshima) moved out of the Milne Inlet study area or anchored at Ragged Island was 38°T (angular SD 40°T); narwhal headings were not randomly distributed (Z = 381.25, P < 0.001; Figure 28D). This suggests that on average narwhals were headed towards the northeastern portion of Tremblay Sound. See also Table D-2 which presents data for each photographic survey.
FIGURE 28. Headings of narwhals in Tremblay Sound during 22 and 30 August 2015, for periods A. Before the passage of southbound vessels (n = 26), B. During the passage of southbound vessels (n = 761), C. During the passage of northbound vessels (n = 205), and D. After the passage of southbound vessels (n = 628). Figures are based on sightings from photographic surveys.
Page 59 Marine Mammal Aerial Surveys 2015
4.5.2.4 Density Estimates and Shipping Activity An availability bias of 3.18 (SE = 0.038) was used to correct narwhal density estimates in this study. This value is from a recent satellite tagging study of narwhals in Admiralty Inlet and Tremblay Sound fitted in August (Watt et al. 2015). Based on a sample of photographic data from Survey 5, Replicate 4, the perception bias for narwhals was calculated as 0.969 with a SE of 0.239 (31, 1, 1 for shared sightings, first reviewer, and second reviewer, respectively). Based on Figure 29, only sightings that occurred within 200 m of the trackline were used in the perception bias calculation. Based on pooled sighting data for narwhal sightings with DISTANCE estimates from the surveys, the detection probability estimate or ƒ(0) value was 1.30 with an estimated effective survey strip width of 769.0 m. The half-normal model was selected as the best model during analysis. The estimated detection function for narwhals is shown in Figure 29. Narwhal density estimates were calculated for each survey replicate and geographic stratum with available photo data. Results are presented below, organized by survey. Survey 3: 18 August.—Narwhals were present in Milne Inlet in relatively low numbers during Survey 3 (Figure 30A and 30B). Two ore carriers (M/V Golden Brilliant, M/V Nordic Orion) were located at the Ragged Island anchorage site during the survey. During Replicate 2 (Before Vessel), the ore carrier M/V Golden Ice was anchored at Milne Port. During Replicate 4
FIGURE 29. Estimated narwhal detection function based on photographic survey data.
Page 60 Marine Mammal Aerial Surveys 2015
A. Survey 3, Replicate 2 B. Survey 3, Replicate 4 Flight time: 1358−1443 h Flight time: 1929−2014 h (Before Vessel) (Before/During Vessel)
FIGURE 30. Narwhal sightings in Milne Inlet during Survey 3 (18 August 2015): A. Replicate 2, Before the passage of a northbound vessel, and B. Replicate 4, Before/During the passage of a northbound vessel (M/V Golden Ice; bottom left in B.). The vessel trackline connecting the position of the vessel at the start and end of the replicate gives an approximate trackline.
(Before/During Vessel), the M/V Golden Ice entered the southern portion of the study area heading in a northerly direction. Narwhal densities were low both Before and During the vessel passage—0.11 individuals/km2 and 0.46 individuals/km2 during Replicates 2 and 4, respectively (Table 17). Tremblay Sound was surveyed once on 18 August, but it was not selected to be reviewed in the photo analysis. Aerial observers did observe large numbers (100s) of narwhals in Tremblay Sound during the survey. Six small local boats (likely some repeat sightings) were observed during the survey, although no hunting activity was observed.
TABLE 17. Density estimates of narwhals in Milne Inlet during Survey 3 (18 August 2015). Estimates were generated from the DISTANCE software (Thomas et al. 2010) using the CDS module and are corrected for availability, perception, and detection biases.
No. of No. of De nsity 95% Confidence Interval Shipping 2 Area/Replicate Sightings Individ. (indiv./km ) CV Lower Upper Activity Milne Inlet Replicate 2 8 8 0.11 0.59 0.24 0.49 Before
Replicate 4 17 35 0.46 0.73 0.81 2.61 Before/During
Page 61 Marine Mammal Aerial Surveys 2015
Survey 4: 22 August.—Narwhals were present in Tremblay Sound during Survey 4 (Figure 31A and 31B). During the survey, two ore carriers (M/V Golden Brilliant, M/V Golden Saguenay) were anchored at Ragged Island. During Replicate 1 (During Vessel), the ore carrier M/V Nordic Odyssey was at the northern end of Milne Inlet headed north. During Replicate 2 (After Vessel), the Nordic Odyssey had been out of the study area for about 1.7 hours and was > 43 km east of Tremblay Sound. High densities of narwhals (22.71 and 28.20 individuals/km2) were observed in Tremblay Sound during Replicate 1 and 2, respectively (Table 18). Narwhal sightings appeared to be less clumped during Replicate 2. Milne Inlet was surveyed (flight time: 1412−1457 h) just prior to Replicate 2 and no narwhals were observed (Table 18). Six replicates were surveyed in Milne Inlet from 0814 h to 1456 h and aerial observers recorded three sightings in Mine Inlet during the survey. One small boat was observed during the survey, although no hunting activity was observed.
A. Survey 4, Replicate 1 B. Survey 4, Replicate 2 − Flight time: 1306 1316 h Flight time: 1501−1512 h (During Vessel) (After Vessel)
FIGURE 31. Narwhal sightings in Tremblay Sound during Survey 4 (22 August 2015): A. Replicate 1 During the passage of a northbound vessel, and B. Replicate 2 After the passage of a northbound vessel (M/V Nordic Odyssey; bottom left in A.).
Page 62 Marine Mammal Aerial Surveys 2015
TABLE 18. Density estimates of narwhals in Milne Inlet and Tremblay Sound during Survey 4 (22 August 2015). Estimates were generated from the DISTANCE software (Thomas et al. 2010) using the CDS module and are corrected for availability, perception and detection biases.
No. of No. of De nsity 95% Confidence Interval Shipping 2 Area/Replicate Sightings Individ. (indiv./km ) CV Lower Upper Activity Milne Inlet Replicate 6 0 0 0.00 − − − After Tremblay Sound Replicate 1 273 417 22.71 0.25 8.55 60.29 During
Replicate 2 358 537 28.20 0.25 10.37 76.66 After
Survey 5: 30 August.—The ore carrier M/V Golden Saguenay transited south through the study area (from 0833 h to 1125 h) on its way to Milne Port prior to the commencement of Survey 5 (at 1326 h). The ore carrier M/V Nordic Oshima entered the study area from the north at 1340 h and anchored on the south side of Ragged Island at ∼1458 h. During the survey, a project ore carrier (M/V Nordic Olympic) remained at the Ragged Island anchorage site. Narwhals were present in Milne Inlet during Survey 5 with most sightings located on the northwest side of Milne Inlet (Figure 32A-C). In Milne Inlet, narwhal densities were relatively low during Replicate 1 (0.16 individuals/km2) and increased during Replicate 2 and Replicate 3 (1.04 individuals/km2 and 6.58 individuals/km2, respectively; Table 19). Narwhals were also present in Tremblay Sound during Survey 5 with a large herd located in the center of the fjord and a smaller herd located at the north end of the fjord (Figure 33A-C). In Tremblay Sound, narwhal densities were relatively low during Replicate 1 (5.95 individuals/km2) and increased substantially during Replicate 2 and Replicate 4 (Table 19). The highest density of narwhals was observed in Tremblay Sound during Replicate 2 (80.50 individuals/km2) when the M/V Nordic Oshima was in the Milne Inlet study area (Table 19). Two hunting boats (likely a repeat sighting) were observed during the survey at the south end of Milne Inlet North.
TABLE 19. Density estimates of narwhals in Milne Inlet and Tremblay Sound during Survey 5 (30 August 2015). Estimates were generated from the DISTANCE software (Thomas et al. 2010) using the CDS module and are corrected for availability, perception and detection biases.
No. of No. of De nsity 95% Confidence Interval Shipping 2 Area/Replicate Sightings Individ. (indiv./km ) CV Lower Upper Activity Milne Inlet Replicate 1 10 12 0.16 0.93 0.02 1.64 During Replicate 2 64 79 1.04 1.02 0.09 12.68 After Replicate 3 275 497 6.58 1.03 0.52 83.55 After Tremblay Sound Replicate 1 27 67 5.95 0.29 2.63 13.47 Before Replicate 2 823 1663 80.50 0.25 29.68 218.33 During Replicate 4 449 838 40.70 0.25 15.01 110.35 After
Page 63 Marine Mammal Aerial Surveys 2015
A. Survey 5, Replicate 1 B. Survey 5, Replicate 2 − Flight time: 1339−1426 h Time: 1610 1657 h (During Vessel) (After Vessel)
C. Survey 5, Replicate 3 Time: 1701−1747 h (After Vessel)
FIGURE 32. Narwhal sightings in Milne Inlet during Survey 5 (30 August 2015): A. Replicate 1 During the passage of a southbound vessel, B. Replicate 2 After the passage of a southbound vessel, and C. Replicate 3 After the passage of a southbound vessel (M/V Nordic Oshima; bottom left in A.). The vessel entered Milne inlet from the north and anchored on the south side of Ragged Island.
