Canadian Technical Report of Fisheries and Aquatic Sciences 3103 2014

Canadian Technical Report of Fisheries and Aquatic Sciences 3103 2014
Canadian Technical Report of
Fisheries and Aquatic Sciences 3103
2014
BIVALVE AQUACULTURE AND EELGRASS (ZOSTERA MARINA) COVERAGE ON A BAY-WIDE
SCALE UTILIZING BATHYMETRIC LIDAR AND AERIAL PHOTOGRAPHY
by
M. Niles, A. Locke, T. Landry, T. Webster1, K. Collins1, G. Robichaud, A. Hanson2,
M. Mahoney2, H. Vandermeulen3, S. Doiron4 and M.-J. Maillet5
Fisheries and Oceans Canada
Science Branch, Gulf Region
Moncton, NB
E1C 9B6
1
Applied Geomatics Research Group, Centre of Geographic Science, Nova Scotia Community College,
Middleton, NS B0S 1M0
2
Environment Canada, Canadian Wildlife Service (EC-CWS), Sackville, NB E4L 1G6
3
Fisheries and Oceans Canada, Science Branch, Maritimes Region, St Andrews, NB E5B 1W8
4
Fisheries, Agriculture and Aquaculture, Shippagan, NB E8S 1H9
5
Fisheries, Agriculture and Aquaculture, Bouctouche, NB E4S 2T2
© Her Majesty the Queen in Right of Canada, 2014.
Cat. No. Fs 97-6/3103E-PDF
ISBN 978-1-100-24901-8 ISSN 1488-5375 (online version)
Correct citation for this publication:
Niles, M., Locke, A., Landry, T., Webster, T., Collins, K., Robichaud, G., Hanson, A., Mahoney, M.,
Vandermeulen, H., Doiron, S. and Maillet, M.-J. 2014. Bivalve aquaculture and eelgrass coverage on
a bay-wide scale utilizing bathymetric lidar and aerial photography. Can. Tech. Rep. Fish. Aquat. Sci.
3103:xiv + 82.p.
TABLE OF CONTENTS
LIST OF FIGURES ............................................................................................................................................ vi
LIST OF TABLES ............................................................................................................................................... x
ABSTRACT .......................................................................................................................................................xii
RÉSUMÉ ...........................................................................................................................................................xiii
1.
2.
INTRODUCTION ........................................................................................................................................ 1
1.1
EELGRASS ........................................................................................................................................... 1
1.2
AQUACULTURE ................................................................................................................................. 1
1.3
RELATIONSHIPS BETWEEN EELGRASS AND AQUACULTURE .............................................. 2
1.4
STUDY OBJECTIVES ......................................................................................................................... 2
METHODS ................................................................................................................................................... 3
2.1
AREA OVERVIEW .............................................................................................................................. 3
2.2
BATHYMETRY.................................................................................................................................... 3
2.2.1
Bathymetric Lidar........................................................................................................................... 3
2.2.2
DEM Processing and Analysis ....................................................................................................... 7
2.2.3
Supplementary Bathymetry............................................................................................................. 8
2.2.4
DEM Depth Validation ................................................................................................................... 9
2.2.5
Orthophoto Mosaics ....................................................................................................................... 9
2.3
EELGRASS ......................................................................................................................................... 10
2.3.1
FPI Eelgrass From Orthophoto Frames ...................................................................................... 10
2.3.2
Environment Canada and DFO ................................................................................................... 10
2.3.3
Eelgrass Coverage Per Bay ......................................................................................................... 11
2.3.4
Eelgrass Coverage at 1 m Depth Intervals .................................................................................. 11
2.4
2.4.1
AQUACULTURE ............................................................................................................................... 11
Aquaculture: Biomass Calculations – Boat Surveys........................................................................ 12
iv
Aquaculture: Biomass and Depth Calculations – Orthophoto Mosaics .......................................... 13
2.4.2
3.
2.4.2.1
Oyster Collector Lines .............................................................................................................. 13
2.4.2.2
Floating Oyster Bags (Single and Double) .............................................................................. 14
2.4.2.3
Sub-Surface Oyster Cages ........................................................................................................ 16
2.4.2.4
Oyster String Cages .................................................................................................................. 16
2.4.2.5
Oystergro Cages ....................................................................................................................... 18
2.4.2.6
Dark Sea Oyster Cages ............................................................................................................. 19
2.4.2.7
Mussel Longlines ...................................................................................................................... 20
RESULTS ................................................................................................................................................... 22
3.1
BATHYMETRY.................................................................................................................................. 23
3.1.1
Surface Area of DEM Depth Intervals ......................................................................................... 23
3.1.2
Total Area Comparisons .............................................................................................................. 23
3.1.3
Bathymetric Intervals, DEMs and CSRs ...................................................................................... 25
3.1.4
DEM Comparison/Validation....................................................................................................... 34
3.2
EELGRASS ......................................................................................................................................... 34
3.2.1 Total Eelgrass Area .......................................................................................................................... 38
3.2.2 Overlaps with No Data ...................................................................................................................... 38
3.3.3 Eelgrass Coverage per Bay ............................................................................................................... 39
3.3
4.
AQUACULTURE GEAR AND BIOMASS – BOAT-BASED SURVEYS ...................................... 48
3.3.1
Aquaculture Gear and Biomass – Orthophoto Mosaics .............................................................. 50
3.3.2
Biomass – Comparison Between Boat-based Surveys and Orthophoto Mosaics ........................ 53
DISCUSSION ............................................................................................................................................ 68
4.1
BATHYMETRY.................................................................................................................................. 68
4.2
EELGRASS ......................................................................................................................................... 69
4.3
AQUACULTURE ............................................................................................................................... 69
v
5.
CONCLUSIONS ........................................................................................................................................ 70
5.1
BATHYMETRY.................................................................................................................................. 70
5.2
EELGRASS ......................................................................................................................................... 70
5.3
AQUACULTURE ............................................................................................................................... 71
5.4
RECOMMENDATIONS ..................................................................................................................... 71
6.
ACKNOWLEDGEMENTS ....................................................................................................................... 72
7.
REFERENCES ........................................................................................................................................... 72
APPENDIX A: EELGRASS METHODOLOGY .............................................................................................. 76
1.
DFO- EELGRASS FIELD SURVEY CRITERIA .................................................................................. 76
2.
DFO - EELGRASS TOWFISH SURVEY METHODOLOGY (H. VANDERMEULEN) .................... 76
3. ENVIRONMENT CANADA- CANADIAN WILDLIFE SERVICE (EC-CWS) (MATT MAHONEY)
CLASSIFICATION........................................................................................................................................ 77
3.1 EC METADATA ................................................................................................................................ 77
APPENDIX B: ADDITIONAL INFORMATION ON LIDAR TECHNOLOGY ............................................ 81
APPENDIX C: LIST OF ABBREVIATIONS ................................................................................................... 82
vi
LIST OF FIGURES
Figure 1: Study area on the Eastern shore of New Brunswick and Eastern tip of Prince Edward Island.
Background map is a shaded relief terrain model with red polygons denoting bays where bathymetric lidar
and orthophotos were acquired. Green polygons denote bays for which a bathymetric lidar survey was
planned but not completed. .................................................................................................................................. 5
Figure 2: Airborne bathymetric lidar uses green and N1R laser. The green laser penetrates the water column to
two Secchi depths and measures the timing and intensity of the returned laser pulse. Source:
http://optech.ca/pdf/Brochures/SHOALS2007.pdf. ............................................................................................. 6
Figure 3: Orthophoto mosaic in St Mary’s Bay, Prince Edward Island showing poor aerial photo quality. The
southern edges of the individual orthophotos are affected by glint; dark patches near the center of the image
are likely cloud shadow. ..................................................................................................................................... 10
Figure 4: Collector lines in Bouctouche Bay, New Brunswick. ........................................................................ 14
Figure 5: Collector lines in Bouctouche Bay, NB seen in the orthophoto mosaic from September, 2011. ....... 14
Figure 6: Orange lines are interpreted and digitized collector lines in Bouctouche Bay, NB. .......................... 14
Figure 7: Floating oyster bags with white buoys; more are seen in the distance. .............................................. 15
Figure 8: A double-wide line of floating oyster bags. ........................................................................................ 15
Figure 9: Single-wide floating bags in Tracadie Bay, NB in the FPI orthophoto mosaic from Sept., 2011. ..... 15
Figure 10: Double-wide floating bags in Richibucto Bay, NB in the FPI orthophoto mosaic from Sept., 2011.
............................................................................................................................................................................ 15
Figure 11: Dark lines are sub-surface oyster cages in the Tracadie Bay, NB orthophoto mosaic from Sept.,
2011. ................................................................................................................................................................... 16
Figure 12: Orange lines are interpreted and digitized sub-surface oyster cages in the Tracadie Bay, NB
orthophoto mosaic from Sept., 2011. ................................................................................................................. 16
Figure 13: Oyster string cages. ........................................................................................................................... 17
Figure 14: The white end buoys of the oyster string cages are barely visible in the left orthophoto; the lines
digitized between the buoys in the image on the right represent the oyster string cages in the Caraquet Bay,
NB orthophoto mosaic from Sept., 2011. ........................................................................................................... 17
Figure 15: Oystergro cages in Bouctouche Bay, NB. ....................................................................................... 18
Figure 16: Upside-down Oystergro cages. ......................................................................................................... 18
vii
Figure 17: Oystergro cages in Tracadie Bay, NB with 10-12 cages per line spaced 3 m apart; average line
length is 43 m. Image is FPI orthophoto mosaic from Sept., 2011. ................................................................... 18
Figure 18: Oystergro cages in Caraquet Bay, NB with 6 cages visible per line spaced 6 m apart; average line
length is 30 m. .................................................................................................................................................... 19
Figure 19: Digitized Oystergro cages seen in the Caraquet Bay, NB orthophoto mosaic from Sept., 2011. .... 19
Figure 20: Dark sea oyster cages. ....................................................................................................................... 19
Figure 21: Typical design of a Prince Edward Island longline mussel farming setup from Drapeau et al. 2006.
............................................................................................................................................................................ 20
Figure 22: Mussel longlines in St Mary’s Bay, PEI. The red box shows submerged lines with white end buoys
visible; the yellow box shows floating longlines. .............................................................................................. 21
Figure 23: Left: Mussel longline buoys in St Mary’s Bay, PEI, seen in orthophotos from FPI, Sept. 2011.
Right: Digitized mussel longlines. ..................................................................................................................... 21
Figure 24: The extent of the Digital Elevation Model (DEM) shows the region, referred to as Richibucto that
was collected by FPI. For biomass and eelgrass calculations, Richibucto has been divided into three smaller
regions: Aldouane, Bedec, and Richibucto. The polygonal outlines were used to actually subdivide
Richibucto, as the FPI coastline outlines are smaller than the DEM. ................................................................ 22
Figure 25: Caraquet DEM and lidar-derived 1 m depth intervals. All depths are CGVD28. ............................ 26
Figure 26: Caraquet 1 m depth intervals on elevation colour-shaded relief (CSR) image. ............................... 26
Figure 27: Miscou DEM and lidar-derived 1 m depth intervals. All depths are CGVD28. ............................... 27
Figure 28: Miscou 1 m depth intervals on elevation colour-shaded relief (CSR) image. .................................. 27
Figure 29: Tracadie DEM and lidar-derived 1 m depth intervals. All depths are CGVD28.............................. 28
Figure 30: Tracadie 1 m depth intervals on elevation colour-shaded relief (CSR) image. ................................ 28
Figure 31: Richibucto DEM and lidar-derived 1 m depth intervals. All depths are CGVD28. ......................... 29
Figure 32: Richibucto 1 m depth intervals on elevation colour-shaded relief (CSR) image. ............................ 29
Figure 33: Aldouane DEM and lidar-derived 1 m depth intervals. All depths are CGVD28. ........................... 30
Figure 34: Aldouane 1 m depth intervals on elevation colour-shaded relief (CSR) image. .............................. 30
Figure 35: Bedec DEM and lidar-derived 1 m depth intervals. All depths are CGVD28. ................................. 31
Figure 36: Bedec 1 m depth intervals on elevation colour-shaded relief (CSR) image. .................................... 31
viii
Figure 37: Bouctouche DEM and lidar-derived 1 m depth intervals. All depths are CGVD28. ....................... 32
Figure 38: Bouctouche 1 m depth intervals on elevation colour-shaded relief (CSR) image. ........................... 32
Figure 39: St Mary’s DEM and lidar-derived 1 m depth intervals. All depths are CGVD28. ........................... 33
Figure 40: St Mary’s Bay 1 m depth intervals on elevation colour-shaded relief (CSR) image. ....................... 33
Figure 41. Caraquet FPI lidar bottom reflectance and eelgrass boundary. Arrows point to areas in the DEM
with No Data, areas where the eelgrass polygon overlaps the DEM, and where the eelgrass polygon overlaps
an area of No Data in the DEM. ......................................................................................................................... 39
Figure 42: Map showing the division of Shippagan Bay and Tracadie Bay, New Brunswick into smaller
regions (tributaries). ........................................................................................................................................... 40
Figure 43: Miscou FPI eelgrass boundary and EC/DFO eelgrass polygons. Background image is elevation
colour-shaded relief (CSR). ................................................................................................................................ 42
Figure 44: Miscou FPI lidar bottom reflectance and eelgrass boundary. ........................................................... 42
Figure 45: Tracadie FPI lidar bottom reflectance and eelgrass boundary. ...................................................... 43
Figure 46: Tracadie FPI eelgrass boundary and EC/DFO eelgrass polygons. Background image is elevation
colour-shaded relief (CSR). ................................................................................................................................ 43
Figure 47: Richibucto FPI lidar bottom reflectance and eelgrass boundary. ..................................................... 44
Figure 48: Richibucto FPI eelgrass boundary and EC/DFO eelgrass polygons. Background image is elevation
colour-shaded relief (CSR). ................................................................................................................................ 44
Figure 49: Bouctouche FPI lidar bottom reflectance and eelgrass boundary..................................................... 45
Figure 50: Bouctouche FPI eelgrass boundary and EC/DFO eelgrass polygons. Background image is elevation
colour-shaded relief (CSR). ................................................................................................................................ 45
Figure 51: St Mary’s Bay FPI lidar bottom reflectance and eelgrass boundary. ............................................... 46
Figure 52: Shippagan Bay EC/DFO eelgrass polygons. .................................................................................... 46
Figure 53: Neguac Bay EC/DFO eelgrass polygons. ......................................................................................... 47
Figure 54: Cocagne Harbour EC/DFO eelgrass polygons. ................................................................................ 47
Figure 55: Tabusintac Bay EC/DFO eelgrass polygons. .................................................................................... 48
Figure 56: Biomass of oysters (or mussels for Prince Edward Island) per bay area (tons per hectare) for 20072009 (left) and 2011 (right). ............................................................................................................................... 50
ix
Figure 57: Number of oyster bags (or mussel socks for Prince Edward Island) per lease area (number per
hectare) for 2007-2009 (left) and 2011 (right). .................................................................................................. 50
Figure 58: Caraquet orthophoto mosaic (FPI, Sept. 2011), aquaculture leases and interpreted aquaculture. .... 54
Figure 59: Tracadie orthophoto mosaic (FPI, Sept. 2011), aquaculture leases and interpreted aquaculture. .... 55
Figure 60: Richibucto, Aldouane, and Bedec orthophoto mosaic (FPI, Sept. 2011), aquaculture leases and
interpreted aquaculture. ...................................................................................................................................... 56
Figure 61: Bouctouche orthophoto mosaic (FPI, Sept. 2011), aquaculture leases and interpreted aquaculture.
