slocum gliders: robust and ready

Slocum Gliders: Robust andReady

Oscar Schofield,Josh Kohut, and David AragonCoastal Ocean Observation LabInstitute of Marine and Coastal SciencesRutgers UniversityNew Brunswick, New Jersey 08540

Liz CreedCoastal Ocean Observation LabInstitute of Marine and Coastal SciencesRutgers UniversityNew Brunswick, New Jersey 08540and OASIS, Inc.Lexington, Massachusetts 02421

Josh Graver, Chip Haldeman,John Kerfoot, and Hugh RoartyCoastal Ocean Observation LabInstitute of Marine and Coastal SciencesRutgers UniversityNew Brunswick, New Jersey 08540

Clayton Jones and Doug WebbWebb Research CorporationEast Falmouth, Massachusetts 02536-4441

Scott GlennCoastal Ocean Observation LabInstitute of Marine and Coastal SciencesRutgers UniversityNew Brunswick, New Jersey 08540

Received 6 June 2006; accepted 5 April 2007

• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •

Journal of Field Robotics 24(6), 473–485 (2007) © 2007 Wiley Periodicals, Inc.Published online in Wiley InterScience (www.interscience.wiley.com). • DOI: 10.1002/rob.20200

Buoyancy driven Slocum Gliders were a vision of Douglas Webb, which Henry Stommelchampioned in a futuristic vision published in 1989. Slocum Gliders have transitionedfrom a concept to a technology serving basic research and environmental stewardship.The long duration and low operating costs of Gliders allow them to anchor spatial timeseries. Large distances, over 600 km, can be covered using a single set of alkaline batteries.Since the initial tests, a wide range of physical and optical sensors have been integratedinto the Glider allowing measurements of temperature, salinity, depth averaged currents,surface currents, fluorescence, apparent and inherent optical properties. A command/control center, entitled Dockserver, has been developed that allows users to fly fleets ofgliders simultaneously in multiple places around the world via the Internet. Over the last2.5 years, Rutgers Gliders have logged 27 056 kilometers, and flown 1357 days at sea.Gliders call into the automated Glider Command Center at the Rutgers campus via sat-ellite phone to provide a status update, download data, and receive new mission com-mands. The ability to operate Gliders for extended periods of time are making them thecentral in situ technology for the evolving ocean observatories. Off shore New Jersey Glid-ers have occupied a cross shelf transect and have documented the annual variability inshelf wide stratification on the Mid-Atlantic Bight and the role of storms in sediment re-suspension. The sustained data permits scientists to gather regional data critical to ad-dressing if, and how, the oceans are changing. © 2007 Wiley Periodicals, Inc.

1. INTRODUCTION

For centuries, oceanographers have relied on dataand observations gathered from ships during cruisesof limited duration. This expeditionary research ap-proach has resulted in major advances in understand-ing global ocean circulation, the energy associatedwith mesoscale circulation, plate tectonics, globalocean productivity, and climate-ocean coupling.These and many other successes have expanded ourview of Earth and ocean processes, and have demon-strated a need for sampling strategies spanning tem-poral and spatial scales that are not effectively carriedout using ships. To address this observational gap,the scientific community has consistently called forthe development of the capability to maintain a con-tinuous sampling and monitoring presence in theocean �National Research Council, 2003, Schofieldand Tivey, 2005�.

Our inability to sample the oceans coherently isone of the central problems for oceanography. It canintroduce biases, compromising our understandingof many ocean processes. For example, the role thatlarge storms play in enhancing the productivity in theoligotrophic �nutrient-poor� open ocean was not ap-preciated until they were sampled by fortuitouslytimed open-ocean cruises �Glover et al., 1988� and thedeployment of bio-optical moorings that providemeasurements when ships could not remain at sea�Dickey et al., 1998�. Incorporating storm-inducedchanges in phytoplankton growth helped reconcilethe widely disparate estimates in global ocean pro-

ductivity based on either discrete biology rate mea-surements �Platt, 1984� or bulk chemical budgets�Shulenberger and Reid, 1981; Jenkins and Goldman,1985�. Similar examples exist from all disciplines inoceanography. Therefore, efforts have focused on thedevelopment of a capability to have a sustained andcontinuous presence in the ocean. The relevance ofthis approach is validated by the successes of the fewavailable oceanic time-series data sets that have docu-mented the importance of high-frequency, annual,and decadal processes �McGowan et al., 1998; Karland Lukas, 1996; Michaels and Knap, 1996�.

