development, operation, and results ... - tabs.gerg.tamu.edu · the texas automated buoy system...

Development, Operation, and Results From theTexas Automated Buoy System

LESLIE C. BENDER III, NORMAN L. GUINASSO, JR., JOHN N. WALPERT, LINWOOD L. LEE III,ROBERT D. MARTIN, ROBERT D. HETLAND, STEVEN K. BAUM, AND MATTHEW K. HOWARD

The Texas Automated Buoy System (TABS) is a coastal network of moored buoys

that report near–real-time observations about currents and winds along the Texas

coast. Established in 1995, the primary mission of TABS is ocean observations in the

service of oil spill preparedness and response. The state of Texas funded the system

with the intent of improving the data available to oil spill trajectory modelers. In its 12

years of operation, TABS has proven its usefulness during realistic oil spill drills and

actual spills. The original capabilities of TABS, i.e., measurement of surface currents

and temperatures, have been extended to the marine surface layer, the entire water

column, and the sea floor. In addition to observations, a modeling component has

been integrated into the TABS program. The goal is to form the core of a complete

ocean observing system for Texas waters. As the nation embarks on the development

of an integrated ocean observing system, TABS will continue to be an active participant

of the Gulf of Mexico Coastal Ocean Observing System (GCOOS) regional association

and the primary source of near-surface current measurements in the northwestern

Gulf of Mexico. This article describes the origin of TABS, the philosophy behind the

operation and development of the system, the resulting modifications to improve the

system, the expansion of the system to include new sensors, the development of TABS

forecasting models and real-time analysis tools, and how TABS has met many of the

societal goals envisioned for GCOOS.

INTRODUCTION

On 8 June 1990 the Norwegian supertankerMega Borg, loaded with 41 million gallons

of Angolan crude, exploded and caught firewhile lightering its cargo about 60 nautical milessouth of Galveston, Texas (Scholz and Michel,1992). Four crewmen lost their lives, and the fireraged for days until it was extinguished. Eventu-ally 5.1 million gallons of oil were released intothe Gulf of Mexico. A climatology of oceancurrents available at the time, together with winddata, suggested that the oil would be drivenonshore by the winds and downcoast (toward thesouthwest) by the coastal current. Ultimatelandfall was expected to occur around CorpusChristi. Counter to the usual June climatology(Cochrane and Kelly, 1986), the coastal currentswere running up the Texas coast and the oil wascarried northeast into Louisiana waters. Roughly50% of the light crude oil burned and 25%evaporated. Responders used skimmers andbooms and applied dispersants to recover andcontrol the remaining oil. Fortunately the off-shore nature of the spill and the limited fauna inthe region limited the natural resource damage(Helton and Penn, 1999).

During the first few hours of an oil spill criticaldecisions regarding the logistics of protectionand cleanup operations must be made by the

spill-response management team. An effectiveresponse requires immediate information aboutwind and current velocity conditions to quicklyevaluate the trajectory, fate, and potential impactof the spilled material; information that was notavailable in 1990. In 1991 the Texas legislaturepassed the Texas counterpart of the federal OilPollution Act of 1990, the Oil Spill Preventionand Response Act. This act designated the TexasGeneral Land Office (GLO) as the lead stateagency for preventing and responding to oilspills in the marine environment. In 1994 theGLO implemented plans for an operationalsystem of instrumented buoys off the Texascoast, to be known as the Texas Automated BuoySystem (TABS). The purpose of the buoy systemwas to protect Texas coastal waters by providingtimely, accurate observations of winds andcurrents (Kelly et al., 1998; Guinasso et al.,2001; Martin et al., 2005) for use in spill responseoperations. The GLO funded, from its CoastalProtection Fee, the Geochemical and Environ-mental Research Group (GERG) at Texas A&MUniversity to design, build, and operate a systemof moored, telemetering current meter buoysusing off-the-shelf technology. GERG, workingwith Woods Hole Group (WHG) of East Fal-mouth, MA, designed the buoys to measurecurrent velocity at a fixed depth of about 6 feetbelow the surface using an electromagnetic

Gulf of Mexico Science, 2007(1), pp. 33–60

E 2007 by the Marine Environmental Sciences Consortium of Alabama

current sensor and transmit the data to shore ona regular schedule via the existing offshorecellular telephone network. In early 1995, lessthan 9 mo after receiving the contract, GERGdeployed the first five buoys using this technol-ogy. In March 1996 TABS experienced its firstmajor test with the barge Buffalo 292 oil spill(Lehr, 1997). In its first 10 yr of operation therehave been 20 major spills in which NationalOceanic and Atmospheric Administration(NOAA) personnel have worked with and con-sulted the TABS data (Martin et al., 2005).

The primary mission of TABS is to providenear–real-time data when a spill occurs. Howev-er, the GLO recognized from the inception ofthe project that three factors would form TABSinto an effective public resource as well. Thus,the GLO supports research to improve thereliability, operational range, and versatility ofthe TABS buoys; it insists that all TABS data beimmediately disseminated through a user-friend-ly Internet website; and it encourages otherscientific research projects to build on the TABSresources. To that end, the buoys have beencontinuously improved since the original designto incorporate new technology, lessons learnedin the field, and expanding mission goals. Fromits inception in 1995, when the concept of a user-friendly Internet was just beginning to emerge,the buoy observations have been made availableto the GLO and the general public on theInternet. In 1998 a modeling component wasadded to the TABS program with the develop-ment and implementation of the PrincetonOcean Model (POM), adapted to performsimulations on the Texas shelf. In 2002 theRegional Ocean Modeling System (ROMS) wasimplemented for the Texas shelf, but with a gridthat covered the entire Gulf of Mexico. In orderto complement the numerical models, a statisti-cally based methodology for achieving optimalnowcasts of the shelf-wide circulation was startedin 2003. Also in 2003 real-time analysis of thedaily observations was included, which pro-vides the user with quality controlled oceano-graphic, meteorological, and engineering prod-ucts. (see http://tabs.gerg.tamu.edu/Tglo/RTA//RTA_index.html)

Today the TABS buoy network consists of 10actively monitored sites, eight along the coastand two on either side of the Flower GardenBanks National Marine Sanctuary. The eightcoastal sites are funded by the GLO: one nearSabine Pass, two off Galveston, one midwaybetween Freeport and Corpus Christi, two offCorpus Christi, and two off Brownsville. Thestate of Texas has funded TABS at a level ofabout $700,000 per year in fiscal years 2002–

2007. The two Flower Garden Banks sites arefunded separately (a yearly average of $350,000from 2001 to 2006) by an oil industry consor-tium, but are operated as part of the TABSprogram. The GLO-supported inshore sites offof Galveston and Corpus Christi have beenoccupied continuously since 2 April 1995.Figure 1 shows the locations and Table 1 liststhe coordinates of the 10 actively monitoredsites, as well as the discontinued sites.

TABS was, to the best of our knowledge, thefirst offshore observing system in the Gulf ofMexico. The Texas Coastal Ocean ObservingNetwork (TCOON; http://lighthouse.tamucc.edu/TCOON/HomePage) began earlier withthree stations in 1991 and has expanded tomore than 40 stations today, but it focuses onwater level on the coast and inshore waters.The Physical Oceanographic Real-Time System(PORTS; http://tidesandcurrents.noaa.gov/ports.html) has been operational in Tampa Baysince 1990–1991 and in Galveston Bay/HoustonShip Channel since 1996–1997. The Wave-Cur-rent-Surge Information System for CoastalLouisiana (WAVCIS; http://wavcis.csi.lsu.edu/)is operated by the Coastal Studies Institute ofLouisiana State University. It began with its firststation (CSI 13) in 1998 (Zhang, 2003); todaythere are six operational stations in water depthsranging from 5 m to 21 m. The Coastal OceanMonitoring and Prediction System (COMPS;http://comps.marine.usf.edu/), operated bythe University of South Florida, was implemen-ted in 1997 for the West Florida Shelf (Merz,2001). It consists of a real-time array of bothoffshore buoy and coastal stations (Weisberg etal., 2002). A comprehensive list of all theobserving systems that are part of the Gulf ofMexico Coastal Ocean Observing System(GCOOS) is provided at http://ocean.tamu.edu/GCOOS/System/insitu.htm.

The purpose of this article is to present anoverview of the development, operation, andresults of running the TABS operational coastalobserving system on the Texas shelf for the past12 yr. The paper is organized in six sections:Development, Field Operations, Data Manage-ment, Modeling, Achievements, and Conclu-sions. The Development and Field Operationssections provide a review of the development,capabilities, and operational experience of theTABS system. The section on Data Managementdescribes the measures used to retrieve, store,and quality control the observations, the stepsused for the real-time analysis of the qualitycontrolled observations, and the steps taken todisseminate the data and the products. Thesection on Modeling discusses the development

34 GULF OF MEXICO SCIENCE, 2007, VOL. 25(1)

and implementation of the numerical andstatistical models used to complement andenhance the buoy observations. The section onAchievements highlights a few of the moresignificant results of the TABS system, includinghow TABS has worked with NOAA during oilspills and how we have endeavored to meet thesocietal goals of the Integrated Ocean ObservingSystem. Finally, the Conclusions section sum-marizes the article and outlines the futuredevelopment of TABS.

DEVELOPMENT

The mandate of the TABS system is to pro-vide high-quality, near-surface current mea-surements. At its most basic level, each TABSbuoy records vector-averaged currents at a fixeddepth of 1.8 m below the surface, does so every

30 min, and transmits the current speed anddirection to shore once every 2 hr. In order toaccomplish this mission, the buoy consists offour principal subsystems: the oceanographicand meteorological sensors, the communicationslink, a solar-powered electrical system, and thebuoy flotation structure (Chaplin and Kelly,1995). Until recently the flotation structurefor all TABS buoys was a spar design. Aspar buoy provides a stable platform for makinghigh-quality, low-noise current measurementsbecause it does not respond to high-frequencywaves like the more common discus buoy usedby the National Data Buoy Center (NDBC).A spar buoy does not have the reserve buoyancy,power, and payload capacity that a discus buoyhas, and this has placed acceptable constraintson the versatility and operational range of thebuoys.

Fig. 1. Map displayed on the TABS Internet home page showing the location of the TABS buoys as well as theNOPP and LSU buoys, the NOAA CMAN weather stations, and the NOAA NDBC weather buoys that are linked tothe TABS Internet home page. Bathymetry contours are shown for the 20-, 50-, 200-, 2,000-, and 3,500-m depths.

BENDER ET AL.—TEXAS AUTOMATED BUOY SYSTEM 35

TABS I.—The original spar buoys, designatedas TABS I and first deployed in 1995, weredesigned for the near-shore coastal environmentand were intended to obtain near-surface cur-rents and water temperature. Urethane Tech-nologies, Inc. of Denham Springs, LA, fabricatedthe buoy with a flotation package of closed-cell,cross-linked, polyethylene foam with a polyure-thane fabric-reinforced skin. A Marsh McBirney,Inc. (MMI) electromagnetic two-axis 585-currentsensor was used to measure water velocities.Woods Hole Group (WHG), in addition toassisting with the design, manufactured theoriginal computer system that ran the buoy.The cellular telephone network operated byPetrocom for the offshore oil industry providedthe means for the near–real-time observations.The buoy was equipped, as are all buoys, with anintegral radar reflector in the upper mast anda Coast Guard–approved amber night flashinglight. A schematic of the buoy in its present formis shown in Figure 2. The system and samplinginformation of the TABS I buoy is detailed inTable 2, which lists the measurements made byeach buoy type, the sensors used, the elevation ofthe sensors, the sampling time, the averaginginterval, and the telemetry acquisition frequency.