............................................................................................................................................................................ 57
Figure 62: St Mary’s Bay, PEI orthophoto mosaic (FPI, Sept. 2011) and interpreted aquaculture. .................. 58
Figure 63: Shippagan Bay, NB aquaculture lease areas and EC-interpreted aquaculture polygons. ................. 59
Figure 64: Caraquet Harbour DEM, FPI eelgrass and interpreted aquaculture. ................................................ 61
Figure 65: Tracadie Bay DEM, FPI and EC/DFO eelgrass, and interpreted aquaculture. ................................. 62
Figure 66: Richibucto DEM, FPI and EC/DFO eelgrass, and interpreted aquaculture. ..................................... 63
Figure 67: Aldouane DEM, FPI and EC/DFO eelgrass, and interpreted aquaculture. ....................................... 64
Figure 68: Bedec DEM, FPI and EC/DFO eelgrass, and interpreted aquaculture. ............................................ 65
Figure 69: Bouctouche DEM, FPI and EC/DFO eelgrass, and interpreted aquaculture. ................................... 66
Figure 70: St Mary’s Bay DEM, FPI eelgrass, and interpreted aquaculture. ..................................................... 67
Figure 71: The turquoise image is the orthophoto mosaic; the greyscale image is the Digital Elevation Model
(DEM) (m, CGVD28). Orange symbols are depth values from Canadian Hydrographic Service (CHS) chart
402401 adjusted from Chart Datum to CGVD28 as outlined in Table 3 and Section 2.1.4. Aquaculture is
shown as yellow lines. Arrows indicate No Data values in the DEM. ............................................................... 68
x
LIST OF TABLES
Table 1. Depth interval definitions. ...................................................................................................................... 8
Table 2: Existing bathymetry data (X) for each bay and each data type. X* indicates there is a digitized chart
but insurmountable issues prevented the charts from being imported into ArcGIS software. ............................. 8
Table 3: Vertical Datum corrections used to convert to CGVD28. Details on GRS80- HT2 can be obtained
here: http://www.nrcan.gc.ca/earth-sciences/products-services/land-geodetic-survey/geodetic-tools/5199. ..... 9
Table 4: Summary of Environment Canada (EC) and Department of Fisheries and Oceans (DFO) eelgrass data
collection methods and years. ............................................................................................................................ 11
Table 5. Approximate dimension, surface area and oyster bag equivalence of each aquaculture gear type. For
oyster gear, dimension is length × width × height. For mussel socks, dimension is length × radius. ............... 13
Table 6: Surface area of DEM depth intervals (ha). Sum Depth Area is the sum of the area in each depth
interval, except 0-1 m interval. ........................................................................................................................... 23
Table 7: Area of Bays shown in Figure 1, known as Whole Area. Although not all the bays in Figure 1 were
surveyed, the area of the polygon provided by FPI was calculated to be used for various calculations in the
Results section. ................................................................................................................................................... 24
Table 8: Difference1 represents the area of the bay that was not detected by the lidar. %Difference1 shows the
difference relative to Whole Area, the closest approximation available for the correct area of each bay. ........ 25
Table 9: Sources of data for DEM depth comparison. Negative values indicate that zcomparison was deeper than
zDEM. ................................................................................................................................................................... 34
Table 10: Eelgrass surface area (ha) per 1 m depth interval. See Appendix A for eelgrass quality definitions.36
Table 11: Comparisons of different calculations of eelgrass polygons. Total Polygon Area=sum of eelgrass
area per depth interval; Whole Area=areas in Table 7, Figure 1; Summed Total = Sum Eelgrass Area;
Difference2 = difference between Total Polygon Area and Summed Total; No Data = % of eelgrass polygon
that is No Data= Difference2/Total Polygon Area. ............................................................................................ 37
Table 12: Total eelgrass area in hectares and percent eelgrass coverage for each bay from FPI (%eelgrass = %
of bay that has eelgrass = Total Polygon Area in Table 11/Whole Area in Table 6) and/or EC/DFO data. ..... 41
Table 13: Aquaculture lease area, lease area percentage of bay, aquaculture gear area (two dimensional),
bivalve biomass (BM), bivalve biomass per bay area and bags per lease area for each bay for year
corresponding to eelgrass surveys (2007-2009). All oyster aquaculture. .......................................................... 49
Table 14: Aquaculture lease area, lease area percentage of bay, aquaculture gear area (two dimensional),
bivalve biomass (BM), bivalve biomass per bay area and bags per lease area for each bay for year
xi
corresponding to FPI (lidar) eelgrass surveys (2011). All the New Brunswick calculations are based on oyster
aquaculture and St. Mary’s Bay PEI is based on mussel aquaculture. ............................................................... 49
Table 15. Aquaculture gear spacing and line length and percentage of total for each gear type in each bay
calculated from orthophoto mosaics................................................................................................................... 52
Table 16: Orthophoto mosaics total biomass per bay, boat-based survey total biomass. Difference2=
orthophoto total biomass – boat-based survey total biomass, and Difference2=100*Difference2/boat-based
survey total biomass. .......................................................................................................................................... 53
Table 17. Mean depth (m), depth ranges, and number of points used (n) of aquaculture leases (gear) for each
bay. ..................................................................................................................................................................... 60
Table 18: DFO Habitat field survey criteria. ...................................................................................................... 77
xii
ABSTRACT
Eelgrass (Zostera marina) is well established in many bays of the Southern Gulf of St Lawrence, but it is
experiencing some declines here and elsewhere in Atlantic Canada. Eelgrass plays an important biophysical
role in the health of ecosystems and therefore is recognized as an ecologically significant species. Current
methods to map eelgrass involve a combination of field samples and traditional remote sensing consisting of
aerial photographs and satellite imagery, which can be inaccurate, costly, and laborious. The main goal of this
study was to use aerial photography and a remote sensing method called bathymetric lidar to map eelgrass
distribution, determine water depth, and estimate bivalve aquaculture biomass for six bays in the southern
Gulf of St Lawrence. Lidar-detected depths ranged from 0 to 9 m. The lidar-derived depths agreed to within
~1 m with other sources of bathymetry data such as Canadian Hydrographic Service (CHS) charts and echo
sounder data. The lidar was unable to collect data for between 7% and 46% of the bay area where the water
was either too deep (e.g., in channels) or too turbid for the lasers to penetrate to the bottom. Since the areal
distribution of different depth ranges is important to determine the capacity of a bay to support eelgrass, the
lidar dataset was coupled to a Geographic Information System (GIS) dataset. Results indicated that the
seabed area for layers of 1 m depth intervals was inversely proportional to depth.
Aerial photographs were obtained during the lidar survey flights and used to map eelgrass coverage in the six
surveyed bays. Additional eelgrass classifications were obtained from interpretations by Environment Canada
(EC) from a variety of sources and field observations by DFO. Eelgrass coverage per bay was calculated in a
GIS and the results from the different methods compared. Lidar survey and EC/DFO estimates of eelgrass
coverage tended to agree. In absolute terms eelgrass coverage ranged from 374 ha (St Mary’s Bay) to 4500 ha
(Caraquet Bay). In relative terms eelgrass coverage ranged from 26% (St Mary’s Bay) to 94% (Bedec Bay) of
the total bay area. Eelgrass coverage per 1 m depth interval decreased with depth and was negligible at depths
greater than 4 m as was expected given reduced light penetration with depth.
Aerial photo (orthomosaic) quality was influenced by cloud shadow, light glint, and surface roughness.
Overall however, aerial photos were useful for identifying aquaculture buoys and gear type. The digitized
aquaculture lines were used to estimate bivalve biomass based on buoy spacing and biomass per buoy for each
aquaculture gear type. In five out of six bays compared, the values for total biomass estimated from aerial
photos were within 10% of boat-based survey estimates. In relative terms, oyster lease area ranged from
0.25% (Tracadie South) to 8.73% (Bedec) of the total bay area. In PEI, mussel leases in St. Mary’s Bay
occupied 22.67% of the total bay area.
Bathymetric maps, eelgrass maps and aquaculture maps for several oyster bays in New Brunswick and a
mussel bay in Prince Edward Island were constructed. They allowed the calculation of depth areas, eelgrass
coverage and aquaculture biomass. These results will feed a concurrent study to assess the statistical
relationship between bivalve aquaculture and eelgrass density on a bay-wide scale.
xiii
RÉSUMÉ
La zostère (Zostera marina) est répandue dans plusieurs baies du sud du golfe du St Laurent mais connait
certains déclins dans ces baies et ailleurs dans les provinces Atlantiques. Surtout connu pour son rôle
biophysique important dans la santé des écosystèmes, elle correspond aux critères d’une espèce d’importance
écologique EIE . Les méthodes utilisées pour cartographier la zostère consistent d’une combinaison
d’échantillonnage sur le terrain et de méthodes par télédétection telles les photos aériennes et images satellites
qui peuvent être coûteuses et laborieuses. Un des buts de cette étude est de cartographier l’étendu spatiale de
la zostère dans six baies du sud du golfe du St Laurent à l’aide de photos aériennes et d’une méthode par
télédétection appelé lidar bathymétrique. Les photos aériennes ont aussi été utilisées pour estimer
l’infrastructure de l’aquaculture des bivalves à partir de laquelle la biomasse a été estimée. Les profondeurs
détectées par le lidar variaient entre 0 et 9 m. Les profondeurs dérivées du lidar concordaient à 1 m près aux
profondeurs provenant d’autres sources telles les cartes du Service Canadien Hydrographique (SCH) et les
sondes acoustiques. Une grande partie des baies a été évaluée pour la profondeur à l’exception de certaines
données manquantes dans 7 à 46% de la surface des baies. Ces manques s’expliquent par des cas où l’eau
était soit trop profonde (chenal) ou soit trop trouble pour que le lidar pénètre jusqu’au fond marin. La
répartition spatiale des différentes profondeurs est importante pour étudier la capacité de support d’une baie
par rapport à la zostère, d’où l’importance de l’information bathymétrique de cette étude. En général, la
surface occupée par intervalle de profondeur de 1 m a diminuée avec la profondeur.
Lors des relevés lidar, des photos aériennes ont été obtenues simultanément et utilisées pour cartographier la
zostère dans les six baies évaluées. Des cartes de répartition de la zostère ont aussi été obtenues par les
interprétations d’Environnement Canada (EC) provenant d’une variété de sources et de données de terrain du
Ministère des Pêches et des Océans (MPO). La superficie de la zostère a été calculée dans un système
d’information géographique (SIG) pour chaque baie. Les résultats de chaque méthode ont été comparés. En
général, les estimations de superficies de la zostère des relevés lidar et d’EC/MPO concordaient. En termes
absolus, la surface occupée par la zostère variait de 374 ha (St Mary’s Bay) à 4500 ha (Baie de Caraquet). En
termes relatifs, la surface de zostère variait entre 26% (St Mary’s Bay) et 94% (Bedec) de la surface totale de
la baie. La surface occupée par la zostère fut calculée pour chaque intervalle de profondeur. Cette surface
diminuait avec la profondeur et était négligeable à des profondeurs de plus de 4 m, ce à quoi on pouvait
s’attendre vu la diminution de la pénétration de la lumière avec la profondeur.
La qualité des photos aériennes (orthomosaics) capturées lors du relevé lidar était influencée par l’ombrage
des nuages, le brillant de lumière et par la rugosité de la surface. En général, les photos aériennes étaient
efficaces dans l’identification des bouées et des engins de culture. Les filières d’aquaculture ont été utilisées
pour estimer la biomasse des bivalves en se basant sur des informations de la distance entre les bouées et de la
biomasse par bouée pour chaque type d’engin. Dans cinq des six baies, les valeurs de biomasse totale par baie
estimé à partir des photos aériennes concordaient à moins de 10% de différence aux estimations par relevé en
bateau. En termes relatifs, la surface occupée par les baux d’huitre variait entre 0.25% (Tracadie Sud) et
8.73% (Bedec) de la surface totale de la baie. À l’Île-du-Prince-Edward, les baux de moules dans la baie St.
Mary’s occupait 22.67% de la surface totale de la baie.
xiv
Des cartes de bathymétrie, de zostère et d’aquaculture de bivalve ont été construites pour plusieurs baies du
Nouveau-Brunswick et une de L’Île-du-Prince-Édouard. Ceci a permis le calcul de surface de profondeur, de
zostère et de biomasse de bivalves cultivés. Ces résultats serviront dans une étude connexe à évaluer la
relation entre l’aquaculture des bivalves et la superficie occupée par les herbiers de zostère à l’échelle de la
baie.
1
1. INTRODUCTION
1.1 EELGRASS
Eelgrass is established in beds on the sandy intertidal and subtidal flats in many of the bays and
protected harbours of the southern Gulf of St Lawrence. Eelgrass is an important primary
producer and serves several essential biophysical functions within the ecosystem, including
providing shelter and protection to many organisms (seaweed, invertebrates, fish), and acting as
a source of food to others, such as migratory aquatic birds (DFO, 2005). Eelgrass filters the
water column, stabilizes sediments in the nearshore marine environment, and acts as a shoreline
buffer (DFO, 2009).
Growth of eelgrass depends on light penetration into the water column, thus its depth distribution
is limited by water clarity. It is intolerant of anoxic (low oxygen) and eutrophic (excess nutrient)
conditions. Nutrient loading of 30 kg of nitrogen per hectare per year has been associated with
losses of 80% to 96% of eelgrass bed area (DFO, 2009). The excess nutrients in the water fuel
phytoplankton growth and reduce water clarity. Because of the link between ecosystem health
and eelgrass abundance, eelgrass can be considered a measure of ecosystem health, where
healthy ecosystems have an abundance of eelgrass, and stressed ecosystems have less, or
declining eelgrass (Lee et al., 2004; McKenzie, 2008; Kennish and Fertig, 2011; Washington
State DNR, 2011).
1.2 AQUACULTURE
Blue Mussels (Mytilus edulis) and Eastern Oysters (Crassostrea virginica) make up the majority
of the bivalve aquaculture industry in New Brunswick and Prince Edward Island. In 2006
mussels and oysters ranked behind salmon as the second and third most dominant aquaculture
category in Canada, and Canada ranked 12th globally in the production of both mussels and
oysters (Canadian Aquaculture Industry Alliance, 2011). In 2010, PEI produced 21x103 tonnes
of bivalve aquaculture worth $30 million, more than any other province in Canada; New
Brunswick produced 976 tonnes of bivalve aquaculture worth $5.6 million (Statistics Canada,
2010).
Mussels are farmed using the longline system, wherein mussel seeds are collected and placed in
mesh sleeves or socks which are then suspended in the water column until the mussels reach
market size. In PEI, this cycle takes between 18 and 24 months.
There is more variation in oyster aquaculture systems and infrastructure than mussel systems.
After seed collection, oysters can be suspended in the water column in a number of different
styles of cages or bags (Doiron, 2008).
2
1.3 RELATIONSHIPS BETWEEN EELGRASS AND AQUACULTURE
Eelgrass and bivalve aquaculture are both present in bays in the southern Gulf of St Lawrence.
Aquaculture can have both positive and negative effects on eelgrass health, density, and growth
(Tallis et al., 2009). Both oysters and mussels are filter feeders, and the main positive impact of
aquaculture on eelgrass is the improved water transparency caused by increased filtration, which
allows greater light penetration (Tallis et al, 2009).
Eelgrass may also be negatively affected by shading from aquaculture, and scour damage from
boats and anchors (Skinner et al. 2013).
