Evaluation of arctic multibeam sonar data quality using nadir crossovererror analysis and compilation of a full-resolution data product

Ashton F. Flinders n, Larry A. Mayer, Brian A. Calder, Andrew A. ArmstrongCenter for Coastal and Ocean Mapping, Joint Hydrographic Center, University of New Hampshire, Durham, NH 03824, USA

a r t i c l e i n f o

Article history:Received 10 June 2013Received in revised form7 February 2014Accepted 8 February 2014Available online 21 February 2014

Keywords:Arctic OceanCanada BasinChukchiCrossover analysisMultibeamECS

a b s t r a c t

We document a new high-resolution multibeam bathymetry compilation for the Canada Basin andChukchi Borderland in the Arctic Ocean – United States Arctic Multibeam Compilation (USAMBC Version1.0). The compilation preserves the highest native resolution of the bathymetric data, allowing for moredetailed interpretation of seafloor morphology than has been previously possible. The compilation wascreated from multibeam bathymetry data available through openly accessible government and academicrepositories. Much of the new data was collected during dedicated mapping cruises in support of theUnited States effort to map extended continental shelf regions beyond the 200 nm Exclusive EconomicZone. Data quality was evaluated using nadir-beam crossover-error statistics, making it possible to assessthe precision of multibeam depth soundings collected from a wide range of vessels and sonar systems.Data were compiled into a single high-resolution grid through a vertical stacking method, preserving thehighest quality data source in any specific grid cell. The crossover-error analysis and method of datacompilation can be applied to other multi-source multibeam data sets, and is particularly useful forgovernment agencies targeting extended continental shelf regions but with limited hydrographiccapabilities. Both the gridded compilation and an easily distributed geospatial PDF map are freelyavailable through the University of New Hampshire's Center for Coastal and Ocean Mapping (ccom.unh.edu/theme/law-sea). The geospatial pdf is a full resolution, small file-size product that supports interpretation ofArctic seafloor morphology without the need for specialized gridding/visualization software.

& 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Advances in multibeam echosounding (MBES) and navigationtechnology, along with decreased summer sea ice extent and therecognition of potential economic, scientific and geopolitical advan-tages, have led to increasing acquisition of MBES data in the ArcticOcean over the past decade. These new data have provided criticalinsights into the evolution of the Arctic Basins (Lawver et al., 2011),the nature of deep water circulation (Björk et al., 2007), oceanicmixing processes (Nghiem et al., 2012), and the history of ice in theArctic Ocean (Jakobsson et al., 2010). With the commissioning of theice-breaking vessel United States Coast Guard Cutter (USCGC) Healyin 1999 and its operation as a multibeam sonar-equipped platformfor Arctic science, the quantity of Arctic MBES data has increaseddramatically. A large portion of the MBES bathymetry collected bythe USCGC Healy was done so as part of the United States effort tomap regions beyond the 200 nm Exclusive Economic Zone that may

be considered “extended continental shelf” (ECS) under Article 76 ofthe Convention on the Law of the Sea (Mayer et al., 2010; UNCLOS,1982). The University of New Hampshire's (UNH) Center for Coastaland Ocean Mapping (CCOM) and the National Oceanic and Atmo-spheric Administration–UNH Joint Hydrographic Center (JHC) havetaken a lead role in this effort, with seven dedicated ECS cruisesaboard the USCGC Healy, three in collaboration with the GeologicalSurvey of Canada and the Canadian Coast Guard Ship (CCGS) Louis S.St-Laurent (see http://ccom.unh.edu/theme/law-sea for detailedcruise reports). Much of the data collected by the USCGC Healyand other vessels are now in the public domain, available throughgovernment and academic repositories. The availability of dataallows for the creation of high-resolution MBES compilations.Foremost among these compilations has been the InternationalBathymetric Chart of the Arctic Ocean – IBCAO (Jakobsson et al.,2012). Although IBCAO provides an indispensable representation ofArctic bathymetry, its large scope and incorporation of single-beam/spot-sounding sources inhibit it from preserving the highest resolu-tion of the MBES data – a critical need for detailed interpretation ofECS regions.

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/cageo

Computers & Geosciences

http://dx.doi.org/10.1016/j.cageo.2014.02.0030098-3004 & 2014 Elsevier Ltd. All rights reserved.

n Corresponding author.E-mail address: a_flinders@gso.uri.edu (A.F. Flinders).

