Revision of depth contours and stratification of the West Greenland Bottom Trawl survey for Northern shrimp. Technical Report No. 56

S Technical report no. 56, 2004 Pinngortitaleriffik, Greenland Institute of Natural Resources Revision of depth contours and stratification of the West Greenland Bottom Trawl Survey for Northern Shrimp 2 Title: Revision of depth contours and stratification of the West Greenland Bottom Trawl survey for Northern shrimp Author(s): Kai Wieland and Per Kanneworff Serial title and number: Technical report no. 56 Publisher: Pinngortitaleriffik, Greenland Institute of Natural Resources Date of publication: 1st July 2004 Translation: Sofia Geisler and Per Kanneworff Cover photo: Carsten Egevang ISBN: 87-91214-08-4 ISSN: 1397-3657 Contact address: The report is only available in electronic format. You can download a PDF-file of the report at this homepage http://www.natur.gl/publikationer/tekniske rapporter It is possible to achieve a print of the report here: Pinngortitaleriffik, Greenland Institute of Natural Resources P.O. Box 570 3900 Nuuk Greenland Phone: +299 36 12 00 Fax: +299 36 12 12 E-mail: [email protected] www.natur.gl 3 Revision of depth contours and stratification of the West Greenland Bottom Trawl Survey for Northern Shrimp by Kai Wieland and Per Kanneworff Technical report no. 56, 2004 Pinngortitaleriffik, Greenland Institute of Natural Resources 4 Contents Eqikkaaneq...................................................................................................................................... 5 Resumé ............................................................................................................................................ 5 Abstract ............................................................................................................................................5 Introduction .................................................................................................................................... 6 Material and Methods ................................................................................................................... 6 Data source and selection ................................................................................................... 6 Geostatistical analysis.......................................................................................................... 7 Results and Discussion ............................................................................................................... 11 Bathymetry ......................................................................................................................... 11 Survey stratification .......................................................................................................... 11 References ..................................................................................................................................... 14 Acknowledgements ..................................................................................................................... 14 Appendix 1: Selected input data and computed depth contours ...................................................... 15 Appendix 2: Stratum limits, depth stratification and areas for the shrimp survey in 2004 .......... 23 5 Eqikkaaneq Kitaata sineriaani rejenik aalisakkanillu ukiumoortumik Pinngortitaleriffiup misissuiffigisartagaata annersaani paasissutissat itisuumi uuttortaanermit pisut naammaginartumik tutsuiginassusillit suli pigineqanngillat. Taamaattumik paasissutissat itisuumi kilisannermit aammalu Paamiut misissuinerani angallavigineqartuni pissarsiat immikkut suliarineqaratik katiterneqarput. Nalilersuinermi uani paasissutissanut tunngavigineqartunut tapiliunneqarput qaammataasaq atorlugu itisuumi uuttortaanernit paasisat. Kitaani kilisalluni Pinngortitaleriffiup misissuisarnerani tunngaviusup immap naqqata qaleriiaarneranut tunngasup nalileqqinnissaanut itissutsip isikkuata suliarinissaanut periuseq suut tigussaasut oqimaaqatigiinnerannik oqariartuutitalik atorneqarpoq. Sumiiffik misissuiffigineqartoq tassaavoq avannarpasissutsip 59-p 73-llu akornanniittoq aammalu Qeqertarsuup Tunuani/Vaigatimi sinerissap qanittua. Resumé I hovedparten af det område langs Grønlands vestkyst, som Grønlands Naturinstitut dækker med årlige reje-fiskesurveys, har der ikke hidtil været tilgængelige bathymetriske oplysninger med tilstrækkelig pålidelighed. Dybdedata, opsamlet fra trawlstationer og fra sejlruterne under R/V Paamiuts surveys, er derfor kompileret. Til støtte for disse data er dybdedata fra