The central technology for in situ ocean time se-ries data, other than ships, has been moorings whichcan provide high frequency time series; however, thecosts of the moorings and their associated mainte-nance limit the number that can be deployed. Thisconstrains our ability to describe spatial complexity,even when an array of moorings is deployed. For ex-ample, three-dimensional Observation System Simu-lation Experiments �OSSEs� have demonstrated that aprohibitively large number of moorings are requiredto define the cross-shelf fluxes of organic carbon offthe northern coast of California �Mosian et al., 2005�.Additionally, physical-chemical-biological interac-tions are extremely complex and often seem to con-spire to “out smart” even the most experienced ob-servationalist �Figure 1�. Thus, if sampling is theproblem, the solution is to get observations over theappropriate time and space scales. Additionally,given the costs, the observational systems need to be“scalable” to the process of interest �Rudnick and

474 • Journal of Field Robotics—2007

Journal of Field Robotics DOI 10.1002/rob

Perry, 2003�. Mobile platforms are undergoing expo-nential development and are transitioning into obser-vational tools �Rudnick and Perry, 2003�. One autono-mous platform that is rapidly becomingindispensable are Gliders. Gliders, as currently con-figured, were first detailed in Doug Webb’s lab bookon 2/8/86 �Webb, 1986� as a novel instrument ap-proach and publicized in 1989 by Henry Stommel’sview of a futuristic smart fleet of instruments �Stom-mel, 1989�. It has taken some time to bring these con-cepts to reality, yet gliders are steadily earning theirreputation as a high-endurance sensor platform.More importantly, this class of long-range and rela-tively low-cost autonomous underwater vehicle�AUV� is making affordable adaptive sampling net-works a reality. We will review our experience withSlocum Gliders and will demonstrate how they offerthe potential improvement in our capability to ob-

serve the oceans. Additionally, we will also highlighthow Gliders will benefit many different users.

2. THE SLOCUM WEBB GLIDER PLATFORM

A number of different gliders have been developedand are being used by many organizations; however,for this paper we will only discuss the field effortsconducted by Rutgers University �RU� and Webb Re-search Corporation �WRC�. We emphasize that thesuccesses of this group are matched by other groupsat other institutions. Our “take home” message is thatGliders are a robust technology capable of anchoringlarge field campaigns. The Glider used by this groupis the Webb Slocum Glider �Figure 2�.

The Slocum Glider is a 1.8 m long torpedo-shaped, winged AUV. It maneuvers through the

Figure 1. A sea surface temperature �SST� map off the New Jersey coast in May 2006. During this period the LagragianTransport and Transformation Experiment �LaTTE� was conducted. As part of that field effort, a series of moorings �blackcircles� were deployed to capture the nearshore coastal and Hudson Canyon cross shore buoyant plume jets that aretypical for this time of year. The deployment was based on several years of process studies prior to year 2006. The currentsand winds this year, however, resulted in the coastal jet detaching from the coast and passing between the fixed mooringarrays. The jet spread offshore after missing the offshore moorings. This illustrates the potential pitfalls of relying on fixedEulerian grids on continental shelves.

Schofield et al.: Sloccum Gliders: Robust and Ready • 475


ocean at a forward speed of 20–30 cm/s in asawtooth-shaped gliding trajectory, deriving its for-ward propulsion by means of a buoyancy change andsteering by means of a tail fin rudder. The altimeterand depth sensor enable preprogrammed samplingof the full water column. The primary vehicle navi-gation system uses an on-board GPS receiver coupledwith an attitude sensor, depth sensor, and altimeter toprovide dead-reckoned navigation, with backup po-sitioning and communications provided by an Argostransmitter. Two-way communication with the ve-hicle is maintained by RF modem or the global sat-ellite phone service Iridium. All antennas are carriedwithin the tail fin that is raised out of the water whenthe vehicle is commanded to surface at some prede-termined interval. Operational endurance, utilizingalkaline batteries, is 25 to 50 days, depending on sen-sor payload and sampling regimes. Horizontal dis-tance traveled averages 24 km per day. The vehicle isoperational in 5 to 200 m of water depth and can beoptimized for 30, 100, and 200 m operation with se-lect gearboxes. Newly developed is the Slocum 1 km�1000 m depth rated� vehicle which is still undergo-ing its initial testing phase. A more complete descrip-tion of this vehicle class, including other forms, canbe found in Davis et al. �2003�.