Computers & Geosciences 66 (2014) 228–236

We evaluate the quality of MBES data sets collected in theArctic Ocean – essential for their potential use in an ECS submis-sion – and present a subset of these data in a manner thatpreserves the highest level of spatial resolution. This newlycompiled data set, the United States Arctic Multibeam Compilation(USAMBC Version 1.0), with a maximum spatial resolution of 40 m,is available both as a gridded bathymetric data set and a stand-alone geospatial pdf.

2. Methods

2.1. Multibeam data sources

MBES data were compiled from publicly available repositories(Figs. 1 and 2, Tables 1 and A1), specifically the Japan Agency forMarine-Earth Science and Technology (JAMSTEC) which operatesthe R/V Mirai (8 cruises), and U.S. holdings from the National

Canada Basin

Chukchi BorderlandFigure 5

Figure 6

180˚

160˚W

140˚W

120˚W

100˚W

70˚N

70˚N

500 km

−4500 −3000 −1500 0Depth − USAMBC V1.0 (m)

−4500 −3000 −1500 0Depth − IBCAO V3.0 (m)

Fig. 1. A Polar stereographic map (center meridian 01W, true scale 751N) showing the coverage area and bathymetry of the newly compiled United States Arctic MultibeamCompilation (USAMBC V1.0), overlain over the International Bathymetric Chart of the Arctic Ocean (IBCAO V3.0) in gray (Jakobsson et al., 2012). Multibeam bathymetry nadircrossover error analysis for the region is shown in detail in Fig. 2. The majority of data are from dedicated ECS cruises operated in cooperation with the University of NewHampshire's Center for Coastal and Ocean Mapping (CCOM). The two red boxes outline regions of closer detail shown in Figs. 5 and 6. (Inset, left corner) An expanded view ofthe region, with the USAMBC study area outlined in red.

A.F. Flinders et al. / Computers & Geosciences 66 (2014) 228–236 229

Geophysical Data Center (NGDC) for the USCGC Healy (non-ECS, 17cruises excluding transits), R/V Marcus G. Langseth (1 cruise) andthe R/V Nathaniel B. Palmer (1 cruise). The majority of these datawere unprocessed MBES depth soundings. The Ocean MappingGroup at the University of New Brunswick provided processeddata from the CCGS Amundsen upon request (43 cruises). Thesedata were supplemented with processed and cleaned MBES datacollected during ECS dedicated mapping cruises aboard the USCGC

Healy (8 cruises), and publicly available from the University of NewHampshire's Center for Coastal and Ocean Mapping.

2.2. Data processing

The lack of reliable uncertainty estimates for deep-watermapping, combined with the high level of acoustic backgroundnoise while operating aboard an active ice-breaking vessel,

180˚

160˚W

140˚W

120˚W

100˚W

70˚N

70˚N

500 km

1%5%

15%

25%

50%

Err

or a

s %

Wat

er D

epth

0 40 80 120 160 200Multibeam Nadir Crossover Error (m)

−4500 −3000 −1500 0Depth − IBCAO V3.0 (m)

USCGC Healy (ECS-EM122)USCGC Healy (ECS-SB2112)

USCGC Healy (EM122)USCGC Healy (SB2112)

R/V Marcus G. LangsethR/V Nathaniel B. Palmer

CCGS Amundsen (EM300)CCGS Amundsen (EM302)R/V Marai

Fig. 2. A Polar stereographic map showing multibeam bathymetry nadir-depth crossover errors for the USAMBC V1.0, over the Chukchi Cap, Canada Basin and northernportion of the Alaskan continental shelf. Absolute depth differences between overlapping MBES segments appear as colored circles, with the size of the circle proportional tothe difference scaled by mean water depth. MBES coverage for each segment is shown in black under their respective trackline, color coded by data source.

A.F. Flinders et al. / Computers & Geosciences 66 (2014) 228–236230

required a processing approach heavily based on physical inspec-tion of the returned depth soundings. Although automated meth-ods were generally limited, bathymetry filters were applied to alldata to reject depth soundings shallower than 50 m and deeperthan 5000 m, as well as any sounding that had a beam-to-beamslope greater than 851 (CARIS, 2007). Data collected duringdedicated ECS cruises aboard the USCGC Healy (Table A1) wereprocessed at sea in near real-time. Processing involved visualinspection of each sonar swath and removal of erroneous returns(swath-editing), often based on interpretation of seafloor mor-phology. The complete data set was visually inspected, andvariations between overlapping area-based regions were mini-mized by removing conflicting depth soundings (area-based edit-ing) using the Caris HIPS & SIPS processing environment. Cleaneddata were exported from Caris HIPS & SIPS as generic sensorformat (GSF) files.