satellithøjdemåling også medtaget i denne analyse. Geostatistike metoder er anvendt til konstruktion af dybdekonturer med henblik på revision af den stratifikation, der ligger til grund for udførelsen af institutionens trawlsurveys ved Vestgrønland. Det undersøgte område dækker de udenskærs områder mellem 59° og 73°N og indenskærsområdet Disko Bugt / Vaigat. Abstract Reliable information on the bathymetry of the West Greenland shelf has been missing for a substantial part covered regularly by surveys conducted by the Greenland Institute of Natural Resources. Therefore, depth data recorded at trawls stations and along cruise tracks during surveys with RV Paamiut have been compiled. Additionally, depth information derived from satellite altimetry has been included in the analysis. Geostatistical tools were used to provide depth contours for a revision of the stratification applied in the West Greenland Bottom Trawl Survey for northern shrimp. The study area covered West Greenland offshore waters between 59 and 73°N and the inshore waters in the Disko Bay /Vaigat area. 6 Introduction The Greenland Institute of Natural Resources has conducted annual stratified-random bottom trawl surveys in the distribution area of northern shrimp (Pandalus borealis) at West Greenland since 1988. The area coverage of the survey was gradually extended in the early years, and since 1993 the entire West Greenland shelf from 59°15’ to 72°30’N and the inshore waters in the Disko Bay and Vaigat area has been covered. The main target species is northern shrimp and the majority of the trawl stations are allocated to depths between 150 and 600 m according to the depth preference of this species. Depths shallower than 150 m are included in the survey area south of 68°50’N (northern limit of NAFO division 1B) to collect additional information on the distribution of demersal fish species. The survey strata in areas deeper than 150 m correspond to major latitudinal divisions of the West Greenland shrimp fishing grounds (Carlsson et al. 2000). The offshore region south of 69°30’N and west of 48°15’W have been stratified applying four depth ranges: 150-200 m, 200-300 m, 300-400 m and 400-600 m. Available information did not allow a depth stratification in the remaining survey area, and the offshore region north of 69°30’N (west and northwest off Disko Island) and east of 48°15’W (Julianehåb Bay) as well as the inshore area (Disko Bay and Vaigat) were subdivided in a number of small geographical units instead (Fig. 1, Kanneworff & Wieland 2001). The depth stratification of the offshore survey area south of 69°30’N and west of 48°15’W was based on depth contours derived from geological surveys. However, some substantial deviations between these depth contours and the information obtained from the echo sounder during the shrimp surveys were found. This made it regularly necessary to move trawl stations several nautical miles away from the planned position in order to maintain the original depth allocation. In addition to information from trawl stations, continuous depth recordings along cruise tracks have become available from several surveys conducted by the Greenland Institute of Natural Resources with RV Paamiut since 2000. These data have been analyzed in the present study together with results from satellite altimetry in order to provide revised depth contours and to modify the stratification of the shrimp survey. Material and methods Data sources and selection Depth at start and end of trawl operations were extracted from the institute’s database. The data set contained shrimp surveys conducted in the years 1988 to 2003 covering depths from 25 to 600 m and surveys for Greenland halibut in the years 1997 to 2003, which covered mainly depths between 400 and 1500 m in the area between 62°30’ and 72°45’N. Position (GPS) and depth recordings along cruise tracks were made from RV Paamiut during the shrimp surveys in the years 2000 to 2003 and the Greenland halibut survey in 7 2003. However, depth data were not available for the entire survey periods because the echo sounder was switched off or the settings were not monitored properly during steaming in some cases. From the raw data, which were recorded in time intervals of 2 min, unreliable records were removed. False echo soundings were identified in particular belonging to two situations: - constant values occurring over a longer period when the echo sounder was “locked” or the actual depth was out of the selected range; - sudden jumps to a high value with more a less constant increments when the echo sounder lost bottom contact and