The mission duration of a glider is largely a func-tion of the number of sensors and the water depth.The largest power drain in the glider involves the op-eration of the pump and, therefore, the battery life isshortest in shallow seas. Despite the shortened bat-tery life, deployments last over three weeks, provid-

ing the scientist usually several thousand verticalcasts. The increase in data quickly justifies the costs ofmaintaining Gliders for sustained observations �Fig-ure 3�. The operational costs for Gliders include tech-nician time, costs for deployment/recovery, batteries,and Iridium phone charges. Based on standard dailycosts for a range of research vessels �deep water, me-dium, small coastal vessel�, the operational costs ofGliders are economical �Figure 3�. The typical costs ofoperating the deep-ocean and coastal class researchvessels exceed the cost of operating single glider de-ployed for a full multi-week mission. The costs ofsmaller research vessels exceed a glider after threedays. One technician can operate several gliders sothe increased costs associated with operating mul-tiple gliders reflect increased deployment/recoverycosts, batteries, and Iridium charges. Given this thecosts of medium research vessel will exceed operat-ing a fleet of six gliders in about four days. Gliderswill never replace ships, but populating the oceanswith Gliders will allow ships to use their time wiselyas they will know when and where to sample theocean. This will allow the ship time to be used tospend its time at sea testing/deploying new instru-ments and conducting experiments.

Figure 2. Two Slocum Gliders cruising underwater off-shore Florida.

Figure 3. The costs in United States dollars for maintain-ing ships and Gliders at sea. Ship costs represent the av-erage daily charge which varies with ship class based onaverages in the year 2005. The glider costs include the ex-penses of deploying, maintaining, and recovering a Gliderat sea.



The value of the Glider surveys will increase asthe sensors available for Webb Gliders expands. Themain bottleneck for integrating sensors is minimizingtheir size and power consumption. Several physical�temperature, conductivity� and optical measure-ments �backscatter, chlorophyll a, and colored dis-solved organic fluorescence� are now standard mea-surements. These sensors are now complementedwith a range of optical sensors which include inher-ent optical sensors �http://www.wetlabs.com�, mul-tispectral radiometers �http://www.satlantic.com/�;and hyperspectral absorption meters �Kirkpatrick etal., 2003; Schofield et al., 2007�. Taken together, back-scatter and attenuation measurements provide infor-mation on the concentration and composition of theparticles present �Boss et al., 2001; Twardowski et al.,2001�. Measurements of absorption provide informa-tion about phytoplankton taxa, detritus, and coloreddissolved organic matter �Chang & Dickey, 1999,Schofield et al., 2004, Glenn et al., 2004�. These con-centration measurements are complemented withrate measurements provided by oxygen sensors. Ad-ditionally, a kinetic chlorophyll a fluorometer is beingminiaturized to allow the glider to make measure-ments of the photosynthetic efficiencies which pro-vides information on the physiological status of phy-toplankton �Falkowski and Kolber, 1993, Kolber andFalkowski, 1993�. This is not an exhaustive list of theavailable sensors that have been flown on Webb Glid-ers by other research groups. Regardless, the largenumber of sensors and their associated power re-quirements is beyond the capacity of a single gliderand, therefore, fleets of gliders will need to be coor-dinated to provide a complete biogeochemical pic-ture of the ocean.

3. COMMAND AND CONTROL „C2…

Operating a fleet of gliders necessitates an automatedcommand and control �C2� system in order to opti-mize glider missions to resolve the temporal and spa-tial patterns of the process of interest. This requiresthe C2 to be flexible and adaptable as the environ-ment is constantly evolving. We have been construct-ing a C2 system for a fleet of Webb Gliders; however,the system is scalable to allow the incorporation of anumber of data inputs, allowing the fleet to make in-telligent goal oriented decisions that feeds back intodynamic adaptive resource allocation �Figure 4�. Thesoftware package allows information from a scientist,

the glider itself, other sensing systems such as highfrequency radar, satellites, or additional gliders, tooptimize a particular glider’s flight characteristics orwaypoints. New mission directives are automaticallyuploaded to the glider during surfacing and theglider begins its new sampling regime or waypointbearing. Optimization can be done for features like,but not limited to, currents, tides, thermoclines, andhaloclines. Deployments can also allow ground-truthing of satellite imagery. Data are automaticallypulled from the vehicle and made available for webbased presentation.