1.4 STUDY OBJECTIVES
The status of eelgrass as an Ecologically Significant Species (ESS) (DFO 2009) emphasizes the
importance of eelgrass conservation to the health of estuarine ecosystems. For effective
conservation of this resource, an efficient means of mapping is required, keeping time and
financial considerations in mind. One method of mapping eelgrass is through boat-based surveys
equipped with video cameras, sidescan sonar and echo sounders. DFO has used this technique
successfully for years (Vandermeulen, 2011). However, piecing together these data is laborious
and costly, and distribution and abundance of eelgrass based on these surveys is considered to be
underestimated and, in some cases, several decades old since these surveys are not done on a
regular schedule (DFO, 2009). Interpretation of remote sensing methods such as aerial photos
and satellite images has the potential to improve eelgrass distribution mapping, although tide
state, water clarity, season and sea surface conditions can affect the detectability of eelgrass by
remote sensing techniques. An Environment Canada report (2011) used a combination of a
variety of remote sensing techniques and several different field samples to classify eelgrass;
ground-truthing indicated the authors classified eelgrass correctly between 77.2% and 96.7% of
the time. This integrated technique shows promise, but a single effective mapping technique still
remains to be proven.
There is evidence to suggest that eelgrass beds in many locations, including the southern Gulf of
St Lawrence, are declining (Hanson, 2004; Locke, 2005; AMEC, 2007; DFO, 2009). According
to a 2009 report by DFO, declines of 30% to 95% were reported in some locations of the
Maritime Provinces on inter-annual scales ranging from 2 to 20 years. The authors suggest
possible reasons for these declines in eelgrass distribution include eutrophication, disturbance
(uprooting and grazing) by invasive green crab, human activities, and environmental changes. In
Prince Edward Island, eutrophication and nutrient enrichment of bays and estuaries is
contributing to reductions in eelgrass distribution and threatening its persistence (Schmidt et al
2012).
The goals of this study, funded by the Program for Aquaculture Regulatory Research (PARR),
are to use aerial photography and a remote sensing method called bathymetric lidar to obtain
bivalve aquaculture biomass estimates and to map eelgrass distribution with depth for nine bays
3
in the southern Gulf of St Lawrence. The aerial photos and bathymetric lidar was analysed in a
Geographic Information System (GIS) in order to answer questions such as: How much
aquaculture biomass is present in each bay? How much eelgrass is present in each bay? How
deep does the eelgrass grow? What depth has the most eelgrass?
Data from 2007 to 2009 was compiled from previous work in order to present all available data
for a concurrent study on the relationship between eelgrass distribution and bivalve aquaculture
biomass.
2. METHODS
2.1 AREA OVERVIEW
This study was conducted in St Mary’s Bay, Prince Edward Island and several bays on the
eastern coast of New Brunswick, in the southern Gulf of St Lawrence. The bays in the southern
Gulf are characterized by shallow depths and a mixed geological environment of sandstone, mud
and rock (AMEC, 2007). Barrier islands, sand bars, and low coastal plains are common
morphological features (DFO, 2005), and tidal range is between 2.0 m and 2.5 m (CHS, 2012).
2.2 BATHYMETRY
2.2.1 Bathymetric Lidar
The primary source for depth and elevation data was bathymetric lidar, collected in September
2011 by Fugro Pelagos Inc. (FPI). Five major bays in New Brunswick and one bay in PEI were
surveyed (Fig 1). Airborne bathymetric lidar is a remote sensing technology that is used to
acquire elevation information about the Earth’s surface. A lidar system is comprised of three
technologies: GPS (Global Positioning System); an IMU (Inertial Measurement Unit); and a
laser ranging system (Flood and Gutelius, 1997; Liu, 2008).
The GPS was used to determine the geographic position of the aircraft in three dimensions. The
IMU was used to measure the altitude of the aircraft (roll, pitch and heading) (Liu, 2008). The
roll, pitch, and heading were accurately measured to allow for the correction of the motion of the
aircraft by computer software (Flood and Gutelius, 1997). The laser ranging system transmits a
laser pulse towards the Earth’s surface, and records the time delay between the transmission of
the laser pulse and its return. In terrestrial mapping lidar systems each emitted laser pulse can
have up to four returns encoded with the GPS, IMU and range data (Liu, 2008). The laser pulses
are directed across a swath with an oscillating mirror. Researchers such as Flood and Gutelius
(1997) and Wehr and Lohr (1999) provide a general description and overview concerning
4
airborne lidar technology and the principles behind it. Terrestrial airborne lidar uses NIR Laser,
with a wavelength typically at 1064 nm.
Bathymetric lidar follows the same theory as traditional lidar, but includes an additional laser to
penetrate the water column and return information on the ocean floor (Figure 2). FPI employs the
SHOALS-1000T acquisition system. The SHOALS system is discussed in detail in Guenther et
al. (2000), along with extensive background theory on bathymetric lidar. The remainder of this
section on general bathymetric lidar principles has been extracted from FPI’s survey report
(2011).
The laser output is infrared (1064 nm) with a frequency doubled green wavelength (532 nm) in a
single beam. The infrared wavelength is used to detect the water surface and does not penetrate
the air/water interface. The green wavelength penetrates through the water and detects the
seafloor. The green wavelength also generates red energy (645 nm) in the water column. This
by-product is known as Raman scattering and is another method used to detect the sea surface.
Distances from the surface and seafloor are calculated using the speed of light, index of
refraction in water, and the times of the laser pulse returns recorded by the receivers.
FPI collected bathymetric lidar between September 11 and 20, 2011. The SHOALS-1000T was
operated to achieve an IHO Order 1b category of survey coverage and accuracy. This was
achieved by combining a 5 m x 5 m spot spacing (flying at 400 m altitude and speed-overground of approximately 160 knots) with a 100% coverage plan. Planned line spacing provided
30 m of sidelap. The survey was flown with sufficient options made available to the airborne
operator to devise a best plan of the day for climatic and water quality considerations, such that
successful data collection was possible in both shallow and deep regions of the area or in areas
with known turbidity issues at various states of the tide and or wind direction and strength. A
laser pulse of 1 kHz was used for all topography and bathymetry data collection. Data received
by the airborne system were continually monitored for data quality during acquisition operations.
Display windows show coverage and information about the system status. In addition, center
waveforms at 5 Hz rate are shown in the display. All of this information allowed the airborne
operator to assess the quality of data being collected.
FPI produced a continuous water to land Digital Elevation Model (DEM) for each bay surveyed
using a combination of auto-processing algorithms and manual data inspection and editing. The
auto-processing algorithms obtained inputs from the raw data and calculated a height, position
and confidence for each laser pulse. This process, using the default environmental parameters,
also performed an automated first cleaning of the data, rejecting poor land and seafloor
detections. Other SHOALS specific tools, such as swapping a sounding that was falsely
recognized as land to water, were used inside Fledermaus by experienced data analysts. In the
shallower nearshore margins, the Shallow Water Algorithm (SWA) for bottom detection was
used to recover the bathymetry values and to allow, where valid returns permitted, a seamless
5
join with the topographic data obtained on the specific missions for which these data were
collected.
Bottom reflectance data were also recorded by the lidar system. In addition to recording the two
way travel time of the green laser pulse through the water column, the intensity or amplitude of
the reflected laser pulse off the seabed is also recorded. The intensity of this pulse will vary
depending on bottom type, thus having the potential to assist in mapping the seabed cover type,
i.e. sand vs. eelgrass. The reflectance data is represented as a grey scale image, similar to an air
photo. However, it represents the reflected energy from the green laser and not the sunlight. The
reflectance data were not part of the original Canadian Hydrographic Service (CHS) contract
deliverable, but were a specific request for the purpose of evaluating the potential of bathymetric
lidar for bottom type mapping.
Figure 1: Study area on the Eastern shore of New Brunswick and Eastern tip of Prince
Edward Island. Background map is a shaded relief terrain model with red polygons
denoting bays where bathymetric lidar and orthophotos were acquired. Green polygons
denote bays for which a bathymetric lidar survey was planned but not completed.
6
Figure 2: Airborne bathymetric lidar uses green and N1R laser. The green laser penetrates
the water column to two Secchi depths and measures the timing and intensity of the
returned laser pulse. Source: http://optech.ca/pdf/Brochures/SHOALS2007.pdf.
7
2.2.2 DEM Processing and Analysis
Digital Elevation Models (DEMs) and Colour Shaded Relief (CSR) maps were created for each
surveyed bay using the lidar data. The digital elevation model (DEM) heights provided by FPI
were referenced to the GRS-80 ellipsoidal model of the earth. To convert the elevations to
orthometric heights a geoidal separation value derived from NRCan’s Geodetic Survey HT2 was
used and subtracted from the ellipsoidal heights to reference them to the Canadian Geodetic
Vertical Datum 1928 (CGVD28). Details on GRS80- HT2 can be obtained here:
http://www.nrcan.gc.ca/earth-sciences/products-services/land-geodetic-survey/geodetictools/5199.
The DEMs can be used to derive a variety of other map layers including slope, aspect (land
facing orientation), as well as maps that can be used to improve interpretation of the bathymetry.
Shaded relief maps have been constructed from the DEMs where the terrain was illuminated
from the northwest at a 45 degree angle. In order to enhance the subtle relief present in the study
areas, a five times vertical exaggeration has been applied to the terrain. The shaded relief maps
are viewed and interpreted as greyscale maps which highlight the local relief, but do not depict
the actual elevation (i.e. a slope in the valley will look the same as a slope at the top of a hill).
Another series of map products that have been constructed consist of colour shade relief (CSR)
maps. Colours have been assigned to the elevations based on the DEMs and merged with the
shaded relief maps. The colourized elevation is then merged with the shaded relief that gives the
terrain and bathymetry texture and enhances the information that can be interpreted from the
map.
CSR maps have been built for all six DEM datasets and are presented in Section 3.1.3.
Elevations less than 0 m CGVD28 (bathymetry) are coloured in shades of blue; elevations
greater than 0 m CGVD28 (topography) are coloured green, yellow and red. Note that CGVD28
is approximately equal to mean sea level (MSL). These maps are qualitative and are designed for
use as a backdrop to other information within the GIS. The CSR images have been converted
into a compressed georeferenced format, JPEG 2000, which is compatible with most GIS
systems.
The surface area of each 1 m depth interval was estimated from the DEM using ArcMAP in the
following manner. First, depth values in the DEM were rounded up to the nearest integer. This
converted every pixel of the raster into an integer, sorting the DEM into 1 m depth intervals,
defined in Table 1. Then the number of pixels in each depth interval was multiplied by the size
of each pixel (5 m x 5 m) to arrive at a value for surface area for each 1 m depth interval.
8
Table 1. Depth interval definitions.
1 m Depth (z) Interval Polygons
Label
Definition (depth range in m)
0<z<1
1m
-1 < z <= 0
0m
-2 < z <= -1
-1 m
-3 < z <= -2
-2 m
-4 < z <= -3
-3 m
-5 < z <= -4
-4 m
-6 < z <= -5
-5 m
-7 < z <= -6
-6 m
-8 < z <= -7
-7 m
-9 < z <= -8
-8 m
-10 < z <= -9
-9 m
2.2.3 Supplementary Bathymetry
Several sources of depth data were available to compare with the lidar bathymetry (Table 2) such
as echo soundings, depth measurements and digitized CHS bathymetric charts. Unfortunately
there were problems importing the charts for all bays except Caraquet and Richibucto into
ArcMap.
Table 2: Existing bathymetry data (X) for each bay and each data type. X* indicates there
is a digitized chart but insurmountable issues prevented the charts from being imported
into ArcGIS software.
Bay
Caraquet Bay, NB
Miscou Harbour, NB
Tracadie Bay, NB
Richibucto Harbour, NB
Bouctouche Bay, NB
Cocagne Harbour, NB
Shippagan Bay, NB
St Mary’s Bay, PEI
Bathymetric
Lidar (FPI)
X
X
X
X
X
X
Echo
Soundings
(DFO)
Field Samples
(incl. depth)
(DFO)
X
X
X
X
Digitized
Chart (CHS)
X
X*
X*
X
X*
X
X
X*
9
2.2.4 DEM Depth Validation
The lidar-derived depths of the DEM were validated by comparing them to existing depths in a
particular bay (depth data available per bay shown in Table 2). A common vertical datum was
necessary for the comparison to be useful. Table 3 shows the values used to bring all depth data
into CGVD28, the standard vertical datum.
Table 3: Vertical Datum corrections used to convert to CGVD28. Details on GRS80- HT2
can be obtained here: http://www.nrcan.gc.ca/earth-sciences/products-services/landgeodetic-survey/geodetic-tools/5199.
Data
Lidar DEM
Echo Soundings (DFO)
Field Samples (incl. depth)
(DFO)
Digitized Chart (CHS)
Original Vertical Datum
Heights relative to ellipsoid
GRS80
Assumed MSL
Assumed MSL
Conversion to CGVD28
GRS80 – HT2
Chart Datum 2000
Caraquet: 0.9 m
Richibucto: 0.5 m
Richibucto and St Mary’s: 0 m
Miscou and Tracadie: 0 m
The echo soundings and field samples did not have a complete meta-data record, so there was no
indication of whether or not the elevations had been compensated for tidal elevations, and data
were not time-stamped. Tidal range in the southern Gulf of St Lawrence is relatively small (2 to
2.5 m), so the assumption of MSL as the vertical datum for the echo soundings and field samples
will not introduce extreme error if incorrect.
ArcMAP was used to extract points from the DEM at the location of a discrete depth value. Once
all data were converted to CGVD28, a simple comparison of depths provided insight into the
accuracy of the DEM.
2.2.5 Orthophoto Mosaics
Aerial photographs were captured by FPI coincident with the lidar collection. A DuncanTech
DT4000 digital camera was mounted in a bracket at the rear of the lidar sensor and used to
acquire one 24-bit, 4 megapixel color photo per second. The photos were post-processed by FPI
into orthophoto mosaics. Table 2 shows the six bays that have orthophoto mosaics (all bays that
have lidar coverage have orthophoto mosaics). The priority for FPI during the flights was lidar
collection, and as a result, some of the aerial photos have quality issues including glint, cloud
shadow, and surface roughness of the sea (Fig.3).
10
Figure 3: Orthophoto mosaic in St Mary’s Bay, Prince Edward Island showing poor aerial
photo quality. The southern edges of the individual orthophotos are affected by glint; dark
patches near the center of the image are likely cloud shadow.
2.3 EELGRASS
2.3.1 FPI Eelgrass From Orthophoto Frames
Areas identified as having eelgrass coverage were delineated by FPI with enclosing polygons
interpreting primarily the orthophoto frame imagery prior to mosaicking. Results relied on
bottom visibility (due to water clarity and depth) and the FPI Data Analyst’s interpretation. The
individual aerial photo frames were used to identify eelgrass to overcome glint and shadow
issues with the orthophoto mosaic imagery, an example of which can be seen in Fig.3.
Groundtruthing for this 2001 data was conducted in St. Mary’s Bay only.
2.3.2 Environment Canada and DFO
Environment Canada classified eelgrass in eight bays in New Brunswick using a combination of
remote sensing images and field sampling (Table 4). Aerial photography and satellite imagery
were used to classify eelgrass according to quality, or coverage; the classification schemes are
defined in Appendix A. Five of the bays followed Environment Canada’s field survey criteria
that classify eelgrass as good, medium, or poor quality/absent eelgrass, while Shippagan Bay has
used DFO’s classification of dense, moderate, thin, or exposed eelgrass. Cocagne and Tabusintac
Harbours included a polygon that represents eelgrass presence, but does not indicate quality.
Aerial photography and satellite image classification were object-oriented, and conducted using
eCognition Developer software (Mahoney, 2011).
