The Hudson Bay Lithospheric Experiment - University of Bristol

BASTOW ET AL.: HUBLEThe Hudson BayLithospheric ExperimentI D Bastow, J-M Kendall, A M Brisbourne, D B Snyder, D Thompson,D Hawthorn, G R Helffrich, J Wookey, A Horleston and D Eatondescribe the motivation for – and successful operation of –a remote seismic survey in Arctic Canada.1: Fields of sea ice melt in northeasternCanada’s Hudson Bay in this false-colourAqua MODIS image acquired in July 2003. Theice in this image appears bright turquoise,while liquid water is black. Land is green andbrown, and clouds are white and light blue.Hudson Bay is the second largest bay in theworld, encompassing 1 230 000 km 2 , but isrelatively shallow with an average depth ofabout 100 m. (Jacques Descloitres, MODISRapid Response Team, NASA/GSFC)A&G 6.21

BASTOW ET AL.: HUBLEGeologists can usually interpret the rocksthey encounter on Earth in the light oftectonic and volcanic processes presentlyoperating at the plate boundaries. Thisapproach works well for relatively young rocks(Phanerozoic: younger than 550 million yearsold), but for the older rocks that formed duringPrecambrian times (more than 550 million yearsold), the “plate tectonic” assumption must ultimatelybreak down. Processes operating on theyounger, hotter Earth would have been quite differentto those we see today. Gathering detailedevidence preserved deep within the plates in theancient cores of the continents (“shields”), isthus essential to our understanding of the earlyEarth. This can be achieved using data fromdense seismograph networks, but building andmaintaining them in remote areas is both logisticallyand financially challenging; innovativestation and equipment designs are required todeliver the success enjoyed in gentler climes.The Hudson Bay Lithospheric Experiment(HuBLE), a recent UK–Canadian venture inArctic Canada, has addressed these issues inorder to place fundamental constraints on Earthstructure beneath the Canadian Shield. Theresulting data provide a tantalizing hint as to theprocesses that operated on the youthful Earth.Beneath the shieldsOver the past 25 years seismologists have studiedthe internal structure of the Earth beneathregions of tectonic interest using data from distantearthquakes recorded by dense networks ofseismograph stations. These experiments haveyielded high-resolution images of crust andmantle structure that have advanced considerablyour understanding of the internal structureand dynamics of the Earth. However, asbudgets tighten and the target regions for suchendeavours become increasingly remote, maintainingthis level of success becomes ever morechallenging. Shields, the diamond and mineralrichancient cores of the continents that haveremained stable for billions of years since theirformation, can be particularly awkward customersin this regard. The motivation to understandtheir geology is high but our database ofconstraints remains relatively small. There isan issue of interpretation in shield regions too:most relatively young (

BASTOW ET AL.: HUBLEalso financially and logistically prohibitive, soinnovative station designs are required.HuBLEbatteriespower regulatorseismometerrecording unitWith the goal of constraining better the reasonfor the existence of Hudson Bay, and the Precambrianprocesses that shaped the Canadianshield, the Hudson Bay Lithospheric Experiment(HuBLE) was deployed in the summer of 2007by personnel from the University of Bristol, incollaboration with the Geological Survey ofCanada. Nunavut, the homeland of the Inuit,is the most sparsely populated region ofCanada; centres of population andinfrastructure are few and farbetween. Wherever possible,seismograph stations weredeployed in safe compounds,such as airports and weatherstations with mains powersupply, and in small communitysettlements in Nunavut.Elsewhere, vast tracts of wildernessmeant that remote, independentlypowered recording sites had tobe designed within the financial limitations ofthe project. Transport to these locations was bychartered light aircraft with large tundra tyresto permit landing and take-off from relativelyflat and well-drained glacial deposits (figure 4).Figure 4 shows a completed HuBLE seismographstation in northern Hudson Bay. Eachsite was equipped with a Güralp CMG-3TDbroadband seismometer, recording at 40 samplesper second. Güralp DCM data recorderswere used at the stations, which were poweredby up to six solar panels (providing 100–140 Wsatellite modem4: HuBLE-UK remote station construction. 20–40 W solar panels on a steel frame recharge3 100 Ah batteries that power the seismometer and recording equipment. The GPS antennaprovides continuous accurate timing information. Remote communication with the stations isvia an Iridium satellite modem antenna that was scheduled to operate twice weekly. All fieldequipment is provided by NERC’s seismic equipment pool, SEIS-UK (box 1).