Dockserver, written in Java, is a self deployedLinux based communications center. It handles all ve-hicle traffic and acts as a data repository �Figure 4�.When a glider surfaces, it calls in via RF modem orIridium, is answered by the Dockserver that is incharge of logging the surface dialog of the instru-ment, and can perform any number of scripts thatmay be waiting in prioritized mailboxes. The glideritself is menu driven in its surface dialog. Thesescripts, written in one’s language of choice, are typi-cally text recognition drivers that engage the glider toperform particular functions, such as requesting adata transfer, uploading new mission parameters, orhold the glider at the surface for direct user interac-tion. Inherent is the ability to send email notificationof actions taken or specific text strings read. The sys-tem can handle multiple vehicles in separate win-dows that one can toggle between. Given the appro-

Figure 4. The Glider thought process including the on-board and onshore components that allow for remotecontrol.



priate level of access, one can engage the Dockserverfrom any internet hub in the world and multiple ter-minals can view the process simultaneously. This hasallowed WRC and COOL to collaboratively fly Glidermissions despite that the groups are in different lo-cations.

The data retrieved from the glider, as well as thesurface dialog, is available for file transfer protocol�ftp� for data visualization. A health monitor parsesengineering data from the surface dialog text that aglider sends and publishes a web-based GUI inter-face, providing the first layer of data visualization of

Figure 5. The web based control screen that monitors glider health and allows the user to plot new waypoints that willbe uploaded to the Glider when connected to Dockserver.

Figure 6. An alongshore Glider deployment offshore New Jersey during the Lagrangian Advection Transport and Trans-formation Experiment �LATTE�. The Glider was directed to zig-zag in and out of the southerly flowing buoyant coastal jet�delineated by the low salinity values�. The depth averaged and surface currents estimated by the Glider are alsopresented.



environmental and engineering data. Shown are suchitems as battery voltage, vehicle map position, inter-nal vacuum, and depth-averaged currents �Figure 5�.The data files themselves are transferred in a compactbinary format converted and renamed by the Dock-server and then available by ftp to any user. Here thedata may be quality controlled and are served to theweb in various forms. In addition, the data may berewritten into additional forms for assimilation intonumerical models �OI assimilation� as has been doneon a regular basis by the Naval Oceanographic Officeduring field exercises in 2005 and 2006. This modulealso provides an archive link to recall previous data.These capabilities make the Glider C2 an integralcomponent of the operations command center �Fig-ure 4�. All of the data collected are presented andfreely available to the public via web access. Giventhe existing basics of C2 provided, Rutgers has ap-plied tools to enable several forms of adaptive sam-pling and dynamic control. The most recent case was

during the Langragian Transport and TransformationExperiment �LaTTE�, which was focused on charac-terizing the flow of the Hudson river out onto theMid-Atlantic Bight. The river plume is a highly dy-namic feature that is difficult to track. Gliders wereused to track the location of the plume neashore. Us-ing the Dockserver mailbox system, glider salinitydata was used to assess whether the glider was in orout of the plume and then new way points were au-tomatically provided by either by the Glider analyz-ing its own data or by accepting direction from shore-side scientists. The goal was to develop the ability ofthe Glider to move in and out of the low salinityplume as it transected south along the coast of NewJersey �Figure 6�.

4. GLIDER DEPLOYMENTS

With the web-based C2 and the global Iridium cov-erage, it is now possible to deploy and remotely con-

Figure 7. A map of where Gliders have been deployed globally by the Rutgers group since October 2003. This footprintwill be expanded in the coming months with deployments planned for Antarctica, Alaska, and the Caribbean.



trol fleets of Gliders all over the world �Figure 7�. Wehave deployed Gliders over the last three years span-ning waters offshore Australia, Europe, NorthAmerica, and Antarctica. In all these deployments theC2 was always maintained from the COOL opera-tions center in New Jersey. The Iridium communica-tion also allows shore-side scientists to maneuverGliders through treacherous costal conditions whichinclude areas of fishing debris, ship traffic, and moor-ings �Figure 8�. During the last three years, our grouphas flown 99 missions, which represent over 27 056kilometers and 1357 days at sea. This effort representsa single University group, and given the Glider’s C2capabilities it represents a technology that can be costeffectively scaled up by larger institutions and/or

federal agencies. To illustrate some of the Glider re-sults, we will highlight the “Endurance Line” whichis a time series transect being occupied by the WebbGlider for the last three years.