DFO conducted visual field surveys of four of the bays (Table 4) during the same season in
which the aerial photography had been captured. The field samples were used to assist in photo
classification and to assess accuracy. For the remaining four bays, field data were collected along
11
transects in the bays using a differential GPS positioned towfish holding a video camera. More
details on methodology can be found in Appendix A.
Table 4: Summary of Environment Canada (EC) and Department of Fisheries and Oceans
(DFO) eelgrass data collection methods and years.
Region
Miscou Hbr, NB
Tracadie Bay, NB
Richibucto Hbr, NB
Bouctouche Bay, NB
Shippagan Bay, NB
Neguac Bay, NB
Cocagne Hbr, NB
Tabusintac Bay, NB
St. Mary’s Bay, PE
Aerial
Quickbird
photography satellite
EC
EC
2009
2009
2007
2009
2007
2009
2008
2008
Field survey
Video camera
DFO
DFO
2009
2009
2007
2009
2007
2009
2008
2008
2011
Classification
categories*
GQ, MQ, PQ
GQ, MQ, PQ
GQ
GQ, MQ, PQ
DE,ME,TE,EE
GQ, MQ
EP
EP
EP
*Classification categories: GQ = Good Quality, MQ = Medium Quality, PQ = Eelgrass absent/Poor Quality; DE = Dense
Eelgrass, ME = Moderate Eelgrass, TE = Thin Eelgrass, EE = Exposed Eelgrass; EP = Eelgrass Presence.
2.3.3 Eelgrass Coverage Per Bay
Eelgrass coverage per bay was calculated using ArcMAP for both FPI and EC eelgrass coverage
by simply summing the area of each eelgrass polygon. A surface area field was computed for the
FPI eelgrass polygons using the ArcMAP tool “Zonal Geometry as Table” because, unlike the
EC eelgrass data, the FPI data did not include a surface area field. For EC eelgrass, the
calculation was done separately for each class of eelgrass, Good Quality, Medium Quality, etc.
2.3.4 Eelgrass Coverage at 1 m Depth Intervals
To calculate the surface area of eelgrass beds in each depth interval, the DEM was rounded up to
the nearest integer, as with the area per depth calculation above. Elevations greater than 1 m
were rejected to avoid unnecessary computations on land while including the complete intertidal
zone, and then the DEM raster was converted to polygons. Next the eelgrass polygons were
converted into rasters. The ArcMAP tool, “Zonal Statistics as Table” was used to calculate the
area of the eelgrass class within each “zone”, or depth interval polygon. This procedure was
followed for both the FPI and the EC eelgrass, for each lidar-surveyed bay.
2.4 AQUACULTURE
Data on oyster and mussel aquaculture was compiled from boat surveys conducted by the New
Brunswick Department of Fisheries, Agriculture and Aquaculture (DFAA) for each bay in NB
and by DFO for St. Mary’s Bay in PEI, respectively. Data was collected for each year in which
12
an eelgrass survey was conducted (see Table 4) for a particular bay, 2007, 2008, 2009 or 2011.
In 2011, aquaculture data was also acquired from the orthophoto mosaics obtained in September,
2011 for the bays covered during the lidar survey.
The aquaculture data consisted of lease area and the gear type (e.g. oyster strings, cages, and
floating bags), dimensions (Table 5), area and amount present and the estimated biomass in each
bay. Lease area was obtained from the appropriate leasing agency (NB leasing and DFO leasing).
In this report it is defined as the area occupied by suspension bivalve aquaculture leases where
sales have been reported in the last year. Lease area has been stable since 2007 except for
Miscou which has not had any reported sales since 2011. The percent of the bay occupied by
leases was calculated by dividing the total lease area in a bay by the whole area of the bay (Table
7) multiplied by 100.
For the 2011 aquaculture dataset, the two sources of data (boat surveys and orthophoto mosaics)
were compared in terms of aquaculture gear distribution and biomass in each bay.
2.4.1 Aquaculture: Biomass Calculations – Boat Surveys
The biomass was estimated from boat surveys conducted by DFAA and DFO for each bay in NB
and PEI, respectively. For NB oysters, all gear present in a bay was tallied and then each gear
type was converted to a standard oyster bag equivalent (Table 5) to obtain the total number of
bags for each bay. Biomass per bag was estimated to be 6.04 kg by Comeau et al. (2006) as
described in their equation:
bm = ∑
MT × DT × PT
=
6.04 kg
where bm is the mass of one oyster bag, T is size category from 1 to 4 based on shell height. For
each category, M is the average weight of one oyster, D is the average number of oysters
contained in each bag and P is the percentage present in a typical lease. For PEI mussels,
biomass was obtained by multiplying the number of socks by an average weight of 7.6 kg per
sock (Drapeau et al. 2006). Aquaculture gear area for each bay is the sum of each gear type
surface area multiplied by the total number of gear units of that type.
13
Table 5. Approximate dimension, surface area and oyster bag equivalence of each
aquaculture gear type. For oyster gear, dimension is length × width × height. For mussel
socks, dimension is length × radius.
Gear Type
Floating Oyster Bag
Sub-Surface Oyster Cages
Oyster String Cages
OysterGro Cages
Dark Sea Cages
Oyster Table
Mussel Sock
Gear dimension
(m)
0.8 × 0.4 × 0.1
0.8 × 0.4 × 0.4
1.8 × 0.9 × 0.6
1.47 × 0.91 × 0.15
0.6 × 0.6 × 0.09
0.8 × 1.2 × 0.2
1.83 × 0.07
Gear surface area
(m2)
0.32
0.32
1.62
1.34
0.37
0.96
0.015
Oyster bag equivalent
1
4
16
6
10
6
n/a
2.4.2 Aquaculture: Biomass and Depth Calculations – Orthophoto Mosaics
For 2011, biomass was also calculated by interpreting and measuring line length and spacing on
the orthophotos and using information on cage dimensions and distribution, and biomass per bag.
Biomass per bag was estimated to be 6.04 kg as described in Comeau et al. (2006) and Comeau
2013.
In bays for which a DEM and digitized aquaculture data exist, the average depth at the
aquaculture gear (longlines) was calculated in ArcMAP using the extraction spatial analyst tool.
2.4.2.1 Oyster Collector Lines
Collector lines are temporarily strung out between August and October outside aquaculture lease
areas to collect oyster spat on various structures such as Chinese hats (Fig.4). The spat are used
as seed to begin next season’s aquaculture crop. Spat biomass was not calculated. Figure 5 shows
collector lines in Bouctouche Bay as they appear in the orthophoto. Figure 6 shows the
interpreted and digitized lines.
14
Figure 4: Collector lines in Bouctouche Bay, New Brunswick.
Figure 5: Collector lines in Bouctouche
Bay, NB seen in the orthophoto mosaic
from September, 2011.
Figure 6: Orange lines are interpreted
and digitized collector lines in Bouctouche
Bay, NB.
2.4.2.2 Floating Oyster Bags (Single and Double)
Floating oyster bags were found in all bays except Bedec Bay (the eastern section of Richibucto
Bay) in both single and double width lines. Each shallow vexar oyster bag is approximately 0.85
m × 0.40 m × 0.10 m (Doiron, 2006) and represents about 6.04 kg of oyster biomass (Comeau et
al., 2006). Figure 7 shows a vexar bag in the foreground with white buoys and Figure 8 shows
submerged floating bags with black buoys. Single lines are one bag wide and have 50 bags per
line (Figure 9); double lines are two bags wide and have 100 bags per line (Figure 8 and Figure
10).
15
Figure 7: Floating oyster bags with white
buoys; more are seen in the distance.
Figure 8: A double-wide line of floating
oyster bags.
Figure 9: Single-wide floating bags in
Tracadie Bay, NB in the FPI orthophoto
mosaic from Sept., 2011.
Figure 10: Double-wide floating bags in
Richibucto Bay, NB in the FPI
orthophoto mosaic from Sept., 2011.
Biomass is estimated based on line length and line width: single or double. Figure 8 shows that
spacing of floating bags was very consistent at 1 bag/m for single-wide, and 2 bags/m for
double-wide lines:
Lines were interpreted and digitized from the orthophotos in a GIS, their lengths were calculated,
and then used in the above equation as line length. Each line was classed as either single or
16
double wide, such that # bags/m was either 1 bag/m or 2 bags/m for single and double lines,
respectively, for use in the above equation.
2.4.2.3 Sub-Surface Oyster Cages
Sub-surface oyster cages, present only in Tracadie Bay, NB, consist of four bags stacked
vertically (6.04 kg/bag). Biomass was calculated based on line length and cage spacing as
estimated from the orthophotos (Figure 11, Figure 12), as follows:
Figure 11: Dark lines are sub-surface
oyster cages in the Tracadie Bay, NB
orthophoto mosaic from Sept., 2011.
Figure 12: Orange lines are interpreted
and digitized sub-surface oyster cages in
the Tracadie Bay, NB orthophoto mosaic
from Sept., 2011.
2.4.2.4 Oyster String Cages
Oyster string cages contain between ~3200 and ~4500 oysters (60 kg and 84 kg) cemented in
clusters of three to strings (Fig. 13). There are 8-12 cages per line.
17
Figure 13: Oyster string cages.
Figure 14: The white end buoys of the oyster string cages are barely visible in the left
orthophoto; the lines digitized between the buoys in the image on the right represent the
oyster string cages in the Caraquet Bay, NB orthophoto mosaic from Sept., 2011.
Biomass was estimated for each bay based on cage spacing and line length. Oyster string cages
were easy to interpret in the orthophotos (Fig. 14), allowing an accurate measurement of spacing
and length. The following equation was used to estimate biomass:
An estimate for cage spacing of 1 cage per 4 m was made for Tracadie and Caraquet Bays (the
only bays that contained oyster string cages), and each line of oyster string cages was measured
individually. In Tracadie Bay each cage contained 60 kg of oyster biomass (kg/cage = 60 kg);
Caraquet cages contained 84 kg (kg/cage = 84 kg).
18
2.4.2.5 Oystergro Cages
Oystergro cages contain six bags of oysters per cage (three wide and two deep), with each bag
the equivalent of approximately 6.04 kg of oysters. They are suspended by two long, dark and
narrow floats (Figure 15 and Figure 16). In the bays surveyed, there were two typical setups: a
longer line with shorter cage spacing and more cages per line (Figure 17), and a shorter line with
fewer cages spaced farther apart (Figure 18). Oystergro cages were present in all six of the
surveyed bays.
Figure 15: Oystergro cages in Bouctouche
Bay, NB.
Figure 16: Upside-down Oystergro cages.
Figure 17: Oystergro cages in Tracadie Bay, NB with 10-12 cages per line spaced 3 m
apart; average line length is 43 m. Image is FPI orthophoto mosaic from Sept., 2011.
19
Figure 18: Oystergro cages in Caraquet
Figure 19: Digitized Oystergro cages seen
Bay, NB with 6 cages visible per line
in the Caraquet Bay, NB orthophoto
spaced 6 m apart; average line length is
mosaic from Sept., 2011.
30 m.
Biomass was calculated based on cage spacing and line length as follows:
An estimate for cage spacing was made for each bay, and each line of Oystergro cages was
measured individually (Figure 19).
2.4.2.6 Dark Sea Oyster Cages
Dark sea cages (Figure 20) can have many levels (ten is a typical number) each containing the
equivalent of one 6.04 kg bag. Lines tend to be 30 m long with 7 or 10 cages per line. Digitized
lines were 30 m on average, so the lower estimate of cages per line was used in the following
biomass calculation:
Figure 20: Dark sea oyster cages.
20
2.4.2.7 Mussel Longlines
Mussel longlines, present only in St Mary’s Bay, PEI, are a single line with several drop lines,
socks, or sleeves, containing mussels and extending to 1.83 m deep (Figure 21). In St Mary’s Bay
the lines were on average 122 m long, with 0.45 m between socks, and 7.6 kg of mussels per sock
(Drapeau et al., 2006), such that biomass was calculated based on measured line length as follows:
Mussel longlines were interpreted from the orthophotos to be both floating on the surface as well
as submerged below the surface, where each had white end marker buoys (Fig. 22). The floating
longlines seen in Figure 23 show an example of longlines with and without digitized aquaculture;
the left image, when compared to the previous figure, shows how the longline appearance can
differ across the bay.
Figure 21: Typical design of a Prince Edward Island longline mussel farming setup from
Drapeau et al. 2006.
20
21
Figure 22: Mussel longlines in St Mary’s Bay, PEI. The red box shows submerged lines with
white end buoys visible; the yellow box shows floating longlines.
Figure 23: Left: Mussel longline buoys in St Mary’s Bay, PEI, seen in orthophotos from FPI,
Sept. 2011. Right: Digitized mussel longlines.
21
22
3. RESULTS
Section 3.1 contains results of the DEM surface area calculations, and figures of the DEM and
CSR with 1 m depth intervals for each bay. In Section 3.2 tables present results of eelgrass area per
depth for each bay, and total eelgrass area per bay. Figures of the DEMs with FPI eelgrass
coverage and CSRs with EC/DFO eelgrass coverage are presented. Aquaculture results are
presented in Section 3.3 in tables containing information on aquaculture spacing, line length,
biomass of each aquaculture type, etc., for each bay. Figures show orthophotos with digitized
aquaculture lines colour-coded by aquaculture type.
The results for Richibucto Bay are presented as three smaller regions: Aldouane, Bedec, and
Richibucto (Figure 24). The DEM provided by FPI was for the entire region, which has been
referred to as Richibucto to this point in the report, but to present and discuss the eelgrass and
aquaculture results it has been subdivided.
Figure 24: The extent of the Digital Elevation Model (DEM) shows the region, referred to as
Richibucto that was collected by FPI. For biomass and eelgrass calculations, Richibucto has
been divided into three smaller regions: Aldouane, Bedec, and Richibucto. The polygonal
outlines were used to actually subdivide Richibucto, as the FPI coastline outlines are smaller
than the DEM.
22
23
3.1 BATHYMETRY
3.1.1
Surface Area of DEM Depth Intervals
Table 6 shows the results of the DEM depth interval surface area calculations described in Section
0 (depth interval definitions are shown in Table 1). The total area, Sum Depth Area, is the sum of
the surface area of the DEM depth intervals (0 to -10 m). Caraquet Bay has the greatest lidardetected total surface area (3614 ha) while Aldouane has the smallest (410 ha). Aldouane also
reported the shallowest depths (3 m), while Caraquet, Tracadie, and St Mary’s reported the deepest
depths (~9 m); lidar penetration did not exceed 10 m in any bay. In general, depth interval areas
decreased with depth, such that the -1 – 0 m interval had the largest area, and the deepest interval
had the smallest area. The exception to this trend was the shallowest interval (0 – 1 m), which had
less area than the -1 – 0 m interval.
Table 6: Surface area of DEM depth intervals (ha). Sum Depth Area is the sum of the area in
each depth interval, except 0-1 m interval.
Surface area of DEM depth intervals (ha)
Caraquet
274
1144
818
548
366
267
197
133
103
35
Miscou
226
575
549
642
309
80
4.8
7.3
3.0
0.11
Tracadie
883
2264
173
62
27
21
13
4.7
0.24
0. 13
Richibucto
172
675
302
78
30
28
26
26
5.4
0. 46
Aldouane
70
282
128
0. 83
Bedec
155
341
229
35.6
0. 29
Bouctouche
196
1048
800
730
74
18
2.2
0. 54
St Mary’s
7400
34300
29400
32000
25300
24100
10900
3900
-10 – -9 m
-9 – -8 m
-8 – -7 m
-7 – -6 m
-6 – -5 m
-5 – -4 m
-4 – -3 m
-3 – -2 m
-2 – -1 m
-1 – 0 m
0–1m
Depth intervals
2.2
Sum
Depth
Area
3614
2170
0.0002
2565
1170
410
605
2673
1800
1000
450
1632
3.1.2 Total Area Comparisons
The area of the polygons provided by FPI is summarized in Table 7, and is referred to as the Whole
Area of each bay, in contrast to the area detected by the lidar. Table 8 shows some comparisons
between Sum Depth Area (the sum of the surface area of DEM intervals from Table 6 and Whole
Area.