‘‘ Theremoteness of theHuBLE network wasa hindrance during thefield campaign, buta blessing for thedata analysispower) and three 100 Ah deep-cycle batteries.Each remote site was equipped with an Iridiumsatellite modem that provided state-of-healthdata from the stations. Data retrieved using themodems included: station up-time; digitizer,seismometer, and recorder baud rates; externalhard disk usage; seismometer levelling information;GPS timing information; power supply.Using modems over the Iridium networkprovides pole-to-pole global coverage of bothshort message and short data burst services.For the CMG-DCM, this allowed the UK basestation to pull off weekly reports of thestate of health of the remote system.Where problems were diagnosed,’’GPS antennasolar panelslow-latency two-way communicationfor reconfigurationof the remote systems wasalso used via simple terminalinteraction. All seismicstation equipment are maintainedby SEIS-UK (see box1), part of the Natural EnvironmentResearch Council’s geophysicalequipment facility in Leicester.Where major physical damage to a station(such as cable damage from polar bear attacks)was deduced from the satellite modem data, theservicing team was able to focus its visits onthe problem stations, leaving those operatingsmoothly for the following year. This strategysaved thousands of pounds in plane chartertime. Arguably the greatest benefit of the satellitemodem system, however, was the abilityto carry out instrument configuration repairsremotely from the UK. In the summers of 2008and 2009, for example, two stations suffered1: SEIS-UKSEIS-UK – the UK’s Seismic EquipmentFacility based within the University ofLeicester – is one of three nodes that makeup the NERC Geophysical EquipmentFacility (GEF). It maintains a large anddiverse pool of seismic instrumentationand associated field equipment for onshorerecording of both earthquakes andcontrolled seismic sources. The equipmentis available for use free-of-charge toUK-based academics via a single loanapplication and subsequent peer review.To date the majority of SEIS-UK projectshave been basedoutside the UK, fromEthiopia to HudsonBay, and haveinvolved substantialcollaboration withnon-UK academics and institutions. Thefacility supports around 12 field projectsa year, providing expertise and trainingin the use of the field equipment andassociated data management systems.A processing system for continuousseismic data is fully supported, includingearthquake detection and location.In-house servers with more than 20 Tb ofstorage provide a rapid and convenientroute to data processing while removingthe need for users to maintain expensiveand time-consuming computing facilitiesof their own. SEIS-UK also manages thearchiving of data at the IRIS DMC fromwhere data become publicly available.winter power-outages from which they could notrecover. The seismometer and recording module(figure 4) had defaulted to different communicationbaud rates but, by logging into the stationremotely, the necessary corrections to therecording parameters were made. An additionalbenefit of the satellite modem systems was theability to unlock and reorient seismometers thathad moved significantly or ceased to operate –common in spring-time when shallow permafrostin the Arctic can begin to melt. Carryingout station repairs in this way, as opposed towaiting until the summer service runs, improveddata yield from the experiment by ~20%. Thecost of running the modems, the equivalent oftwo or three transatlantic flights per year, wasexcellent value for money.Scientific insights from CanadaWhile the remoteness of the HuBLE networkwas a hindrance during the field campaign, itwas a blessing for the data analysis. The qualityof seismic signals from seismic stations installedso close to basement rock, far away fromA&G 6.23

BASTOW ET AL.: HUBLE2: Receiver functions(a)S waveP wavePpPs PsPs Pp Ps+PpSsthree-componentseismic stationMohoReceiver functions are time series, computedfrom three-component seismograms, whichshow the relative response of Earth structurenear the receiver (Phinney 1964). Thewaveform is a composite of P-to-S convertedwaves that reverberate in the structurebeneath the seismometer (figure 5a). Theythus carry valuable information aboutvelocity discontinuities in the Earth suchas the crust–mantle boundary (the Moho).A radial receiver function (figure 5b) candirect P wave(b)PsPpPsPpSs+PsPs5: Receiver function analysis of velocity structure beneath a seismic station. (After Ammon 1991)PsSsbe computed by deconvolving the verticalcomponent seismogram from radial (SV)component (and the same procedure for thetangential component, SH). In the frequencydomain this can be written:H(w) = R(w) / Z(w) (1)where w is the angular frequency 2f.Z(w) and R(w) are the Fourier transformsof the vertical and radial seismograms.H(w) is the Fourier transform