2.3. Crossover-error analysis

The reliability of all data used in the compilation was assessedby calculating nadir depth differences between crossing multi-beam segments (Fig. 2). A running three-ping average of the nadir-beam depth was used to downsample the cleaned MBES data foreach individual cruise (Eq. (1)) using MBsystem (Caress andChayes, 2006), and crossover error (COE) analyses were performedusing the Generic Mapping Tools x2sys package (Wessel, 2010),described below. Averaging the nadir-beam depth over multiplepings (Eq. (1)) helped ensure that the depth would not be subjectto erratic/erroneous returns in the COE analyses. Similarly, ping-averaging reduced the total number of data points thereby redu-cing total computation time, without a loss of error resolution. Thenumber of pings used in the depth average need not be uniformbetween multiple cruises or multiple depths. As shallower depthsresult in more closely group along-track returns, a depth depen-dent ping-averaging may be more computationally beneficial inthe future. The UNIX time (in integer seconds) of data collectionwas also extracted from the clean MBES data (“U” option), andused in subsequent vessel speed discriminations

mblist �I $input_file_list �OXYZU �P34$output_file ð1ÞThe x2sys package is a set of tools designed to detect intersec-

tions among tracks in 2-D Cartesian or geographic coordinates, andevaluate crossover errors (COEs) at these intersections (Wessel, 2010).The package implements the general line intersection algorithm ofSedgewick (1990) to find crossover locations, and observations atthe intersections were based on a linear interpolation of nearbypoints. X2sys additionally allows for the analysis of COE's todetermine appropriate linear models of systematic correctionsfor each track, and application of these corrections to eliminate

crossover discrepancies from the final 2-D compilations. Theanalysis relies on the premise that a substantial portion of errorbetween intersecting tracks has a systematic origin, which cantherefore be resolved. For MBES data these errors are likely due toerroneous navigation, unaccounted tidal effects, or incorrect soundspeed calculations (causing refraction). Nadir-beam (near vertical)soundings will generally have the shortest range through thewatercolumn, and are thereby affected the least by these errors.While COE analyses have been commonly applied to trackline data(gravity, magnetics, single-beam sonar) its application to swathdata (MBES) has been limited – as there are no longer lines ofintersection but areas of overlap. Extraction of nadir depths fromthe MBES data allowed the direct application of the x2sys COEroutines using the standard available x2sys package (see includedexample code). X2sys was used solely to identify intersecting tracksand calculate nadir-beam COEs at these intersections, and in aneffort to preserve the bathymetric depths, was not used to applycorrections for crossover discrepancies.

As many of the cruises departed from the same locations andtraveled overlapping courses before leaving the continental shelffor deeper water, it was necessary to remove the statistical biastowards the numerous shallow water crossings (Fig. 2). Similarbiases resulting from erroneous navigation information and/ortimes where a vessel was maintaining an approximate fixedposition, causing the shiptrack to wander over its own coursewithin a small area, were also removed. As such, the analyses wereconstrained to crossing MBES segments with nadir depths deeperthan 250 m and at vessel speeds greater than two knots. Analyseswere performed individually for each cruise, to characterize acruise's MBES internal consistency (Table A1), between differentcruises aboard the same vessel with the same MBES system(Table 1), and for the compiled data set as a whole (Table 1).

2.4. Grid compilation

MBES cruises were subdivided into similar data-quality groupsbased on the vessel of operation, the sonar system aboard, andwhether the cruise objective was ECS related (Figs. 2–4,Tables 1 and A1). Separate MBES grids (quality group grids) werecreated for each group using a “swath-angle” beam-footprintbased gridding algorithm (CARIS, 2007). This algorithm assigns aweight to each sounding, with the sounding's contribution to thegrid surface varying with the sounding's grazing angle. Soundingsfrom larger grazing angles (near vertical beams) have been shownto be of higher quality (Calder and Mayer, 2003). This weightingbecame critical in areas with adjacent or overlapping MBESsegments, particularly when soundings were from multiplesources. Beams with a grazing angle between 901 and 751 were

Table 1MBES sources, crossover statistics, and compilation weighting.