The design has been continuously improvedsince the original TABS I buoy went to sea.During the first 5 yr of the program, the design

work was subcontracted to WHG, but beginningabout 7 yr ago (2000) the design work wastransferred in-house. From the beginning ofthe TABS program, all of the assembly, wiring,system upgrades, and maintenance on the buoyshas been done at GERG’s facilities at Texas A&MUniversity. In 2001 the hull was redesigned toutilize structural aluminum alloys to make thebuoy more robust and serviceable. The top end-cap of the buoy was also redesigned to takeadvantage of the increased hull diameter andnewer antenna designs, which enabled theantenna to be mounted inside the protectivecovers of the buoy. The new tops were equippedwith 10,000-psi bulkhead connectors for allcables to provide hull integrity and increasesurvivability should the buoy become submergedduring collision or storm. This modification wasthe result of lessons learned in the field whenflooding of the mast occasionally occurredthrough the cable glands. After these changes,there have been no broken antennas on a TABSI buoy nor have any of the buoys flooded.

Major changes in the TABS I buoys werealso made in the current sensor, the onboardcomputer system, and the communication link.After a few years of operations, many of theMarsh McBirney sensors developed saltwaterleaks that affected data availability. These sensorshave all been replaced with a single-point, vector-

TABLE 1. Locations of TABS buoys.

Buoy Depth Latitude (N) Longitude (W) Date first deployed

Aa 40 feet 29u31.9509 93u48.7339 21 June 1995B 63 feet 28u58.18509 94u54.9669 2 April 1995Ca 72 feet 28u48.5499 94u45.1269 2 April 1995D 60 feet 27u55.939 96u48.4609 31 May 1995Ea 126 feet 27u20.2989 97u06.0009 31 May 1995F 79 feet 28u50.1539 94u14.1319 22 Feb. 1996Ga 41 feet 29u33.9859 93u28.0969 11 March 1997Hb 110 feet 27u52.4069 96u33.3679 4 June 1997J 68 feet 26u11.3009 97u03.0409 13 May 1998K 204 feet 26u13.0109 96u29.9309 13 May 1998Lc 270 feet 28u02.5009 94u07.0009 20 April 1998Mc 186 feet 28u11.5009 94u11.5009 20 April 1998Nc 345 feet 27u53.3829 94u02.2229 20 April 1998Pd 66 feet 29u10.0009 92u44.2509 15 Aug. 1999R 32 feet 29u38.6439 93u38.3869 27 July 1998Sa 72 feet 28u26.2069 95u48.6749 19 Feb. 1999W 73 feet 28u20.0869 96u01.3289 28 Nov. 2001Ve 90 feet 27u54.0189 93u37.2609 23 Jan. 2002

a These buoy locations have been discontinued. Data are available in the website archive.b The buoy H site was reoccupied on 27 Aug. 2005, after being discontinued in 1998.c These buoys were operated by a project funded by the NOPP, Office of Naval Research through Dynalysis of Princeton. Funding ended in CY1999

and operations ceased. N was resurrected as part of the FGBJIP in 2002.d This buoy was operated by a project funded by the Minerals Management Service through Louisiana State University. Funding ended in CY1999 and

operation has ceased.e Buoys N and V are operated on behalf of a consortium of oil companies operating in the vicinity of the Flower Garden Banks National Marine

Sanctuary.


averaging, acoustic Doppler sensor manufacturedby Aanderaa Data Instruments AS of Norway, theDoppler Current Sensor (DCS) 3900R and DCS4100R. This acoustic sensor is significantly lesssusceptible to fouling (see Fig. 3) than the MMIsensor and has proven to be very reliable in thefield (Walpert et al., 2001). The 4100R is a newgeneration of the 3900R sensor, designed so thatthe electronics are housed outside the Durotongplastic material that encapsulates the Dopplerceramics. This change was made by the manufac-turer in mid-2005 to address the problem offailure in the DCS tilt sensors. In conjunction withthe change to the DCS current sensor, the systemelectronics for the TABS I were re-designed. Thenewly designed electronics made use of a singleRemote System Manager (RSM)/daughterboardcombination and eliminated three electronicboards from the system. The new system was alsodesigned to allow the attachment of ancillarysystems such as the Seabird MicroCat C/T sensors.

Digital satellite communications are now avail-able at costs less than the original offshorecellular telephone service first used in the TABSI buoy. All TABS buoys now use the Qualcomm

GSP-1620 Packet Data Modem, which uses theGlobalstar satellite network as the primarycommunication link. The Globalstar Corpora-tion provides the satellite data-link service,utilizing a constellation of 48 low-earth orbitsatellites that can transfer data at a rate of9,600 bps. This communications link is fasterand more reliable than the cellular system usedby the original TABS buoys. The average datatransmission success rates have increased tomore than 97%, whereas individual buoys havehad long stretches, i.e., months, in which thetransmission rate is 99.9%.

The power system for the TABS I buoy imposesconstraints on the number and type of sensorsand onboard systems that can be accommodated.The 6-inch interior hull diameter of the TABS Ispar buoy provides a physical limitation to thesize of the instrument compartment and the areaavailable on the mast for solar panels. Conse-quently each TABS I buoy contains two 12V DCgel cell batteries, each with 144 watt hourscapacity at full charge and six 10-watt multi-crystalline silicon solar panels made by BP Solar.Even in winter the solar panels are capable of

Fig. 2. Schematic of TABS I (far right), TABS II (middle two: leftmost with downward-looking RDI ADCP andAanderaa DCS 4100R and rightmost with Aanderaa DCS 4100R only) and 3-m discus buoys (far left) fabricated atGERG. The older style Windsonic anemometers are depicted on the two TABS II buoys. Buoys are to scale.


TA

BL

E2.

TA

BS

bu

oy

syst

eman

dsa

mp

lin

gp

aram

eter

s.

Mea

sure

men

tsT

AB

SI

TA

BS

IIa

Oce

ano

grap

hic

Cu

rren

tsp

eed

YY

Cu

rren

td

irec

tio

nY

YSe

awat

erte

mp

erat

ure

YY

Met

eoro

logi

cala

Win

dsp

eed

NY

Win

dd

irec

tio

nN

YA

irte

mp

erat

ure

NY

Rel

ativ

eh

um

idit

yN

YA

irp

ress

ure

NY

Sen

sors

Oce

ano

grap

hic

Cu

rren

ts:

spee

d,

dir

ecti

on

,b

uo

yo

rien

tati

on

and

tilt

Aan

der

aaD

CS

3900

RA

and

eraa

DC

S41

00R

Seaw

ater

tem

per

atu

reb

Aan

der

aaD

CS

3900

RA

and

eraa

DC

S41

00R

AD

CP

–O

pti

on

alA

cou

stic

mo

dem

–O

pti

on

alC

on

du

ctiv

ity

and

tem

per

atu

re–

Aan

der

aaC

on

du

ctiv

ity,

turb

idit

y,an

dtr

ansm

isso

met

er–

Op

tio

nal

Met

eoro

logi

cal

Win

dsp

eed

and

dir

ecti

on

–G

ill

Inst

rum

ents

Win

dO

bse

rver

IIU

ltra

son

icA

nem

om

eter

Mo

del

1390

Bu

oy

tilt

com

pen

sati

on

for

win

ds

–H

on

eyw

ell

HM

R33

00D

igit

alC

om

pas

so

ra

Shae

vitz

Acc

uSt

arII

Du

alA

xis

Cli

no

met

erB

uo

yo

rien

tati

on

for

win

ds

–E

ith

era

KV

HC

100

com

pas

so

ra

Ho

ney

wel

lH

MR

3300

dig

ital

com

pas

sA

irte

mp

erat

ure

–E

ith

era

Ro

tro

nic

sM

P10

1Ao

ran

RM

You

ng

4134

2VC

ina

rad

iati

on

shie

ldR

elat

ive

hu

mid

ity

–R

otr

on

ics

MP

101A

Air

pre

ssu

re–

Vai

sala

PT

B10

0ASi

teel

evat

ion

sea

leve

lse

ale

vel

Sen

sor

elev

atio

nO

cean

ogr

aph

icC

urr

ents

1.8

mb

elo

wsi

teel

evat

ion

1.8

mb

elo

wsi

teel

evat

ion

Wat

erte

mp

erat

ure

1.8

mb

elo

wsi

teel

evat

ion

1.8

mb

elo

wsi

teel

evat

ion

Tem

per

atu

re/

con

du

ctiv

ity

No

ne

1.5

mb

elo

wsi

teel

evat

ion

Met

eoro

logi

cal

Win

ds

–3.

4m

abo

vesi

teel

evat

ion

Air

tem

per

atu

re–

3.4

mab

ove

site

elev

atio

nR

elat

ive

hu

mid

ity

–3.

4m

abo

vesi

teel

evat

ion

Air

pre

ssu

re–

3.4

mab

ove

site

elev

atio

nG

PS

No

ne

Gar

min


Mea

sure

men

tsT

AB

SI

TA

BS

IIa

Mai

nte

lem

etry

Qu

alco

mG

SP-1

620

Pac

ket

Dat

aM

od

emu

tili

zin

gth

eG

lob

alst

arL

EO

sate

llit

eco

mm

un

icat

ion

sn

etw

ork

Qu

alco

mG

SP-1

620

Pac

ket

Dat

aM

od

emu

tili

zin

gth

eG

lob

alst

arL

EO

sate

llit

eco

mm

un

icat

ion

sn

etw

ork

Bac

kup

tele

met

ryc

Syst

emA

RG

OS

Syst

emA

RG

OS

Acq

uis

itio

nfr

equ

ency

d

Fo

ral

lse

nso

rsE

very

2h

rb

egin

nin

gat

0000

GM

TE

very

2h

rb

egin

nin

gat

0000

GM

TSa

mp

lin

gti

me

Fo

ral

lse

nso

rsE

very

30m

in,

star

tin

go

nth

eh

ou

ran

dh

alf-

ho

ur

Eve

ry30

min

,st

arti

ng

on

the

ho

ur

and

hal

f-h

ou

r

Ave

ragi

ng

inte

rval

d

Oce

ano

grap

hic

Cu

rren

tse

5m

inav

erag

eo

fp

ing

dat

ata

ken

on

ceev

ery

seco

nd

,ti

ltan

db

uo

yo

rien

tati

on

sim

ult

aneo

usl

yst

rob

edat

1H

z

5m

inav

erag

eo

fp

ing

dat

ata

ken

on

ceev

ery

seco

nd

,ti

ltan

db

uo

yo

rien

tati

on

sim

ult

aneo

usl

yst

rob

edat

1H

z

Seaw

ater

tem

per

atu

reIn

stan

tan

eou

s:ta

ken

atth

een

do

fth

e5

min

sam

pli

ng

per

iod

Inst

anta

neo

us:

take

nat

the

end

of

the

5m

insa

mp

lin

gp

erio

dM

eteo

rolo

gica

lW

ind

sf–

10m

inav

erag

eo

f25

ms

sam

ple

dd

ata,

exce

pt

for

win

dgu

sts

wh

ich

isth

em

axim

um

spee

dre

cord

edin

the

10m

inA

irte

mp

erat

ure

–In

stan

tan

eou

s:ta

ken

atth

een

do

fth

e10

min

sam

pli

ng

per

iod

Rel

ativ

eh

um

idit

y–

Inst

anta

neo

us:

take

nat

the

end

of

the

10m

insa

mp

lin

gp

erio

dA

irp

ress

ure

–In

stan

tan

eou

s:ta

ken

atth

een

do

fth

e10

min

sam

pli

ng

per

iod

aO

nly

the

TA

BS

IIb

uo

yis

cap

able

of

carr

yin

ga

met

eoro

logi

cal

pac

kage

.W

ind

spee

dan

dd

irec

tio

n,

air

tem

per

atu

re,

and

bar

om

etri

cp

ress

ure

are

alw

ays

mea

sure

d.