Rutgers initiated a long term time series to moni-tor the hydrography of the mid-Atlantic Bight usingWebb Gliders that transect from the nearshoreelectro-optic sea floor cabled observatory �LEO� tothe edge of the continental shelf ��150 km offshore,Figures 9 and 10� in October 2003. During the periodthe gliders have traveled 10 090 km on this transectline and logging 487 days at sea offshore New Jersey.The interannual data shows the erosion of late au-tumn shelf stratification by storms and surface cool-ing. This autumn period is followed by a rapid cool-ing during the month of December. The wintermonths �November through April� are associatedwith a well-mixed water column with the MAB coldpool representing one of the warmer water masses onthe shelf. Numerous storms keep the water columnwell-mixed in the winter and associated with thesestorms are dramatic benthic resuspension events�Figures 10 and 11�. These resuspension events aresome of the most prominent particle transport eventson the shelf and would not be visible by satellites asthey occur during cloudy stormy atmospheric condi-tions. The timing and spatial extent of these resuspen-sion events suggest both storm induced �Styles andGlenn, 2002� and buoyancy instabilities �Gargett etal., 2004� contribute to the sediment resuspension.These spatial physical and bio-optical time series onthe MAB will provide critical spatial data in a sus-tained manner which data will be critical to under-standing the processes underlying the changing wa-ter masses in the MAB �Mountain and Taylor, 1998;Mountain, 2003�.

The success of maintaining the Glider time serieshas convinced operational entities within the state ofNew Jersey that these platforms are robust and canserve their operational needs. The New Jersey StateDepartment of Environmental Protection and UnitedStates Environmental Protection Agency have pro-posed to purchase two Webb Gliders in 2006 whichwill be dedicated to monitoring ocean water qualityfor the State of New Jersey. The MAB is experiencingsignificant anthropogenic nutrient inputs �30% of theU.S. population lives along its coastline� associatedwith runoff from land, direct ocean discharges fromsewage treatment plants along most of coastal NewJersey, atmospheric deposition, and estuarine exportassociated with large urban centers. While the initial

Figure 8. Glider deployments in May of 2006 offshore thecoast of New Jersey. The red lines indicate ship traffic inand out of New York Harbor. Red boxes indicate zones offishing debris and old wrecks which must be avoided. Theblack circles indicate moorings. The blue and red blacklines indicate the paths taken by two gliders that weredeployed during the LaTTE field effort. Green triangle in-dicates the starting position of the Glider, and the red tri-angle indicates the final point of the glider mission when itwas recovered. Glider progress was continually updatedby shore-side scientists.



Figure 9. A spatial time series of temperature offshore the coast of New Jersey that has been occupied since October 2003.The glider transects illustrate the Gliders capability to provide a sustained presence on the continental shelves.

Figure 10. A spatial time series of optical backscatter offshore the coast of New Jersey that has been occupied sinceOctober 2003.



ideas focused on anthropogenic loading of organicmaterial, recent data suggests that complex interac-tions between the seafloor topography and summerupwelling also contribute to the low dissolved oxy-gen �DO� along the NJ shore along the Southern coastof New Jersey �Schofield et al., 2002, Glenn andSchofield, 2003, Glenn et al., 2004�.

Upwelling along NJ is a complex interaction ofatmospheric forcing, bottom topography, local �tensof kilometers� and mesoscale �hundreds of kilome-ters� circulation �Song et al., 2001�. While upwellingand oxygen depletion appears to be spatially linked,the sequence of events that drive the intensity of up-welling and the corresponding recurrent low dis-solved oxygen �DO� zones is unclear. This will re-quire a regional perspective. Water quality samplingby the state of New Jersey is designed to resolve near-shore anthropogenic driven declines in low DO usingfixed time and location station sampled by ships andhelicopters. These surveys are not designed to resolve

contributions and the interactions between the up-welling and anthropogenic driven declines in lowDO. This requires an enhanced in situ spatial presencewhich can be filled by Webb Gliders capable oftransecting along the coast of New Jersey twice amonth. These surveys will fill the spatial gaps in thedata and will provide real-time data back to shorethat can be used to direct the ship based sampling.The transition of Gliders to fulfill operational needsof state and federal regulatory agencies represents thebest evidence that these platforms are robust.