Repository Vessel Sonar Cruises Distance (km) Crossings Δ�Nadir (m/% w.d.) sa (m/% w.d.) Weight

CCOM/ECS USCGC Healy EM122 3 29 234 380 5/0.4 13/1.2 1CCOM/ECS USCGC Healy SB2112 5 34 313 971 7/0.6 22/1.6 2NGDC USCGC Healy EM122 1 5461 12 10/1.6 18/2.4 2NGDC USCGC Healyb SB2112 29 99 936 34 071 10/3.0 56/14.7 3NGDC R/V Marcus G. Langseth EM122 1 7077 17 7/1.3 11/1.6 2NGDC R/V Nathaniel B. Palmer EM120 1 3289 35 37/3.9 70/7.0 4JAMSTEC R/V Mirai SB2112 8 49 922 1617 21/4.8 104/24.2 4OMG CCGS Amundsen EM302 21 30 984 2318 2/0.6 5/1.7 2OMG CCGS Amundsen EM300 22 38 530 386 2/1.1 3/1.6 3

Total 91 298 746 40 870 8/2.4 32/11.3

a Calculated as the root of the mean square difference.b Includes available transit-cruises.

A.F. Flinders et al. / Computers & Geosciences 66 (2014) 228–236 231

given a weight of 1.0, with the weight linearly decreasing to 0.01 asthe angle with the sea floor decreased to 151 and below. Thismethod of preliminary grid construction ensured higher weightwas given to soundings from the inner part of a MBES swath ratherthan to outer beam soundings from adjacent segments, regardlessof the cruise from which the data was collected.

Quality group grids were then assigned a quality value (1–4),reflecting the group's overall MBES data reliability and uncertainty(Fig. 3 and Table 1). A quality value of one represented the mostreliable data, four the least reliable. ECS cruises operated aboardthe USCGC Healy were assigned a value of one or two, dependingon the sonar system used to collect the data – one for theKongsberg Maritime EM122 system and two for the older L3SeaBeam (SB) 2112 system. These data represent the most reliable

data available, shown by their low mean MBES nadir crossovervalues (Δ�Nadir). Data from these sources consistently have morethan 50% of the crossover errors with o0:25% the mean nadirwater depth (w.d., Fig. 3). The low uncertainty of the ECS cruises isdue to the at-sea real-time processing and quality assuranceinvolved in their data collection (Fig. 3). A quality value of twowas assigned to all non-ECS data collected aboard the USCGC Healywith its newer EM122 MBES system, as well as data from the R/VMarcus G. Langseth and the CCGS Amundsen with it's KongsbergMaritime EM302 system. Data collected aboard these two vessels,with older MBES systems (SB2112 and EM300, respectively) wereassigned a value of three. The R/V Mirai and R/V Nathaniel B. Palmerboth showed particularly high depth uncertainties (410% w.d.;Table 1), and were assigned a quality value of four.

0 %

20 %

40 %

60 %

Crossover Error (m)

012345

R/V Marai

0 %

20 %

40 %

60 %

0 25 50 75 100 0 25 50 75 100

Crossover Error (m)

012345

Compiled Data Set

0 %

20 %

40 %

60 %012345

CCGS Amundsen (EM302)

0 %

20 %

40 %

60 %

0 25 50 75 100 0 25 50 75 100

012345

CCGS Amundsen (EM300)

0 %

20 %

40 %

60 %

Freq

uenc

y

012345

R/V Marcus G. Langseth

0 %

20 %

40 %

60 %

0 25 50 75 100 0 25 50 75 100

012345

R/V Nathaniel B Palmer

0 %

20 %

40 %

60 %012345

USCGC Healy (EM122)

0 %

20 %

40 %

60 %

0 25 50 75 100 0 25 50 75 100

012345

USCGC Healy (SB2112)

0 %

20 %

40 %

60 %012345

Crossover Error (As % Mean Water Depth)

USCGC Healy (ECS−EM122)

0 %

20 %

40 %

60 %

0 25 50 75 100 0 25 50 75 100

012345

Crossover Error (As % Mean Water Depth)

USCGC Healy (ECS−SB2112)

Fig. 3. Crossover-error (COE) statistics grouped by similar vessel, sonar system aboard, and whether the cruise objective was ECS related. The red histograms show COE's inmeters, while the blue histograms show COEs as a function of mean water depth. The quality group shown in Table 1 were created from these subgroups. The bottom right-hand histogram shows the COE histograms for treating all sources as one data set.