23
24
Equation 1: Definitions of surface areas used to determine proportion of bay detected by
Lidar.
Whole Area = areas in Table 7, area of polygons in Figure 1
Sum Depth Area = sum of DEM surface area contours in Table 5
Difference1 = Whole Area – Sum Depth Area
Table 7: Area of Bays shown in Figure 1, known as Whole Area. Although not all the bays in
Figure 1 were surveyed, the area of the polygon provided by FPI was calculated to be used
for various calculations in the Results section.
Bay Name
Area (ha)
Caraquet
Miscou
Tracadie
Richibucto
Aldouane
Bedec
Bouctouche
Shippagan
Tabusintac
Cocagne
Neguac
6674
3225
3123
1588
692
649
3015
8115
3104
2009
3813
Difference1 represents the area of the bay that was not detected by lidar, and ranged from ~3100
ha (Caraquet) to less than 100 ha (Bedec), with an average difference of ~800 ha. %Difference1
highlights the difference relative to the total surface area of each bay. For example, Difference1 =
282 ha and %Difference1 = 41% for Aldouane. This means that the lidar did not penetrate all the
way to the bottom in 41% of Aldouane; equivalently, the lidar did detect the bottom in 59% of
Aldouane. %Difference1 ranged from 7% in Bedec to 46% in Caraquet, with an average value of
26%. These calculations assume Whole Area is the best approximation of the correct area of each
bay.
24
25
Table 8: Difference1 represents the area of the bay that was not detected by the lidar.
%Difference1 shows the difference relative to Whole Area, the closest approximation
available for the correct area of each bay.
Difference1 (ha)
3059
Caraquet
1055
Miscou
559
Tracadie
Richibucto 418
282
Aldouane
44
Bedec
Bouctouche 343
3.1.3
%Difference1 (%)
46
33
18
26
41
7
11
Bathymetric Intervals, DEMs and CSRs
Figure 25- Figure 40 show 1 m bathymetric intervals on the DEM and CSR for each bay. The
intervals, derived from the DEM, emphasize the slope of the bottom and the distribution of depth
in each bay. The variation in the percentage of lidar coverage between bays evident in the figures
can be explained by water clarity and the depth of the bay: the laser penetration is limited to 2-3
Secchi depths. Secchi depth is a measure of water transparency, and is typically shallow in muddy
or phytoplankton-dense water, and deep in clear water in which the light encounters no obstacles.
In clear water the lidar is able to reach the bottom of these shallow bays, but in bays that have low
transparency or are clear but deep, gaps in the lidar are evident.
In Caraquet, 46% of the area of Caraquet was not detected by the lidar (Table 8, Figure 25) either
because the water was muddy or cloudy, or too deep. In contrast, the absence of large holes in the
DEM in Figure 29 suggests that water was clear in Tracadie Bay during the lidar survey and did
not exceed 10 m, and Table 8 indicates that only 18% of the area of Tracadie was not detected by
the lidar. Further discussion on lidar depth penetration in Caraquet is found in Section 4.1.
25
26
Figure 25: Caraquet DEM and lidar-derived 1 m depth intervals. All depths are CGVD28.
Figure 26: Caraquet 1 m depth intervals on elevation colour-shaded relief (CSR) image.
26
27
Figure 27: Miscou DEM and lidar-derived 1 m depth intervals. All depths are CGVD28.
Figure 28: Miscou 1 m depth intervals on elevation colour-shaded relief (CSR) image.
27
28
Figure 29: Tracadie DEM and lidar-derived 1 m depth intervals. All depths are CGVD28.
Figure 30: Tracadie 1 m depth intervals on elevation colour-shaded relief (CSR) image.
28
29
Figure 31: Richibucto DEM and lidar-derived 1 m depth intervals. All depths are CGVD28.
Figure 32: Richibucto 1 m depth intervals on elevation colour-shaded relief (CSR) image.
29
30
Figure 33: Aldouane DEM and lidar-derived 1 m depth intervals. All depths are CGVD28.
Figure 34: Aldouane 1 m depth intervals on elevation colour-shaded relief (CSR) image.
30
31
Figure 35: Bedec DEM and lidar-derived 1 m depth intervals. All depths are CGVD28.
Figure 36: Bedec 1 m depth intervals on elevation colour-shaded relief (CSR) image.
31
32
Figure 37: Bouctouche DEM and lidar-derived 1 m depth intervals. All depths are CGVD28.
Figure 38: Bouctouche 1 m depth intervals on elevation colour-shaded relief (CSR) image.
32
33
Figure 39: St Mary’s DEM and lidar-derived 1 m depth intervals. All depths are CGVD28.
Figure 40: St Mary’s Bay 1 m depth intervals on elevation colour-shaded relief (CSR) image.
33
34
3.1.4 DEM Comparison/Validation
The bathymetry values of the DEM were compared to various other sources of bathymetry data,
from CHS charts to echo soundings. The data available for these comparisons are shown in Table 2
and results of the comparisons are shown in Table 9. Negative values in the table indicate that
zcomparison was deeper than zDEM. The mean difference between the DEM and the other bathymetry
data was less than 1 m, and standard deviations were less than -1.5 m. In the following equation,
zcomparison is depth from the comparison source, e.g. CHS Chart, echo sounding, etc.
∆z =zcomparison - zDEM
The greatest ∆z was -8.46 m in Richibouctou Bay, and the mean ∆z between the echo sounder
data and the DEM was 0.02 m. Vertical datum information for echo soundings and field samples
were missing, and MSL was assumed. Conversions from CD to CGVD28 for CHS charts are
shown in Table 3.
Table 9: Sources of data for DEM depth comparison. Negative values indicate that zcomparison
was deeper than zDEM.
Compariso Number Minimum
Maximum
Mean ∆z
St. Dev.
n source
of points ∆z (m)
∆z (m)
(m)
of ∆z (m)
(zcomparison) used
CHS Chart 7
-1.49
3.72
0.79
1.49
Caraquet
402401
EC eelgrass 85
-4.79
1.91
-0.09
0.93
Miscou
pts
EC eelgrass 101
-2.04
0.06
-0.24
0.47
Tracadie
pts
DFO echo
36438
-4.31
2.06
-0.61
0.47
Richibucto*
sounder
CHS Chart 103
-8.46
6.30
0.02
1.22
Richibucto*
490902
*: referring to entire region of Richibucto rather than the sub-divided Richibucto, Aldouane,
and Bedec.
3.2 EELGRASS
The presence of eelgrass in each bay was determined by FPI using the orthophoto frames and used
to define eelgrass polygons, as described in Section 2.3.1. EC/DFO eelgrass classification is based
on a combination of field samples, satellite imagery, aerial photos, and in situ photos summarized
in Table 4 and outlined in Section 2.3.2.
Figures 41 to 52 show FPI and EC/DFO eelgrass coverage for each bay. Comparisons between FPI
and EC eelgrass surface area must be made with caution due to differences in data collection
methods and temporal differences. Additionally, EC PQ classification is somewhat ambiguous, as
it contains Poor Quality or Absent Eelgrass (see APPENDIX A: ), making it difficult to compare to
the FPI classification which indicates only a presence of eelgrass and no quality indicator.
34
35
Estimations of eelgrass surface area per depth interval for FPI and EC classifications are presented
in Table 10. In general, eelgrass surface area decreased with depth after 0 m for all bays and all
classifications. Eelgrass area per depth was proportional to area per depth (Table 6), such that in
Miscou, where the -2 m interval had more surface area than the 0 m interval, there was more
eelgrass at -2 m than at 0 m. Eelgrass extended to the deepest depth measured by the lidar in half of
the bays (Caraquet, Richibucto, Aldouane, and Bedec). Nonetheless, 95% of eelgrass area was
found at depths less than 4m. Table 11 presents secondary calculations for eelgrass coverage;
parameters used are defined below:
Equation 2: Definitions of eelgrass surface area comparisons.
Total Polygon Area=sum of eelgrass area per depth interval defined in Table 1
Whole Area=areas in Table 7, Figure 1
Summed Total = Sum Eelgrass Area per Depth
Difference2 = difference between Total Polygon Area and Summed Total
No Data = % of eelgrass polygon that is No Data= Difference2/Total Polygon Area
%eelgrass = % of bay that has eelgrass = Total Polygon Area/Whole Area
35
36
Table 10: Eelgrass surface area (ha) per 1 m depth interval. See Appendix A for eelgrass
quality definitions.
Eelgrass surface area (ha)
Bay name
Caraquet
Miscou
Tracadie
Tracadie
North
Richibucto
Aldouane
Bedec
Bouctouche
St Mary’s
Depth intervals
FPI
FPI
EC GQ
EC MQ
EC PQ
FPI
1m
59
74
14
30
13
265
0m
733
351
312
59
35
1673
-1 m
499
224
398
82
18
90
-2 m
339
90
503
26
14
2.2
-3 m
275
24
235
0.045
19
0.03
EC GQ
EC MQ
EC PQ
FPI
EC GQ
FPI
EC GQ
FPI
EC GQ
FPI
EC GQ
EC MQ
EC PQ
FPI
13
84
244
5.6
2.9
7.7
0. 90
11
2.8
6.0
0.3
0. 9
3.3
0. 5
361
601
656
552
292
249
143
289
174
733
376
134
149
101
0.38
0. 36
10
257
113
123
43
226
181
392
199
284
27
121
0.07
13
61
13
0. 23
0.055
35
29
35
5.2
516
183
109
36
-4 m
246
8.9
36
-5 m
195
-6 m
132
-7 m
103
20
2.6
6.4
2.4
0.005
1.9
6.4
2.0
0.022
5
0. 7
0.58
0. 55
0. 14
0. 38
0.017
0.15
0. 28
0. 10
1.15
0. 43
3.5
2.6
35
3.6
0.048
-8 m
26
0. 15
0.017
0.11
-9 m
0. 38
37
Table 11: Comparisons of different calculations of eelgrass polygons. Total Polygon
Area=sum of eelgrass area per depth interval; Whole Area=areas in Table 7, Figure 1;
Summed Total = Sum Eelgrass Area; Difference2 = difference between Total Polygon Area
and Summed Total; No Data = % of eelgrass polygon that is No Data= Difference2/Total
Polygon Area.
Bay name
Caraquet
Miscou
Tracadie
Tracadie
North
Richibucto
Aldouane
Bedec
Bouctouche
St Mary’s
Shippagan
Neguac
Cocagne
Tabusintac
FPI
Total Polygon Area
(ha)
4448
FPI
909
771
138
15
EC GQ
EC MQ
EC PQ
2001
233
365
1499
197
131
502
36
235
25
15
64
FPI
2177
2030
147
7
EC GQ
EC MQ
EC PQ
FPI
EC GQ
FPI
EC GQ
FPI
EC GQ
FPI
EC GQ
EC MQ
EC PQ
FPI
EC DE
EC ME
EC TE
EC EE
EC GQ
EC MQ
EC
EC
376
711
955
1164
1030
597
440
609
467
1180
595
1124
853
374
1286
1435
543
293
1295
580
935
1326
375
686
932
885
425
379
187
561
387
1167
580
938
608
370
1
25
22
279
605
218
252
49
80
13
14
186
245
4
0
4
2
24
59
36
57
8
17
1
2
17
29
1
37
2607
Difference2
(ha)
1841
% No
Data
41
Summed Total (ha)
38
3.2.1 Total Eelgrass Area
The largest Total Polygon Area was FPI’s estimate of eelgrass in Caraquet Bay (~4500 ha). The
largest area classified as good quality by Environment Canada (EC QG) was in Miscou (~2000
ha). In Richibucto, Aldouane and Bedec, the FPI and EC GQ estimates for eelgrass agreed very
well, to within ~150 ha. In Miscou and Bouctouche, the FPI eelgrass area was less than the sum of
EC GQ and EC MQ. In Tracadie, the FPI estimate was conducted for the northern and southern
sections of the bay while Environment Canada estimated eelgrass in the northern section only.
This northern section had a greater proportion of poor quality/absent eelgrass than medium or good
quality. In Bouctouche eelgrass classification was dominated by medium quality, followed by
poor quality/absent and good quality.
Eelgrass surface areas for bays not surveyed by lidar are also included in Table 11. In Shippagan
Bay, there was more dense and moderate eelgrass cover (EC DE and EC ME) than thin or exposed
eelgrass (EC TE and EC EE). In Neguac, there was twice as much good quality eelgrass as there
was medium quality.
3.2.2 Overlaps with No Data
The Summed Total defined in Equation 2 and shown in Table 11 is the sum of the eelgrass area at
each depth interval. This value differs from Total Polygon Area because the eelgrass polygons
overlap areas of No Data in the DEM, and the eelgrass per depth interval is not calculated for areas
in the DEM of No Data. Difference2 is the difference between Total Polygon Area and Summed
Total, and %No Data represents what percentage of the eelgrass polygon overlaps sections of the
DEM that are No Data. For Caraquet (Figure 41), Difference2 was equal to ~1800 ha, meaning that
there are 1800 ha of eelgrass polygon that overlaps No Data values in the DEM, representing 41%
of the total area of Caraquet (Table 11). This is relevant because it means that the estimates of
eelgrass area per depth are only as accurate as the DEM. For example, 500 ha of EC GQ eelgrass in
Miscou overlaps with No Data values in the DEM, meaning that an area of 500 ha is potentially
“missing” from the eelgrass surface area per depth calculations.
38
39
Figure 41. Caraquet FPI lidar bottom reflectance and eelgrass boundary. Arrows point to
areas in the DEM with No Data, areas where the eelgrass polygon overlaps the DEM, and
where the eelgrass polygon overlaps an area of No Data in the DEM.
3.3.3 Eelgrass Coverage per Bay
As was done in the case of Richibucto Harbour, the results for Shippagan Bay and Tracadie Bay
are presented in smaller regions (Figure 39). The area of eelgrass presence is calculated from FPI
and EC/DFO data for each bay. The percent of each bay that has eelgrass presence is estimated by
% eelgrass in Table 12. This calculation is not based on the DEM, but is a simple calculation of
Total Polygon Area / Whole Area of a bay (Equation 2). Bedec has the highest coverage, indicating
that 94% of Bedec has some eelgrass presence, according to the FPI classification. St Mary’s has
the least amount of eelgrass presence, according to FPI, at 26%. According to EC classifications,
74% of St-Simon North is covered with either good or medium quality eelgrass, while only 27% of
Shippagan South has good or medium quality EC eelgrass presence.
39
40
Inlet Shippagan
North
South
Miscou Hbr
Shippagan Bay
Caraquet Hbr
North
Tabusintac Bay
New Brunswick
South
Neguac Bay
Tracadie Bay
Aldouane
0
40
80
Bouctouche Bay
Richibouctou
Bedec
kilometres
Cocagne Bay
Richibouctou Hbr
Figure 42: Map showing the division of Shippagan Bay and Tracadie Bay, New Brunswick
into smaller regions (tributaries).
40
41
Table 12: Total eelgrass area in hectares and percent eelgrass coverage for each bay from
FPI (%eelgrass = % of bay that has eelgrass = Total Polygon Area in Table 11/Whole Area in
Table 6) and/or EC/DFO data.