Page 64 Marine Mammal Aerial Surveys 2015
A. Survey 5, Replicate 1 B. Survey 5, Replicate 2 Flight time: 1326−1334 h Flight time: 1435−1446 h (Before Vessel) (During)
C. Survey 5, Replicate 4 Flight time: 1755−1805 h (After Vessel)
FIGURE 33. Narwhal sightings in Tremblay Sound during Survey 5 (30 August 2015): A. Replicate 1 Before the passage of a southbound vessel, B. Replicate 2 During the passage of a southbound vessel, and C. Replicate 4 After the passage of a southbound vessel (M/V Nordic Oshima; bottom left in B.). The vessel anchored on the south side of Ragged Island.
Page 65 Marine Mammal Aerial Surveys 2015
Survey 7: 4 September.—Three ore carriers transited in at least a portion of the study area on 4 September. The M/V Golden Brilliant transited through the study area; it entered from the south at 1027 h and exited the study area at ∼1310 h. The M/V Nordic Oshima left anchor from the south side of Ragged Island at 1353 h enroute to Milne Port; it exited the study area at ∼1608 h. The M/V Golden Ruby entered the study area from the north at ∼1652 h and anchored at Ragged Island at ∼1800 h. During the survey, two ore carriers (M/V Nordic Odin, M/V Golden Opportunity) remained at the Ragged Island anchorage. All eight replicates during Survey 7 (starting at 0952 h and ending at 1731 h) overlapped with part or all of a vessel transit in the study area. Tremblay Sound was not surveyed on 4 September in order to focus more effort in Milne Inlet. Narwhals were most abundant in the southern portion of Milne Inlet during Survey 7 (Figures 34 and 35). Narwhal densities were generally higher during the first four replicates (2.28, 4.33, 2.81, and 6.35 individuals/km2, respectively) compared to the last four replicates (1.85, 2.98, 1.12 and 0.23 individuals/km2, respectively; Table 20). Narwhal density was highest during Replicate 4 (6.35 individuals/km2) when an ore carrier was at the north end of Milne Inlet at its farthest point (while still in the study area) from the narwhal herd (Table 20; Figure 34D). There was a noticeable shift in the narwhal herd location from around the middle of the southern portion of the Milne Inlet study area (narwhal sightings were on average ∼2 km from shore) during the first four replicates to areas close to the shoreline near northern Bruce Head (narwhal sightings were on average ∼500 m from shore) during the last four replicates (Table 20; Figures 34 and 35). In addition, narwhal group size approximately doubled during the last four replicates versus the first four replicates (Table 20). During the survey, aerial observers noted large numbers of narwhals in Fairweather Bay (west of Stephens Island) which were generally not captured in the photographs. Nineteen small boats were observed during the survey. There was one sighting of a small boat in close proximity to the narwhal herd during the first four replicates and seven sightings of small vessels (probably many repeat sightings) that appeared to be hunting during the last four replicates. A narwhal carcass was also observed during Replicate 7 in close proximity to the narwhal herd.
TABLE 20. Density estimates of narwhals in Milne Inlet during Survey 7 (4 September 2015). Density estimates were generated from the DISTANCE software (Thomas et al. 2010) using the CDS module and are corrected for availability, perception and detection biases. Narwhal group size and distance from shore are also provided.
Shipping Mean Mean De nsity 95% Confidence Interval No. of No. of Activity Group Distance from Area/Replicate Sightings Individ. (indiv./km2) CV Lower Upper Size Shore (m) Milne Inlet Replicate 1 106 164 2.28 0.88 0.25 20.99 Before/During 1.5 2603 Replicate 2 179 298 4.33 0.76 0.62 30.08 During 1.7 2182 Replicate 3 131 199 2.81 0.77 0.40 19.91 During 1.5 2961 Replicate 4 263 469 6.35 0.83 0.77 52.68 During 1.8 1930 Replicate 5 35 136 1.85 0.86 0.21 16.07 During 3.9 549 Replicate 6 59 219 2.98 0.62 0.64 13.91 During 3.7 445 Replicate 7 21 81 1.12 0.57 0.31 4.12 During/After 3.9 476 Replicate 8 5 17 0.23 1.06 0.02 2.89 After/During 3.4 541
Page 66 Marine Mammal Aerial Surveys 2015
A. Survey 7, Replicate 1 B. Survey 7, Replicate 2 Flight time: 0952−1038 h Flight time: 1041−1123 h (Before/During Vessel) (During Vessel)
C. Survey 7, Replicate 3 D. Survey 7, Replicate 4 Flight time: 1127−1210 h Flight time: 1214−1300 h (During Vessel) (During Vessel)
FIGURE 34. Narwhal sightings in Milne Inlet during Survey 7, Replicates 1−4 (4 September 2015): A. Replicate 1 Before/During the passage of a northbound vessel, B. Replicate 2 During the passage of a northbound vessel, C. Replicate 3 During the passage of a northbound vessel and D. Replicate 4 During the passage of a northbound vessel (M/V Golden Brilliant; bottom left in A.).
Page 67 Marine Mammal Aerial Surveys 2015
A. Survey 7, Replicate 5 B. Survey 7, Replicate 6 Flight time: 1416−1502 h Flight time: 1506−1553 h (During Vessel) (During Vessel)
C. Survey 7, Replicate 7 D. Survey 7, Replicate 8 Flight time: 1556−1643 h Flight time: 1646−1732 h (During/After Vessel) (After/During Vessel)
FIGURE 35. Narwhal sightings in Milne Inlet during Survey 7, Replicates 5−8 (4 September 2015): A. Replicate 5 During the passage of a southbound vessel, B. Replicate 6 During the passage of a southbound vessel, C. Replicate 7 During/After the passage of a southbound vessel (M/V Nordic Oshima; bottom left in A.) and D. Replicate 8 After/During the passage of a southbound vessel (M/V Golden Ruby).
Page 68 Marine Mammal Aerial Surveys 2015
4.5.3 Generalized Non-linear Mixed Model The final data matrix contained 10,260 photographs. Only 2.1% of these photos contained narwhals. On average, there were 0.2 narwhals per photograph (range = 0−115). For the final model specification, the plot of observed versus predicted values indicated random fluctuation around an unbiased prediction with reasonable prediction intervals (Appendix C; Figure C-3). The moving sum residual plot further indicated the model specification was appropriate (Appendix C; Figure C-4). Independence across random subjects indicated success in partitioning residual variance (Table 21). Significantly more narwhals were observed when there was a land barrier between a photograph location and a vessel trackline (Table 22; Figure 36). The interaction of vessel activity (Before, During and After or BDA) and photograph CPA distance to the vessel trackline was highly significant (Type III P < 0.0001). More specifically, the number of narwhals increased in photographs farther from the trackline During vessel passages as compared to Before and After a vessel passage (Table 23; Figure 37).
TABLE 21. Covariance parameter estimates and tests of significance for the GNLMM of the photographic data.
Likelihood ratio test Parameter Covariance structure Subject Estimate SE (P-value) Intercept Variance components Date × Replicate 2.80 1.23 <0.0001 Scale for NegBin 61.28 5.14
TABLE 22. P-values for fixed effects included in the GNLMM of the photographic data.
Fixed effect Type III P-value Land barrier between photo and vessel 0.0498 Before, During, or After vessel transit (BDA) 0.1252 Spline function for photo distance from vessel track line (Spline) <0.0001 BDA_×_Spline <0.0001
Page 69 Marine Mammal Aerial Surveys 2015
FIGURE 36. Predicted number of narwhals (individuals per photograph) versus whether the photograph was blocked from the vessel by a land barrier (variable LandBarr described in the methods) based on least square means output from the generalized nonlinear mixed model (GNLMM). Ratio of means for “Yes” versus “No” was 4.2; Type III P = 0.0498. Values in parentheses reflect 90% confidence limits.
TABLE 23. Weighted mean distances (m) between photographs and the closest point of approach (CPA) Before, During, and After vessel passages. Estimates are given both for observed values and for output from the GNLMM. Mean distances are weighted by the number of narwhals observed/estimated for each photograph.
Weighted Mean Difference from Vessel Activity Ratio to "Before" CPA Distance "Before" Based on observed values Before 4545 0 1.00 During 7775 3230 1.71 After 4633 88 1.02 Based on GNLMM marginal means Before 7291 0 1.00 During 11048 3757 1.52 After 7560 269 1.04
Page 70 Marine Mammal Aerial Surveys 2015
FIGURE 37. Output for the fixed effect interaction of the categorical variable, BDA (Before, During, After vessel transit), and the continuous variable, PhotoCPA, used in the generalized nonlinear mixed model (GNLMM). Type III P < 0.0001 for the BDA×PhotoCPA interaction. Dashed lines represent 90% confidence limits.