5. PROBLEMS ENCOUNTERED DURING GLIDERMISSIONS

To date, Rutgers has flown 99 glider missions aroundthe world with each mission presenting its own chal-lenges and surprises. When a Glider encounterstrouble, it initiates an automated mission abort. Au-

Figure 11. Resuspension of particles during the passage of the remnants of Hurricane Ivan as it passed over the Glider.�A� The depth averaged and surface currents calculated from the Glider flight path. �B� The extreme stratification of theMid-Atlantic is not disrupted however there is sufficient energy to lift particulate matter from the sea floor to thepycnocline. The backscatter data at 532 nm is proportional to the particulate concentration in the watercolumn. Thebackscatter to attenuation ratio provides information on both the composition and particle size distribution of the materialpresent. The high values present during the storm would suggest that the resuspended material has significant contribu-tions of inorganic particles.



tomated mission aborts are set up in such a way thatthe severity of the problem determines the glider’s re-sponse. The most common aborts are relatively mildand only briefly interrupt the mission. These includeglider stalls, behavior errors, and communication in-terrupts. In the event of a mild abort, the glider startsa new segment of its user-defined mission and con-tinues toward the desired waypoint. On its next sur-facing, the abort cause and glider status are sent to thepilots over the iridium link. At that point the pilotscan determine their next course of action. Less com-mon are the more severe aborts like device errors andleak detects. Device errors are commonly due to mo-tor or pump malfunctions �fin motor moving toslow�, low battery voltage, or sensor dropouts �badaltimeter hits, pressure sensor faults, etc.�. Leaks areless common now than in the initial years because ofadvances in several components of the glider includ-ing the air bladder. If one of these more severe abortsoccur, the glider immediately breaks out of the mis-sion and heads to the surface to notify the pilots. Atthat point the pilots on shore can download engineer-ing data to determine the severity of the problem andpotentially get a fix or setup for a recovery. In our ex-perience, it is crucial to be able to adjust the gliderbehavior or mission based on device problems thatmay occur. With this flexibility a pilot can adjust theglider’s behavior so that it will either continue on itsmission or, if flight is not possible, remain in contactuntil the next possible recovery window. In order tominimize the occurrence of these types of aborts, anextensive checkout lists has been constructed to en-sure gliders are properly ballasted and prepared foreach mission. These sheets verify that all communi-cations systems are operating, the pumps and motorsare moving without faults, and that software versionshave been tested and approved for use in the field.During the summer of 2006 we flew over 20 missionsin the Atlantic and Pacific Oceans. No glider was de-ployed without first satisfying all the criteria on thecheckout sheet.

Of the 99 missions we have conducted, a total offour vehicles have been lost. These losses were due tosituations beyond the control of the glider software orhardware. Three of the vehicles were lost in severeweather, hurricanes and tropical storms in which thevehicles were thrown into nearby rocks and obstruc-tions. Another was lost in a shipping lane markingthe entrance and exit into New York Harbor. Giventhis, we now get as much environmental informationabout a planned deployment location prior to de-

ployment, especially in coastal waters. These includeknown stationary hazards like surface and sub-merged buoys, shipping lanes, gas, and oil rigs. Also,we familiarize ourselves with tidal bore as well as ar-eas in which there is a high probability a glider willencounter a surface ship, including AIS data streamsare used. Using these methods, more recent missionsflown in tights spots near the entrance to New YorkHarbor and Liverpool Bay and flights into TropicalStorms Ernesto and Beryl have resulted in unprec-edented oceanographic datasets with no loss ofequipment.

6. THE FUTURE

Internationally there is call for building an integratedglobal ocean observing system to enable both scien-tific research and the needs of society. Much of theplanning for these networks, which began almost adecade ago, largely relied on fixed moorings and/orsea floor cabled networks; however, these initialplans are now rapidly incorporating Glider technol-ogy to provide a regional spatial footprint �Figure 12�.