Bay
Caraquet, NB
Miscou, NB
St-Simon Inlet, NB
St-Simon North, NB
St-Simon South, NB
Shippagan South, NB
Tracadie North, NB
Tracadie South, NB
Tabusintac, NB
Néguac, NB
Aldouane, NB
Richibucto Hbr, NB
Bedec, NB
Bouctouche, NB
Cocagne, NB
St Mary’s, PEI
EC/DFO
eelgrass
(ha)
2007
EC/DFO
eelgrass
(ha)
2008
EC/DFO
eelgrass
(ha)
2009
2234
FPI
eelgrass
(ha)
2011
4448
929
648
613
668
761
1087
1418
759
1326
1875
440
1030
467
1719
597
1164
609
1156
935
266
41
EC/DFO
FPI
%eelgrass %eelgrass
2007-2009 2011
69
63
74
70
27
50
43
49
64
65
72
57
47
67
29
66
78
86
73
94
38
26
42
Figure 43: Miscou FPI eelgrass boundary and EC/DFO eelgrass polygons. Background image is
elevation colour-shaded relief (CSR).
Figure 44: Miscou FPI lidar bottom reflectance and eelgrass boundary.
42
43
Figure 45: Tracadie FPI lidar bottom reflectance and eelgrass boundary.
Figure 46: Tracadie FPI eelgrass boundary and EC/DFO eelgrass polygons. Background
image is elevation colour-shaded relief (CSR).
43
44
Figure 47: Richibucto FPI lidar bottom reflectance and eelgrass boundary.
Figure 48: Richibucto FPI eelgrass boundary and EC/DFO eelgrass polygons. Background
image is elevation colour-shaded relief (CSR).
44
45
Figure 49: Bouctouche FPI lidar bottom reflectance and eelgrass boundary.
Figure 50: Bouctouche FPI eelgrass boundary and EC/DFO eelgrass polygons. Background
image is elevation colour-shaded relief (CSR).
45
46
Figure 51: St Mary’s Bay FPI lidar bottom reflectance and eelgrass boundary.
Figure 52: Shippagan Bay EC/DFO eelgrass polygons.
46
47
Figure 53: Neguac Bay EC/DFO eelgrass polygons.
Figure 54: Cocagne Harbour EC/DFO eelgrass polygons.
47
48
Figure 55: Tabusintac Bay EC/DFO eelgrass polygons.
3.3 AQUACULTURE GEAR AND BIOMASS – BOAT-BASED SURVEYS
Aquaculture data from the boat-based surveys is presented in Tables 13 and 14. Table 13 relates to
the 2007 to 2009 period that corresponds to the years in which the EC/DFO eelgrass surveys were
conducted while Table 14 provides data for 2011 only when the FPI eelgrass survey was
conducted.
For the 2007 to 2009 period, biomass was highest for St Simon South (358t) and lowest for
Shippagan South (8t) (Table 13). Biomass per area (total biomass per bay (t) / Whole Area (ha)
(Table 11)), gives an indication of aquaculture density per bay. St Simon South, Bedec and
Aldouane contained the most biomass per area with 0.37, 0.29 and 0.23 t/ha, respectively, while
the other bays held less than 0.1 t/ha.
In 2011, mussel aquaculture in St Mary’s Bay, PEI had an order of magnitude more biomass than
oyster aquaculture in NB bays (Table 14). Of the oyster bays, Bedec had the greatest biomass and
Miscou had none.
Considering only the 2011 dataset, mussel aquaculture in St Mary’s Bay had an order of magnitude
more biomass per bay than the oyster aquaculture bays with 2.85t/ha. Bedec and Aldouane
continued to show high oyster biomass per area in 2011, relative to the other oyster bays studied,
with 0.43 and 0.29 t/ha, respectively. In all years, the highest observed density of oyster bags per
leased area was 1077 bags/ha in Miscou (Figure 57).
48
49
Table 13: Aquaculture lease area, lease area percentage of bay, aquaculture gear area (two
dimensional), bivalve biomass (BM), bivalve biomass per bay area and bags per lease area
for each bay for year corresponding to eelgrass surveys (2007-2009). All oyster aquaculture.
Bay
Miscou, NB
St-Simon Inlet, NB
St-Simon North, NB
St-Simon South, NB
Shippagan South, NB
Tracadie North, NB
Tabusintac, NB
Néguac, NB
Aldouane, NB
Richibucto Hbr, NB
Bedec, NB
Bouctouche, NB
Cocagne, NB
%
2011 (2009) lease
lease area
area
(ha)
(8.33)
16.20
30.08
71.16
13.89
89.56
58.02
75.17
50.99
56.82
56.63
44.56
67.03
0.26
1.56
3.62
7.44
0.49
4.14
1.87
1.97
7.37
3.58
8.73
1.48
3.34
2007-2009
gear area
(ha)
20072009
BM
(t)
0.20
0.05
0.45
1.89
0.04
0.75
0.33
0.79
0.36
0.29
0.70
0.79
0.18
54
85
83
358
8
109
71
199
160
86
190
152
79
20072009
BM/ bay
area
(t/ha)
0.02
0.08
0.10
0.37
0.00
0.05
0.02
0.05
0.23
0.05
0.29
0.05
0.04
2007-2009
bags/ lease
area (ha)
1077
864
458
833
94
201
203
438
520
250
556
563
194
Table 14: Aquaculture lease area, lease area percentage of bay, aquaculture gear area (two
dimensional), bivalve biomass (BM), bivalve biomass per bay area and bags per lease area
for each bay for year corresponding to FPI (lidar) eelgrass surveys (2011). All the New
Brunswick calculations are based on oyster aquaculture and St. Mary’s Bay PEI is based on
mussel aquaculture.
Bay
2011
lease area
(ha)
Caraquet, NB
Miscou, NB
Tracadie North, NB
Tracadie South, NB
Aldouane, NB
Richibucto Hbr, NB
Bedec, NB
Bouctouche, NB
St. Mary’s , PEI
35.68
0
89.56
2.42
50.99
56.82
56.63
44.56
233.00
%
lease
area
0.68
0.26
4.14
0.25
7.37
3.58
8.73
1.48
22.67
2011
gear area
(ha)
2011
BM
(t)
0.17
0.00
0.41
0.01
0.96
0.53
1.03
0.77
0.57
50
0
94
3
198
124
278
209
2833
49
2011
BM/bay
area
(t/ha)
0.01
0.00
0.04
0.00
0.29
0.08
0.43
0.07
2.85
2011
bags/ lease
area (ha)
232
0
174
207
644
361
813
777
1599
50
Miscou
Caraquet
Miscou
Caraquet
Shippagan
Shippagan
St Simon
St Simon
Tracadie
Tracadie
Tracadie South
Tracadie South
Biomass per bay area (t/ha)
Biomass per bay (t/ha)
2011
2007-2009
Tabusintac
Néguac
Tabusintac
Néguac
0.37
0.185
0.037
2.9
1.45
0.29
Aldouane
Aldouane
Richibouctou
Bedec
Richibouctou
Bedec
Bouctouche
Bouctouche
Cocagne
Cocagne
St Marys
St Marys
0
40
0
80
40
80
kilometers
kilometers
Figure 56: Biomass of oysters (or mussels for Prince Edward Island) per bay area (tons per
hectare) for 2007-2009 (left) and 2011 (right).
Miscou
Caraquet
Miscou
Caraquet
Shippagan
Shippagan
St Simon
St Simon
Tracadie
Tracadie South
Tracadie
Tracadie South
Bags per lease area
2007-2009
Tabusintac
Néguac
550
1,700
850
110
170
Aldouane
Aldouane
Richibouctou
Bedec
Richibouctou
Bedec
Bouctouche
Bouctouche
Cocagne
Cocagne
St Marys
0
40
Bags/socks per lease area
2011
Tabusintac
Néguac
1,100
St Marys
80
0
kilometers
40
80
kilometers
Figure 57: Number of oyster bags (or mussel socks for Prince Edward Island) per lease area
(number per hectare) for 2007-2009 (left) and 2011 (right).
3.3.1
Aquaculture Gear and Biomass – Orthophoto Mosaics
Information derived from orthophoto mosaics on aquaculture gear types, line spacing and length
for each bay is presented in Table 15. In the table, buoy/cage spacing means distance between
buoys or cages within a line; line spacing indicates mean distance between lines; line length is
mean line length (m). Orthophoto mosaics showing digitized aquaculture lines and lease areas are
shown in Figure 58 through Figure 63. A total of 3618 aquaculture lines were digitized from the
orthophoto mosaics in 2011. The majority of them held Oystergro cages in NB while they
represented mussel longlines in St Mary’s Bay, PEI. Lines ranged from 28 to 90 m in length
depending on gear type and bay. Line spacing varied between 6 to 30 m but overall averaged 10 m.
For mussel longlines in PEI, spacing varied between 4.5 and 10 m.
50
51
Oystergro cages tended to be evenly spaced and laid out in a grid-like manner, whereas oyster
strings were more irregularly spaced. Floating bags were evenly spaced and did not appear in the
large, grid-like groups characteristic of Oystergro, but appeared in smaller clusters throughout the
bays.
In Caraquet Bay oyster aquaculture is composed of collector lines, single floating bags, oyster
string and dark sea cages, but mostly Oystergro cages.
Single floating bags represent the greatest proportion of aquaculture type in Tracadie Bay (25%,)
while Oystergro cages represent the least (7%). Most of the aquaculture in Tracadie is found in the
northern section of the bay (Figure 59).
In Richibucto Oystergro cages made up 67% of the aquaculture. Seventy percent of the aquaculture
in Aldouane was double floating bags. Oystergro cages are the only aquaculture gear present in
Bedec. In Richibucto and Aldouane the aquaculture is located mostly in the nearshore, but in
Bedec the cages appear to be more centered in the bay (Figure 60).
Bouctouche contains only collector lines and Oystergro cages, clustered in a cove in the north-west
part of the Bay (Figure 61). Oystergro cages make up 86% of the aquaculture.
Longline mussel aquaculture in St Mary’s Bay had evenly spaced lines in grid-like arrays.
Average line length was 122m, but Figure 62 shows that lines ranged from ~ 90m to ~160m.
51
52
Table 15. Aquaculture gear spacing and line length and percentage of total for each gear type
in each bay calculated from orthophoto mosaics.
Caraquet Bay
Buoy/cage spacing
Line spacing (m)
Line length (m)
Percent of lines (%)
Tracadie Bay
Buoy/cage spacing
Line spacing (m)
Line length (m)
Percent of lines (%)
Richibucto
Buoy/cage spacing
Line spacing (m)
Line length (m)
Percent of lines (%)
Aldouane
Buoy/cage spacing
Line spacing (m)
Line length (m)
Percent of lines (%)
Bedec
Buoy/cage spacing
Line spacing (m)
Line length (m)
Percent of lines (%)
Bouctouche Bay
Buoy/cage spacing
Line spacing (m)
Line length (m)
Percent of lines (%)
St Mary’s Bay
Buoy/cage spacing
Collector Floating
Lines
Bag
Double
Floating
Bag
Single
String
Cages
Oystergro
Cages
Dark
Sea
Cages
N/A
15
33
30
1 bag/m
15
83
7
1 cage/4 m
30 - 50
28
18
1 cage/6 m
8
30
30
30 /line
10
90
15
2 bags/m
10
38
25
1 bag/m
10
47
33
1 cage/4 m
6 - 30
38
25
1 cage/4 m
30
38
7
2 bags/m
10
42
15
1 bag/m
8-10
38
17
1 cage/5 m
8-12
65
67
2 bags/m
8-10
43
70
1 bag/m
8
44
4
1 cage/3 m
8-12
55
26
N/A
10-30
55
10
Mussel
socks
1 cage/4 m
8-10
44
100
N/A
N/A
58
13
1 bag/m
variable
38
1
1 cage/4 m
10
39
86
1
sock/0.45m
4.5-10
122
100
Line spacing (m)
Line length (m)
Percent of lines (%)
52
53
3.3.2
Biomass – Comparison Between Boat-based Surveys and Orthophoto Mosaics
Boat-based surveys provided estimates of total aquaculture biomass per bay (Tables 13 and 14).
Totals are based on the number of lines per bay, with lines having a standard number of cages or
bags at 6.04 kg/bag. The number of lines per bay was determined by visual counts in each bay. The
boat-based survey estimates assume that each gear type (e.g. strings, floating bags) has a standard
line length, in contrast to the orthophoto mosaics method, in which biomass depends on line length
for most aquaculture gear types. Differences between boat-based survey Total Biomass and
orthophoto mosaics Total Biomass in Table 16 reflect that difference in calculation, but also reflect
the effectiveness of interpreting and identifying aquaculture gear from orthophoto mosaics.
Boat-based survey Total Biomass values agree well with the totals estimated from orthophoto
mosaics. In every bay except Bedec and Bouctouche Total Biomass was less than boat-based
survey Total Biomass. Assuming boat-based survey Total Biomass is the correct value of biomass
per bay, a value for %Difference2 is calculated as follows:
Difference2 = orthophoto mosaics Total Biomass - boat-based survey Total Biomass
%Difference2 = 100* (Difference2 / boat-based survey Total Biomass)
In Bedec and Bouctouche %Difference2 was less than 1%, and was less than 10% for Tracadie,
Richibucto and Aldouane. In Caraquet and St Mary’s Bay, %Difference2 was 29% and 28%,
respectively. There was an average of 5t difference between orthophoto mosaics Total Biomass
and boat-based survey Total Biomass for the oyster bays.
Table 16: Orthophoto mosaics total biomass per bay, boat-based survey total biomass.
Difference2= orthophoto total biomass – boat-based survey total biomass, and
Difference2=100*Difference2/boat-based survey total biomass.
Bay name
Caraquet
Tracadie
North
Richibucto
Aldouane
Bedec
Bouctouche
St Mary's
Orthophoto Total
Biomass (t)
35.57
94.10
Boat-based Total
Biomass (t)
49.94
97.38
Difference2
(t)
-14.37
-3.28
%Difference2
(%)
29
3
114.54
189.99
280.34
211.13
2116.57
123.77
198.29
278.21
209.02
2833
-9.23
-8.30
2.13
2.11
-814
7
4
1
1
28
53
54
Figure 58: Caraquet orthophoto mosaic (FPI, Sept. 2011), aquaculture leases and interpreted
aquaculture.
54
55
Figure 59: Tracadie orthophoto mosaic (FPI, Sept. 2011), aquaculture leases and
interpreted aquaculture.
55
56
Figure 60: Richibucto, Aldouane, and Bedec orthophoto mosaic (FPI, Sept. 2011),
aquaculture leases and interpreted aquaculture.
56
57
Figure 61: Bouctouche orthophoto mosaic (FPI, Sept. 2011), aquaculture leases and
interpreted aquaculture.
57
58
Figure 62: St Mary’s Bay, PEI orthophoto mosaic (FPI, Sept. 2011) and interpreted
aquaculture.
58
59
Figure 63: Shippagan Bay, NB aquaculture lease areas and EC-interpreted aquaculture
polygons.
3.3.3 Eelgrass and Aquaculture and Depth
Figure 64 through Figure 70 show the bays for which a DEM, eelgrass polygons, and
aquaculture data exist. This excludes Miscou since there was no aquaculture present in Miscou
during September 2011. In Caraquet (Figure 64) and St Mary’s Bay (Figure 70) all the
aquaculture is outside of the eelgrass polygons. In Richibucto, Aldouane, and Bedec the majority
of the aquaculture exists within eelgrass polygons (Figure 66-Figure 68). (Note that for these
bays the only eelgrass classification was good quality.) In Tracadie and Bouctouche, aquaculture
is present only in areas classified as poor quality eelgrass or eelgrass absent (Figure 65 and
Figure 69). (See APPENDIX A: Eelgrass Field Survey Criteria).