Page 71 Marine Mammal Aerial Surveys 2015
5 DISCUSSION The 2015 open-water season marked the first year of Operational shipping for the ERP of the Mary River Project. Baffinland’s Marine EEM Plan outlined the approach for monitoring of marine mammals and emphasized the importance of documenting potential effects of shipping on narwhals during this first year of increased shipping activity (SEM and LGL 2016). As such, the aerial survey program was modified in 2015 relative to 2013 and 2014 when the emphasis was on baseline data collection (Elliott et al. 2015; Thomas et al. 2015). Extensive aerial surveys of the northern shipping route were conducted in conjunction with photographic surveys of Milne Inlet and Tremblay Sound—two key summering areas for narwhals. The text below offers interpretation of the findings and highlights the uncertainties associated with documenting narwhal response to shipping in the study area. 5.1 Narwhal Relative Abundance and General Distribution 5.1.1 Observed Densities and General Distribution Observed narwhal densities (i.e., numbers of individuals/km2; uncorrected for biases) were calculated for each year Baffinland conducted extensive aerial surveys and serve as an indicator of relative abundance in 2015, 2014 and 2013 (as well as 2007 and 2008). As in previous years, in 2015, observed narwhal densities varied considerably across geographic strata and two-week survey periods (early August to mid-September in 2015). During the three surveys in August 2015, highest narwhal densities were observed in key summering areas—i.e., Milne Inlet South (2.21, 0.87, 4.93 individuals/km2), Milne Inlet North (0.60, 1.23, 0.14 individuals/km2), Koluktoo Bay (1.90, 0.20, 0.89 individuals/km2) and Tremblay Sound (0.91, 2.61, 18.26 individuals/km2). Based on the GNLMM results of the extensive aerial survey data, the overall spatial-temporal trend in narwhal abundance has been consistent across 2013, 2014 and 2015—narwhal numbers peak in Milne Inlet, Koluktoo Bay and Tremblay Sound from mid to late August. By the last extensive aerial survey conducted in mid-September 2015, narwhals had mostly left their primary summering areas and had moved into Eclipse Sound West. This apparent departure time from primary summering areas was generally consistent with the pattern during aerial surveys conducted in 2013 and 2014 but in 2013 there were still relatively high numbers of narwhals in Koluktoo Bay and Milne Inlet South during mid-September (Figure 23; Elliott et al. 2015; Thomas et al. 2015). By late September 2013, narwhals had moved into Eclipse Sound East. In late September and early October 2014, for unknown reasons, relatively few narwhals were observed in the entire study area with small numbers observed in Tay Sound, Tremblay Sound, Eclipse Sound and Navy Board Inlet (Thomas et al. 2015). By mid-October 2014, narwhal migration out of the fjords was essentially complete and relatively high densities of narwhals were recorded in Pond Inlet (i.e., 1.27 and 0.33 individuals/km2 during Replicate 1 and Replicate 2, respectively) with many narwhals travelling in an easterly direction. Similarly, in October 2013, narwhals were primarily observed in Pond Inlet headed out of the study area towards their wintering grounds in Baffin Bay and northern Davis Strait.
Page 72 Marine Mammal Aerial Surveys 2015
5.1.2 Narwhal Movements within the Study Area Extensive aerial survey findings have consistently shown that there is much fine scale movement by groups of narwhals among various areas of Milne Inlet and adjacent fjords, both from day to day and over a longer time interval. In 2013 and 2014, the number of narwhals varied substantially between replicates within each Survey Period even though in most instances the second replicate was surveyed within two days of the first replicate (only a single survey was flown in each area in each period in 2015; replicates were not flown). This variability between replicates within the same survey period indicates that narwhals exhibit as a minimum, localized movements within a 24-hour period and their often highly clumped distribution increases the likelihood that sightings can be missed by observers surveying systematic transects. There was also much variation in the numbers of narwhals observed on a given day during the shore-based study of narwhals at Bruce Head, Milne Inlet in 2013, 2014 and 2015 (Thomas et al. 2014; Smith et al. 2015, 2016). The reasons for these fine-scale movements of narwhals within their summering area are uncertain but could possibly be related to natural factors like tide, response to killer whales, feeding activity and/or anthropogenic activities like hunting and shipping. 5.1.3 Annual Variation Narwhal density within the study area varies considerably from year-to-year based on aerial surveys completed in 1978-79, 1994, 1996, 2002, 2004, 2007, 2008, 2013, 2014 and this study (Koski and Davis 1979, 1980; Richard et al. 1994, 2010; Baffinland 2012; Doniol-Valcroze et al. 2015b; Elliott et al. 2015; Thomas et al. 2015). Depending on the geographic strata (Eclipse Sound, Milne Inlet, or smaller fjords) within the study area, the density may vary from year-to-year by a factor that ranges between 2 and 85 times (e.g., Koski and Brandon 2012). The reasons for these wide fluctuations in narwhal numbers from year to year have not been well- studied to date. Possible reasons include the timing of aerial surveys, observer bias and whether or not aerial surveys captured the location of large herds of narwhals on a given survey day. Also, some narwhals in the Eclipse Sound complex move to adjacent areas within a given summering period and do not necessarily return to the same summering area from year to year. Satellite tag data show that some narwhals move between Eclipse Sound and Admiralty Inlet within a given year (Watt et al. 2012). Over 40% of the 12 narwhals tagged in Eclipse Sound travelled west to Admiralty Inlet and two occurred in the summering range for the Somerset Island stock (Watt et al. 2012). Interestingly, the only narwhal to retain its tag beyond a year, did not return to Eclipse Sound where it was tagged in August 2010, it instead travelled past Eclipse Sound and remained in Admiralty Inlet from late July to early October when the tag stopped transmitting (Watts et al. 2012). If enough narwhals move between Eclipse Sound and Admiralty Inlet, it could explain some of the fluctuations in narwhal numbers within and between years. This was, in part, the rationale provided by Doniol-Valcroze et al. (2015b) for such different abundance estimates for the Eclipse Sound narwhal stock in 2002−2003 versus 2013. This natural variation must be accounted for when evaluating the effects of shipping on narwhals.
Page 73 Marine Mammal Aerial Surveys 2015
5.2 Narwhals and Shipping The AIS shipping data and overall aerial survey findings clearly demonstrate that narwhals in the study area have been regularly exposed to large vessel shipping even in the absence of vessels associated with the Mary River Project. However, most of the large vessel shipping activity in Milne Inlet is associated with the Mary River Project and in 2015, included primarily ore carrier traffic. Narwhals in the study area would have been exposed to sounds from these various types of large vessels, as well as smaller vessels (without AIS data), including those engaged in hunting. As previously noted, the photographic and extensive aerial surveys were designed to complement each other. The photographic survey allowed for examination of changes in the distribution and abundance of narwhals in key summering areas during a ship transit relative to periods Before, During and After the ship passage. From these data, the objective was to not only to detect potential changes in narwhal densities but also to quantify the duration of possible effects and how far from large vessels narwhals responded. In addition, the precision of the photographic approach allowed for investigation of more “subtle” changes in behaviour of narwhals (orientation, group size) during periods with and without large vessel shipping. In contrast, the extensive aerial surveys were designed to determine whether the overall spatial- temporal pattern of narwhal distribution and abundance in the study area changed and whether narwhal relative abundance within specific geographic strata changed in response to large vessels. 5.2.1 Extensive Aerial Surveys Results from the statistical model analyses (GNLMM) of the 2013, 2014 and 2015 extensive aerial data indicated that there was a statistically significant difference in narwhal densities relative to the number of large vessels present in the study area on the days that aerial surveys were conducted (Table 12). A consistent decrease in narwhal densities for a given geographic strata occurred as the number of large vessels increased from zero to greater than two vessels. However, the only statistically significant (i.e., P-value <0.05) result among pairwise comparisons for this shipping level variable was during periods when there were no large vessels
present versus periods when more than two vessels (i.e., three to five vessels) were present. There were about 11× more narwhals when no vessels were present compared to when more than two vessels were present. Despite this finding, and given the increase in vessel activity during 2015, the model term Year was not statistically significant, which indicates there was no substantial differences in the overall density of narwhals in 2013, 2014 and 2015. The GNLMM of the extensive aerial survey data included an average of 0.9, 0.6 and 2.3 large vessels per geographic stratum in each of 2013, 2014 and 2015, respectively (Appendix D, Table D-1; only vessels present on the day of the aerial surveys were included). These results suggest that while narwhals may be responding to large vessel transits by exhibiting temporary displacement and/or changes in behaviour that reduce sighting probability, large-scale decreases in their density and spatial-temporal distribution were not apparent. This is consistent with findings of the shore- based study of narwhals (Smith et al. 2016) and the photographic survey results discussed below.