A.F. Flinders et al. / Computers & Geosciences 66 (2014) 228–236232

Quality group grids were merged into a final bathymetriccompilation using a vertical stacking method. The quality valueassigned to each group described the position of a grid in thestacked compilation (one at the surface, four at the base, Fig. 4).The method made use of GMT's “grdmath” routine to iterativelycombine two grids using an exclusive disjunction (XOR) logicaloperation (Wessel and Smith, 1991). The exclusive disjunctioncombined two grids (a primary and secondary), using the cell valueof the secondary grid only if there was no value in the primary gridcell. The operation resulted in a compiled grid where the primarygrid overlaid the secondary grid. The exclusive disjunction was firstperformed on quality group grids three and four. This intermediateproduct was then combined with quality group grid two, and thenin turn quality group one, ensuring only the highest quality sourceavailable would be used in any grid cell (Fig. 4).

The vertical stacking method allowed the highest quality datato be preserved, without contamination from overlapping lowsignal-to-noise or low-resolution MBES soundings (Fig. 4). Similargrids were generated for sounding density and depth standarddeviation. All grids were created using a Polar Stereographicprojection, with a true scale at 751N, and a cell spacing of 40 m.While much of the SeaBeam data does not support a 40 m cellspacing, this oversampling was necessary to preserve other high-resolution data sources, particularly in shallow waters. The finalcompiled gridded data sets were converted to both netCDF gridsand high-resolution geospatial PDFs for distribution.

3. Results and discussion

3.1. Internal consistency through crossover analysis

MBES depth uncertainties (Hare et al., 1995; Lurton andAugustin, 2010) tend to increase in deeper water, and thereforecruises spending more time in deep water have larger crossovererrors and variations. This depth scaling is minimized by lookingonly at nadir MBES depth (opposed to outer beam depths), andfurther mitigated by examining the crossover errors as a functionof water depth (Fig. 3 and Table 1). Low nadir-depth crossovererrors and variations are seen for all MBES groupings, particularlyamong the dedicated ECS cruises, where the mean difference is

NCompiled Grid

Group 4

Group 3

Group 2

Group 1

Fig. 4. Visualization of the Generic Mapping Tools gmtmath exclusive-disjunctionbased vertical stacking method. Group 1 contains the highest quality multibeamdata sources, Group 4 the least.

Fig. 5. A detailed comparison of resolution differences provided by the USAMBC V1.0 (B) compared to IBCAO V3.0 (A) for a subset of the region shown in Fig. 1 (incorporatingmultiple data sources). While IBCAO V3.0 (A) provides an indispensable digital representation of Arctic bathymetry, its large scale and more coarse resolution mask thehigher supported native resolution of the MBES data.

A.F. Flinders et al. / Computers & Geosciences 66 (2014) 228–236 233

less than 1% of water depth (Table 1). Data collected aboard the R/VMirai and R/V Nathaniel B. Palmer showed particularly large nadir-depth crossover errors, and were hence assigned the lowestweighting factor in the compilation. The compiled data set,consisting of 91 cruises and approximately 298 000 km of track-line multibeam data, had a mean nadir-depth crossover error of2.4% of water depth, calculated from more than 40 000 crossings.

3.2. Comparison to IBCAO

The International Bathymetric Chart of the Arctic Ocean (IBCAO)provides an unparalleled portrayal of the entire Arctic seafloor. Itsmost recent iteration (V3.0) takes advantage of many of the recentMBES cruises compiled here (Table A1). While IBCAO Version3.0 uses an improved gridding algorithm, resulting in higher spatialresolution, the final gridded compilation is still limited to 500 mspacing (Jakobsson et al., 2012). While its extensive coverage makesit an indispensable tool for planning ECS mapping cruises, its coarseresolution makes it less useful for direct ECS related interpretation.Figs. 5 and 6 show areas comprised of source data from the R/VMarai and USCGC Healy (SB2112; both ECS and non-ECS), with thearea in Fig. 5 additionally supplemented with a small portion ofdata from the R/V Marcus G. Langseth. Both IBCAO V3.0 and theUSAMBC V1.0 use these same multibeam data sources in theircompilations – with the newer USAMBC V1.0 not containing anyadditional sources than IBCAO V3.0.