Figure 12. A sea surface temperature satellite image forthe Eastern seaboard. A proposed ocean observatory net-work for the East Coast United States to study the interac-tion between the Gulf Stream and the Labrador Currentsystems. The lines with arrows indicate proposed Glidertransects that will be occupied for time series measure-ments. The red lines indicate the Glider lines that are nowgoing to be occupied by research institutions in Massachu-setts and New Jersey.



Given the recent advances, we believe that oceanog-raphy is on the verge of fulfilling the vision laid outby Stommel over a decade ago. The ability of ocean-ographers to maintain a sustained spatial subsurfacepresence will, in our opinion, usher in a new era forthe marine sciences.

ACKNOWLEDGMENTS

The evolution of the Glider technology could nothave evolved with out the generous support of theOffice of Naval Research and the National OceanPartnership Program. Additionally, the support ofthe National Science Foundation Coastal Ocean Pro-gram’s support of the Lagrangian Transport andTransformation Experiment is acknowledged. Fi-nally the generous support of Rutgers Universityand State of New Jersey was an absolute key to thiswork.

REFERENCES

Boss, E., Twardowski, M.S., & Herring, S. �2001�. Shape ofthe particulate beam attenuation spectrum and its in-version to obtain the shape of the particulate size dis-tribution. Appl. Opt. 40, 4885–4893.

Chang, G.C., & Dickey, T.D. �1999�. Partitioning in situspectral absorption by use of moored spectralabsorption-attenuation meters. Appl. Opt. 38,3876–3887.

Davis, R.E., Eriksen, C.E., & Jones, C.P. �2003�. Technologyand Applications of Autonomous Underwater Ve-hicles, Autonomous Buoyancy Driven UnderwaterGliders, G. Griffiths �Ed� Taylor & Francis, London,pp. 37–58.

Dickey, T., Frye, D., McNeil, J., Manov, D., Nelson, N.,Sigurdson, D., Jannasch, H., Siegel, D., Michaels, T., &Johnson, R. �1998�. Upper-ocean temperature responseto Hurricane Felix as measured by the Bermuda Test-bed Mooring. Mon. Weather Rev. 126, 1195–1201.

Falkowski, P.G., & Kolber, Z. �1993�. Estimation of phy-toplankton photosynthesis by active fluorescence.ICES Mar. Sci. Symp. 197, 92–103.

Gargett, A., Wells, J., Tejada-Martinez, E., & Grosch, C.E.�2004�. Langmuir supercells: A mechanism for sedi-ment resuspension and transport in shallow seas.Science 306, 1925–1928.

Glenn, S.M., & Schofield, O. �2003�. Observing the oceansfrom the COOLroom: Our history, experience, andopinions. Oceanogr. 16, 37–52.

Glenn, S.M., Arnone, R., Bergmann, T., Bissett, W.P., Crow-ley, M., Cullen, J., Gryzmski, J., Haidvogel, D., Kohut,J., Moline, M.A., Oliver, M., Orrico, C., Sherrell, R.,Song, T., Weidemann, A., Chant, R., & Schofield, O.

�2004�. The Biogeochemical impact of summertimecoastal upwelling in the Mid-Atlantic Bight. J. Geo-phys. Res. 109, C12S02,

Glover, H.E., Prézelin, B.B., Campbell, L., Wyman, M., &Garside, C. �l988�. A nitrate-dependent Synechococcusbloom in surface Sargasso Sea water. Nature�London� 331, 161–163.

Jenkins, W.J., & Goldman, J.C. �1985�. Seasonal oxygen cy-cling and primary production in the Sargasso Sea.Journal of Marine Research 43, 465–491.

Karl, D.M., & Lukas, R. �1996�. The Hawaii Ocean Time-series �HOT� program: background, rationale andfield implementation. Deep-Sea Res., Part II 43,129–156.

Kirkpatrick, G.J., Orrico, C., Moline, M.A., Oliver, M., &Schofield, O. �2003�. Continuous hyperspectral ab-sorption measurements of colored dissolved organicmaterial in aquatic systems. Appl. Opt. 42, 6564–6568.