Page 74 Marine Mammal Aerial Surveys 2015
5.2.2 Photographic Surveys Photographic surveys of Milne Inlet and Tremblay Sound were conducted for the first time in 2015. Four photographic surveys were conducted from 18 August to 4 September 2015. Most of the 4545 km of survey effort (~90%) occurred along four transects in Milne Inlet North. Analyses of photos were limited to about 60.9% (11,103 photos reviewed) and 71.4% (999 photos reviewed) of the overall effort in Milne Inlet North and Tremblay Sound, respectively. During two of the four surveys (Table 15) it was possible to collect data Before, During and After a vessel passage (i.e., 4 September, M/V Golden Brilliant and M/V Nordic Oshima; 30 August, M/V Nordic Oshima). Surveys on the other two days were limited to Before/During (18 August; M/V Golden Ice) and During/After (22 August; M/V Nordic Odyssey). The categorization of when a large vessel was considered Before, During or After was based on whether the vessel was within or outside of the photographic survey study area and the timing of its passage. It is recognized that at least some and perhaps many narwhals in Milne Inlet (and perhaps part of Tremblay Sound) were likely exposed to sound (albeit at reduced levels) from these large vessels as they approached (Before) and exited the study area (After). 5.2.2.1 Narwhal Orientation The orientation of narwhals was significantly different (P < 0.001 in all cases except two) from a random distribution, regardless of whether a large vessel was present in the study area and regardless of whether narwhals were located in Milne Inlet or Tremblay Sound. To put this more clearly, during photographic surveys, most narwhals were headed in a particular direction and not headed in random directions. Examining narwhal orientation relative to shipping activity needs to be interpreted carefully given that orientation relative to vessel presence will depend on
where narwhals are located in the study area relative to the location of the vessel as well as the direction of travel of the vessel. As such, specific results of the orientation analyses relative to shipping activity in Milne Inlet are discussed below. 5.2.2.2 Narwhal Group Size An overall analysis of narwhal group size in Milne Inlet during periods Before, During and After ship passage was not significantly different (P = 0.126). Average group sizes in Milne Inlet were 1.5, 2.0 and 1.8 narwhals Before, During, and After a vessel passage, respectively. In Tremblay Sound narwhal group size in periods During and After vessel passage was not significantly different (P = 0.321); there were too few sightings Before a vessel passage to test statistically. Average group sizes in Tremblay Sound were 2.5, 1.9 and 1.7 narwhals Before, During and After a vessel passage, respectively. Although overall group size in Milne Inlet did not appear to be affected by the presence of large vessels, narwhal group size did double on 4 September as narwhals were observed to move closer to shore in apparent response to vessel activity. 5.2.2.3 Narwhal Distribution and Relative Abundance Milne Inlet.—Results from the GNLMM of the photographic data indicated that During large vessel (i.e., ore carrier) transits through Milne Inlet, the number of narwhals increased significantly in areas farther from the vessel trackline relative to periods Before and After a large vessel transit. This indicates that narwhals exhibited at least temporary avoidance of large vessel
Page 75 Marine Mammal Aerial Surveys 2015
transits and/or exhibited changes in behaviour closer to the ship trackline that affected their detection in photographs. The data indicated that on average, narwhals were ~1.5 times farther from a large vessel trackline During its passage than Before and After its passage through the Milne Inlet study area. However, based on the model findings, narwhal avoidance of large vessels seems temporary. The pattern of narwhal abundance relative to distance from the vessel trackline was similar both Before and After a large vessel passage (Figure 37). These results should be interpreted with caution given that there were only two photographic surveys of Milne Inlet with high numbers of narwhals present and during which there were data for periods Before, During and After a large vessel passage. Also, during these two photographic surveys (i.e., on 30 August and 4 September), there was small boat and hunting activity in the study area. The GNLMM also indicated that more narwhals were present in areas separated from the vessel trackline by a land barrier (i.e., Stephens Island). This is likely a spurious finding given the limited number of photographic surveys and that during the 4 September photographic survey narwhals were predominantly located on the west side of Stephens Island, opposite of the vessel trackline to the east of the island, before the vessel transit—it is likely narwhals happened to be located in that area on the day the survey was flown. Narwhals have frequently been observed west of Stephens Island during previous aerial surveys (e.g., Baffinland 2012; Elliott et al. 2015; Thomas et al. 2015). The photographic survey conducted on 4 September contributed the most data to the GNLMM and the observed findings support the modelling results. Narwhals were initially located west and northwest of Stephens Island in relatively large numbers (Figure 34A–D and Figure 35A,B) and as the photographic survey progressed throughout the day (from ~1000 h to 1730 h) and ore carriers transited past the location of the narwhals, narwhals moved closer to the shoreline, increased group size, and then moved south past Bruce Head (and also into Fairweather Bay). A decrease in narwhal densities in the photographic study area coincided with the shore-based study team located ‘around the corner’ at Bruce Head recording 447 narwhals swimming south at a fast rate of speed. The ore carrier Nordic Oshima was transiting south through the shore-based study area (i.e., the Stratified Study Area) at this time (Smith et al. 2016). It is uncertain how long this response to an ore carrier transit lasted.
Tremblay Sound.—Given the limited data collected in Tremblay Sound, particularly during periods Before large vessels transited through Milne Inlet, it is not possible to determine whether narwhals moved into Tremblay Sound to avoid large vessels in Milne Inlet. However, there are indications from two photographic surveys that narwhals in Tremblay Sound possibly responded to ore carriers. On 22 August, large numbers of narwhals were recorded in the centre of Tremblay Sound (Figure 31) both During (417 individuals) and After (537 individuals) an ore carrier (M/V Nordic Odyssey) had passed through the Milne Inlet study area. During the vessel passage just south of the entrance to Tremblay Sound, the narwhal herd appeared more clumped compared to ~ 2 hours after the vessel had exited Milne Inlet North (Figure 31). On average, narwhals were orientated in a southerly direction during the vessel passage and in a north- easterly direction after the vessel passage. This suggests narwhals in Tremblay Sound may have changed their orientation and were possibly headed out of that fjord after the vessel passed through Milne Inlet. It is uncertain how sound from vessels transiting along the northern
Page 76 Marine Mammal Aerial Surveys 2015
shipping route through Milne Inlet propagates into Tremblay Sound. Oceans North deployed an acoustic recorder about half way down the fjord in 2015 and data from those recorders could help to address this question when they become available. Data collected on 30 August also indicated that narwhals in Tremblay Sound possibly responded to large vessel traffic in Milne Inlet. When the fjord was surveyed During an ore carrier (M/V Nordic Oshima) transit, two large herds (totalling 1663 individuals) were recorded—one in the centre and one at the entrance of Tremblay Sound (Figure 33B). On the subsequent survey replicates After the ore carrier had anchored south of Ragged Island, the narwhal herd that was located centrally in Tremblay Sound moved to the northeast (Figure 33C), and the herd that was located at the entrance to Tremblay Sound presumably moved into Milne Inlet North (Figure 32B and 32C). These movements out of Tremblay Sound are supported by narwhal orientation data. On average, narwhal headings in Tremblay Sound were distributed randomly Before, in a southwestern direction During, and in a northeastern direction After the M/V Nordic Oshima transited to Ragged Island. There was an extensive aerial survey flown the next day (31 August) and large numbers of narwhals were distributed throughout Tremblay Sound (558 individuals) with relatively few narwhal sightings occurring in Milne Inlet North. The 13 narwhals in Milne Inlet North were located south of the Tremblay Sound entrance (Figure 21). During this extensive survey, an ore carrier (M/V Nordic Olympic) transited from Ragged Island to Milne Port (Figure 10). Influences of Other Anthropogenic Activities.—Complicating the interpretation of the photographic data is the presence of small motorized boats and hunting activity in the study area. For instance, it seems quite likely that hunting activity influenced narwhal distribution and abundance during the photographic survey on 4 September. Increased hunting activity later on
that day coincided with reduced narwhal numbers. Narwhals have been observed to respond to hunting (more specifically shooting events) by swimming away, diving and increasing their swim speed (Thomas et al. 2014; Smith et al. 2015, 2016). It is important to account for small boats and hunting activity in analyses of narwhal behaviour. 5.2.3 Summary
The extensive and photographic aerial survey program and data analysis approach allowed for detection of statistically significant differences in narwhal relative abundance and distribution during periods with and without shipping—demonstrating that the study design has effectively met monitoring objectives. Results from both the extensive and photographic surveys indicate that narwhal numbers are reduced during periods with large vessel activity. It is uncertain how these statistically significant differences translate into biological significance for narwhals. However, there were no detectable changes in the spatial-temporal pattern of narwhal occurrence in their summering areas and no significant changes in their relative abundance from year-to- year. There are also indications that narwhal avoidance of large vessels may be temporary. We recognize that sample size was limited, particularly for the photographic surveys, and results were potentially influenced by factors like small boat traffic and hunting in the study area. In addition, there remain important questions about how far away from a vessel narwhals respond, the duration of avoidance, as well as whether narwhals habituate to repeated ship passages.