Geological features that are lost in the 500 m resolution of IBCAOV3.0 are easily identifiable in USAMBC V1.0 (Figs. 5 and 6). Marinevalleys and sediment channels spanning the shoaler portions of theChukchi Borderland are seen clearly in the USAMBC, while they aresmoothed over in IBCAO (Fig. 5). Similar detail is lost along theslopes of the continental shelf (Fig. 6), where slumped sediment,derived from up slope, has created a pervasive wave-like seafloorfabric. Comparisons have been intentionally made only where bothIBCAO and USAMBC have the same multibeam data coverage,emphasizing the difference in product resolution and not sourcedata availability. Data gaps in the USAMBC (Figs. 5 and 6) resultfrom USAMBC being a data-only compilation, compared to IBCAOwhich interpolates between areas where data is not available.Similarly, as IBCAO creates a low-pass smoothed interpolated sur-face, small discrepancies between adjacent multibeam data trackswill appear less prominent. While USAMBC preserves the high-resolution seafloor, it similarly preserves these discrepancies. WhileUSAMBC will never replace the extent of IBCAO, it does allow formore in-depth geological interpretation of seafloor morphology,making it a highly useful companion to IBCAO.

3.3. Chukchi borderland

The largest potential for a U.S. Extended Continental Shelf in theArctic Ocean is in the area of the Chukchi Borderland. The ChukchiBorderland occupies a roughly rectangular area 600 by 700 km, some4% of the Arctic Ocean (Mayer et al., 2002; Mayer, 2003), juttingnorthward between eastern Siberia and western Alaska, north of theChukchi Sea (Fig. 1). Comprised predominantly of a tightly clusteredgroup of generally N–S trending topographic highs, the area forms anatural prolongation from the Chukchi Shelf north of Alaska (Hall,1990). The area was identified early in the ECS project as an areawhere the existing database of bathymetric data was too sparse tosupport a well-defended ECS submission (Mayer, 2003) and was thusthe focus of significant mapping efforts.

Low nadir-depth crossover errors are seen throughout the ChukchiBorderland, particularly on the eastern high-sloping transitionfrom the relatively shallow Chukchi Cap to the abyssal Canada Basin(o1:5% w:d:, o50 m, Figs. 2 and 3). The low difference in thistransition region is surprising given that five different MBES sources

cross the slope (Fig. 2). Similarly low nadir-depth crossover differencesare seen in the northern-most region of the Chukchi Borderland wherethe MBES compilation consists of only ECS collected data (Fig. 2).

4. Conclusion

Our new multibeam bathymetry compilation provides the high-est spatial resolution currently available for the Arctic seafloor inthe Canada Basin and Chukchi Borderland. As new data becomesavailable it should continued to be incorporated into both full-coverage (e.g., IBCAO) as well as multibeam-only (e.g., USAMBC)data products. Both the gridded compilation of USAMBC and aneasily distributed geospatial PDF are freely available through theUniversity of New Hampshire's Center for Coastal and OceanMapping (http://ccom.unh.edu/theme/law-sea). This pdf is a fullyresolvable, small file-size product that provides easy access forinterpretation of Arctic seafloor morphology without the need forspecialized gridding/visualization software.

Fig. 6. A second comparison of the resolution differences provided by the USAMBCV1.0 (B). The USAMBC high-resolution compilation allows for in-depth geologicalinterpretation of depositional sediment waves.

A.F. Flinders et al. / Computers & Geosciences 66 (2014) 228–236234

Acknowledgments

Thank you to P. Johnson and J. Beaudoin of CCOM for help withdata management and processing. J. Hughes Clark and I. Churchprovided access and support for data collected aboard the CCGSAmundsen. Our continued thanks to P. Wessel and the entire GMTcommunity. A very special thank you to the USCGC Healy for allowingthe authors to sail with them and without whom this work would

not be possible. C.L. Boomer provided critical last minute support.Work funded by NOAA Grant NA10NOS4000073.

Appendix A

Source data and individual crossover statistics are shown inTable A1.

Table A1Source data and individual crossover statistics. An expanded listing of multibeam source data and crossover statistics for individual cruise legs. Crossover statistics werecalculated only for data greater than 250 mwater depth, at vessel speeds greater than two knots. Distances include all tracklines greater than 50 mwater depth. The numberof crossings in an individual cruise has a tendency to be exaggerated when cruise lines are nearly co-located; resulting in multiple crossings. U.S. Extended Continental Shelfspecific cruises are highlighted in bold.

Vessel Cruise Year Sonar Distance (km) Crossings Δ�Nadir s Weight Repository

(m) (% w.d.) (m) (% w.d.)