Th

ein

stru

men

tso

nth

em

etp

acka

gear

ein

terc

han

geab

le;

con

seq

uen

tly

ah

um

idit

yse

nso

ris

op

tio

nal

.b

Th

eA

and

eraa

curr

ent

sen

sor

pro

vid

esan

inte

gral

seaw

ater

tem

per

atu

rese

nso

r.c

Cu

rren

tve

loci

tyan

dse

awat

erte

mp

erat

ure

on

ly,

no

net

dat

a.d

Inth

eev

ent

of

anin

cid

ent

we

hav

eth

eab

ilit

yto

incr

ease

the

acq

uis

itio

nfr

equ

ency

and

sam

pli

ng

tim

e.e

Cu

rren

tsar

eti

ltco

mp

ensa

ted

and

ori

ente

dto

mag

net

icn

ort

ho

nb

oar

dth

eb

uo

ys.

fW

ind

sar

eti

ltco

mp

ensa

ted

and

ori

ente

dto

mag

net

icn

ort

ho

nb

oar

dth

eb

uo

y.

TA

BL

E2.

Co

nti

nu

ed.


fully recharging the batteries on a sunny day. Atfull charge the buoy can operate for 45 d inovercast skies when little if any charging occurs.

TABS II.—In 1997, after a year-and-a-half ofsuccessful field operations with the TABS Imodel, the GLO directed GERG to develop animproved and more capable TABS buoy. GERGworked with manufacturers to design and builda ‘‘second-generation’’ version of the spar buoy,known as TABS II. The TABS II was originallydesigned with four major enhancements: 1)operation in regions with poor or no cellularphone coverage using the Westinghouse HS1000satellite telephone system (Globalstar was notavailable at the time and Geostationary Opera-tional Environmental Satellite (GOES) did notprovide two-way communications that was need-ed); 2) an increased size of the flotation packagefor deployment in water depths greater than100 ft (30 m); 3) an ARGOS satellite data trans-mission system that is automatically activated ifthe primary communication link fails, and; 4)a Climatronics TAC Met meteorological packageto measure wind speed and direction. Since theinitial TABS II buoys were designed and success-

fully deployed, several modifications have beenmade to improve the reliability, robustness, anddata quality of the TABS buoy (Magnell et al.,1998). The original TABS II design, which in-corporated the MMI current sensor, was upgradedto the Aanderaa DCS 3900R and DCS 4100R inconjunction with the change made in the TABS Icurrent sensor. The original meteorological pack-age was redesigned to use a Gill acoustic windvelocity sensor and now includes sensors tomeasure air temperature, humidity, and baromet-ric pressure. In 2001 the TABS II design wasfurther modified to incorporate a downward-looking acoustic Doppler current profiler (ADCP)in addition to the surface measurement with theAanderaa DCS velocity sensor.

Beginning in 2001 the original Westinghousesatellite telephone system was replaced with theQualcomm satellite data modem (GSP-1620),which operates on the Globalstar satellite datanetwork. The greatest drawback of the use of theWestinghouse system on a spar buoy was itstuned 37-inch antenna. The antenna was thegreatest failure point of the buoy because of itsinconsistent tuning response and its vulnerabilityand exposure to damage. The Qualcomm data

Fig. 3. Barnacle encrusted Aanderaa DCS 3900 sensor recovered from site R after a 6-mo deployment. Thevelocity data were acceptable in spite of this level of fouling.


modem made use of a much smaller antenna (4-inch diameter by 2.5-inch height), reported dataat a higher rate, and required half the power andhalf the space of the Westinghouse system. Newradio frequency cables from Times MicrowaveSystems were incorporated to improve the datatransmission through the bulkhead fittings.These were eventually bypassed by locating thedigital modem in a water-tight housing on thetop of the buoy. The first system was deployed onbuoy ‘‘N’’ and went in the water on 23 Jan. 2002.Since then the modem has been extremelyreliable (working even when buoys are lying ona ship’s aft deck while in transit to be deployed)and has been incorporated as the primarycommunications link on all TABS buoys.

All TABS buoys now have a redundant, in-dependent, communications link based on theService ARGOS satellite system. ARGOS providesboth location information and data-collectionservice worldwide using three polar-orbitingsatellites. The communication link for two ofthe ARGOS satellites is such that data can be sentby the buoy only while a satellite is passingoverhead, whereas the third satellite, launchedin Oct. 2006, has two-way capability. Eachsatellite makes six to eight passes per day. Thetimes of the passes are predictable, but notevenly spaced. The data transfer rate is4,800 bps, but the message length can only be32 bytes. Consequently only the surface currentsand battery voltage data can be included in themessage. Although limited, this system is reli-able, consumes little power, can be equippedwith a short, easy to waterproof antenna (impor-tant for improving the robustness of the buoy tovandalism), and is economical. The computersoftware of all the TABS buoys recognizes whenthe primary communications system is notfunctioning and automatically switches to theARGOS mode. Operation of the ARGOS backupsystem is enabled automatically once every 10 d.

In 2004 and 2005, GERG fabricated four newTABS II buoys based on the original TABS II hulldesign, but with an all new electronics andcomputer system of our own design in lieu of theoriginal WHG design. The newest buoys have anintegrated high-resolution temperature and con-ductivity cell and a GPS system as standardsensors. They also have the capability of accept-ing additional sensors such as ADCPs, turbiditysensors, transmissometers, and acoustic modemsto retrieve data from bottom-mounted instru-mentation such as wave gauges or upward-looking ADCPs. A schematic of the buoy in itspresent form is shown in Figure 2. The systemand sampling information of the TABS II buoyare detailed in Table 2, which lists the measure-

ments made by each buoy type, the sensors used,the elevation of the sensors, the sampling time,the averaging interval, and the telemetry acqui-sition frequency.

GERG considers software as one of the criticalcomponents of any observing system. The buoycontroller software needs to be extremely robustand capable of diagnosing and repairing prob-lems when possible and sending diagnosticinformation ashore when errors or faults aredetected. Errors that cannot be corrected auton-omously on the buoy have to be repairable froma shore base via telemetry. As part of the newTABS II buoys, GERG designed and developeda new buoy controller based on the PrometheusPC-104 computer system manufactured by Di-amond Systems Corporation. The computer usesTiny Linux as the operating system and systemprograms that are written in Perl and C. ThePrometheus is a small footprint computeroperating at 100 MHz with multitasking ability,powerful computational ability, large storagecapacity, and relatively low power consumption.The computer is interfaced to three proprietaryboards developed at GERG: an ADCP powersupply board that provides a clean 54 volts forADCP operation; a sensor interface board thatenables two-way communication with all thedigital sensors as well as control over analogsensors; and a 12-channel power switch boardthat turns power on and off to each individualsensor according to program requirements. Thisprovides the operator with total control overeach sensor schedule and provides the ability toremotely change the schedule or samplingregime whenever required. A Windows-basedgraphical user interface (GUI) that interfaceswith the Linux-based software on the buoy makesit possible for technicians without a Unix back-ground to effectively communicate, set up, makechanges, and test the buoys either remotely viasatellite, through the existing WIFI system, orthrough a hardwired monitor port.

Of the six buoys GERG fabricated using thePC-104–based controllers, only one has failed orreset in the field. The one failure occurred whena hard disk on board the buoy filled with imagefiles due to a malfunctioning instrument. Part ofthe reason for the success of these buoys is thatthe software running on the PC-104 is robust andself-diagnosing. The buoy software was engi-neered in modular form to enable easy sensoror system updates to be uploaded via satellite,hardwire, or over the integrated WIFI system.The buoy software monitors its own operationand reports any problems in the form ofdiagnostic files that are transmitted with thedata. The software monitors the system voltage,


system current, charge and discharge rates of thebatteries and is capable of running sensordiagnostics on individual sensors at any timethey are required.

The power system for the TABS II buoy mustmeet greater demands than the TABS I buoy,primarily because of the increased number andtype of sensors that are accommodated and thenew PC-104 computer. The 22-inch interiordiameter of the TABS II spar buoy providesa larger instrument compartment than a TABS Ias well as greater buoyancy to carry morebatteries and instruments. Consequently, eachTABS II buoy contains sixteen 12V DC gel cellbatteries, each with 144 watt hr capacity at fullcharge and nine 10-watt amorphous silicon solarpanels capable of generating a conservative200 watt hr per d in full sun. During normaloperation the buoy uses approximately140 watt hr in a 24-hr period. At an operatingpower of 5 watts, the PC-104 is the major powerconsumer. Efforts have been made to allow thecomputer to sleep when not needed. At fullcharge the buoy is capable of operating allsensors for at least 16 d in overcast skies whenlittle if any charging occurs. The charge/discharge current and the overall system currentof the buoy are continuously monitored andreported by the controller. Should it becomenecessary, individual sensors may be shut downto conserve power.

Three-meter discus buoy.—The reserve buoyancyof a TABS II buoy (Fig. 2) is approximately 750pounds, whereas the reserve buoyancy of thenewer 3-m discus is on the order of 7,300 pounds.This gives the 3-m buoy the capability to operatein much deeper depths and during higher seastates than the TABS II buoy. It also provides forthe capability to support far greater powerbudgets that the TABS I and II designs. Beginningin 2002, NOAA funded research to build anddeploy an in situ optical early warning system todetect harmful algal blooms on the Texas coast.This provided the impetus to design, fabricate,and outfit a 3-m discus buoy around a Flow-Camcytometer, an instrument capable of imagingindividual microscopic phytoplankton associatedwith harmful algal blooms (Campbell et al.,2007). The buoy was also equipped to operatea variety of subsurface and surface sensors thatfulfilled the TABS mission as well as additionalsensors, some of which required large powersupplies and continuous operation, includingnutrient analyzers, acoustic modems, and direc-tional wave accelerometers. The PC-104 computerwas first designed and built for the 3-m discusbuoy and then adapted for the TABS II buoys, as

we have discussed. The buoy was built, deployed,and, despite numerous technical challenges,showed that the concept of detecting harmfulalgal blooms is technically feasible. This led to theopportunity to design and fabricate 3-m discusbuoys for the University of Southern Mississippi(USM). The first buoy was deployed in theMississippi Sound in Nov. 2004 and surviveda close pass by Hurricane Katrina.

Pilot studies.—In 2005, tests were conducted ofan instrument package deployed on the bottomthat showed it was feasible to a) use an upward-looking, bottom-mounted ADCP to measurenear–real-time waves and currents and b) trans-fer that data to an overhead TABS II buoy withacoustic modems (Bender et al., 2006). Moreimportantly, the test demonstrated that it ispractical to use the TABS buoys as focal pointsfor making sea floor in situ oceanographicmeasurements, particularly for light, nutrients,particles, and dissolved oxygen. The ongoingdevelopment of the 3-m discus buoys will enableadditional ancillary sensors such as directionalwave measurement, flow cytometry, and nutrientsensors. We have embarked on a program tofabricate and operate the fourth TABS buoytype, a 2.25-m discus hull design. This hulldesign will have significantly more reservebuoyancy, as well as additional sensors andcapabilities including directional waves, but willbe capable of deployment from smaller vessels.

FIELD OPERATIONS

Deployment of the first TABS I buoys began 2April 1995 at Sites B and C, followed bydeployments at Sites D and E on 31 May andSite A on 21 June. Since then, sites have beenadded, sites have been removed, and some siteshave been relocated based on experience andoperational requirements. As of Aug. 2007 thereare 10 active locations: B, D, F, H, J, K, N, R, W,and V in water depths ranging from 10 m to105 m and eight discontinued sites: A, C, E, G, L,M, P, and S. Sites R, D, F, and W are monitoredwith a TABS I buoy, and B, J, K, N, and V aremonitored with TABS II buoys. The map (Fig. 1)and table of locations (Table 1) give the posi-tions occupied by the TABS buoys since theinception of the project. The solid circles inFigure 1 show the present buoy locations; thesmall diamonds show the discontinued (ar-chived) sites. When a TABS buoy is moved toa new location it is given a new designator letter,and when a buoy is removed from service itsdesignator letter is retired. Thus, the data setassociated with a letter is from a single location.