Page 77 Marine Mammal Aerial Surveys 2015
6 ACKNOWLEDGEMENTS We thank the following people and companies for their assistance in planning and implementing the aerial survey program. Robert Aglak, Cain Mucktar, Ikey Milton, Daniel Ootoova, Peter Oolateeta and Elijah Panipakoocho (from Pond Inlet) assisted us as observers during our surveys in 2015. LGL Limited observers were Robert (Ted) Elliott and Tannis Thomas. Unaalik Aviation (Kenn Borek Aviation) provided us with excellent contract staff, pilots and base managers who made the difficult job of completing a program around vagaries of the weather successful. Specifically, we thank: Joe Consaul, Brian Crocker and Wally Dobchuk (contract and logistics); the captains Jon Sipko, Monica Dauenhauer and Philip Amos; the first officers Reagan Schroeder and Kelsey Kushneryk; the Iqaluit base managers Bill Mcdonald and Joan Griffin; the Iqaluit base engineer Creig Statz. Rita Webb (manager) and Cyndi Morton (assistant manager) of the Sauniq Hotel in Pond Inlet assisted us during our stay in Pond Inlet including providing room/accommodation, troubleshooting problems with the rooms when they arose, providing us contact information for people in the community and especially for driving us back and forth to the airport with all our equipment. The board members of the Mittimatalik Hunters and Trappers Organization (HTO) made sure our project plan was reviewed and they provided us with a letter of support for the field work. Jan Slavicek of J.S. Micro Products in Winnipeg fabricated the DSLR camera frame from a DFO design. Throughout this aerial survey program and during previous Baffinland work various DFO researchers have provided valuable information and insight about the marine mammals in the study area, these researchers included: Steve Ferguson, Blair Dunn, and Bernard LeBlanc. Oliver Curran, Linda Chepyha, Joe Krimmerdjuar and Joe Tigullaraq (all of Baffinland) have provided support for the project. We thank the members of Baffinland’s Marine Environment Working Group (MEWG) for
supporting this research and directing some of the ideas regarding the conduct of the survey. Bill Koski of LGL provided invaluable advice on the design and conduct of these surveys as well as reviewing early drafts of the report. Rolph Davis of LGL reviewed the report. Joseph Beland and Andrew Davis of LGL assisted with the photo review.
Page 78 Marine Mammal Aerial Surveys 2015
A. B.
C. D.
Aerial Survey Crew for the four extensive surveys: A. Survey Period 1 - Left to right: Kelsey Kushneryk (FO), Cain Mucktar (Pond Inlet), Daniel Ootoova (Pond Inlet) and Jon Sipko (Cpt). Missing from photo Ted Elliott and Tannis Thomas; B. Survey Period 2 - Left to right: Ted
Elliott, Reagan Schroeder (FO), Monica Dauenhauer (Cpt) and Tannis Thomas. Missing from photo Cain Mucktar (Pond Inlet) and Daniel Ootoova (Pond Inlet); C. Survey Period 3 – Kelsey Kushneryk (FO), Ikey Milton (Pond Inlet), Tannis Thomas, Ted Elliott, Elijah Panipakoocho (Pond Inlet) and Phil Amos (Cpt); and D. Survey Period 4 - From left to right: Kelsey Kushneryk (FO), Tannis Thomas, Robert Aglak (Pond Inlet), Peter Oolateeta (Pond Inlet), Ted Elliott and Phil Amos (Cpt).
Page 79 Marine Mammal Aerial Surveys 2015
7 LITERATURE CITED Arab, A., M.L. Wildhaber, C.K. Wikle and C.N. Gentry. 2008. Zero-inflated modeling of fish catch per unit area resulting from multiple gears: application to channel catfish and shovelnose sturgeon in the Missouri River. N. Am. J. Fish. Manage. 28: 1044–1058. Asselin, N.C. and P.R. Richard. 2011. Results of narwhal (Monodon monoceros) aerial surveys in Admiralty Inlet, August 2010. DFO Can. Sci. Advis. Sec. Res. Doc. 2011/065. iv + 26 p. Baffinland (Iron Mines Corporation). 2012. Appendix 8A-2 in Mary River Project Final Environmental Impact Statement. Mary River Project Environmental Field Studies Marine Mammal Baseline 2007−2010. Prepared by LGL Limited, King City, On and North/South Consultants Inc., Winnipeg, MB for Baffinland Iron Mines Corporation, Toronto, ON. 168 p. + appendices. Boveng, P.L., J.L. Bengtson, D.E. Withrow, J.C. Cesarone, M. A. Simpkins, K. J. Frost and J. J. Burns. 2003. The abundance of harbor seals in the Gulf of Alaska. Mar. Mamm. Sci. 19: 111–127. Buckland, S.T., D.R. Anderson, K.P. Burnham, J.L. Laake, D.L. Borchers and L. Thomas. 2001. Introduction to Distance Sampling: Estimating Abundance of Biological Populations. Oxford University Press, New York, NY. 432 p. Campbell, R.R., D.B. Yurick and N.B. Snow. 1988. Predation on narwhals, Monodon monoceros, by killer whales, Orcinus orca, in the eastern Canadian arctic. Canadian Field-Naturalist 102(4):689–696. Caughley, G. 1977. Sampling in aerial survey. J. Wildl. Manage. 41(4): 605−615. Cosens, S.E. and L.P. Dueck. 1986. Responses of migrating narwhal and beluga to icebreaker traffic at the Admiralty Inlet ice-edge, N.W.T. in 1986. pp. 39–54 In W.M. Sackinger and M.O. Jeffries (eds.) Port and ocean engineering under Arctic conditions, Vol. 2. University of Alaska Fairbanks, Fairbanks, AK. COSEWIC (Committee on the Status of Endangered Wildlife in Canada). 2004. COSEWIC Assessment and Update Status Report on the Narwhal Monodon monoceros in Canada. Committee on the Status of Endangered Wildlife in Canada, Ottawa, Ont. vii + 50 p. COSEWIC (Committee on the Status of Endangered Wildlife in Canada). 2009. COSEWIC Assessment and Update Status Report on the Bowhead Whale Balaena mysticetus, Bering-Chukchi-Beaufort Population and Eastern Canada-West Greenland Population, in Canada. Committee on the Status
of Endangered Wildlife in Canada, Ottawa, Ont. vii + 49 p. Davis, R.A. and W.R. Koski. 1980. Recent observations of the bowhead whale in the eastern Canadian High Arctic. Rep. Int. Whal. Comm. 30:439−444. Davis, R.A., W.R. Koski, W.J. Richardson, C.R. Evans and W.G. Alliston. 1982. Distribution, numbers and productivity of the Western Arctic stock of bowhead whales in the eastern Beaufort Sea and Amundsen Gulf, summer 1981. Rep. from LGL Ltd., Toronto, Ont., for Sohio Alaska Petrol. Co. [now BP Alaska], Anchorage, AK, and Dome Petrol. Ltd., Calgary, Alb. (co-managers). 134 p. Summarized as SC/34/PS20, Int. Whal. Comm., Cambridge, UK, 1982, 13 p. DFO (Fisheries and Oceans Canada). 2012. Advice on Total Allowable Landed Catch for the Baffin Bay Narwhal Population. DFO Can. Sci. Advis. Sec. Sci. Advis. Rep. 2012/021. 11 p. DFO. 2014. High Arctic Cetacean Survey: Post-survey Summary. DFO, Winnipeg. 9 p.
Page 80 Marine Mammal Aerial Surveys 2015
Dietz, R., M.P. Heide-Jørgensen, P.R. Richard and M. Acquarone. 2001. Summer and fall movements of narwhals (Monodon monoceros) from northeastern Baffin Island towards northern Davis Strait. Arctic 54:244−261. Doniol-Valcroze, T., J.F. Gosselin, D. Pike, J. Lawson, N. Asselin, K. Hedges and S. Ferguson. 2015a. Abundance estimates of narwhal stocks in the Canadian High Arctic in 2013. DFO Can. Sci. Advis. Sec. Res. Doc. 2015/060. v + 36 p. Doniol-Valcroze, T., J.F. Gosselin, D. Pike, J. Lawson, N. Asselin, K. Hedges and S. Ferguson. 2015b. Abundance estimate of the Eastern Canada – West Greenland bowhead whale population based on the 2013 High Arctic Cetacean Survey. DFO Can. Sci. Advis. Sec. Res. Doc. 2015/058. v + 27 p. Dunn, M.R. 2009. Review and stock assessment of black cardinalfish (Epigonus telescopes) on the east coast North Island, New Zealand. New Zealand Fisheries Assessment Report 2009/39, 55 p. Eberhardt, L.L. 1978. Transect methods for population studies. J. Wildl. Manage. 42(1): 1−31. Elliott, R.E., V.D. Moulton, S.W. Raborn and R.A. Davis. 2013. Hudson Strait Marine Mammal Surveys, 10 March – 2 April 2012. LGL Report No. TA8129-2. Prepared by LGL Limited, King City, ON for Baffinland Iron Mines Corporation, Toronto, ON. 87 p. Elliott, R.E., S. Raborn, H.R. Smith, and V.D. Moulton. 2015. Marine mammal aerial surveys in Eclipse Sound, Milne Inlet, Navy Board Inlet, and Pond Inlet, 31 August – 18 October 2013. Final LGL Report No. TA8357-3. Prepared by LGL Limited, King City, ON for Baffinland Iron Mines Corporation, Oakville, ON. 61 p. Fallis, B., W.E. Klenner and J.B. Kemper. 1983. Narwhal surveys and associated marine mammal observations in Admiralty Inlet, Navy Board Inlet and Eclipse Sounds, Baffin Island, N.W.T. during 1974–76. Can. Tech. Rep. Fish. Aquat. Sci. 1211:1−20.