USCGC Healy HLY07TH 2007 Seabeam 2112 1529 39 7 0.7 8 0.9 3 NGDCHLY0202 3148 192 4 1.2 5 1.4HLY0203 7566 1694 7 1.8 54 9.8HLY0204 5834 18 20 4.3 25 5.6HLY0302 2003 3035 18 9 1.5 13 2.3 2 UNH/CCOMHLY0303 3034 7640 7 1.8 14 7.2 3 NGDCHLY03TD 5442 1 0 0.0 0 0.0HLY0402 2004 8277 637 37 15.8 85 40.7HLY0403 10 867 9851 4 1.3 7 1.8HLY0404 13 888 2033 4 1.3 6 1.8HLY0405 6715 61 7 0.5 11 1.0 2 UNH/CCOMHLY04TG 2259 No crossings 3 NGDCHLY0502 2005 1991 27 11 1.7 18 3.1HLY0503(1) 10 115 470 11 0.8 57 3.6HLY05TC 833 No depths greater than 250 mHLY05TD 1055 No depths greater than 250 mHLY0601 2006 477 No depths greater than 250 mHLY0602 1021 87 41 5.8 94 12.6HLY06TG 4913 No depths greater than 250 mHLY06TH 1139 No depths greater than 250 mHLY0702 2007 374 No depths greater than 250 mHLY0703 9399 308 1 0.2 8 0.6 2 UNH/CCOMHLY07TG 964 No depths greater than 250 m 3 NGDC

USCGC Healy HLY07TH 2007 Seabeam 2112 1000 No depths greater than 250 m 3 NGDCHLY0804 2008 1628 19 4 0.6 5 0.7HLY0805 5717 225 11 0.7 33 1.9 2 UNH/CCOMHLY0806 5509 57 6 0.4 10 0.6 3 NGDCHLY08TH 931 No depths greater than 250 mHLY08TI 1100 No crossingsHLY0904 2009 2653 31 13 2.1 33 5.6HLY0905 9447 71 2 0.2 3 0.4 2 UNH/CCOMHLY09TD 1001 No depths greater than 250 m 3 NGDCHLY09TE 1388 No crossingsHLY1002 2010(2) Kongsberg EM122 9414 41 3 0.2 4 0.3 1 UNH/CCOMHLY1003 5461 12 10 1.6 18 2.4 2 NGDCHLY1102 2011 9672 148 1 0.1 5 0.4 1 UNH/CCOMHLY1202 2012 10 148 147 10 0.7 21 1.9

R/V Marai(3) MR99-K05 1999 Seabeam 2112 1879 20 7 1.4 10 2.1 4 JAMSTECMR00-K06 2000 4663 51 10 2.6 25 5.7MR02-K05 2002 8803 68 10 2.8 33 6.9MR04-K05 2004 8900 67 3 0.7 4 1.0MR06-K04 2006 56 No depths greater than 250 mMR08-K04 2008 11147 86 3 1.1 5 1.8MR09-K03 2009 5450 73 2 0.7 3 1.0MR10-K05 2010 9024 156 3 0.6 6 0.9

CCGS Amundsen (multiple) 2003–2008 Kongsberg EM300 30 984 386 2 1.1 3 1.6 3 UNB/OMG

CCGS Amundsen (multiple) 2009–2011 Kongsberg EM302 38 530 2318 2 0.6 5 1.7 2 UNB/OMG

R/V Nathaniel B. Palmer NBP0304A(4) 2003 Kongsberg EM120 3289 36 36 3.8 69 7 4 NGDCR/V Marcus G. Langseth MGL1112(5) 2011 Kongsberg EM122 7077 17 7 1.3 11 2 2 NGDC

Cruise references: (1) Darby, D., Jakobsson, M., Polyak, L., 2005. Icebreaker expedition collects key Arctic sea floor and ice data. In: EOS Transactions, vol. 86(52). AmericanGeophysical Union, pp. 549–556.(2) Mayer, L.A., Armstrong, A.A., Calder, B.R., Gardner J.V., 2010. Seafloor mapping in the Arctic: support for a potential US Extended Continental Shelf. Int. Hydrogr. Rev. 3, 14–23.(3) Data provided through the JAMSTEC Data Research System for Whole Cruise Information (DARWIN). http://www.godac.jamstec.go.jp/darwin/e.(4) Downey, N.J., Stock, J.M., Clayton, R.W., Cande, S.C., 2007. History of the Cretaceous Osbourn spreading Center. J. Geophys. Res. B: Solid Earth 112(4).(5) Coakley, B., Ilhan, I., 2011. Abstract T33A-2365: Chukchi Edges Project—geophysical constraint on the history of the Amerasia Basin. In: American Geophysical Union Fall Meeting.