of the receiverfunction.3: Shear-wave splittingSeismic anisotropy – the directionaldependence of seismic wavespeed – canresult from the alignment of minerals suchas olivine in the crust and mantle, andthe preferential alignment of fluid. In ananisotropic medium, a shear-wave willsplit into two orthogonal components,one travelling faster than the other. Thisphenomenon is known as shear-wavesplitting. Analysis of shear-wave splittingis common amongst seismologists (e.g.Silver and Chan 1991) because the splittingparameters (the orientation of the fastshear wave) and t (the lag-time betweenthe fast and slow shear wave, see figure6) can be related readily, for example, tomantle flow (e.g. Fouch et al. 2000), thepreferential alignment of magma (e.g.Blackman and Kendall 1997, Bastow et al.2010), pre-existing fossil anisotropy frozendeep in plates (e.g. Helffrich 1995, Bastowet al. 2007), or any combination thereof.sources of cultural noise, has been extremelyhigh. This, in turn, has enabled us to place fundamentalnew constraints on the deep structureof the Earth beneath this region.A receiver function (see box 2) study of crustalstructure (Thompson et al. 2010) providedfresh insight into the processes that shaped andformed Earth’s oldest crust. Formation of theCanadian Shield is likely to have evolved fromprocesses characterized by a hot ductile regimeduring the Paleoarchean, to those more closelyresembling modern day-style plate tectonicsby the Paleoproterozoic. An SKS shear-wavesplitting (see box 3) study of seismic anisotropy(Bastow et al. 2011) provided further evidencein support of this view, with strong anisotropicfabrics preserved deep in the North Americanplate. These recorded the 1.8 billion-year-oldmountain-building event called the Trans-Hudson Orogen that geologists have speculatedwas of similar scale and nature to the currentTibetan–Karakoram–Himalayan orogen ofAsia (e.g. Hoffman 1988, St-Onge et al. 2006).As well as shedding light on the ancient lithosphericprocesses that shaped northern Canada,our seismic data are capable of addressing thegeodynamic puzzle of why Hudson Bay (thelargest intracratonic basin in the world) existsat all. One hypothesis for the existence of theBay is that a down-welling in the underlyingmantle is dragging the Hudson Bay basin downbelow sea level. Using receiver function (box 2)analysis to measure the thickness of the mantletransition zone, however, we have shown thatthere is no evidence for a thermal anomaly –hot or cold – beneath the region. This impliesthat whatever the reason for the existence of theBay, it is almost certainly confined to the uppermantle (Thompson et al. 2011 in press). Ouranalysis of shallow crustal structure beneaththe region also indicates that the Hudson Baybasin may owe its existence, at least in part, tocrustal thinning (Pawlak et al. 2010).The innovative seismograph station designsemployed during the HuBLE experiment showclearly that successful seismic experiments canbe conducted in even the harshest and remotestof environments. The challenge now is thus toexplore seismically Earth’s more remote regions,as well as the sunnier climes where our databaseof constraints is increasingly well established. ●I D Bastow (ian.bastow@bristol.ac.uk), J-MKendall, G R Helffrich, D A Thompson, J Wookey,A Horleston, School of Earth Sciences, Universityof Bristol, UK; A M Brisbourne, D Hawthorn,SEIS-UK, Geology Department, Universityof Leicester, Leicester; D Eaton, Departmentof Geoscience, University of Calgary, Alberta,Canada; D B Snyder, Geological Survey ofCanada, Ottawa, Ontario, Canada.Acknowledgments. HuBLE-UK was supportedby NERC grant NE/F007337/1, with logisticalsupport from the GSC, CNGO, SEIS-UK and FirstNations communities of Nunavut. J Beauchesneand J Kendall provided invaluable assistance in thefield. IB is funded by the Leverhulme Trust.ReferencesAmmon CBull. Seis. Soc. Am. 81(6)Bastow I et al.Geophys. Res. Lett. 34(5)Bastow I et al.Geology 39(1)6: Shear-wave splitting in an anisotropicmedia. When the two pulses reach aseismometer, seismologists can measurethe differential time and polarization of theenergy. These parameters can then be usedto characterize the anisotropy.Bastow I et al.Geochem. Geophys. Geosyst.Q0AB10Blackman D and J KendallPhil. Trans. R. Soc.Lond. 355Cawood P et al. GSA Today 16(7)Eaton D and F DarbyshireTectonophysicsEaton D J et al. EOS Transactions 86Fouch M A et al. J. Geophys Res. 105Furnes H et al. Science 315(5819)Hamilton WGSA Today 13(11)Helffrich G J. Geophys. Res. 100(B9)Hoffman PAnnual Review of Earth and PlanetarySciences 16Hopkins M et al. Nature 456Pawlak A et al. Geophys. J. Int. 184Phinney R A J. Geophys. Res. 69Silver P and G Chan J. Geophys. Res. 96St-Onge M et al. Tectonics 25 (4)Stern R Geology 33(7)Thompson D A Earth Planet. Sci. Lett. 297Thompson D A et al.Earth Planet. Sci. Lett. 3126.24 A&G