Finley, K. J. and E.J. Gibb. 1982. Summer diet of the narwhal (Monodon monoceros) in Pond Inlet, northern Baffin Island. Can. J. Zool. 60: 3353−3363. Finley, K.J., and G.W. Miller. 1982. The 1979 hunt for narwhals (Monodon monoceros) and an examination of harpoon gun technology near Pond Inlet, northern Baffin Island. Report SC/33/SMIO to the International Whaling Commission. 12 p. Finley, K.J. 1990. Isabella Bay, Baffin Island: An important historical and present-day concentration area for the endangered bowhead whale (Balaena mysticetus) of the eastern Canadian Arctic.
Arctic 43:137−152. Finley, K.J., G.W. Miller, R.A. Davis and C.R. Greene. 1990. Reactions of belugas, Delphinapterus leucas, and narwhals, Monodon monoceros, to ice-breaking ships in the Canadian high arctic. Can. B. Fish. Aquat. Sci. 224:97–117. Finley, K.J. 2001. Natural history and conservation of the Greenland whale, or bowhead, in the Northwest Atlantic. Arctic 54:55−76. GC (Government of Canada). 2016. Species at Risk Public Registry: A to Z Species Index. Available at: http://www.sararegistry.gc.ca/sar/index/default_e.cfm. Accessed in March 2016. Heide-Jørgensen, M.P., R. Dietz, K.L. Laidre and P. Richard. 2002. Autumn movements, home ranges and winter density of narwhals (Monodon monoceros) tagged in Tremblay Sound, Baffin Island. Polar Biol. 25:331–341. Heide-Jørgensen, M.P., R. Dietz, K.L. Laidre, P. Richard, J. Orr and H.C. Schmidt. 2003. The migratory behaviour of narwhals (Monodon monoceros). Can. J. Zool. 81:1298–1305.
Page 81 Marine Mammal Aerial Surveys 2015
Heide-Jørgensen, M.P., K.L. Laidre, M.V. Jensen, L. Dueck and L.D. Postma. 2006. Dissolving stock discreteness with satellite tracking: bowhead whales in Baffin Bay. Mar. Mamm. Sci. 22:34−45. Heide-Jørgensen, M.P., K. Laidre, D. Borchers, F. Samarra and H. Stern. 2007. Increasing abundance of bowhead whales in West Greenland. Biology Lett. 3: 577−580. Heide-Jørgensen, M.P., K.L. Laidre, M.L. Burt, D.L. Borchers, T.A. Marques, R.G. Hansen, M. Rasmussen and S. Fossette. 2010. Abundance of narwhals (Monodon monoceros L.) on the hunting grounds in Greenland. Mammalogy 91(5): 1135–1151. Heide-Jørgensen, M.P., N.H. Nielsen, R.G. Hansen, H.C. Schmidt, S.B. Blackwell and O.A. Jørgensen. 2015. The predictable narwhal: satellite tracking shows behavioural similarities between isolated subpopulations. J. Zool. 297:54−65. IMO (International Marine Organization). 2014. AIS transponders. Accessed 14 November 2014, http://www.imo.org/OurWork/Safety/Navigation/Pages/AIS.aspx. IWC (International Whaling Commission). 2008. Report of the Scientific Committee. Annex F. Report of the sub-committee on bowhead, right and gray whales. J. Cetac. Res. Manage. 10 (Supplement):150-166. IWC. 2009. Report of the Scientific Committee. Annex F. Report of the sub-committee on bowhead, right and gray whales. J. Cetac. Res. Manage. 11 (Supplement):169−192. Kim, K.H. and A.C. Conrad. 2015. Acoustic monitoring near Koluktoo Bay, Milne Inlet, 2014. Greeneridge Rep. 511-1. Rep. from Greeneridge Sciences Inc. (Santa Barbara, CA) for Baffinland Iron Mines Corporation (Oakville, ON). 12 p. Kim, K.H. and A.C. Conrad. 2016. Acoustic monitoring near Koluktoo Bay, Milne Inlet, 2015: Methods and preliminary results. Greeneridge Rep. 522-1. Rep. from Greeneridge Sciences Inc. (Santa Barbara, CA) for Baffinland Iron Mines Corporation (Oakville, ON). 10 p.
Kingsley, M.C.S., H.J. Cleator and M.A. Ramsay. 1994. Summer distribution and movements of narwhals (Monodon monoceros) in Eclipse Sound and adjacent waters, north Baffin Island, N.W.T. Medd. om Grønl. Biosci. 39:163−164. Koski, W.R. and R.A. Davis. 1979. Distribution of marine mammals in northwest Baffin Bay and adjacent waters, May–October 1978. Rep. from LGL Ltd., Toronto, ON., for Petro-Canada Exploration Inc., Calgary, AB. 305 p. Koski, W.R. 1980. Distribution and migration of marine mammals in Baffin Bay and eastern Lancaster Sound, May–July 1979. Rep. from LGL Ltd., Toronto, ON., for Petro-Canada Exploration Inc., Calgary, AB. 317 p. Koski, W.R. and R.A. Davis. 1980. Studies of the late summer distribution and fall migration of marine mammals in NW Baffin Bay and E Lancaster Sound, 1979. Rep. from LGL Ltd., Toronto, ON., for Petro-Canada Exploration Inc., Calgary, AB. 214 p. Koski, W.R., R.A. Davis, G.W. Miller and D. E. Withrow. 1993. Reproduction. Pages 239-274, Chapter VII. In: J.J Burns, J.J. Montague and C.J. Cowles (eds.), The Bowhead Whale. Special Publication No. 2, The Society of Marine Mammalogy. Lawrence, KS. 787 p. Koski, W.R. and R.A. Davis. 1994. Distribution and numbers of narwhals (Monodon monoceros) in Baffin Bay and Davis Strait. Medd. om Grønl. Biosci. 39:15−40. Koski, W.R., M.P. Heide-Jørgensen and K.L. Laidre. 2006. Winter abundance of bowhead whales, Balaena mysticetus, in Hudson Strait, March 1981. J. Cetac. Res. Manage. 8:139−144.
Page 82 Marine Mammal Aerial Surveys 2015
Koski, W.R. and J.R. Brandon. 2012. A planning tool to estimate population size, confidence intervals and Potential Biological Removal estimates that would be obtained assuming different survey effort in sub-areas of beluga, narwhal and bowhead whale summering areas. Presentation at the DFO planning workshop, Winnipeg, Manitoba, 5-7 November 2012. Laidre, K.L., M.P. Heide-Jørgensen, M.L. Logsdon, R.C. Hobbs, P. Heagerty, R. Dietz, O.A. Jørgensen and M.A. Trebel. 2004. Seasonal narwhal habitat associations in the high Arctic. Mar. Biol. 145:821–831. Laidre, K.L. and M.P. Heide-Jørgensen. 2005. Winter feeding intensity of narwhals (Monodon monoceros). Mar. Mamm. Sci. 21(1): 45−57. LGL and Greeneridge. 1986. Reactions of beluga whales and narwhals to ship traffic and ice-breaking along ice edges in the eastern Canadian High Arctic: 1982-1984. Environ. Stud. 37. Indian & Northern Affairs Canada, Ottawa, ON. 301 p. Lin, D. Y., L. J. Wei and Z. Ying. 2002. Model-checking techniques based on cumulative residuals. Biometrics 58: 1−12. Lubbock, B. 1937. The arctic whalers. Brown, Son & Ferguson, Glasgow, Scotland. 483 p. Magnusson, W.E., G.J. Caughley and G.C. Grigg. 1978. A double-survey estimate of population size from incomplete counts. J. Wildl. Manage. 42(1):174−176. Mansfield, A.W., T.G. Smith and B. Beck. 1975. The narwhal, Monodon monoceros in eastern Canadian waters. J. Fish. Res. Bd. Can. 32(7): 1041−1046. Marcoux, M., M. Auger-Méthé and M.M. Humphries. 2009. Encounter frequencies and grouping patterns of narwhals in Koluktoo Bay, Baffin Island. Polar Biol. 32:1705-1716. McLaren, P.L. and R.A. Davis. 1982. Winter distribution of arctic marine mammals in ice-covered waters of eastern North America. Rep. from LGL Ltd., Toronto, ON., for Petro-Canada
Explorations Inc., Calgary, AB. 151 p. Minami, M., C.E. Lennert-Cody, W. Gao and M. Roman-Verdesoto. 2007. Modeling shark bycatch: the zero-inflated negative binomial regression model with smoothing. Fish. Res. 84: 210–221. Moore, S.E. and R.R. Reeves. 1993. Distribution and movement. p. 313-386 In J.J. Burns, J.J. Montague, and C.J. Cowles (eds.), The bowhead whale. Spec. Pub. No. 2. Soc. Mar. Mammal., Lawrence, KS. 787 p. NAMMCO 2010. Report on the Joint NAMMCO/JCNB Scientific Working Group – narwhal. In:
NAMMCO Annual Report 2009. NAMMCO. Tromsø, Norway. 291-296. Nerini, M.K., H.W. Braham, W.M. Marquette and D.J. Rugh. 1984. Life history of the bowhead whale, Balaena mysticetus (Mammalia: Cetacea). J. Zool. 204: 443–468. Priest, H. and P.J. Usher. 2004. The Nunavut Wildlife Harvest Study, August 2004. Iqaluit: Nunavut Wildlife Management Board. 862 p. Remnant, R.A. and M.L. Thomas. 1992. Inuit traditional knowledge of the distribution and biology of High Arctic narwhal and beluga. Unpublished report by North/South Consultants Inc. Winnipeg, MB. vii + 96 p. Reeves, R., E. Mitchell, A. Mansfield and M. McLaughlan. 1983. Distribution and migration of the bowhead whale, Balaena mysticetus, in the eastern North American Arctic. Arctic 36:5-64. Richard, P., P. Weaver, P. Dueck and L. Barber. 1994. Distribution and numbers of Canadian high arctic narwhals, (Monodon monoceros) in August 1984. Medd. om Grønl. Biosci. 39:41-50.