The mean depth for oyster aquaculture gear in New Brunswick bays ranged from 0.48 to 2.1 m
(Table 17). Mussel longlines in St. Mary’s Bay PEI were found at an average depth of 4.16 m.
59
60
Table 17. Mean depth (m), depth ranges, and number of points used (n) of aquaculture
leases (gear) for each bay.
Bay name
Tracadie Bay NB
Aldouane Richibouctou NB
Bedec Richibouctou NB
Richibouctou Hbr NB
Bouctouche Bay NB
St Marys Bay PEI
Mean Water Depth (m) at
aquaculture leases*
0.48
1.13
1.75
1.15
2.1
4.16
Range
N
0.04 - 0.90
0.12 - 1.62
1.05 - 2.14
0.48 - 2.51
0.13 - 2.50
1.12 - 8.6
1103
426
1224
745
1214
1932
*According to lidar bathymetry 2011 and only where lidar penetrated to the bottom.
60
61
Figure 64: Caraquet Harbour DEM, FPI eelgrass and interpreted aquaculture.
61
62
Figure 65: Tracadie Bay DEM, FPI and EC/DFO eelgrass, and interpreted aquaculture.
62
63
Figure 66: Richibucto DEM, FPI and EC/DFO eelgrass, and interpreted aquaculture.
63
64
Figure 67: Aldouane DEM, FPI and EC/DFO eelgrass, and interpreted aquaculture.
64
65
Figure 68: Bedec DEM, FPI and EC/DFO eelgrass, and interpreted aquaculture.
65
66
Figure 69: Bouctouche DEM, FPI and EC/DFO eelgrass, and interpreted aquaculture.
66
67
Figure 70: St Mary’s Bay DEM, FPI eelgrass, and interpreted aquaculture.
67
68
4. DISCUSSION
4.1 BATHYMETRY
This bathymetric lidar survey was one of the first in Atlantic Canada. Bathymetric lidar was an
effective method to collect a seamless dataset of land elevations and bathymetry. However, the
technique is known to be limited by water clarity. Evidence of this is seen in Figure 71, where
the lidar did not detect the bottom in 46% of Caraquet Bay (Table 8). The CHS chart values are
relatively shallow (mean value -2.6 m, Figure 71) and suggest that water clarity limited the
penetration of the lidar and not depth. A Secchi depth of between -0.4 m and -0.6 m would result
in the lidar penetrating to -1.2 m, the shallowest CHS value in the DEM No Data region. Figure
71 also shows the orthophoto mosaics showing through the gaps in the DEM, and digitized
aquaculture, which was minimal in Caraquet. It is difficult to conclude from the orthophotos if
the water in the regions of No Data was particularly cloudy.
Figure 71: The turquoise image is the orthophoto mosaic; the greyscale image is the Digital
Elevation Model (DEM) (m, CGVD28). Orange symbols are depth values from Canadian
Hydrographic Service (CHS) chart 402401 adjusted from Chart Datum to CGVD28 as
outlined in Table 3 and Section 2.2.4. Aquaculture is shown as yellow lines. Arrows indicate
No Data values in the DEM.
68
69
Overall, the DEM depths agreed well with the other sources of bathymetry data. The mean ∆z
was -0.58 m and the standard deviation was 1.18 m. The negative mean value indicates that, on
average, the DEM depths were shallower than the depths they were being compared to.
4.2 EELGRASS
Eelgrass coverage was estimated in two different ways in this study: FPI interpreted eelgrass
using individual orthophoto frames, while EC/DFO used a combination of techniques (aerial and
satellite photos, field samples, etc.). The EC/DFO eelgrass had the added value of giving quality
indicators to the eelgrass polygons, in contrast to FPI’s coverages which indicated only eelgrass
presence. Some quantitative comparisons of FPI and EC/DFO eelgrass area found that for the six
bays in which comparisons were possible, half had good agreement between FPI and EC/DFO
eelgrass area, in two FPI estimated a smaller eelgrass area than EC/DFO, and in one bay FPI’s
eelgrass area was larger than EC/DFO’s.
4.3 AQUACULTURE
The interpretation of aquaculture gear from aerial photographs was, overall, successful. The
interpreted aquaculture lines agreed well with the information from boat-based field surveys on
the quantity and type of aquaculture gear present in each lease.
In most cases the lines of buoys were easy to see in the orthophoto mosaics, and in some cases
(Oystergro cages) it was possible to discern individual buoys from the photos. Sub-surface oyster
aquaculture was more difficult to identify, and it was very challenging to identify dark sea cages.
This is a probable explanation for the 30% difference between biomass estimates in Caraquet
Bay, where a large proportion of the aquaculture was known to be dark sea cages. Oystergro was
one of the simplest types of aquaculture to identify. In Bedec and Bouctouche, 100% and 86% of
the total aquaculture was Oystergro, respectively, and estimates for total biomass agreed to
within 1% of field survey estimates. This indicates that the interpretation of Oystergro from the
orthophotos was successful.
Water surface roughness and photo glint introduced great variability into the orthophoto mosaics.
The most severe case of this was in St Mary’s Bay, where it was a simple task to identify
aquaculture in some sections of the bay, but in other sections nearly impossible, even though all
the aquaculture was mussel longlines. This difficulty in orthophoto interpretation is a likely
cause for the 28% difference between the field survey estimate for total biomass in St Mary’s
Bay and the total estimate using the orthophotos. Since the lidar was given top priority for data
collection during the survey flights and the aerial photos were a secondary priority, photos were
taken whether or not ideal aerial photography conditions were present.
69
70
Biomass was estimated from the interpreted aquaculture lines based on information provided by
field surveys on each gear type and the estimated biomass per bag (Comeau, 2013). In some
cases the number of bags or cages per meter was standard for a particular aquaculture gear (e.g.
floating bags), but in some cases an estimate of cage spacing was made using the orthophoto
mosaics (e.g. oyster string cages).
5. CONCLUSIONS
Maps have been constructed for bathymetry, eelgrass coverage and bivalve aquaculture gear.
These maps could be useful for marine spatial planning (ie. site selection for bivalve aquaculture
leases) and decision making in the protection of fish habitat (eelgrass). To better understand the
relationship between eelgrass coverage and bivalve aquaculture, the data resulting from this
study will feed a concurrent study to assess the statistical relationship between the two at a baywide scale (Locke et al. 2014, In Prep.).
5.1 BATHYMETRY
Bathymetric lidar data was obtained for six bays in the southern Gulf of St Lawrence: five in
New Brunswick and one in PEI. Seamless DEMs and CSR maps were constructed for each bay
using the lidar data. Maximum lidar-detected depths ranged from ~-3 m in Aldouane to ~-9 m in
several bays. Using a GIS, the surface area of 1 m depth intervals was calculated for each bay.
The area of depth intervals generally decreased with depth. The percentage of area of each bay
that was not detected by the lidar ranged from 7% in Bedec to 46% in Caraquet. Lidar detection
was limited by water clarity and depth. The lidar-derived depths agreed to within ~1 m with
other sources of bathymetry data such as CHS charts and echo sounder data.
5.2 EELGRASS
Aerial photographs were obtained during the lidar survey flights and used to map eelgrass
coverage in the six surveyed bays. Additional eelgrass classifications, interpreted from a variety
of sources ranging from aerial and satellite imagery to field samples, were produced by EC/DFO
between 2007 and 2009. Eelgrass coverage per bay was calculated in a GIS for the FPI and
EC/DFO eelgrass classifications. The greatest area of eelgrass coverage was the FPI estimate for
Caraquet (~4500 ha); the smallest area of eelgrass was the FPI estimate for St Mary’s Bay
(266ha). The bay with the highest proportion of eelgrass was Bedec (94% coverage according to
the FPI classification); the bay with the least amount of eelgrass was St Mary’s Bay. FPI and
EC/DFO estimates of eelgrass coverage tended to agree spatialy. The surface area of eelgrass in
each depth interval was calculated and showed that eelgrass surface area decreased with depth.
A secondary product of the lidar survey was reflectance data, which provides information on
seabed bottom type. The reflectance data did not prove to be effective t for mapping eelgrass
70
71
coverage on a bay-wide scale, but were found to be useful as a secondary eelgrass mapping tool,
when uncertainties in eelgrass interpretation from aerial photographs were encountered.
5.3 AQUACULTURE
Aquaculture infrastructure was interpreted from FPI orthophoto mosaics generated using aerial
photos captured during the lidar survey. Interpretation depended on photo quality and was
limited by cloud shadow, light glint, and surface roughness. Overall, the interpretation of
orthophoto mosaics was an effective way to digitize and identify aquaculture buoys and
infrastructure. In five out of six bays with oyster aquaculture, the values for total biomass
estimated in this study were within less than 10% of field survey estimates of total biomass per
bay.
The digitized aquaculture lines were used to estimate biomass for each bay based on information
on buoy spacing and biomass per buoy for each aquaculture gear type. Collector lines, double
and single floating oyster bags, sub-surface oyster strings, oyster string cages, Oystergro cages,
dark sea oyster cages, and mussel longlines were identified in the bays. St Mary’s Bay was the
only bay with mussel aquaculture, and it also contained an order of magnitude more bivalve
biomass per hectare and lease area percentage than any of the bays with oyster aquaculture in
New Brunswick.
5.4 RECOMMENDATIONS
Bathymetric lidar bottom reflectance ultimately was not investigated as a tool for mapping
bottom type in this study. To improve upon the utility of intensity data one could develop and
evaluate an algorithm to classify bottom type, specifically eelgrass, following Brennan and
Webster (2006). This method would require the lidar point cloud, and is a method of classifying
surfaces derived from lidar using a segmentation and rule-based object-oriented classification
approach. Point cloud is a vector based structure where each point has its XYZ coordinates and
some attributes.
The gaps in the DEMs caused by the depth penetration limitations of the lidar, particularly in the
channels, could be supplemented with additional bathymetric datasets to produce a complete
seamless terrain model for each bay. The acquisition of the individual orthophoto frames to
construct an improved mosaic may help to resolve issues encountered with shadow and glint in
the orthophotos.
A secondary product of the lidar survey was reflectance data, which provides information on
seabed bottom type. The reflectance data was not used for mapping eelgrass coverage on a baywide scale as it was delivered after the aerial photos were interpreted; however it was found to be
useful as a secondary eelgrass mapping tool, especially when uncertainties in the eelgrass
interpretation from aerial photographs was encountered. To improve upon the utility of intensity
data, it would be beneficial to develop and evaluate an algorithm to classify bottom type,
71
72
specifically eelgrass, similar to what Brennan and Webster (2006) have done for land cover
mapping from lidar data. This method would require the lidar point cloud, and is a method of
classifying surfaces derived from lidar (lidar plus elevation based surfaces) using a segmentation
and rule-based object-oriented classification approach.
Another source of potentially useful remote sensing images for eelgrass distribution mapping is
the Coastal Band (400 – 450 nm) of the Worldview 2.0 Satellite. This band supports vegetation
identification and analysis, and supports bathymetric studies based upon its chlorophyll and
water penetration characteristics (Satellite Imaging Corporation, 2012). At 0.5 m, resolution of
the Worldview 2.0 satellite exceeds the 2.5 m resolution Quickbird satellite images used in this
study.
Recommendations for any future bathymetric lidar projects include having a team in the field
during the lidar surveys to measure Secchi depths and do echo soundings with appropriate tidal
elevation adjustments to accurately validate the lidar elevation data. Having these data available
to indicate the transparency and depth of the survey area would resolve ambiguity surrounding
cases when lidar did not penetrate to the bottom.
6. ACKNOWLEDGEMENTS
This project was funded by the Program for Aquaculture Regulatory Research (PARR). The
authors would like to thank Brad Firth and Dereck Mills at DFO for providing essential
information on aquaculture lease occupations and spacing. A big thanks to Jean-Francois Mallet,
John Davidson, André Nadeau, Jonathan Hill, Aaron Ramsay and Garth Arsenault for providing
eelgrass field data and validation. Many thanks are also given to Jean-Francois Mallet, Venitia
Joseph and Chantal Leger for their contribution in analyzing the numerous hours of underwater
video. We also thank Erica Watson for data compilation and Luc Comeau for his help on oyster
aquaculture biomass analysis.
7. REFERENCES
AMEC Earth and Environmental. 2007. Northumberland Strait Ecosystem Overview Report
(File No. TE61035). Moncton, NB: Fisheries and Oceans Canada. http://www.glf.dfompo.gc.ca/folios/00165/docs/rapport-amec-report-eng.pdf.
Brennen, R. and Webster, T.L. 2006. Object Oriented Land Cover Classification of lidar Derived
Surfaces. Can. J. Remote Sens. 32: 167-172.
Canadian Aquaculture Industry Alliance. 2010. Aquaculture in Canada: Production and Markets.
Retrieved on February 1, 2012: http://www.aquaculture.ca/files/production-markets.php
72
73
Canadian Hydrographic Service. 2012. Tides, Currents and Water Levels. Retrieved on February
1, 2012: http://www.tides.gc.ca/english/WaterLevelsAtYourFingerTips.shtml
Comeau, L.A., Arsenault, E.-J., Doiron S., et Maillet, M.-J. 2006. Évaluation des stocks et
densités ostréicoles au Nouveau-Brunswick en 2005. Rapp. tech. can. sci. halieut. aquat.
2680.
Comeau, L.A. 2013. Suspended versus bottom oyster culture in eastern Canada: Comparing
stocking densities and clearance rates. Aquaculture 410-411 :57-65.
Dealteris, J.T., Kilpatrick, B.D., and Rheault, R.B. 2004. A comparative evaluation of the habitat
value of shellfish aquaculture gear, submerged aquatic vegetation and a non-vegetated
seabed. J. Shellfish Res. 23: 867-874.
http://www.crmc.ri.gov/aquaculture/riaquaworkinggroup/habitat_value_shellfish%20aqu
agear.pdf.
DFO. 2009. Does Eelgrass (Zostera marina) meet the criteria as an ecologically significant
species? (Canadian Science Advisory Secretariat Science Advisory Report 2009/018).
http://www.dfo-mpo.gc.ca/CSAS/Csas/Publications/SAR-AS/2009/2009_018_e.pdf
DFO. 2005. The Gulf of St. Lawrence - A Unique Ecosystem. (Cat. No. FS 104-2/2005).
http://www.glf.dfo-mpo.gc.ca/e0006090
Doiron, S. 2006. Manuel de référence de l’ostréiculteur. Ministère de l’Agriculture, des Pêches et
de l’Aquaculture, Nouveau-Brunswick.75p.
Drapeau, A., Comeau, L.A., Landry, T., Stryhn, H., and Davidson, J. 2006. Association between
longline design and mussel productivity in Prince Edward Island, Canada. Aquaculture
261:879–889.
Flood, M., and Gutelius, B. 1997. Commercial Implications of Topographic Terrain Mapping
Using Scanning Airborne Laser RaDAR. Photogramm. Eng. Remote Sens. 4: 327-366.
Fugro Pelagos, Inc. 2011. Canadian Hydrographic Service- Charting. Airborne Bathymetric lidar
Survey. Report of Survey. (Reference No. – F5129-10-2001). San Diego, CA.
Guenther, G.C., Cunningham, A.G., LaRocque, P.E., and Reid, D.J. 2000. Meeting the Accuracy
Challenge in Airborne lidar Bathymetry. Proceedings of EARSeL-SIG-Workshop lidar,
Dresden, Germany, June 16-17, 2000. 27pp.