A.F. Flinders et al. / Computers & Geosciences 66 (2014) 228–236 235

References

Björk, G., Jakobsson, M., Rudels, B., Swift, J.H., Anderson, L.G., Darby, D.A., Backman,J., Coakley, B., Winsor, P., Polyak, L., Edwards, M., 2007. Bathymetry and deep-water exchange across the central Lomonosov Ridge at 881–891N. Deep-SeaRes. I 54, 1197–1208.

Calder, B.R., Mayer, L., 2003. Automatic processing of high-rate, high-densitymultibeam echosounder data. Geochem. Geophys. Geosyst. 4, 1–24.

CARIS, 2007. HIPS & SIPS 7.0 Users Guide.Caress, D.W., Chayes, D.N., 2006. MB-System: Mapping the Seafloor. ⟨http://www.

mbari.org/data/mbsystem and http://www.ldeo.columbia.edu/res/pi/MB-System⟩.General Assembly of the United Nations, 1982. United Nations Convention on the

Law of the Sea.Hall, J.K., 1990. Chukchi borderland, The Arctic Ocean Region. Geol. North Am. 50,

337–350.Hare, R., Godin, A., Mayer, L.A., 1995. Depth and position error budgets for

multibeam echosounding. Int. Hydrogr. Rev. 72, 37–69.Jakobsson, M., Nilsson, J., O'Regan, M., Backman, J., Löwemark, L., Dowdeswell, J.A.,

Mayer, L., Polyak, L., Colleoni, F., Anderson, L., Björk, G., Darby, D., Eriksson, B.,Hanslik, D., Hell, B., Marcussen, C., Sellen, E., Wallin, A., 2010. An Arctic Oceanice shelf during MIS 6 constrained by new geophysical and geological data.Quat. Sci. Rev. 29, 3505–3517.

Jakobsson, M., Mayer, L.A., Coakley, B., Dowdeswell, J.A., Forbes, S., Fridman, B.,Hodnesdal, H., Noormets, R., Pedersen, R., Rebesco, M., Schenke, H.W.,Zarayskaya, T., Accettella, A.D., Armstrong, A., Anderson, R.M., Bienhoff, P.,Camerlenghi, A., Church, I., Edwards, M., Gardner, J.V., Hall, J.K., Hell, B.,Hestvik, O.B., Kristoffersen, Y., Marcussen, C., Mohammad, R., Mosher, D.,

Nghiem, P., Pedrosa, M.T., Travaglini, P.G., Weatherall, P., 2012. The internationalbathymetric chart of the arctic ocean (IBCAO) version 3.0. Geophys. Res. Lett.39, L12609.

Lawver, L.A., Gahagan, L.M., Norton, I.O., 2011. Palaogeographic and tectonicevolution of the Arctic region during the Palaeozoic. In: Spencer, A.M., Embry,A.F., Gautier, D.L., Stoupakova, A.V., Sorensen, K. (Eds.), Arctic PetroleumGeology, Memoirs 35. Geological Society, London, pp. 61–77.

Lurton, X., Augustin, J., 2010. A measurement quality factor for swath bathymetrysounders. J. Ocean. Eng. 35 (4), 852–862.

Mayer, L.A., 2003. U.S. Law of the Sea Cruise to Map the Foot of the Slope and 2500-m Isobath of the U.S. Arctic Ocean Margin, Barrow to Barrow, Cruise Report.University of New Hampshire, Center for Coastal and Ocean Mapping/JointHydrographic Center, pp. 1–47.

Mayer, L.A., Armstrong, A.A., Calder, B., Gardner, J.V., 2010. Seafloor mapping in theArctic: support for a potential U.S. extended continental shelf. Int. Hydrogr. Rev.3, 14–23.

Mayer, L.A., Martin, J., Armstrong, A.A., 2002. The Compilation and Analysis of DataRelevant to a U.S. Claim Under United Nations Law of the Sea Article 76: APreliminary Report. University of New Hampshire, Center for Coastal and OceanMapping/Joint Hydrographic Center, pp. 1–75.

Nghiem, S.V., Clemente-Coln, P., Rigor, I.G., Hall, D.K., Neumann, G., 2012. Seafloorcontrol on sea ice. Deep-Sea Res. II 77–80, 52–61.

Sedgewick, R., 1990. Algorithms in C. Addison-Wesley, Reading, MA p. 657.Wessel, P., 2010. Tools for analyzing intersecting tracks: the x2sys package. Comput.

Geosci. 36, 348–354.Wessel, P., Smith, W.H.F., 1991. Free software helps map and display data. EOS

Trans. AGU 72 (441).

A.F. Flinders et al. / Computers & Geosciences 66 (2014) 228–236236