Page 83 Marine Mammal Aerial Surveys 2015
Richard, P.R. 2010. Stock definition of belugas and narwhals in Nunavut. DFO Can. Sci. Advis. Sec. Res. Doc. 2010/022. iv + 14 p. Richard, P.R., J.L. Laake, R.C. Hobbs, M.P. Heide-Jørgensen, N.C. Asselin and H. Cleator. 2010. Baffin Bay narwhal population distribution and numbers: Aerial surveys in the Canadian High Arctic, 2002-04. Arctic 63(1): 85–99. SAS Institute Inc. 2012. SAS Online Doc, Version 9.4. Cary, NC. SEM and LGL. 2016. Mary River Project, Marine Environmental Effects Monitoring Plan (Draft, Vers. 1.1). Prepared by Sikumiut Environmental Management Ltd., St. John’s, NL and LGL Limited, St. John’s, NL for Baffinland Iron Mines Corporation, Oakville, ON. 81 p. Shono, H. 2008. Application of the Tweedie distribution to zero-catch data in CPUE analysis. Fish. Res. 93: 154–162. Smith, H.R., J.R. Brandon, P. Abgrall, M. Fitzgerald, R.E. Elliott, and V.D. Moulton. 2015. Shore-based monitoring of narwhals and vessels at Bruce Head, Milne Inlet, 30 July – 8 September 2014. Final LGL Report No. FA0013-2. Prepared by LGL Limited, King City, ON for Baffinland Iron Mines Corporation, Oakville, ON. 73 p. + appendices. Smith, H.R., S. Raborn, M. Fitzgerald and V.D. Moulton. 2016. Shore-based monitoring of narwhals and vessels at Bruce Head, Milne Inlet, 29 July – 5 September 2015. LGL Report No. FA0044-1. Prepared by LGL Limited, King City, ON for Baffinland Iron Mines Corporation, Oakville, ON. 73 p. + appendices. Smith, T.G., M.O. Hammill, D.J. Burrage and G.A. Sleno. 1985. Distribution and abundance of belugas Delphinapterus leucas and narwhals Monodon monoceros in the Canadian high Arctic. Can. J. Fish. Aquat. Sci. 42:676-684. Terceiro, M. 2003. The statistical properties of recreational catch rate data for some fish stocks of the northeast U.S. coast. Fish. B-NOAA 101: 653–672.
Thomas, L., S.T. Buckland, E.A. Rexstad, J.L. Laake, S. Strindberg, S.L. Hedley, J.R.B. Bishop, T. A. Marques and K.P. Burnham. 2010. DISTANCE software: design and analysis of distance sampling surveys for estimating population size. J. Appl. Ecol. 47:5-14. DOI: 10.1111/j.1365- 2664.2009.01737.x Thomas, T., P. Abgrall, S.W. Raborn, H. Smith, R.E. Elliott, and V.D. Moulton. 2014. Narwhals and shipping: shore-based study at Bruce Head, Milne Inlet, August 2013. Final LGL Report No.
TA8286-2. Prepared by LGL Limited, King City, ON for Baffinland Iron Mines Corporation, Oakville, ON. 60 p. + appendices. Thomas, T.A., S. Raborn, R.E. Elliott and V.D. Moulton. 2015. Marine mammal aerial surveys in Eclipse Sound, Milne Inlet, Navy Board Inlet, and Pond Inlet, 1 August – 22 October 2014. Final LGL Report No. FA0024-2. Prepared by LGL Limited, King City, ON for Baffinland Iron Mines Corporation, Oakville, ON. 70 p. Ver Hoef, J.M. and P.L. Boveng. 2007. Quasi-Poisson vs. negative binomial regression: How should we model overdispersed count data? Ecology 88: 2766-2772. Vincenty, T. 1975. Direct and inverse solutions on the ellipsoid with application of nested equations. Survey Review 23(176):88-93. Watt, C.A., J. Orr, B. LeBlanc, R. Richard and S.H. Ferguson. 2012. Satellite tracking of narwhals (Monodon monoceros) from Admiralty Inlet (2009) and Eclipse Sound (2010- 2011). DFO Can. Sci. Advis. Sec. Res. Doc. 2012/046. iii + 17 p.
Page 84 Marine Mammal Aerial Surveys 2015
Watt, C.A., M. Marcoux, N.C. Asselin, J.R. Orr and S.H. Ferguson. 2015. Instantaneous availability bias correction for calculating aerial survey abundance estimates for narwhal (Monodon monoceros) in the Canadian High Arctic. DFO Can. Sci. Advis. Sec. Res. Doc. 2015/044. v + 13 p.
Page 85 Marine Mammal Aerial Surveys 2015
APPENDIX A: MAPPING NARWHAL SIGHTINGS FROM PHOTOGRAPHS Camera Calibration The Camera Calibration Toolbox (http://www.vision.caltech.edu/bouguetj/calib_doc/) for Matlab (Windows 64bit, R2015b) was used to determine the radial and tangential distortion for the Nikon lens at a lens length of 35 mm. This toolbox was written for the Matlab software; we used a trial copy of the software to run the toolbox. We photographed 80 images of a checkerboard pattern 40x40 inches square that consisted of black and white checkers 20x20 mm (see Figure A-1). The toolkit is used to automatically extract the corners between adjacent black and white squares. The camera was kept level and we adjusted the angle of the view of the target in 20 increments (3.5 degrees) between 23.5 and 90 degrees orientation to the lens axis. We used a digital inclinometer to position the target as close to these increments as possible. The top of the target was kept level using a carpenter’s level, the angle along the downslope was varied. We used a tripod video head that allowed us to rotate the camera 90 degrees so that in one image the horizontal axis of the image was parallel to the top of the target, in the camera image the checkerboard squares got smaller towards the top of the image. We then rotated the camera so the long axis of the image was now vertical and parallel to the vertical axis of the target, the checkerboard squares in the image got smaller towards the right side of the image. We then mounted the video head and camera upside down by attaching it to the base of the tripod`s vertical bar. We repeated the photographs taken above. With the camera upside down the horizontal axis of the image was parallel to the top or bottom of the target, this caused the checkerboard squares in the images to get smaller towards the bottom of the image. Rotating the upside down camera so the long axis of the image was now vertical and parallel to the vertical axis of the target, the checkerboard squares in the images got smaller towards the left side of the image. An initial calibration is done using the toolbox; these results are then used to refine the calibration such that more precise distortion parameters are estimated.
The camera lens (Nikon D810 with Nikon AF-S Nikkor 24-70 mm 1:2.8G ED lens) has a slight barrel distortion (similar but by no means as severe as a fish eye lens effect, evidenced by a small negative second order radial distortion coefficient). The calculation of the distortion parameters allows us to recalculate the sighting pixel coordinates to remove the effect of distortion in the camera lens. This makes the process of calculating the vertical angle and azimuth to each narwhal sighting much simpler because then we only have to deal with the simpler geometry of the camera platform to the surface of the earth. We measured the field of view for the camera using the 24-70 mm lens set at a lens length of 35 mm. The lens length is the distance from the image plane (indicated on the camera body with a ɸ) to the principal optical point (where the image crosses and inverts itself within the lens elements of the camera). Two tape measures were taped horizontally and vertically on the wall
Page 86 Marine Mammal Aerial Surveys 2015 so they completely span the camera image. The distance was measured from the principal optical point to the center of the camera image. One photograph was taken. We viewed the image to determine the distance on the two tape measures vertically and horizontally. Simple geometry was then used to calculate the horizontal and vertical field of views for the lens. The horizontal and vertical field of views are 51.09011 and 35.53439 degrees, respectively. Note that these are smaller angles than are given on the Nikon website (54.4 and 37.8 for the horizontal and vertical FOV respectively).