Kolber, Z., & Falkowski, P.G. �1993�. Use of active fluores-cence to estimate phytoplankton photosynthesis insitu. Limnol. Oceanogr. 38, 1646–1665.

McGowan, J.A., Daniel, R., Cayan, D.R., & Dorman, L.M.�1998�. Climate-Ocean variability and ecosystem re-sponse in the Northeast Pacific. Science 281, 210–217.

Michaels, A.F., & Knap, A.H. �1996�. Overview of the U.S.JGOFS Bermuda Atlantic time-series study and theHydrostation S program. Deep-Sea Res., Part II 43,157–198.

Moisan, J.R., Miller, A.J., Di Lorenzo, E., & Wilkin, J.�2004�, Modeling and Data Assimilation. In RemoteSensing in Coastal Aquatic Environments. R.L. Miller,C.E. Del Castillo, & B.A. McKee �Eds.�, Springer, NewYork, pp. 229–257

Mountain, D.G. �2003�. Variability in the properties ofShelf Water in the Middle Atlantic Bight, 1977-1999. J.Geophys. Res., �Oceans� 108, 3014–3029.

Mountain, D.G., & Taylor, M.H. �1998�. Spatial coherenceof interannual variability in water properties on theU.S. northeast shelf. J. Geophys. Res., �Oceans� 103,3083–3092.

National Research Council �2003�. Enabling Ocean Re-search in the 21st Century: Implementation of a Net-work of Ocean Observatories, Committee on theImplementation of a Seafloor Observatory Networkfor Oceanographic Research Washington DC, ISBN #0-309-08990-5, 240 pages.

Platt, T. �1984�. Primary productivity in the central NorthPacific: comparison of oxygen and carbon fluxes.Deep-Sea Res., Part A 31, 1311–1319.

Rudnick, D.L., & Perry, M.J. eds. �2003�. ALPS: Autono-mous and Lagrangian Platforms and Sensors, Work-shop Report, 64 pp., www.geo-prose.com/ALPS

Schofield, O., Bergmann, T., Bissett, W.P., Grassle, F., Haid-vogel, D., Kohut, J., Moline, M., & Glenn, S. �2002�.Linking regional coastal observatories to provide thefoundation for a national ocean observation network.IEEE J. Ocean. Eng. 27, 146–154.

Schofield, O., Bergmann, T., Oliver. M., Irwin, A., Kirk-patrick, G., Bissett, W.P., Orrico, C., & Moline, M.A.,�2004�. Inverting inherent optical signatures in the



nearshore coastal waters at the Long Term EcosystemObservatory. J. Geophys. Res. 109, C12S04.

Schofield, O., & Tivey, M. �2005�. Ocean Research Interac-tive Observatory Networks. National Science Founda-tion. 295 pp., www.geo-prose.com/ORION

Schofield, O., Bosch, J., Glenn, S.M., Kirkpatrick, G., Ker-foot, J., Moline, M., Oliver, M., & Bissett, W.P. �2007�.Harmful algal blooms in a dynamic environment:How can optics help the field-going and sample poorbiologist? In Real Time Coastal Observing systems forecosystems dynamics and harmful algal blooms, M.Babin & J.J. Cullen �Eds� UNESCO, Paris. �in press�.

Shulenberger, E., & Reid, J.L., �1981�. The Pacific shallowoxygen maximum, deep chlorophyll maximum, andprimary productivity, reconsidered. Deep-Sea Res.,Part A 28, 901–919.

Song, T., Haidvogel, D.B., & Glenn, S.M. �2001�. Effects oftopography variability on the formation of upwellingcenters off New Jersey: A simple theoretical model. J.Geophys. Res. �Oceans� 106, 9223–9240.

Stommel, H. �1989�. The Slocum mission. Oceanogr. 2,22–25.

Styles, R., & Glenn, S.M. �2002�. Modeling bottom rough-ness in the presence of wave-generated ripples. J.Geophys. Res. �Oceans� 107, C8, 2401–2415.

Twardowski, M.S., Boss, E., MacDonald, J.B., Pegau, W.S.,Barnhard, A.H., & Zanveld, J.R.V. �2001�. A model forestimating bulk refractive index from the optical back-scattering ratio and the implications for understand-ing particle composition in Case I and Case II waters.J. Geophys. Res. �Oceans� 106, 14129–14142.



slocum gliders: robust and ready

Documents