Each buoy is registered with the U.S. CoastGuard as a private aid-to-navigation. This pro-tocol facilitates use of the archive database andsimplifies changes to the U.S. Coast Guard Aidsto Navigation database.

The TABS buoys are intended for oceano-graphic missions of long duration and must beable to reliably withstand storms, hurricanes,fishing pressure, ship collisions, vandalism, andlong periods at sea. In order to accomplish this,the mean time between failure (MTBF) for thebuoy and its components must be well under-stood. Improving the MTBF has been an ongoingprocess of evaluation and modification based onnew technology and lessons learned. Many of thedesign changes implemented over the years weredone to improve the ruggedness of the buoysfollowing failure of a component or subsystem.For example, all hull penetrations are now madewith 10,000-psi rated bulkhead connectors insteadof cable glands. Long whip antennas are nolonger used. When the Globalstar satellite systemcame online and replaced the WestinghouseSatellite telephones, it made it possible to extendthe deployment period to 6–7 mo. As a result ofthese and many other changes, the MTBF ofa buoy is 6 mo. The goal is to continually increasethe MTBF. Experience with deployments at theinshore sites, particularly during the spring andsummer months, continues to show that bio-fouling can become a problem, even for theAanderaa sensor, after only 6 mo. Offshore buoysystems at sites such as K, N, and V do not sufferthe same fouling problems, and the deploymentduration is limited by mooring wear, which takesplace over a 9–10-mo. period.

Storms and hurricanes continue to be one ofthe biggest challenges to improving the MTBF.In July 2003 Hurricane Claudette passed directlyover two TABS II buoys deployed at the FlowerGarden Banks, buoys N and V. High winds andwaves pushed the buoys beneath the surface toa depth (estimated to be about 15 m) where theurethane foam flotation compressed and wasunable to provide enough buoyancy to return to

the surface. The buoys sank. Using side scansonar the buoys were located on the bottom, and6 mo later one of the buoys was grappled for andrecovered. The instrument compartment hadmaintained its water-tight integrity, despite beingin 100 m of water. Relying on internal batteries,the buoy recorded water temperature (Fig. 4)and velocity data for nearly 3 wk while lying onthe bottom. A failure analysis was conductedafter Claudette, and the mooring was found tobe the principle cause of the failure. All TABSmoorings were subsequently examined andredesigned to withstand a category three hurri-cane. In late Sept. 2005 the buoys deployed atsites N and V, with the redesigned moorings,were lost during the close passage of HurricaneRita, a category four hurricane. On 24 Sept. theeye wall of Rita, with 120 mph winds, passed overthe top of buoy R, a TABS I buoy. This buoy wasprobably forced to the bottom as well, butbecause the water depth was less than 15 m itresurfaced afterwards and continued to recordwater temperature and velocity, doing so until itwas recovered 3 wk later. Based on theseexperiences, we have concluded that the TABSII buoy is unlikely to survive a strong categorythree hurricane when moored in waters ap-proaching 100 m in depth. We have embarkedon a program to replace the buoys at N and Vwith the fourth TABS buoy type, a 2.25-m discushull design. This hull design has significantlymore reserve buoyancy, as well as additionalsensors and capabilities, and is capable of beingdeployed from smaller vessels.

Collision damage and vandalism continue tochallenge our best efforts to improve the MTBFof the buoy. Although instances of vandalismand unintended collisions continue, there hasbeen a noticeable reduction in recent years. Weattribute the decrease to increased awareness bythe commercial and charter fishing industry over12 years to the presence and importance of theTABS buoys. In one instance a buoy that hadsustained repeated collision damage was reposi-tioned 7 nautical miles away from its original site,

Fig. 4. Temperature recorded by buoy N during the passage of hurricane Claudette in July 2003. The buoylost buoyancy on 15 July 2003 at 0000 UTC and descended to the bottom, where it continued to record data for3 wk before the battery voltage dropped too low.


and that solved the problem. Some charterfishing fleets in south Texas waters use the TABSdata to organize their fishing trips and then,while they are offshore, keep a watch on thebuoys. GERG occasionally receives calls fromthese charter captains when they notice a buoy isoff location or the data is not up to date. Therewas an instance in 2004 where a buoy had beendragged off location and two GERG-sponsoredservice cruises were unable to find the missingbuoy. GERG subsequently received a phone callfrom a charter captain who had discovered it offlocation and called to report its position. Thebuoy was then recovered.

The quickest, and most reliable, way to servicea TABS buoy is to replace the buoy with a newlyserviced one. The recovered buoy is then broughtback to Texas A&M University for examination,service, and repair. Disassembly of a TABS buoywhile at sea to replace system components can beproblematic because of salt air, spray, and heavyweather. The GLO has provided funding tomaintain several spare TABS I buoys and TABSII buoys, which permits GERG to accomplish mostservice visits by replacing the buoy.

A significant motivation for extending theservice cycle of a buoy is the growing cost of shiptime. During the past several years the pool ofships with the requisite size, speed, liftingcapability, and affordable daily rate structureneeded for the TABS program has shrunk. Giventhe shrinking pool of dedicated ships, findingthe means to service our buoys has becomea challenge. During the last 3 yr, we have useda variety of vessels for TABS buoy operations.Vessels of the size and capabilities of TexasA&M’s 182-foot R/V Gyre and University of TexasMarine Science Institute’s 103-foot R/V Longhornare necessary for launch and recovery of TABSbuoys. Unfortunately the Gyre was retired on 31Aug. 2005 and the Longhorn followed suit a yearlater. At present, there are no dedicated,university-owned, research vessels operating outof Texas ports. The nearest University-NationalOceanographic Laboratory System (UNOLS)research vessel, the Louisiana Universities Ma-rine Consortium (LUMCON) R/V Pelican, ishome ported in Cocodrie, LA, nearly 400nautical miles from our southernmost buoys. Inthe past year we have begun to successfully usevessels of opportunity that we outfit with our ownportable winch, power pack, and A-frame. Wesend an additional person to sea to operate thewinch and budget the cost for shipping the deckmachinery to and from the mobilization site andthe services of a welder and crane operator. Inthe past we chartered a boat and crew to retrievethe TABS II buoy at site J, the southernmost

location. This was a job they had never donebefore and one that had none of our personnelonboard. The buoy was successfully recoveredbut was badly damaged in the process. Thepossibility of having to use boats and crews thatare unfamiliar with the deployment and retrievalof TABS buoys places even more emphasis ondesigning and building a rugged buoy.

DATA MANAGEMENT

Our land-based buoy data systems have threebasic components: communication, data analy-sis, and data dissemination, which we discussbelow.

Communication.—Primary communication withthe TABS buoys is via the Globalstar satellite datanetwork at 9600 baud. The buoys initiate thecommunications link once every 2 hr by placinga call to the standard dialup modem at GERG.The duration of the call is on the order ofa minute or less, during which current speed anddirection, water temperature, meteorologicaldata, and engineering data are transmitted ashexadecimal strings. Full column ADCP profilescan be included as well; the call duration with 30ADCP bins is typically 90 sec or less. Thefrequency of calls on all buoys was increasedduring 2005 from every 3 hr to every 2 hr. Thisprovides the GLO and other users with datacloser to real time. The advantage of Globalstarcompared to GOES is the ability to conduct two-way communications. Because the link is two-way,the buoys can be instructed to transmit datamore frequently in the event of an oil spill orother emergency. No information is lost if thecall is not successful in making a connection or itis dropped prematurely before all the data istransmitted; first, because the computer on thebuoy has an independent internal data archivethat permanently stores all the data, and, second,because the most recent data are stored in anonboard buffer for later retrieval. Memorypointers keep track of what data have beensuccessfully transmitted so no data are lost iftelemetry is lost.

The buoy’s onboard communication buffer issized to hold 6 hr of data, which is uploadedevery 2 hr when primary communications areactive. Once the data are received at GERG, anautomated data collection algorithm checks fordata loss. Any gaps in the telemetered data canthen be filled at the next successful transmission.If the communication buffer on board the buoyfills up, then this is assumed to be an indicationthat the primary communication link is down.The secondary communication link, System


ARGOS, is then initiated. The message size ofSystem ARGOS is limited to 32 bytes, so weassure that the most recent data in the commu-nication buffer have priority over older data.Each message, or burst, contains four sets of half-hourly currents and battery voltages. Duringa satellite overpass, up to seven bursts can beuploaded depending on the duration of the pass(a function of the elevation and azimuth) andthe quality of the transmission link. The intervalbetween satellite passes varies. Because thebuffer is a last-in-first-out type, some older datamay be pushed out of the buffer before they canbe transmitted. However, a given satellite passwill always provide the most recent buoy observa-tions, plus several hours of past observations.Because the interval between passes will rangefrom about 2–4 h, some data gaps do occur. Inthe event that both primary and secondarycommunication fails, the computer on the buoyhas an independent internal data archive thatalways stores all the data. The data can either beaccessed remotely via Globalstar’s two-way link, ifprimary communications are subsequently re-stored, or the data can be downloaded when thebuoy is serviced.

Data analysis.—Once the TABS data are re-ceived in College Station, the analysis of the dataproceeds in two steps: Level I quality control andLevel II quality control. Level I quality control isautomated and begins when the raw data fromthe TABS buoys are received at GERG. The rawdata are transferred to a Linux server where thehexadecimal data are converted to engineeringunits. The second step then removes obviouslyflawed data. Graphical displays are generatedevery hour showing time series plots of thecurrents, water temperature, buoy tilt, andvarious engineering parameters that indicatethe operating status of the buoy. An exampleof the plot displaying currents and watertemperature is shown in Figure 5. Time seriesplots of the meteorological data, winds, airtemperature, and atmospheric pressure aremade available for the TABS II buoys. Oncea day, the quality of the Level I data is reviewedby an experienced oceanographer, who can thenmake further corrections to the data whenneeded. The final quality-controlled Level I dataare then inserted into a database for retrieval byusers.

Level II quality control occurs each morningwhen the Real Time Analysis algorithm automat-ically performs an analysis of the previous 30 d ofLevel I data. The first step is an additional qualitycontrol of the Level I oceanographic, meteoro-logical, and engineering data. Data are flagged

for duplicate values, missing values, duplicatetime stamps, bad time stamps, out-of-chronolog-ical sequence data, and statistical outliers such asspikes or unreasonable physical values. Currentmeter velocities that are identically zero for bothcomponents are flagged. Current speeds thatexceed 150 cm s21 or change by more than35 cm s21 in one time step (30 min) are flagged.The second step replaces the flagged data usinga combination of linear interpolation for smallgaps and a spectral preserving algorithm specif-ically designed at GERG for gaps up to 3 d.Linear interpolation is only used if there are lessthan three consecutive flagged records, i.e., nomore than 90 min. Gaps and flagged data thatare longer than 90 min, but less than 3 d arefilled using a combination of the Lomb-Scargleperiodogram [Lomb (1976) and Scargle (1982)independently developed a robust method foranalyzing the spectral properties of an irregularlyspaced data] to find frequencies of the signifi-cant peaks and a least-squares fit to the data oneither side of the gap. The third step preparesa variety of products to present the Level II dataon the Web at http://tabs.gerg.tamu.edu/tglo/RTA/RTA_index.html. These graphical prod-ucts provide the user multiple views of thequality-controlled oceanographic, meteorologi-cal, and engineering data. Once a day, thequality of the Level II data is reviewed by anexperienced oceanographer and an email reportissued to interested parties. The Level II data arethen made available, through the aforemen-tioned website, for retrieval by users.

The Level II oceanographic data for each buoyis presented as a variety of products, includingvector stick plots of the currents, current roses,scatter plots with the principle componentanalysis over plotted, the tidal analysis, compar-isons to numerical model results, the probabilityof a flow reversal, the water temperature, and thesuccessful quality controlled data return. In everycase, interpolated data is denoted in red, andactual, quality-controlled data is in blue or insome cases (scatter plot) black. Several of theproducts are illustrated here. The current vectorsand current roses are provided in 1-, 2-, 4-, 7-, 14-,and 30-d time slices to accommodate the needsof oil spill managers. An example of a 30-d current stick plot is shown in Figure 6. Thevector velocities are filtered through a 3-hr filterand then through a 40-hr filter to show the long-term currents that control transport. Figure 7 isan example of the tidal analysis product. The 3-hr filtered current stick plot is shown in the toppanel, the synthesized tidal velocity record in themiddle panel, and, in the bottom panel, thedetided velocity record. Finally, at the very


Fig. 5. Level I view of current and water temperature data at buoy H.


bottom of the page, a table of the applicable tidalconstituents used to create the tidal record isshown. The diurnal and semidiurnal tidal con-stituents have been determined from the histor-ical record for each buoy. The nature of the flowon the Texas continental shelf tends to varybetween inertial (circular) and alongcoast (rec-tilinear). The alongcoast variability productpresents one means of visually presenting thisvariability state. It is derived by using a 12-hrsliding window to create a time series of the ratioof the principal component analysis major tominor ellipse axes. This ratio is mapped to a scalethat indicates alongcoast flow when the ratio ismuch greater than one and is inertial as the ratioapproaches one. Figure 8 shows an example.

For those buoys equipped with a meteorologi-cal station, the Level II wind data is presented asvector stick plots, time series of the speed of thewind and the gust, scatter plot, and wind rose. Inaddition, the air temperature, the barometricpressure, and the relative humidity are plotted.The processed winds are presented as 1-, 2-, 4-, 7-,14-, and 30-d wind stick plots and wind roses,similar to that of the currents. The winds aresampled at 0.25 Hz over a 10-min time period.The time series of the 10-min averaged windspeed, as well as the wind gust, which is the

maximum speed recorded in the 10 min, ispresented. See Figure 9 for a typical example.

The Level II engineering data are shown infive different products. They are signal strengthand ping count from the Aanderaa DCS sensor,battery voltage, buoy tilt, and data return. One ofthe products, the battery voltage for each buoy isshown as Figure 10. The unfiltered voltage isshown in the top panel and the 11-hr filteredvoltage in the middle panel. The diurnalvariation in the voltage is a reflection of theamount of solar radiation to which the solarpanels are exposed. Based on a model of theexpected clear sky solar insolation for thelatitude of each buoy, the bottom panel showsthe daily variations in the insolation. By mid-summer the insolation will be 500 W m22. Herewe see no charging due to extensive cloud coverduring the last part of Jan. and the early part ofFeb.

Data dissemination.—The Level I quality con-trolled data are inserted into an archival data-base designed to facilitate the extraction of user-specified subsets. The database is built on mysql,an open source Linux structured query languagedatabase, and on simple flat ascii files. The datahave proven useful for model initialization,

Fig. 6. An example of the RTA product showing the 30-d stick plot of currents at buoy V. The unfiltered data(top), the 3-hr filtered data (middle), and the 40-hr filtered data (bottom). Only 0.14% of the vectors have beeninterpolated (not shown).


model skill assessment, research, and operation-al planning purposes. The GLO has direct accessto this database via FTP over the Internet. Thepublic has access through the World Wide Web(WWW) at http://tabs-os.gerg.tamu.edu/tglo/index.php. A quality controlled data set of alldata collected during the TABS program isavailable on a DODS server at http://tabs.gerg.tamu.edu/DODSdata/. Additionally, TABS me-teorological data from sites B, J, K, N, and V arebranded as NDBC sites 42043, 42044, 42045,42046, and 42047 and formatted for ingest intothe National Data Buoy Center. Efforts arepresently underway to format the TABS currentobservations for NDBC ingest.

The TABS web page provides the user withaccess to a variety of oceanographic and meteo-rological data products. Using their browser, theuser is able to view either the latest data or accessthe database and view archived data. The usercan also download the data for later use. Thisweb presentation has been an integral part of theTABS system since 1996 (Lee et al., 1996). Userscan select a TABS buoy location from the map orfrom text links for those without a graphical webbrowser. For each TABS station the user canchoose to view either a graph of the past 4 d of

data or the data in tabular format. The graphconsists of a ‘‘stick plot’’ of the currents, crossshelf, and along shelf components of the currentand water temperature (see Fig. 5). Data arepresented in both English and metric units.Graphs can be downloaded as either a GIF imageor a postscript file.

Several additional features of the TABS web-site assist in the utilization of the TABS data. Asummary plot provides a stick plot for each buoyusing a common time axis. A status table listsbuoy latitude, longitude, lease block, and waterdepth. The status table also indicates which ofthe buoys have successfully transmitted their dataduring the past 12 hr and contains other in-formation regarding the operational status ofeach buoy. Each buoy page also contains a linkthat allows the user to search the TABS databaseand retrieve data from a buoy for a user-select-able time period. The user can access up to 2 moof data at a time. The results of each databasesearch can be viewed in both graphical andtabular format.

In the summer of 2003 a major power failurecaused a disk hardware failure on the primaryserver that runs and maintains the TABS websiteand data system. Since that time the TABS

Fig. 7. An example of the RTA product showing the tidal velocities at buoy F. The 3-hr filtered data (top), thetidal currents (note the change in vertical scale) (middle), and the detided currents (bottom). The tides are smalleverywhere in the Gulf of Mexico; just 8.4% of the variance is described by the tides at buoy F. Only 0.35% of thevectors have been interpolated (not shown).


Fig. 8. An example of the RTA product showing the along-coast variability at buoy J. This figure is meant toconvey how much of the currents are along-coast versus inertial, where a value of one denotes alongcoast anda value of zero inertial. It is derived from the ratio of the principle component analysis (PCA) major to minorellipse ratio. It is blocky by nature because it evaluates a 12-hr block of currents. The unfiltered data (top), the 3-hrfiltered data (middle), and the 40-hr filtered data where the mean is annotated (bottom). The longer period flowsbecome more and more alongcoast.

Fig. 9. An example of the RTA product showing the wind gust and mean wind speed at buoy V. The unfiltereddata (top), the 3-hr filtered data (middle), and the 40-hr filtered data (bottom).


website, data, and software have been mirroredhourly onto two other machines to ensure re-liability in the case of a hardware or power failure.One of these is a Redundant Array of InexpensiveDisks (RAID) server located at GERG, but ina different building, and the other is a machinelocated in the Department of Oceanography onthe Texas A&M main campus. Both of thesemachines are backed up nightly and the backupsare stored at off-site locations. Both servers atGERG are connected to switches served by re-dundant fiber optic links to the Texas A&MUniversity high-speed backbone. The GERG facil-ity is a node on the University’s Gigapop internetnetwork. An internet ring controller connectsGERG to a loop of controllers through redundantfiber optic paths in such a manner that cutting onefiber optic link will not interrupt internet service.Both GERG servers, separate data communicationsystems, and all networking equipment are sup-ported by uninterruptible power systems. TheTABS website can be supported even with powerfailures of up to a 5-hr duration.

The TABS website also provides access to datafrom the National Data Buoy Center’s buoy andcoastal (CMAN) meteorological data. These dataare obtained directly from NDBC each hour. Weinclude four offshore buoys and two CMANstations, e.g., 42035 located southeast of Galves-

ton; 42019 and 42020, which are east andsoutheast of Port Aransas, respectively; 42038,which is east of the Flower Garden Banks; SRST2near Sabine; and PTAT2 near Port Aransas.These data are updated hourly and presented inboth graphical and tabular formats.

The website also contains a number of links toadditional real-time oceanographic and meteo-rological data. Links to National Weather Servicecoastal and offshore weather forecasts for theGulf of Mexico are provided on the main TABSweb page. Links have been added to modelresults of currents as well as ETA-32 griddedwind forecasts. There are links to the GCOOS,Houston/Galveston PORTS website, TCOON,National Data Buoy Center, Galveston Bay andCorpus Christi Bay Animated Hydrodynamic andOil Spill Model output, Satellite Sea SurfaceTemperature Images from NOAA andJohns Hopkins University, Tampa Bay PORTS,and other relevant sites.

A ‘‘Notice to Mariners’’ is included on theTABS web page to request users avoid contactwith the buoys and report problems if theynotice the buoys off location or if they seedamage. Access to the notice is available on alldata pages as well as the main page.

Analysis of the TABS web server access logsshows that utilization of the TABS website has

Fig. 10. An example of the RTA product showing battery voltage and clear sky insolation for buoy F. Theunfiltered battery voltage (top) and the 11-hr filtered data (middle). Note the distinctive diurnal solar chargingcycle. The bottom panel shows the estimated solar insolation for buoy F’s latitude. The effect of cloud cover is seenat the beginning of Feb.


been increasing since its inception. Peak usageof the TABS website generally occurs in mid-Oct.and then tails off rather sharply. We see this asa reflection of the end of the recreationalboating season and a decrease of usage byboaters. The three largest groups of TABS userscome from the .com, .edu, and .net Internetdomains. The first represents commercial enti-ties primarily from within the United States, thesecond represents educational institutions in theUnited States, and the last are network serviceproviders. However, since some of the majorInternet service providers are in the .comdomain, i.e., AOL, it would appear that themajority of the use of the TABS site is comingfrom the general public.

Noteworthy groups that access the TABS siteinclude users from the Texas State government,specifically the Texas General Land Office, andusers from the U.S. government, including usersfrom NOAA, Minerals Management Service(MMS), United States Coast Guard (USCG), andNASA. Usage by the offshore industry includesmost of the major oil companies. In addition wehave seen usage from 69 foreign countries to date.

MODELING

In 1998 a modeling component was added tothe TABS program with the development andimplementation of the POM adapted to performsimulations on the Texas shelf. In 2002 themodeling was extended with the implementationof the ROMS. It has always been recognized thatthere is a need to estimate the circulation fieldbetween the sparsely located TABS buoys.Whereas the half-hourly temporal coverage ofthe TABS current meters is exceptional, thegeographic coverage, as seen in Figure 1, is toosparse to capture the expected spatial modes ofcirculation on this shelf. On the basis ofhydrographic data primarily collected by theTexas–Louisiana Shelf Circulation and Trans-port Processes Study, Li et al. (1996) examinedthe energetic scales of spatial variability acrossthe Texas–Louisiana continental shelf. Theyfound that the cross-shelf scales over the westernhalf of the shelf are shorter (,15 km) than thosein the eastern and central shelf (,20 km),whereas the along-shelf scales (,35 km) areessentially the same everywhere on the shelf.The difference in the cross-shelf scales wasattributed to the shelf width. These scales areconsiderably smaller that the average 120 kmalong-shelf separation between TABS buoys and70 km across-shelf separation. The minimumalong-shelf separation between buoys is 40 km,and the minimum cross-shelf separation is

55 km. In order to estimate the circulation fieldbetween the sparsely located TABS buoys, twonumerical and one statistical model have beendeveloped and are described below, as are thewinds used to drive the two forecast models.

Princeton Ocean Model.—The original shelfcirculation model, developed and maintainedby Joseph Yip from 1998–2002, consists of a three-dimensional version of the POM adapted toperform simulations on the Texas shelf ona domain extending from 25uN on the Mexicancoast to 85uW at the coastline of Florida. Theoperational POM model is a simplified barotro-pic version that performs a 24-hr surface currentprediction once per day. A data–model compar-ison—performed from April through Dec. 1999of nine near-shore TABS buoys—indicated mod-est skill of the model in predicting the wind-driven circulation.

Regional Ocean Modeling System.—Limitations inthe original POM shelf circulation model led tothe development of a second-generation shelfcirculation model using the ROMS. The develop-ment was started in 2002 and continues today.The ROMS-based circulation model was designedto provide greater maintainability and extensibil-ity than was available with the POM model, as wellas to enable greater flexibility and ease ofmanaging and transforming the simulation mod-el input and output fields. Both the computation-al kernel and the data handling infrastructurewere completely revised for these purposes.

ROMS is a free-surface, hydrostatic, primitiveequation ocean model that uses stretched,terrain-following coordinates in the vertical andorthogonal curvilinear coordinates in the hori-zontal. (See Ezer et al. 2002 and the referencestherein for background information on bothPOM and ROMS.) Computationally, ROMSuses advanced numerical algorithms and soft-ware technology to facilitate efficient simula-tions on single and parallel computer archi-tectures. Scientifically, it contains a variety ofmodular features including high-order advectionschemes; accurate pressure gradient algorithms;several subgrid-scale parameterizations; atmo-spheric, oceanic, and benthic boundary layers;biological modules; radiation boundary condi-tions; and data assimilation. These scientific andcomputational features provide for both an easilymaintained present operational system anda flexible upgrade path for the research anddevelopment of future, improved versions of thesystem. The higher-order advection scheme andthe boundary layer schemes, in terms of mixing,are used; data assimilation is not.


Significant differences between the first andsecond generation systems include:

N The expansion of the computational domainfrom the original POM grid extending fromthe shoreline to the continental shelf toa ROMS grid across the entire Gulf of Mexico.The grid on the shelf is on the order of a fewkilometers

N Four 48-hr predictive simulations per day asopposed to one 24-hr simulation per day withthe original system

N The use of a computer cluster to performparallel simulations of larger domains athigher resolutions in about the same amountof time as the POM simulations

YBR statistical nowcast model.—Efforts to de-velop and refine a statistical circulation model tocomplement the numerical models are under-way. The objective of this endeavor is to demon-strate an effective methodology for achievingoptimal nowcasts of shelf-wide circulation byusing dominant empirical modal patterns ofexisting well-resolved near-surface Surface Cur-rent and Lagrangian Drift Program-I (SCULP-I)surface drifter data fitted to the sparse TABScurrent data. This concept was first explored byYip and Reid (2002) for application to the Texas–Louisiana shelf and was presented at the Oceans2002 Conference on Marine Frontiers shortlyafter the young lead author lost his battle withcancer. Because that paper was well received, wehave worked with Professor Reid to presenta materially expanded version of that study as anappropriate recognition of Yip’s contributions tothe description, data analysis, and dynamics of theTexas–Louisiana shelf circulation. In the YBRmodel (Yip and Reid 2002), empirical orthogonalfunction (EOF) modes are first determined fromdaily average velocity fields derived from theSCULP-I surface drifter data. Ohlmann and Niiler(2005) present a comprehensive analysis of thedrifter measurements made with the near surfacefloats of the SCULP. The SCULP-I subset of thedrifter data are clearly very relevant to the needsof the TABS program, having been deployed inthe northern Gulf of Mexico during a 1-yr period.First the drifters characterize the upper meter ofthe water column, comparable to the depthmeasured by the TABS buoys. Second, thedomain of the data covers the entire Texas shelf,from the Sabine River to Brownsville. These datainclude the two major forcing mechanisms on theTexas shelf, the wind-driven flow in the upperlayer, and the longer term flow driven by weathersystems and freshwater input from rivers, partic-ularly the Mississippi. Both the TABS current data

and the SCULP-I drifter data include tidal andinertial oscillation signals, but these are sup-pressed by employing daily average currents.Furthermore, DiMarco and Reid (1998) haveshown that the tidal signal is weak on the Texas–Louisiana shelf. The drifter velocity data werebinned into a boundary-fitted grid covering theTexas–Louisiana shelf. The bins were comparablein size to the energetic spatial scales of spatialvariability identified by Li et al.(1996). A nowcastof the shelf-wide circulation is made each day(http://tabs.gerg.tamu.edu/Tglo/RTA//RTA_index.html) by using the real-time TABS currentdata to find the amplitudes of the dominantempirical modes, modes first found by analyz-ing the drifter data for EOF spatial patterns. Inthis manner the circulation field between thesparsely located TABS buoys is estimated usinga method quite different from that of a numer-ical circulation model.

Winds.—The readily available meteorologicalobservations and near–real-time forecasts arecollected, archived, and disseminated for use inforcing the POM and ROMS numerical modelsand to the GLO and others for use in spill-response planning. Data are captured from theNational Weather Service, the National DataBuoy Center, and numerical weather modeloutput from the National Centers for Environ-mental Prediction (NCEP).

The Gulf of Mexico NDBC buoy observationsand coastal marine meteorological observationsfrom Gulf-coast first-order airports are ex-tracted from the Global TelecommunicationsSystem (GTS) in near–real-time using UNIDA-TA’s Local Data Manager (LDM) software.Access to the GTS stream is provided by theTexas A&M University Department of Atmo-spheric Sciences. A software program namedZEPHYR converts the data from meteorologicalcodes into convenient tabular listings. Thesedata are used in displays of current conditions,for model-data comparisons, and in the pro-duction of gridded wind fields based onobservations. This collection system is quiterobust and has run with little to no mainte-nance for about a decade.

Maintaining a system to collect NCEP modeloutput on a continuous basis has been morechallenging due to increases in weather modelresolution, forecast time horizons, and file sizesand changes in grid-point locations, host servers,model output file names, and parameter place-ments within files. Some of the maintenanceissues have relatively simple solutions, such asfaster network connections and more disk space.Changes in grid resolution and grid point


locations cause a cascade of work that extendsbeyond the collection systems into the POM andROMS models themselves.

The POM and ROMS modeling systems aredriven by the NCEP NAM forecast model windfields. NCEP’s NAM model was formerly (andperhaps still better known) as the ETA model.We will continue to use ETA here. The ETAmodel is run at NCEP four times per day. Eachnew run is downloaded as it becomes available.The forecast fields represent conditions at 3-hrintervals out to an 80-hr time horizon but wepresently only use fields out to 48 hr. The 17files, collected four times per day, total 5.8 GB/day. The Gulf of Mexico surface wind fields areextracted and made available to the modelers.ETA wind fields and surface currents from POMand ROMS are automatically posted graphicallyto our website and numerically in anotherdirectory for use by NOAA HAZAT teams fortheir use.

An interoperable TABS/modeling system.—Thegoal of the Integrated Ocean Observing System(IOOS) Data Management and CommunicationsPlan is to develop machine-to-machine interop-erable systems, with provisions for data discovery,access, metadata, transport, and archive. Inorder to achieve an interoperable system forthe TABS observations and modeling forecasts,funding was first obtained from the NationalOcean Partnership Program (NOPP). The taskcontinues with funding from the SoutheasternUniversities Research Association (SURA). TheSURA Coastal Ocean and Observing and Pre-diction (SCOOP) program is an Office of NavalResearch and NOAA–funded study designed toimplement the Data Management and Commu-nications Plan.

As part of this work the ROMS program wasconverted to accept input in netCDF format withinternal arrays named and organized accordingto standard formats (COARDS/CF). The outputroutines were also modified to conform to thisinterchange format. With properly constructedURLs, NetCDF files can be moved across thenetwork using OPeNDAP-enabled software aseasily as local files can be accessed. In theory wecould recompile ROMS with OPeNDAP-enablednetCDF libraries, and at run time ROMS couldaccess files directly from the NCEP NOMADservers. However, NOMADS is not yet sufficientlyreliable for our operational system, and issues ofnetwork latency could be a serious problem notbest solved in model code. We will be working oncatalog metadata that will support online brows-ing. This will be particularly useful for establish-ing and maintaining geographic information

systems (GIS) that we are also developing aspart of SCOOP. The GIS system will enableTGLO to rapidly zoom to problem sites andoverlay model, wind, observations, and otherrelevant parameters to give a comprehensiveview of environmental conditions.

POM and ROMS output and the NOAA/ERDLAS server.—TABS and the Texas General LandOffice enjoy an informal, but strong relationshipwith NOAA’s Office of Response and RestorationEmergency Response Division (ERD) (formerlyHazardous Materials Response Division or HAZ-MAT). As a public service we continue tointegrate the General NOAA Oil ModelingEnvironment (GNOME) model into the TABSand TABS modeling system. We have installeda copy of the NOAA PMEL Live Access Server(LAS) for use by NOAA ERD to rapidly acquireand subset the POM and ROMS model outputand ETA wind fields. Alternate methods ofhaving hot-start data sets for GNOME are beingdeveloped by this group so that GNOME will beready to go in a moment’s notice in the event ofa spill.

ACHIEVEMENTS

The primary mission of TABS—to providenear real-time data when a spill occurs—hasbeen met many times. The three-fold collateralgoals envisioned by the GLO to form TABS intoan effective public resource have been success-fully met as well. The reliability, operationalrange, and versatility of the TABS buoys havebeen continually improved, as discussed in thesections on Development and Field Operations,all the TABS data have been disseminatedthrough a user-friendly Internet website asdiscussed in the section on Data Management,and other scientific research projects have beenbuilt on the TABS resources such as modelingand real-time analysis. In this section we elabo-rate further on some of those achievements.

Oil spill response.—Fortunately there have beenno catastrophic oil spills rivaling that of the 1990MegaBorg explosion, but during the major spillsthat have occurred, and the numerous realisticdrills that have been conducted, TABS hasfulfilled its primary mission by providing near–real-time data. In its first 10 yr of operation therewere 20 major spills in which NOAA personnelworked with the GLO and consulted the TABSdata (Martin et al., 2005). There were many less-serious spills in which the TABS data wereconsulted, but such queries were not recordedin the NOAA database. We look at two oil spills,


the Buffalo Marine Barge 292 oil spill of 1996 andthe more recent DBL-152 oil spill of 2005–2006,as examples of the informal relationship that hasdeveloped over the years between the GLO andNOAA. During the Buffalo Marine Barge 292 oilspill the NOAA HAZMAT modeling team andthe GLO’s trajectory modeling team used TABSdata and computer simulations to forecast themovement of the oil to an unprecedented levelof accuracy (Lehr et al., 1997; Martin et al., 1997;Martin et al., 2005). The trajectory modelers didnot have to begin their work with only educatedguesses about the offshore currents. The cur-rents were known within minutes of the spill andwere continuously tracked for 24 d. Midwaythrough the spill TABS data showed the di-rection of the coastal current switching fromupcoast to downcoast. The benefit to cleanupand protection operations allowed IncidentCommand to stand-down an alert to the SabinePass area and refocus efforts down coast a fullday earlier than would have been possible beforeTABS. It also saved an estimated $225,000 incosts for an unnecessary deployment to protectan area no longer at risk.

In Dec. 2005 a TABS II buoy with a surfacecurrent meter and a downward-looking ADCPwas deployed about 30 miles south of Sabine,TX, to assist with tracking subsurface oil from theDBL-152 oil spill (Michel, 2006). Shortly beforemidnight on 10 Nov. 2005 the Integrated TankBarge DBL-152 was in tow from Houston, TX, toTampa, FL, when it struck a submerged oilplatform that had been damaged by HurricaneRita. The tug and barge were approximately55 km south of Cameron, LA, when the collisionoccurred. Eventually 2.7 million gallons of heavyrefined oil were released. Because of the oil’sdensity, it sank to the bottom where it wasperiodically resuspended by storm events. ATABS II buoy with a downward-looking ADCPwas deployed at the spill site to provide data onbottom currents critical to predicting where theoil would be transported.

An example of the spill response community’sacceptance of the TABS concept is the jointindustry project funded by 16 offshore operatorsto maintain two TABS II buoys at the FlowerGarden Banks National Marine Sanctuary. Thesebuoys (see N and V in Fig. 1) provide currentand wind observations to the operators in thevicinity of the Sanctuary in the event they needdata to respond to a spill in this ecologicallysensitive area.

Collateral uses.—The reliability, operationalrange, and versatility of the TABS buoys haveimproved to the point that the buoys have been

successfully used in locations remote from theTexas shelf and for missions beyond that of oilspill response. In 2001 two TABS buoys weredeployed off the Mississippi delta as part of theNorthern Gulf of Mexico Littoral Initiative pro-gram sponsored by the Naval OceanographicOffice (NAVO). In 2001, a TABS I buoyequipped with an Aanderaa DCS 3900R velocitysensor and a TABS II buoy with an ADCP sensorand meteorological station were loaned, with thepermission of the GLO, to the U.S. Navy toprovide meteorological and oceanographic dataduring the recovery operations of the EhimeMaru (Bender et al., 2002a). These buoys weredeployed just offshore of the Honolulu Interna-tional Airport and operated from 18 July 2002 to26 Nov. 2002 when the recovery operations werecompleted. Based in part on the success of thisprogram, a TABS II buoy was purchased byNAVO in 2002 for use at a nationally-importantlocation. This buoy was equipped with anIridium satellite communications system, insteadof the standard Globalstar, and a downward-looking RDI ADCP. GERG-TAMU personneltrained NAVO personnel in the operation andmaintenance of the TABS buoy and assistedthem in creating their own ground station inStennis, MS, to handle data from this buoy.

Gulf of Mexico Coastal Ocean Observing System.—TABS is a charter member of the GCOOS.GCOOS will augment and integrate a sustainedobserving system for the Gulf of Mexico as partof the IOOS (Ocean.US, 2006). GCOOS aims toprovide ocean observations and products neededby users in the region to meet the seven societalgoals of IOOS:

N Detecting and predicting climate variabilityand consequences

N Preserving and restoring healthy marineecosystems

N Ensuring human healthN Managing resourcesN Facilitating safe and efficient marine trans-

portationN Enhancing national securityN Predicting and mitigating coastal hazards.

Since its inception in 1995, TABS has contrib-uted to most of these IOOS goals. The primarypurpose of TABS is to ensure a reliable source ofaccurate, up-to-date information on ocean cur-rents along the Texas coast. The TABS currentmeasurements enable rapid assessment of thefate of oil spills, facilitating efficient remedialefforts to preserve healthy marine ecosystems.Surface current measurements and modelingprovide the basis to predict dispersion of


waterborne contaminants. The TABS oceano-graphic data provide a regional ecologicalclimatology for sea surface temperature for usein assessing ecosystem health. TABS, through itscollection of sustained time series of longduration, provide in situ measurements that aidin the detection and prediction of climaticchange. Today, more than 1.5 million half-hourly current and temperature measurementshave been collected in near real-time. At sites Band D 12 yr of measurements of sea surfacetemperature and currents are available. Thepresent-day TABS system has improved thespatial resolution of measurements in Texasoffshore waters by providing 10 observation sites.TABS has played a significant role in maritimeoperations by providing near–real-time surfacecurrent measurements that improve the effec-tiveness of search, rescue, and emergency re-sponse capabilities. The U.S. Coast Guard usesTABS data following accidents when oil rigworkers are missing or a helicopter disappearsduring an overwater flight to an offshoreplatform. Private mariners also use TABS datato help them safely navigate coastal waters.

Climatology: General.—One of the collateralgoals of TABS is to provide the foundation forscientific research projects. This goal continuesto be successfully met in a number of ways. Wehave many indications from our colleagues thatthese data are being used in teaching andresearch. Early on in the program Crout (1997)and Kelly et al. (1999) used the TABS databasefeatures to facilitate studies comparing currentscalculated from satellite altimetry with thoseobserved by the TABS buoys. Using the first7 yr of TABS data, Bender et al. (2002b) showedthat there is insufficient information to conclu-sively establish if there is a statistically discerniblelink between surface currents and the El NinoSouthern Oscillation (ENSO) or the NorthAtlantic Oscillation (NAO).

We have used the database of currents toconstruct an oceanographic climatology and themonthly historical record for each of the TABSbuoys. The climatology page, http://tabs.gerg.tamu.edu/tglo/Climatology/Climate_index.html,shows a shelf-wide view of the monthly averagedcurrents and the individual current roses foreach buoy site. The historical record, accessiblethroughhttp://tabs.gerg.tamu.edu/tglo/Hindcast/B/2006/Dec/Oceanographic_CurrentStick.html,shows the current stick plot, scatter plot, currentrose, and water temperature for each buoy forevery month since the buoy was first deployed.The historical data for each month can also be

downloaded. If the buoy recorded meteorologicaldata, those products are available as well.

Climatology: Seasonal surface currents.—In coastalregions wind stress is a predominant source ofmomentum. Cochrane and Kelly (1986) andNowlin et al. (1998) showed that there is a highcorrelation between the along-coast wind stressand the along-coast currents on the Texas shelf.Cho et al. (1998) confirmed that the maincirculation over the LATEX shelf is wind driven.The direction of the winds in the Gulf of Mexicois determined by the seasonal position of thehigh-pressure systems (Zavala-Hidalgo, 2003). Inthe fall and winter high-pressure systems movefrom the northwest continental United Statesinto the Gulf generating northeasterly winds inthe western gulf, whereas in the summer theBermuda high and the warming of the conti-nental United States generate southeasterlywinds. During the nonsummer months thenortheasterly winds drive a strong downcoastflow along the inner shelf, while during thesummer the weaker southeasterly winds drivea weaker upcoast flow. Hereafter we definedowncoast (upcoast) as proceeding in thecounterclockwise (clockwise) direction from theAtchafalaya River to Mexico (Mexico to theAtchafalaya), i.e., cyclonically (anticyclonically)along the curved coastline.

As a result of the 1.5 million half-hourlymeasurements of velocity data, we have a statisti-cally reliable description of the mean seasonalsurface currents on the shelf. Figure 11 showsthe mean surface currents for the winter months,from Sept. through May, based on all half-hourlymeasurements available from 1995 to 2005 forthe 10 TABS buoys depicted. A mean downcoastflow is clearly evident, driven by the predominanteasterly winds. The concave shape of the coastcauses the alongshore wind stress to decreasefrom its maximum in the vicinity of buoy R to itsminimum in the vicinity of buoys J and K, wherethe mean currents are weakest. During thesummer the winds are southerly and the condi-tions seen in the winter are reversed. Figure 12shows the mean surface currents for the summermonths, i.e., June, July, and Aug. These meancurrents are based on half-hourly surface currentmeasurements recorded for all the monthly dataavailable from 1995 to 2005 for the 10 TABSbuoys depicted. A mean upcoast flow is clearlyevident.

Hurricane conditions.—Since June 1995 whenthe first TABS buoys were deployed there havebeen eight tropical storms and three hurricanes(Brett, Claudette, and Rita) that have crossed the


Texas shelf. Hurricanes Brett and Katrina havebeen the only major (.category three) land-falling storms; Claudette was a category onestorm. Brett was the first major hurricane to

strike the Texas coast since Hurricane Gerry inOct. 1989. The track of Brett took it to the northof buoy J before making landfall at 0000 UTC on23 Aug. 1999. Before 21 Aug., the surface

Fig. 11. Mean surface currents for the winter (Sept. through May) on the Texas continental shelf. Bathymetrycontours are shown for the 20, 50, and 200 m depths.

Fig. 12. Mean surface currents for the summer (June, July, and Aug.) on the Texas continental shelf.Bathymetry contours are shown for the 20, 50, and 200 m depths.


currents recorded by buoy J were inertiallydominated. As the storm approached from thesoutheast a strong downcoast (where downcoasthas been previously defined as toward Mexico)current was established in response to thedowncoast wind stress. The current speed even-tually peaked at 110 cm s21 as the eye wall madeits closest approach to the buoy around 1600UTC on 22 Aug. After the storm went ashoreover the central portion of Padre Island thecurrents at buoy J reversed to 50 cm s21 upcoastand remained that way for 3 wk. The surfacewater temperature decreased by 2 C as a result ofthe hurricane. Nearly 4 yr later HurricaneClaudette became a category one hurricane justas it made landfall on 15 July 2003. It remaineda tropical storm for 24 hr after making landfall.The track of Claudette took it over buoys N andV and slightly to the north of buoy W. Buoys Nand V recorded peak wind gusts of 56 and 46knots, respectively. As the storm approachedbuoy W from the east, a strengthening down-coast current was recorded by the buoy. Asustained downcoast current of 115 cm s21 wasrecorded for 5 hr as the eye wall made its closestapproach to the buoy and then went ashore.Even after the storm went ashore over MatagordaIsland at 1530 UTC the currents remaineddowncoast for nearly 3 d before reversing toupcoast. The surface water temperature de-creased by 1.5 C as a result of the hurricane.Hurricane Rita was an intense hurricane thatreached category five strength over the centralGulf of Mexico before weakening and comingashore near the Texas/Louisiana border asa category three storm. As it made landfall on24 Sept. 2005, the eyewall of hurricane Ritapassed directly over the top of buoy R. Before 23Sept. the currents were weak and inertiallydominated (see Fig. 13), but as the stormapproached from the southeast a strong down-coast current was established in response to thedowncoast wind stress. The current speed even-tually peaked at nearly 160 cm s21 as the eye wallpassed over the buoy. As the storm went ashorethe winds decreased and the currents quicklyrelaxed, but showed no signs of significantinertial oscillations that might be expected giventhe large and sudden increase in the wind speed.While this seems somewhat surprising, Rita wasfast moving, and the step change in wind speedlasted for less than one inertial period. At buoy F,sustained offshore currents of at least 90 cm s21

were recorded for more than 20 hr until 1400UTC on 24 Sept. The surface water temperatureat buoy R decreased by 3 C and by 2 C at buoy Fas a result of the hurricane. In each hurricane,Brett, Claudette, and Rita, the cyclonic winds

coupled with the curved coastlines to causea nearly identical near-shore current response,a strong downcoast current as the hurricanemakes its approach to the Texas coastline. Up tothe point of landfall this pattern is identical, butafter landfall the current pattern is noticeablydifferent.

CONCLUSIONS

In April 1995, Texas funded the deploymentand operation of a coastal network of near real-time current meters known as TABS. Thefounding mission of TABS was to improve thedata available to oil spill trajectory modelers.Nearly 12 yr later, TABS remains the only systemin the country with the primary mission of oceanobservations in the service of oil spill prepared-ness and response. This mission, coupled withstable GLO funding, has enabled us to improvethe technology and operational range of theTABS buoys, readily disseminate the resultsthrough the web, and fulfill important societaland science goals.

Today TABS forms the core of a regionalocean observing system for Texas waters that canbenefit a great number of research projects andoperational programs for industry, academia,and government. As the nation embarks on thedevelopment of an IOOS, TABS will continue tobe an active participant of the GCOOS regionalassociation and the primary source of near-surface current measurements in the northwest-ern Gulf of Mexico. The lessons learned during12 yr of operations serve as a valuable roadmapfor the operators of new ocean observingsystems.

The underlying theme behind the lessonslearned can be reduced to a few concepts:attention to detail; a highly competent anddedicated staff; stable, long-term funding; andthe flexibility to meet ever new challenges. Forexample, the availability of ships with therequisite size, speed, lifting capability, andaffordable daily structure needed for the TABSprogram has shrunk during the past severalyears. We no longer have the luxury of relying onnearby UNOLS research vessels. This has creatednew challenges for servicing the TABS buoys thatwe have met by chartering vessels and outfittingthe boat with winch, power pack, and A-frame;an endeavor that has been successful. Changes intechnology are relentless, and most provide anopportunity to improve the capability of thebuoys. Other than the basic shape of the hulls,there is little of the TABS buoys today thatoriginally went to sea in 1995. Failures are alwaysdisappointing, and we have had our fair share,


Fig. 13. Surface currents and water temperature at buoy R during the passage of Hurricane Rita. Beginning at1800 in the evening of 22 Sept. 2005 CDT, the temperature begins to drop and the currents increase as the eye ofthe hurricane approaches.


but they generally provide the opportunity to re-examine the design and make constructiveimprovements. Finally, we believe that a primarylesson of TABS is that an academic institution,coupled with a stable source of funding, is fullycapable of running an operational coastalobserving system for the long haul.

It is our intention that TABS continue toprovide operational ocean measurements off theTexas coast. We intend to continue to improvethe reliability of the TABS buoys through testing,field experience, and design modifications andto share that knowledge with the ocean observ-ing community. We are actively working toextend the capabilities of TABS from its original,and ongoing, mission of surface current andtemperature measurement to measurements ofthe water column, the sea floor, and the marinesurface layer. These additions will help increasethe density of offshore meteorological observa-tions and provide the vertical resolution ofcurrents needed for data assimilation into TABSforecast modeling efforts.

ACKNOWLEDGMENTS

We thank Greg Pollock (Deputy Commission-er for Oil Spills, TGLO), Jerry Patterson (Com-missioner, TGLO, 2003–present), David De-whurst (Commissioner, TGLO, 1999–2003),and Garry Mauro (Commissioner, TGLO,1983–1999) for their continued support of theTABS Programs. Frank Kelly played a pivotal roleat GERG during the first 5 yrs of the program.We also would be remiss if we failed to mentionthe marine technicians that have been a criticalpart of TABS over the past 12 yr: Paul Clark, R. J.Wilson, Marty Bohn, Alexey Ivanov, Eddie Webb,Willie Flemings, Andrew Dancer, Chris Schmidt,Chris Cook, Charles Ruckett, Mike Fredericks,Larry Lewis, Cole Markham, Kevin Lamonte,Tony Cocchiarella, Jorge Barrera, and MarcusTrichel. Finally we thank our two anonymousreviewers for their helpful comments. Thispaper does not necessarily reflect the views ofpolicies of Texas A&M University or the TexasGeneral Land Office. Mention of trade names orcommercial products does not constitute a com-mercial endorsement or recommendation foruse.

LITERATURE CITED

BENDER, L. C., S. DIMARCO, N. GUINASSO, J. WALPERT, AND

L. LEE. 2002a. Current observations offshore of PearlHarbor during the Ehime Maru recovery operations.Abstract OS32D-161, 2002 Ocean Sciences Meeting,Honolulu, HI.

———, N. L. GUINASSO JR., J. N. WALPERT, AND L. L. LEE

III. 2002b. Is there decadal scale information in theTexas coastal current? Abstract OS71F-09, AGU 2002Fall Meeting, San Francisco, CA.

BENDER, L. C., III, N. L. GUINASSO JR., AND J. N. WALPERT.2006. A program to test the feasibility of using theTexas automated buoy system to measure wavesimpinging on the Texas coast. Final report preparedfor the Texas General Land Office and the NationalOceanic and Atmospheric Administration underGLO contract no. 04-001, 24 May 2006.

CAMPBELL, L., J. N. WALPERT, AND N. L. GUINASSO JR. Anew buoy-based in situ optical early warning systemfor harmful algal blooms in the Gulf of Mexico. NovaHedwigia, in press.

CHAPLIN, G. F., AND F. J. KELLY. 1995. Surface currentmeasurement network using cellular telephonetelemetry, p. 177–180. In: Proceedings of the IEEEFifth Working Conference on Current Measurement,7–9 Feb. 1995, St. Petersburg, FL.

CHO, K., R. O. REID, AND W. D. NOWLIN JR. 1998.Objectively mapped stream function fields on theTexas–Louisiana shelf based on 32 months ofmoored current meter data. J. Geophys. Res.103:10377–10390.

COCHRANE, J. D., AND F. J. KELLY. 1986. Low-frequencycirculation on the Texas-Louisiana continental shelf.J. Geophys. Res. 91:10645–10659.

CROUT, R. L. 1997. Coastal currents from satellitealtimetry. Sea Tech. 38:33–37.

DIMARCO, S. F., AND R. O. REID. 1998. Characterization ofthe principal tidal current constituents on the Texas-Louisiana shelf. J. Geophys. Res. 103:3093–3109.

EZER, T., H. G. ARANGO, AND A. F. SHCHEPETKIN. 2002.Developments in terrain-following ocean models:intercomparisons of numerical aspects. Ocean Mod-elling 4:249–267.

GUINASSO, JR., N. L., L. C. BENDER, III, L. L. LEE, III, J. N.WALPERT, J. YIP, R. O. REID, M. HOWARD, D. A. BROOKS,R. D. HETLAND, AND R. D. MARTIN. 2001. Observingand forecasting coastal currents: Texas AutomatedBuoy System (TABS), p. 1318–1322. In: OCEANS2001 MTS/IEEE Proceedings, Marine TechnologySociety, Washington DC. 5–8 November 2001.

HELTON, D., AND T. PENN. 1999. Putting response andnatural resource damages in perspective, Paper 114,1999 International Oil Spill Conference. 8–11 March1999.

KELLY, F. J., N. L. GUINASSO JR., L. L. LEE III, G. F.CHAPLIN, B. A. MAGNELL, AND R. D. MARTIN JR. 1998.Texas Automated Buoy System (TABS): A publicresource, p. 103–112. In: Proceedings of the Ocean-ology International 98 Exhibition and Conference,vol. 1, Brighton, UK.

———, L. L. LEE, N. L. GUINASSO, J. N. WALPERT, R. R.LEBEN, AND C. A. FOX. 1999. Shelf-break currents offTexas derived from satellite altimetry versus observa-tions from moored buoys along a TOPEX-POSEI-DON ground track. EOS Trans., AGU 80:204.[Supplement.].

LEE, L. L., III, F. J. KELLY, AND N. L. GUINASSO JR. 1996.Armchair currents using TABS (Texas AutomatedBuoy System). EOS Trans., AGU 76:OS78. [AbstractOS22D-17.].


LEHR, B., D. SIMECK-BEATTY, D. PAYTON, J. GALT, G.WATABAYASHI, R. D. MARTIN, AND R. SOLIS. 1997.Trajectory Prediction for barge Buffalo 292 spill, p.25–33. In: Proceedings of the 1997 International OilSpill Conference, 7–10 April 1997, Fort Lauderdale, FL.

LI, Y., W. D. NOWLIN, JR., AND R. O. REID. 1996. Spatial-scale analysis of hydrographic data over the Texas-Louisiana continental shelf. J. Geophys. Res.101:20595–20605.

LOMB, N. R. 1976. Least-squares frequency analysis ofunequally spaced data. Astrophys. Space Sci.39:447–462.

MAGNELL, B. A., F. J. KELLY, AND R. A. ARTHUR. 1998. Anew telemetering environmental buoy for offshoreapplications, p. 179–197. In: Proceedings of theOceanology International 98 Exhibition and Con-ference, 10–13 March 1998, Brighton, UK.

MARTIN, R. A., F. J. KELLY, Linwood L. LEE III, AND

Norman L. GUINASSO JR. 1997. Texas Automated BuoySystem: real-time currents for oil spill response, p.75–78. In: Proceedings of the 1997 International OilSpill Conference, 7–10 April 1997, Fort Lauderdale,FL.

———, N. L. GUINASSO, JR., L. L. LEE III, J. N. WALPERT,L. C. BENDER, R. D. HETLAND, S. K. BAUM, AND M. K.HOWARD. 2005. Ten years of realtime, near-surfacecurrent observations supporting oil spill response,p. 541–545. In: Proceedings of the 2005 Internation-al Oil Spill Conference, American Petroleum In-stitute, Washington, DC. 15–19 May 2005.

MERZ, C. R. 2001. An overview of the Coastal OceanMonitoring and Prediction System (COMPS). MTS/IEEE 0-933957-28-9 Oceans 2001 Conference Pro-ceedings, 5–8 Nov. 2001, Honolulu, Hawaii. http://comps.marine.usf.edu/conf/cmerpaper.pdf

MICHEL, J. 2006. Assessment and recovery of submergedoil: current state analysis. Prepared for the Researchand Development Center of the US Coast Guard byResearch Planning Inc., of Columbia, South Car-olina.

NOWLIN, W. D., JR., A. E. JOCHENS, R. O. REID, AND S. F.DIMARCO. 1998. Texas-Louisiana Shelf circulation andtransport processes study: Synthesis report, vol. I,technical report OCS Study MMS 98-0035, 502 pp.,U.S. Dept. of the Interior, Minerals ManagementService, Gulf of Mexico OCS Region, New Orleans,LA.

OHLMANN, J. C., AND P. P. NIILER. 2005. Circulation overthe continental shelf in the northern Gulf of Mexico.Prog. Ocean. 64:45–81.

Ocean. U.S. 2006. The First US Integrated OceanObserving System (IOOS) development plan. Areport of the national ocean research leadership

council and the interagency committee on oceanscience and resource management integration. TheNational Office for Integrated and Sustained OceanObservations, Ocean.US Publication No. 9.

SCHOLZ, D., AND J. MICHEL. 1992. The Mega Borg oilspill: chronology and summary of spill responseactivities (chap. 1); and fate of the lost oil (chap. 2)Prepared for the Damage Assessment Center, NOAA,Rockville, MD, 88 pp. + app.

SCARGLE, J. D. 1982. Studies in astronomical time seriesanalysis. II. Statistical aspects of spectral analysis ofunevenly spaced data. Astrophys. J. 263:835–854.

WALPERT, J. N., N. L. GUINASSO JR., L. C. BENDER, AND L. L.LEE III. 2001. The effects of marine fouling on theperformance of a single-point acoustic dopplercurrent sensor mounted on a TABS-II spar buoy,p. 9956–9961. In: OCEANS 2001 MTS/IEEE Pro-ceedings, Marine Technology Society, WashingtonDC. 5–8 Nov. 2001.

WEISBERG, R., R. COLE, R. HE, J. DONOVAN, M. LUTHER, C.MERZ, J. WALSH, AND V. SUBRAMANIAN. 2002. A coastalobserving system and modeling program for theWest Florida Shelf, p. 530–534. In: MTS/IEEE 0-7803-7535-1 Oceans 2002 Conference Proceedings,29–31 October 2002, Biloxi, MS. http://comps.marine.usf.edu/conf/0530-0534.pdf

YIP, K.-J. J. Y., AND R. O. REID. 2002. An estimation ofEkman and geostrophic current over the Texas-Louisiana Shelf, p. 834–840. In: Oceans 2002 MTS/IEEE Conference Proceedings, vol. 2, Biloxi, MS.

ZAVALA-HIDALGO, J., S. L. MOREY, AND J. J. O’BRIEN. 2003.Seasonal circulation on the western shelf of theGulf of Mexico using a high-resolution numericalmodel. J. Geophys. Res. 108:3389, doi:10.1029/2003JC001879.

ZHANG, X. 2003. Design and implementation of anocean observing system: WAVCIS (Wave-Current-Surge Information System) and its application tothe Louisiana coast. Unpubl. Ph.D. thesis, LouisianaState University, 194 pp. http://wavcis.csi.lsu.edu/pubs/xpdis.pdf

(LCB, NLG, JNW, LLL) TEXAS A&M UNIVERSITY,GEOCHEMICAL AND ENVIRONMENTAL RESEARCH

GROUP, 833 GRAHAM ROAD, COLLEGE STATION,TX 77845-9668; (RDM) TEXAS GENERAL LAND

OFFICE, 1700 N. CONGRESS AVENUE, SFA BUILD-

ING, AUSTIN, TX 78701; AND (RDH, SKB, MKH)TEXAS A&M UNIVERSITY, DEPARTMENT OF OCEAN-

OGRAPHY, COLLEGE STATION, TX 77843-3146.Send reprint requests to LCB. Date accepted:August 7, 2007.


development, operation, and results ... - tabs.gerg.tamu.edu · the texas automated buoy system...

Documents