Figure A-1. Checkerboard pattern and camera set-up used for camera calibration.
Mapping Image Pixels We used the radial and tangential distortion parameters coupled with the field of view of the camera and the camera pointing angle to determine the actual, horizontal and vertical angles to each pixel in any camera image. These angles are then used with aircraft altitude, interpolated position and heading of the aircraft at the time the photos were taken to determine the geographic location of each sighting (image pixel) that was extracted as a x-y pixel coordinate from the images by the photo reviewers. In future, it would be beneficial to incorporate an estimate of the aircraft banking angle to better characterize where the camera was pointed instead of assuming the aircraft was flying exactly level nose to tail and wing tip to wing tip.
Page 87 Marine Mammal Aerial Surveys 2015
NMML formulae (http://www.afsc.noaa.gov/nmml/software/excelgeo.php) are used to calculate the distance to the pixel based on the aircraft altitude and angle of the pixels from horizontal (or vertical). The geographic location at each pixel is estimated using Vincenty (1975) formulae (http://www.ngs.noaa.gov/TOOLS/Inv_Fwd/Inv_Fwd.html) for forward calculation of new latitude and longitude given the aircraft latitude longitude, azimuth to the sighting and distance to the sighting. The image size for the Nikon D810 cameras is 7360 by 4912 pixels, for the x axis and y- axis respectively. The distortion coefficients are based on the upper left pixel being at 0,0, and the lower right pixel being at 7359, 4911. Because the image is based on the window (0,0 to 7360,4912), we must account for the half pixel difference that defines the pixel center locations (0.5,0.5 to 7359.5 4911.5).
Page 88 Marine Mammal Aerial Surveys 2015
APPENDIX B: PERCEPTION BIAS CALCULATIONS Based on the method of Magnusson et al. (1978), the estimated number of singles or groups present on the trackline was
(S +B+1)(S +B+1) N̂ = 1 2 -1 (B+1)
Again following Magnusson et al. (1978), the probability that a single observer or for our purpose a single reviewer would detect a singleton or group of marine mammals that is within the trackline is as follows:
(S1+S2+2B) gd(0)= 2N̂ Thus, if a photo reviewer reviews an image from one side of the aircraft, the uncorrected number of marine mammals seen by that reviewer in the trackline should be divided by gd(0) to allow for animals present at the surface but not detected.
The estimated variances of N̂ and gd(0) are required when calculating the confidence limits for the estimated numbers of marine mammals present. Magnusson et al. (1978) provided a formula for var(N̂ ), but this did not allow for the fact that S1, S2, and B are random variables, not constants (D.G. Chapman, pers. comm., 1982). The variances can be approximated using the delta method (Taylor series expansion) as shown below. Raw counts of marine mammals by a single reviewer on one photograph can be corrected for animals present but not detected by dividing by gd(0) ± s.e.( gd(0)).
These gd(0) correction factors are independent of any correction for submerged marine mammals. The ƒ(0) factors, described earlier, account for the reduced probability of detecting a marine mammals as its distance from the trackline increases. The variance of the estimated N is approximated using the delta method (Taylor series expansion) as follows:
∂N̂ 2 ∂N̂ 2 ∂N̂ 2 var(N̂ )≅ var(S )+ var(S )+ var(B) ∂ 1 ∂ 2 ∂ ( S) ( S2 ) ( B)
where the partial derivatives are evaluated at the mean values approximated by S1, S2 and B. Now,
Page 89 Marine Mammal Aerial Surveys 2015
∂N̂ (S +B+1) = 2 ∂ S1 (B+1)
∂N̂ (S +B+1) = 1 ∂ S2 (B+1)
∂N̂ (B+1)[(S +B+1)(S +B+1)]-(S +B+1)(S +B+1) = 1 2 1 2 2 ∂B (B+1)
Assuming S1, S2 and B have approximate Poisson distributions, their variances can be assumed equal to the observed values. Thus,
s.e.(N̂)=√N̂
The variance of the detection probability gd(0) must also be calculated using the delta method because gd(0) is a function of four random variables, S1, S2, B and N. Given that
(S1+S2+2B) gd(0)= 2N̂
Then we have
∂ĝ (0) 2 ∂ĝ (0) 2 ∂ĝ (0) 2 ∂ĝ (0) 2 var(ĝ (0))≅ d var(S )+ d var(S )+ d var(B)+ d var(N̂ ) d ∂ 1 ∂ 2 ∂Β ∂N̂ ( S1 ) ( S2 ) ( ) ( ) again evaluating the partial derivatives at the estimated mean values. Thus,
∂ĝ (0) 1 d = ∂ S1 2N̂
∂ĝ (0) 1 d = ∂ S2 2N̂
Page 90 Marine Mammal Aerial Surveys 2015
∂ĝ (0) 1 d = ∂Β N̂
∂ĝ (0) S +S +2B d = 1 2 2 ∂Ν 2(N̂ )
Assuming that S1, S2 and B have approximate Poisson distributions, and using the previous estimate of var(N̂ ) and the mean values we calculate the var(gd̂ (0)), the corresponding standard error is √var(gd̂ (0)).
Page 91 Marine Mammal Aerial Surveys 2015
APPENDIX C: GNLMM MODEL FIT
FIGURE C-1. Observed (circles) and predicted (red line) number of narwhals for each sample plotted against the sorted rank of predicted values for the extensive survey model. Samples with observed values=0 are excluded. Thin dashed lines represent 90% confidence limits.
FIGURE C-2. Moving sum of residuals plotted against the linear predictor for the number of narwhals (window size=9) for the extensive survey model. Observed values are represented by the red line; blue lines are 20 randomized realizations around the fitted line based on the observed residuals as per Lin et al. (2002). The p-value is from the Kolmogorof-type supremum test and represents the proportion of 1,000 realizations with an absolute residual greater than the largest observed absolute residual.
Page 92 Marine Mammal Aerial Surveys 2015
Figure C-3. Observed (circles) and predicted (red line) number of narwhals for each sample plotted against the sorted rank of predicted values for the photographic survey model. Samples with observed values=0 are excluded. Thin dashed lines represent 90% confidence limits.
FIGURE C-4. Moving sum of residuals plotted against the linear predictor for the number of narwhals (window size=15) for the photographic survey model. Observed values are represented by the red line; blue lines are 20 randomized realizations around the fitted line based on the observed residuals as per Lin et al. (2002). The p-value is from the Kolmogorof-type supremum test and represents the proportion of 1,000 realizations with an absolute residual greater than the largest observed absolute residual.
Page 93 Marine Mammal Aerial Surveys 2015
APPENDIX D: ADDITIONAL TABLES
TABLE D-1. Average number of large vessels present in geographic strata per day for each year (first of August-middle of September) based on extensive survey data modelled with the GNLMM. Parenthetical values indicate range and shading across means is proportionate to magnitude.
Geographic stratum 2013 2014 2015 Eclipse Sound East 1.7 1.4 3.3 (1-4) (0-4) (1-5) Eclipse Sound West 1.1 0.8 2.8 (0-3) (0-3) (1-4) Milne Inlet/Koluktoo Bay 0.4 0.1 3.0 (0-1) (0-1) (0-5) Tremblay Sound 0.2 0.0 0.0 (0-3) (0-0) (0-0)
TABLE D-2. Narwhal orientation and group size calculated for each photographic survey relative to shipping activity.
Shipping Ship Mean Average Heading No. Te st Date Area P-value Activity Direction Group Size Heading (°T) SD (°T) Sightings Statistic 18-Aug Milne Inlet Before North 1.3 229 64 21 5.45 0.003 Mine Inlet During North 4.0 259 31 4 2.98 0.039 22-Aug Tremblay Sound During North 1.5 189 85 205 22.38 <0.001 Tremblay Sound After North 1.5 28 45 284 153.86 <0.001 30-Aug Mine Inlet During South 1.2 345 55 10 3.94 0.015 Mine Inlet After South 1.7 32 31 310 229.87 <0.001
Tremblay Sound Before South 2.5 179 96 26 1.59 0.204 Tremblay Sound During South 2.0 213 92 761 58.33 <0.001 Tremblay Sound After South 1.9 46 35 344 238.02 <0.001 4-Sep Milne Inlet Before North 1.5 309 63 82 24.04 <0.001 Milne Inlet During North/South 2.0 216 71 605 132.16 <0.001 Milne Inlet After South 3.9 235 101 20 0.87 0.422
Page 94