returned close to the initial value within a short period. The remaining data were filtered to a minimum spatial resolution of 0.1 nautical miles along the cruise tracks. A depth grid with a resolution of about 2 km based on an interpolation of satellite altimetry data (see e.g. www. ngdc.noaa.gov/mgg/bathymetry for details on the method) was downloaded from the internet (http://topex.ucsd.edu/cgibin/get_data.cgi). This data set was compared with the survey observations and satellite altimetry data were removed in all areas in which they were in conflict with the echo sounder values. In the Disko Bay and Vaigat area as well as in the area north of 72°N, which was not covered by the satellite altimetry data, additional geological information was used. The study area was divided into nine overlapping parts (Fig. 1) for practical reasons and maps showing the corresponding data sets are given in the appendix 1. The selected depth data were transferred from the original latitude/longitude coordinates to an UTM projection to allow a geostatistical analysis and area calculations in real distance units with Surfer  (Golden Software, Inc., www.goldensoftware.com). Geostatistical analysis A geostatistical analysis usually involves two steps: a) the identification of the spatial structure, e.g. through the calculation and the fitting of a variogram, and b) the use of this structure, e.g. in the application of a linear method of spatial prediction such as kriging (Chilès & Delfiner 1999). In the present study experimental variograms were computed from the sample points as an average over all directions (omnidirectional variogram) according to the classical formula: () ∑ = ∗−+= )( 1 2 )()( )( 1 5.0)( hN iii xZhxZ hN h γ (1) where N(h) is the number of experimental pairs Z(xi + h) - Z(xi) of data separated by the distance h. A distance lag width of 0.1 nautical miles was used and only the survey observations, i.e. depth from the trawl stations and the cruise tracks, were included in the calculation in all cases while the maximum distance considered varied in respect to the extension of the different domains. 8 Figure 1. Division of the study area and survey stratification (shaded area) used in 2000 (total survey area: 118475 km²). 9 Variogram models were fitted using least squares applying a two-component structure, a nugget effect and a spherical component: ()           > ≤< = + −+= rh rh h cc rhrhcch sn sn 0 0 , ,)/(5.0/5.1 ,0 )( 3 γ (2) where cn is the nugget effect representing unresolved small-scale variation and observation error, cs the amplitude of the spherical component, cn + cs the sill of the variogram that represents the maximum level of variability, and r is the range beyond which the data are no longer autocorrelated. Kriging is a linear estimator accounting for the spatial structure and the geometrical configuration. It is optimal in the sense that it is unbiased and it minimizes the variance of the estimation error. In ordinary kriging, the kriging weights i λ , which are attributed to the data points when calculating the expected value at a grid node and which minimize the variance of the estimation error, are found by solving a system of linear equations: 0 ( ) ( ) whatever 1 jij i j i i x xxx i λγ µ γ λ −+= −    =   ∑ ∑ (3) where µ is a Lagrange parameter constraining the weights to satisfy the condition: ∑i λ = 100% = 1 (4) In the present study, depth grids with a resolution of 0.1 nautical miles were computed using ordinary point kriging based on the fitted variogram models and the entire input data sets, i.e. survey observation and satellite altimetry data, for the respective domains. The depth grids were smoothed over a distance range of 0.5 nautical miles around each grid node with a Gaussian low pass filter and contours of 50, 100, 150, 200, 300, 400, 600 and 1000 m were extracted. The 150 m and 600 m contours were used to redefine the outer limits covered by the shrimp survey and the areas of the strata were calculated corresponding to the four depth ranges used previously, i.e. 150-200 m, 200-300 m, 300-400 m and 400-600 m. The depth contours from the nine different domains were finally transferred back to latitude/longitude coordinates and combined for the entire study area. 16 17 18 19 20 21 22 23 Appendix 2 Stratum limits, depth stratification and areas for the shrimp survey in 2004. 24 72°N 71°N 73°N 56°W58°W60°W 54°W 52°W 62°W Depth (m) U2 U3 50 100 150 200 300 400 600 1000 70°N 71°N 56°W58°W60°W 54°W Depth (m) U1 I2 50 100 150 200 300 400 600 1000 25 70°N 69°N 51°W 52°W 53°W54°W Depth (m) I1 I2 50 100 150 200 300 400 600 1000