Hanson, A.R. 2004. Status and Conservation of Eelgrass (Zostera marina) in Eastern Canada.
(Canadian Wildlife Service Technical Report Series #412). Atlantic Region.
http://www.bofep.org/PDFfiles/Eelgrass_Technica_%20Report.pdf
73
74
Kennish, M.J. and Fertig, B. 2012. Application and assessment of a nutrient pollution indicator
using eelgrass (Zostera marina L.) in Barnegat Bay-Little Egg Harbor estuary, New
Jersey. Aquat. Bot. 96: 23-30.
Lee, K.-S., Short, F.T., and Burdick, D.M. 2004. Development of a nutrient pollution indicator
using the seagrass, Zostera marina, along nutrient gradients in three New England
estuaries. Aquat. Bot. 78: 197-216.
http://www.sciencedirect.com/science/article/pii/S0304377003001530
Liu, X. 2008. Airborne lidar for DEM Generation: Some Critical Issues. Prog. Phys. Geog. 32:
31-49.
Locke, A. 2005. Changes In Eelgrass (Zostera marina) In Southern Gulf Of St. Lawrence
Estuaries. Fisheries and Oceans Canada Retrieved on February 1, 2012:
http://www.glf.dfo-mpo.gc.ca/e0006825
Mahoney, M. 2011.Environment Canada Eelgrass Classification Meta-Data Report. Pers. Comm.
McKenzie, L. 2008. Seagrass Educators Handbook. Retrieved on March 1, 2012:
http://www.seagrasswatch.org/Info_centre/education/Seagrass_Educators_Handbook.pdf
Satellite Imaging Corporation. 2012. Worldview-2 Satellite Sensor. Retrieved on March 20,
2012: http://www.satimagingcorp.com/satellite-sensors/worldview-2.html
Schmidt, A.L., Wysmyk, J.K.C., Craig, S.E. and Lotze, H.K. 2012. Regional-scale effects of
eutrophication on ecosystem structure and services of seagrass beds. Limnol. Oceanogr.
57:1389-1402.
Skinner, M., Courtnay, S.C. and McKindsey, C.W. 2013. Reductions in distribution,
photosynthesis, and productivity of eelgrass Zostera marina associated with oyster
Crassostrea virginica aquaculture. Mar. Ecol. Prog. Ser. 486:105-119.
Statistics Canada. 2010. Aquaculture Statistics. (Catalogue no. 23-222-X ).
http://www.aquaculture.ca/files/documents/StatsCanAquacultureData2010.pdf
Tallis, H.M., Ruesink, J.L., Dumbault, B., Hacker, S., and Wisehart, L.M. 2009. Oysters and
aquaculture practices affect eelgrass density and productivity in a Pacific Northwest
estuary. J. Shellfish Res.. 28: 251-261.
Vandermeulen, H. 2011. An echosounder system ground-truthed by towfish data: A method to
map larger nearshore areas. Can. Tech. Report Fish. Aquat. Sci. 2958: v + 21p.
Washington State Department of Natural Resources. 2011. Eelgrass adopted as indicator of
Puget Sound Health. Retrieved on February 1, 2011:
74
75
http://washingtondnr.wordpress.com/2011/02/18/eelgrass-adopted-as-indicator-speciesof-puget-sound-health/
Wehr, A., and Lohr, U. 1999. Airborne Laser Scanning - An Introduction and Overview. J.
Photogramm. Remote Sens. 54: 68-82.
75
76
APPENDIX A: EELGRASS METHODOLOGY
Field data provided by DFO was used along with aerial and satellite images to classify eelgrass
polygons by EC (CWS) (summarized in Table 4).
1. DFO- EELGRASS FIELD SURVEY CRITERIA
Eelgrass boat surveys were conducted in Miscou (2009), Petite-Tracadie (2009), Nequac (2009)
and St Mary’s Bay (2011) using a predetermined random sampling grid covering the entire bays
and estuaries. The sampling design consisted of one randomly located sample within a 300 by
300 meter grid. Samples which were located over land or over deep water were dropped. The
number of available samples in a particular bay was usually larger than the available resources
necessary for full coverage, so the sample locations were randomly sorted and then chosen
sequentially until the maximum number of samples possible with the available resources was
reached. The density, cover and quality of the eelgrass were noted at every station and various
visual observations were also noted.
2. DFO - EELGRASS TOWFISH SURVEY METHODOLOGY (H. VANDERMEULEN)
Eelgrass boat surveys conducted in Saint-Simon (2007), Tabusintac (2008), Richibucto (2007),
Bouctouche (2009) and Cocagne (2008) were made using predetermined bay-wide transects.
Each transect was surveyed using an underwater video camera system (“towfish”) towed on the
side of a boat running at idle speed. Side scan sonar was also deployed but not used for eelgrass
measurements. The video data was stamped with the running time and precise coordinates of the
towfish. A snap shot of the video was then analyzed at approximately every two of minutes
running time. The density, cover and quality of the eelgrass were noted at every station and
various visual observations were also noted:
1.
2.
3.
4.
5.
6.
Transect partitioned (ex, 1a, 1b, 1c....) depending on the length of each transect.
Video is analysed with three main components: cover, density and quality.
Each of these components has a scale.
Video is stopped at 2 minute intervals.
The frame is then analysed for density and quality.
Cover is an overall data which is analysed throughout the 2 minute interval and not just at the
stopped frame like the other 2 components (e.g.: full eelgrass cover, eelgrass cover w/ some
bare patches, some eelgrass patches and bare sand).
The underwater video and photography analysis has been done by four individuals over a three
year period. Thus, the authors would like to mention that there is a possibility of having
interpretative variability in the data due to the subjectivity of analyzing qualitative data. Ideally,
the analysis should have been done by the same individual to eliminate this possibility.
76
77
Table 18: DFO Habitat field survey criteria.
Density
Cover
Quality
1
Dense (cannot see bottom)
Continuous bed
Green, luxurious eelgrass with
very little or no epiphyte load
(leaves visible, epigrazers
common)
2
Average (variable density,
bottom is visible)
Bed with bare patches
Green blades with medium-high
density epiphyte cover (leaves
barely visible) - blades not pulled
down by this cover
3
Sparse to no eelgrass (may
have algal cover)
No continuous bed,
just clumps of eelgrass
separated by bare
sediment
Majority of leaf surface black or
brown OR epiphyte cover so high
no leaf surface visible (bent over
stunted / fragmented blades
covered in algal material) OR
eelgrass remnants with presence
of benthic Ulva or other algal mats
4
Bare bottom, no vegetation,
clear sand / silt, etc.
No eelgrass
No eelgrass
3. ENVIRONMENT CANADA- CANADIAN WILDLIFE SERVICE (EC-CWS) (MATT
MAHONEY) CLASSIFICATION
1. Good Quality Eelgrass: relatively dense, clean, green blades with minimal epiphytes or algal
growth.
2. Medium Quality Eelgrass: predominantly green blades that may have some epiphyte or algal
growth. These stands can be less than or equally dense as Good Quality Eelgrass, but the best
grasses are certainly not as abundant.
3. Eelgrass Absent/Poor Quality: eelgrass is absent, or if it is present it is typically covered with
epiphytes or other algae or dying or dead.
3.1 EC METADATA
Bouctouche
True colour aerial photography at 57 centimetre resolution was collected on September 2, 2009
by Nortek Resources of Thorburn, Nova Scotia (http://www.nortekresources.com/). Image
77
78
classification was conducted using eCognition Developer v. 8 Software, which first segments the
image into spectrally similar units, which were then classified manually. Additionally, the
Department of Fisheries and Oceans (Gulf Region, Moncton, NB) conducted a visual field
survey in the same field season at 688 sites. Two-thirds of these sites were used to assist in
image classification, while the remainder were used to assess accuracy. Eelgrass was classified
correctly 83.7% of the time in a fuzzy accuracy assessment technique, whereby those classes that
were ‘off’ by one class, e.g. Good Quality eelgrass classed as Medium Quality, were given half
credit towards the overall accuracy. Of 187 sites that were within the classification area, 131
were correct, 51 were "one-off", and 5 were incorrect [(131 + (51/2))/ 187 = 0.837].
Cocagne
Visible ortho-rectified aerial photography was used to classify polygons containing eelgrass in
Cocagne Harbour. Field data for image training and validation were collected along transects in
summer 2008 using a dGPS positioned towfish holding a video camera that was later transcribed
as XY geographic points to describe eelgrass presence and a qualitative description of density.
The area was flown for photography on September 24, 2008. eCognition Developer 8 software
was used to segment the imagery, essentially polygons. Polygons were then classified manually
for the presence of eelgrass. Using field data revealed eelgrass presence to be mapped correctly
87.2% of the time.
Miscou
True colour aerial photography at 57 centimetre resolution was collected on August 20th and
24th, 2009 by Nortek Resources of Thorburn, Nova Scotia (http://www.nortekresources.com/).
Image classification was conducted using eCognition Developer v. 8 Software, which first
segments the image into spectrally similar units, which were then classified manually.
Additionally, the Department of Fisheries and Oceans (Gulf Region, Moncton, NB) conducted a
visual field survey in the same field season at 103 sites. From these sites 70% were used to assist
in image classification, while the remainder were used to assess accuracy. Eelgrass was
classified correctly 96.7% of the time (30/31 = 0.967).
Neguac
True colour aerial photography at 57 centimetres resolution was collected on September 2, 2009
by Nortek Resources of Thorburn, Nova Scotia (http://www.nortekresources.com/). Image
classification was conducted using eCognition Developer v. 8 Software, which first segments the
image into spectrally similar units, which were then classified manually. Additionally, the
Department of Fisheries and Oceans (Gulf Region, Moncton, NB) conducted a visual field
survey in the same field season at 126 sites. Two-thirds of these sites were used to assist in
image classification, while the remainder were used to assess accuracy. Eelgrass was classified
correctly 81% of the time in a fuzzy accuracy assessment technique, whereby those classes that
were ‘off’ by one class, e.g. Good Quality eelgrass classed as Medium Quality, were given half
78
79
credit towards the overall accuracy. Of 39 sites that were within the classification area, 27 were
correct, 9 were "one-off", and 3 were incorrect [(27 + (9/2))/ 39 = 0.81].
Richibouctou
Eelgrass classification in Richibucto Harbour, New Brunswick. Derived from a Quickbird
satellite image collected on August 28th, 2007 at as close to low-tide as possible. Quickbird's
ground resolution is 2.4 m. Classification was objected-oriented using Definiens software.
Accuracy was 81.5%. Data used for accuracy and training was collected along transects using a
differential GPS positioned towfish holding a video camera that was later transcribed as XY
points to describe eelgrass presence.
St Simon
Eelgrass classification in Shippagan Harbour, New Brunswick. Derived from a Quickbird
satellite image collected on July 27, 2007 at as close to low-tide as possible. Classification was
object oriented using Definiens software. Data used for accuracy and training was collected
along transects using a differential GPS positioned towfish holding sidescan sonar, and a video
camera that was later transcribed as XY points to describe eelgrass presence.
Tabusintac
Visible orthorectified aerial photography was used to classify polygons containing eelgrass in
Tabusintac Bay. Field data for image training and validation were collected along transects in
summer 2008 using a dGPS positioned towfish holding a video camera that was later transcribed
as XY geographic points to describe eel-grass presence and a qualitative description of density.
The area was flown for photography on September 24, 2008. eCognition Developer 8 software
was used to segment the imagery, essentially into polygons. Polygons were then classified
manually for the presence of eelgrass. Using field data revealed eelgrass presence to be mapped
correctly at a rate of 77.2%.
Tracadie
True colour aerial photography at 57 centimetre resolution was collected on September 2, 2009
by Nortek Resources of Thorburn, Nova Scotia (http://www.nortekresources.com/). Image
classification was conducted using eCognition Developer v. 8 Software, which first segments the
image into spectrally similar units, which were then classified manually. Additionally, the
Department of Fisheries and Oceans (Gulf Region, Moncton, NB) conducted a visual field
survey in the same field season at 101 sites. Approximately two-thirds of these sites were used to
assist in image classification, while the remainder were used to assess accuracy. Eelgrass was
classified correctly 79.3% of the time in a fuzzy accuracy assessment technique, whereby those
classes that were ‘off’ by one class, e.g. Good Quality eelgrass classed as Medium Quality, were
79
80
given half credit towards the overall accuracy. Of 29 sites that were within the classification
area, 18 were correct, 10 were "one-off", and 1 was incorrect [(18 + (10/2))/ 29 = 0.793].
80
81
APPENDIX B: ADDITIONAL INFORMATION ON LIDAR TECHNOLOGY
The products from a bathymetric Lidar survey can easily be used in collaboration to determine
the boundaries of eelgrass in clear water. Each product is explained below:
•
•
Orthoimagery: Georectified imagery has historically been the most used method
of environmental mapping. Most eelgrass can be directly observed in airborne
imagery; occasionally cloud cover can obscure the scene making interpretation
difficult.
•
Lidar Bottom Rugosity: Also known as ‘bottom roughness’, rugosity is a measure
of the variability of the seabed. At the edge of eelgrass the roughness will
increase, allowing for classification.
•
Reflectance Imagery: The strength of the returned bathymetric laser pulse can be
shown as an image where lighter objects reflect more energy and darker objects
reflect less. Since eelgrass is often darker than the bottom, its boundaries can be
clearly delineated from the lidar reflectance image, and aquaculture farms are also
observable in the imagery. The reflectance image is one the most robust ways of
classifying eelgrass using a Lidar bathymetric system.
Fugro Pelagos attempted to evaluate water clarity of the area surrounding the aquaculture
farms by examining the backscatter of the laser waveform as it passed through the water
column. Unfortunately, due to the shallow depths where the aquaculture farms were
located, volume backscatter values returned contained too much uncertainty to accurately
determine water clarity. This was due to a physical limitation of the technology used by
the survey concerning the pulse duration (‘length’) necessary to transmit the amount of
energy required to penetrate the water column to the design depths and return enough of
that energy for the receivers to detect. In the case of the SHOALS system and others like
it, the pulse length is approximately 5ns in order to transmit the energies necessary to
obtain depths in excess of 50 metres. The distance travelled by light in this 5ns
timeframe is about 1.5 m; fortunately detection algorithms have been refined to allow the
system to resolve the points at which the water surface and sea/lake bed are detected to
within 0.2 m. However, the volume backscatter data which is essential to observe in the
determination of water column information is masked between the surface and bottom
returns in very shallow water, or is at best a very small portion of the returning signal
from which measurements and assumptive reasoning is inconsistent. For future
operations, Fugro Pelagos would suggest the utilization of hyperspectral imagery or new
low-power, short-pulse, high density topo-bathymetry systems to increase the likelihood
of water clarity classification in the shallower margins of the areas under study.
81
82
APPENDIX C: LIST OF ABBREVIATIONS
CD: Chart Datum
CGVD28: Canadian Geodetic Vertical Datum 1928
CHS: Canadian Hydrographic Service
CSR: Color-Shaded Relief
CWS: Canadian Wildlife Service (Environment Canada)
DEM: Digital Elevation Model
DFO: Department of Fisheries and Oceans Canada
EC: Environment Canada
FPI: Fugro Pelagos Inc.
GIS: Geographic Information System
GPS: Global Positioning System
GRS-80: Geodetic Reference System 1980
ha: hectare
IMU: International Measurement Unit
MSL: Mean Sea Level
NRCan: Natural Resources Canada
SWA: Shallow Water Algorithm
t: tons
82
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement