4

Remote Sensingof the OceanDavid E. WeissmanProfessorDepartment of Engineering

HOFSTRAHORIZONS

Just by looking at images of theEarth from space, it’s clear that theocean is a significant piece of theEarth’s story. In fact, the globaloceans represent more than 70percent of the Earth’s surface, and

are the primary driver of weather andclimate.

When you open the daily newspa-per or a weekly news magazine, you canfrequently witness the harmful impact ofthe physical forces of Earth’s oceans andatmosphere, and their profound eco-nomic and societal implications. Themore societies become dependent onresources and products from widely sep-arated regions, the more an event thatproduces local devastation, such asHurricane Katrina, most recently, alsoleads to disruptions across the globe.There is never enough predictive abilityand sufficient advance notice, or govern-mental capabilities, to protect all indi-viduals and property from harm. Thecatastrophic hurricane season of 2005demonstrates the value of spacebornesensors that observe atmospheric andoceanic processes to predict natural dis-asters. This demonstration indicates theurgency to both extend these capabili-ties and to improve management of ourland-based facilities. Although scientificconsensus indicates that these currentevents are part of a climatic cycle, thisunusual hurricane season accentuatesconcerns regarding the impendingeffects of human activities that are caus-

ing environmental changes at an acceler-ating rate.

Transitioning From SensorDevelopment to Scientific Analysis

For the purpose of defense andenvironmental applications, I have beenactively involved in radar remote sens-ing of the ocean for more than threedecades. In the early 1970s, I joinedwhat became a group of pioneers whowere developing the field of microwaveradar sensing of the ocean from aircraftand spacecraft. For young engineers andscientists such as ourselves, trained inthe rigorous and precise disciplines ofelectrical engineering and physics, thetransition to geophysical observationswas a bit of a cultural transformation.We would concentrate on producinginstruments that had very carefullydefined specifications and physical capa-bilities, but then realized that what wewere observing was so complicated thatit defied complete characterization.Additionally, we realized that what littlecharacterization information had beenmeasured, changed with location andtime. This is because, when you areobserving the oceans and the atmos-phere, you are dealing with processesthat have substantial variability over awide range of spatial scales, time varia-tions and interactions (heat andmomentum exchange). The best one canaccomplish is to pursue descriptions

with statistical and probabilistic termi-nology. For example, this means thatone must contrast the satellite estimateof the wind as an average over a 25-by-35 km area (footprint), with the meas-urement taken by an instrument on theocean (buoy) that provides the measure-ment at a single point. When discussingits calibration (accuracy), the conven-tional terminology uses the mean andstandard deviations between these ocean“point” estimates and the satellite foot-print estimates. Weather forecasters canbenefit from satellite estimates thatcover an area up to a thousand milesfrom shore, collected during a singleoverpass.

However, oceanographers need afull global view of winds, which cantake the satellite about two days to col-lect. Because the satellite cannot observethe complete globe at every instant, ana-lysts need to qualify their findings withan appropriate list of assumptions andgeographic boundaries. One then createsprograms and investigations that canreduce the uncertainty in the measure-ment of selected parameters, and thatcan improve predictive abilities. Peoplewho must predict weather and oceanevents usually add a “confidence” statis-

Figure 1: An artist’s rendition of the QuikSCAT satellite,which was launched in June 1999. The dish-like object at thebottom is the SeaWinds Scatterometer.

The Need for EnvironmentalObservations

HOFSTRAHORIZONS5

tical parameter, based on the number ofdays ahead one is trying to estimate.

Our initial efforts were to create,explore and evaluate possible technolo-gies using aircraft-based experiments toassess what might be valuable as satel-lite-based sensors, which would ulti-mately be used to observe as many seasurface and air-sea interaction quantitiesas possible. These would include waveheights, wave lengths and wind speed(including the stress on the sea surface).

In the recent decade, I have concen-trated on developing a system for meas-uring wind stress on ocean surfacesusing satellite radar from space. Thismethod of measurement has the abilityto measure sea surface wind stress, orthe force on the seas’s surface created bywind, and also to “correct” these windmeasurements when rain interferes withthe radar across the areas we would liketo observe. Since rain can, at any giventime, exist over about 5 percent of theworld’s oceans, the removal of the inter-ference it causes is a high priority to sci-entists and meteorologists using thesewind measurements for interpretingocean phenomena. Unfortunately, theerroneous wind data is commonly foundin the vicinity of synoptic scale (cover-ing many hundreds of miles) anddynamic atmospheric systems, such ascyclones and frontal boundaries, wheremeasurements of surface winds are mostimportant for scientific and operational(weather prediction) investigations.

Origins of Satellite Remote Sensing

Radar, the “Invention that Changedthe World” [Buderi, 1996], was inventedin the 1930s and played a decisive roleduring World War II. Many scientificleaders at the time stated that, consider-ing the broad scope of its usage on land,sea and air, radar contributed more tothe United States’ World War II victorythan the atomic bomb. Radar quicklybecame indispensable for all militaryand civilian aviation. During thewartime years, radar was a critical factor

in combat operations from aircraft andNavy vessels. Its potential for meteorol-ogy, to precisely locate and measureatmospheric events, was also recognizedand exploited for forecasting weatherchanges to guide mission decisions.

The field of Radar Meteorology thusbegan with World War II. Radar is alsoknown as an “active” sensor because itgenerates and transmits its own electro-magnetic energy. These waves arereflected by whatever medium or surfacethey encounter, and are then returned tothe source where they are stored, meas-ured and interpreted. This capabilityenables the study of the atmosphere inthree dimensions, and the study of landand ocean surfaces under all conditionsand circumstances. The data can bemade available immediately for urgentapplications or later for extended scien-tific investigations. Our current televi-sion-based weather forecasters show usthe color-coded images of rain locations,intensities and their motions.

The field of satellite remote sensingfor earth observations began in April1960 with TIROS-1, the first meteoro-logical observer weighing all of about200 pounds, whose launch was spurredon by the Soviet Union’s Sputnik in

1957. TIROS stands for Television andInfrared Observation Satellite. Congressdelegated to the then U.S. WeatherBureau the responsibility for managingand maintaining these satellites. WhileTIROS-1 consisted of only a TV camerato observe cloud formations, sevenmonths later it was joined by TIROS-2,whose instruments operated in both thevisible and infrared portions of the elec-tromagnetic spectrum. Both TIROS-1and TIROS-2 are designated as “passive”instruments, which only receive theradiation produced by thermal emissionfrom the atmosphere or the Earth’s sur-face, or their reflected sunlight. Theinvention of these instruments was fol-lowed by the invention of dozens of big-ger and better passive instruments, butstill using the same “physics,” that is,visible and infrared frequencies.However, in the early 1970s, anotherinnovation was made; microwave instru-ments operating at much lower frequen-cies were able to demonstrate their abili-ty to “see through” clouds and oftenthrough rain, greatly improving weatherforecasting skills and advancing scientif-ic knowledge. By that time, the NationalOceanic and AtmosphericAdministration (NOAA) was formed

Figure 2: SeaWinds measurement geometry illustrating the antenna beam locations for the different polarizations. Inner beam is horizontal polarization; outer beam is vertical polarization.

SeaWinds orbit track

802 km

nadir track

cross track

outerbeam

innerbeam

1100

km

1245

km

7 km/sec

18 rpm

700 km

900 km

6 HOFSTRAHORIZONS6

and charged by Congress with advanc-ing this field. However, since the 1970s,the prime agency for developing andtesting new microwave Earth sensorshas been the National Aeronautics andSpace Administration (NASA), becauseof its vast resource of scientific and engi-neering manpower, which was createdby the manned space exploration effort.The first Earth observations from space,for scientific investigations, were fromthe NASA Skylab launched by the SpaceShuttle in 1974, for a one-week orbitalmission.

The research community and theproject Science Teams associated witheach satellite instrument consist of abroad variety of specialists. The currentpopulation consists of engineers, physi-cists, oceanographers, meteorologistsand atmospheric scientists. We concen-trate on specific investigations andapplications: some groups focus on test-ing the accuracy of the instrument data,while others develop specific algorithms(mathematical formulas and instruc-tions) to convert the radar data intowind vector and stress estimates anddata products, while still other scientistsuse these data to investigate oceanicprocesses and events. There is bothinterdependence and independenceamong these highly talented and moti-vated investigative subgroups. Often ourstudies involve a combination of two ormore instruments (some may be surface-based), possibly on the same or differentsatellite platforms. Results and accom-plishments promptly sent to scientificpublications rapidly find their way intothe larger international community ofscientists, educators and policy makerswho will make decisions about the lifetimeof current sensors and their successors.

NASA’s Ocean Remote SensingProgram

Part of NASA’s mission is to developan understanding of the total Earth-Sunsystem and the effects of natural andhuman-induced changes on the globalenvironment. Our oceans play a major

role in influencing changes in theworld’s climate and weather, because ofthe continuous exchange of heat, mois-ture, momentum and gases. Collectingand analyzing long-term ocean datafrom satellites is a relatively new field ofexploration. The analysis of remotelysensed ocean data makes it possible tounderstand the ocean in new and excit-ing ways. Exploration is discoverythrough disciplined, diverse observa-tions and the recording of findings.Prior to satellite data, most of what wehave learned about the oceans had comefrom infrequent measurements collectedfrom ships, buoys and drifters. Ship-based oceanographers are limited tosampling the ocean in a relatively smallarea with often a great deal of difficulty,or to sampling in a large area withremarkably poor temporal sampling.Data from ships, buoys and drifters isnot sufficient to characterize the condi-

tions of the spatially and temporallydiverse ocean.

The agency has been observing theoceans from space for more than 25years. NASA launched SEASAT, the firstcivilian oceanographic satellite, on June28, 1978. The satellite carried five com-plementary sensors designed to monitorthe oceans from space. These sensorsincluded: 1) a radar altimeter (lookingstraight down) to measure wave heightsacross the ocean surface; 2) a microwaveradar scatterometer (looking across a700 km swath) to measure wind speedand direction; 3) a scanning multichan-nel microwave radiometer to measuresea surface temperature; 4) a visible andinfrared radiometer to identify cloud,land and water features; and 5) a syn-thetic aperture radar (very high resolu-tion) to monitor the global surface wavefield and polar sea ice conditions.

Figure 3: QuikSCAT image of winds on the surface of the Pacific Ocean on January 8, 2004.

HOFSTRAHORIZONS7

Although a massive short-cir-cuit in its power system ended alldata-taking operations after only105 days, the SEASAT instrumentsprovided as much oceanographicdata as had been acquired by shipsin the previous 100 years! The vari-ables that SEASAT measured in its shortlifetime are some of the most importantfor understanding the ocean and its rolein climate.

After successive NASA missions, incollaboration with the Japanese SpaceDevelopment Agency, led to some limit-ed successes and disappointments, aU.S. satellite dedicated to ocean windvector measurements alone was quicklydeveloped and tested within two years;hence, QuikSCAT was launched in June1999. The radar instrument it carriedwas referred to as the SeaWindsScatterometer (see Figure 1). It scansthe sea surface with a narrow “flash-light” shaped beam, rotating its 30 km“footprint” in a full circle, with 18 revo-lutions per minute (see Figure 2). Itsname, Scatterometer, is derived from thereflection or “backscattering” effect ofmicrowaves that are returned from thesmall sea surface waves that are closelycoupled (“connected”) to the wind vec-tor. One can observe a similar effectfrom a ship on a windy day, when yourback is directed to the setting sun andyou see the “twinkles” of sunlight fromthe short wavelets. Figure 2 shows twoseparate radar beams, one for each ofthe two polarizations (orientation direc-tions) of the electromagnetic waves.

Samples of the global wind vectorsderived from the Scatterometer over atwo-day period can be seen in Figure 3.The swirls here indicate the wind direc-tions, and the color code displays thewind magnitude. This is typical of thedata needed by oceanographers, meteo-rologists and related scientists to studyair-sea interaction, ocean circulation andtheir effects on short-term and long-term weather patterns and climate.However, necessary data can be lostwhen rain exists within the radar beam.The microwave frequency used by this

radar is such that it is very sensitive toprecipitation along the path of the nar-row radar beam. The radar beams repre-sented by the solid lines in Figure 2 aresubject to weakening (attenuation) andreflection by the rain back to the satel-lite (backscattering).

One-time, but important, eventssuch as storms or hurricanes will affectmaritime commerce and coastal regions.These can be incisively observed by theQuikSCAT scatterometer. This is seen inthe Hurricane Rita image of Figure 4. Atthe time this image was recorded, thehurricane had 195 kilometers per hour(120 mph) sustained winds. The imagesdepict wind magnitude in color andwind direction with small “barbs.”

Heavy rain occurred in the northernareas, and this prevented the estimationof wind directions, as can be seen in thisimage, where there are many areas with-out the direction barbs. Within the next24 hours, Rita would surge in power,packing winds near 280 kilometers perhour, becoming the fourth most power-ful storm ever recorded. Accurate sur-face winds are also crucial for the fore-cast of storm surge and coastal tempera-ture change related to the upwelling ofcold sub-surface waters. The stormsurge is the excess sea level that inun-dates coastal communities; its predictionis used in decisions about populationevacuations.

Hofstra’s Recent Contribution –Combining NASA and NationalWeather Service (NWS) Resources

It is possible to make specializedspace-based instruments for the meas-urement of rain. However, because theseinstruments are so far away, they havepoor spatial resolution (they onlyobserve areas larger than individual raincells). Therefore, they cannot resolve thehighly variable, smaller scale 3-D pre-cipitation structure within the largerbeam of the sensor. The optimuminstrument for measuring rain intensityis the ground-based National WeatherService (NWS) next generation radar(NEXRAD). Many of these are stationednear the coasts, and can observe boththe horizontal and vertical structure ofthe rain within each Scatterometerbeam. Their radar reflectivity measure-ments can be converted into the desiredrainrate parameters. These are then usedto calculate the attenuation andbackscattering, which can then be usedto correct the data measured at the satel-lite.

Since 1999, Hofstra, in collabora-tion with colleagues at Florida StateUniversity and the New York City officeof the National Weather Service, hasbeen in involved in a research project todetermine the locations and magnitudesof errors caused by rain within the beam[Weissman et al., 2002]. Scientistsworking on this project have beendeveloping methods for correcting theseerrors, using independent measurementsof the rain from which computation ofthe attenuation and backscatteringamounts can be made, resulting in moreaccurate wind estimates. This willexpand the spatial areas across whichaccurate, useable winds are available toscientists using the data for near-termpredictions of ocean and weather events,and for long-duration climate studies.

The NWS routinely acquires all theNEXRAD data from the entire nationalnetwork of stations, and stores it in theNational Climatic Data Center (NCDC)in Asheville, North Carolinia. However,

Figure 4: Hurricane Rita observed by NASA’s QuikSCAT satellite onSeptember 20, 2005, at 10:29 UTC (6:29 a.m. EST). The swirlsrepresent the wind vectors at the sea surface. The color codeindicates wind magnitude. Missing arrows indicate areas whererain prevents the sensing of winds at the surface.

8 HOFSTRAHORIZONS

the use of this data for the purposes ofthis research involves complex logisticaland algorithmic operations. This capa-bility was developed, starting in 1999,with financial and scientific contribu-tions from the New York City office ofthe NWS. This facility is located inUpton, New York, on the property of theBrookhaven National Laboratory.Scientific Operations Officer Jeff Tonguehas been an active collaborator in theacquisition, processing and applicationof the NEXRAD data to the scatterome-ter studies. Additional support was pro-vided by the NWS Office of Science andTechnology, at its headquarters in SilverSpring, Maryland.

This research project, with the NewYork City NWS office, was the key facil-itator in having numerous NWS officeslocated along the Atlantic and Gulfcoastlines (including Puerto Rico)incorporate the Scatterometer oceanwind measurements into their dailyanalysis of weather conditions for theirlocal areas of responsibility and fore-casts. This was facilitated through mylongtime collaboration with the FloridaState University Center for OceanAtmospheric Prediction Studies, specifi-cally with Drs. Mark Bourassa andJames O’Brien (the director). Theirextensive resources and expertise havecreated the mechanism by which eachcoastal NWS office could use this NASAsatellite-derived near real-time windinformation. The success of this workled to the Navy and National Oceanicand Atmospheric Administration(NOAA) developing their own pathwaysfor bringing these data into use for theirforecasting and analysis. Consideringthe bureaucratic maze and variety ofgovernment agencies and project offices,this partnership is a notable accomplish-ment. The value of this data is height-ened during the hurricane season whenthe NWS offices and the NationalHurricane Center in Miami are strainingto acquire every possible piece of infor-mation about these savage storms.

Hurricane Isabel

One of our (“Hofstra”) researchinvestigations focused on HurricaneIsabel, at the time it crossed onto theNorth Carolina coastline on September18, 2003 [Weissman et al., 2005]. Thisfortuitous observation by the scatterom-eter occurred at about 1600 GreenwichMean Time (GMT). The storm was cen-tered in the swath of the satellite sensor,providing a rich source of scatterometerdata, in addition to the NEXRAD datathat was gathered. The conversion ofthe data files acquired from NCDC into3-D arrays of reflectivity has been a verydemanding task, but handled very effec-tively by my undergraduate researchassistant, Greg Apgar, one of ourElectrical Engineering majors at Hofstra.A sample of the NEXRAD 3-D reflectivi-ty (we use a logarithmic amplitude unit,called dBZ) for a six-minute NEXRADtotal scan for this event, is illustrated in

Figure 5. This representation onlyshows planar slices at the indicated ele-vations and positions.

As this work progressed, we usedthe rain volume measurement to per-form corrections to the Scatterometermeasurements and thereby produce cor-rected wind speed measurements. Thecritical assumptions underlying thesesteps, namely the formulas for convert-ing the NEXRAD measurements intoscatterometer beam parameters, are inthe process of being evaluated. Thiswork is in close collaboration with MarkBourassa of Florida State University.These test cases will continue to extendour understanding of the physicalprocesses inherent in these circum-stances. We will also share our resultsand methodology with NASA’s lead insti-tution for this satellite project, the JetPropulsion Laboratory in Pasadena,California, to enable them to produceimproved global wind vector products.

Figure 5: NEXRAD measurements of the volumetric radar reflectivity (from Morehead City, North Carolina, on September 18, 2003)within Hurricane Isabel. This was collocated with a SeaWinds Scatterometer overpass of Hurricane Isabel. The three horizontalslices were made at altitudes of 0.5, 3.0 and 6.0 km. The elevation plane shows quantized vertical steps of 0.5 km. The eye ofIsabel was located at approximately 35°N, 76°W. The reflectivity scale is logarithmic, in dBZ, and indicated in the colorbar.

HOFSTRAHORIZONS9

David E. Weissman began teaching atHofstra University in 1968, shortly after earn-ing a doctorate in electrical engineering atStanford University, where he was also activelyengaged in leading research on radar remotesensing. He received an undergraduate degreein electrical engineering from New YorkUniversity. Upon arriving at Hofstra, his educa-tion in electrical engineering, combined withhis previous industrial experience, providedProfessor Weissman with an excellent back-ground for offering new leadership to Hofstra’sElectrical Engineering program. HofstraUniversity’s program rapidly earned a reputa-tion as the best electrical engineering programon Long Island for evening students.

Professor Weissman is a professor ofengineering at Hofstra, and teaches intermedi-ate and advanced courses in electronics, elec-tromagnetics and communications. He has notonly taught a wide range of courses, but hasalso created courses for the department.Additionally, he serves as coordinator of theElectrical Engineering program, whenever suchservice is requested.

Professor Weissman has a longstandinghistory of working with and receiving supportfrom NASA. At the beginning of his career atHofstra, summer recess breaks quickly becameopportunities to conduct research-in-residenceat NASA Research Centers such as theElectronics Research Center, Cambridge,Massachusetts, and the Langley ResearchCenter (LaRC), Hampton, Virginia. ProfessorWeissman’s research at NASA was a catalystfor his continued work with that agency, andfor his receipt of annual research grants fromthe LaRC and NASA Headquarters to developand evaluate radar measurements of oceansurface waves and air-sea interaction from air-craft and satellites. In 1975 ProfessorWeissman was awarded a National ResearchCouncil Senior Postdoctoral Fellowship, andspent the next year at the NASA Jet PropulsionLaboratory in Pasadena, California.

Professor Weissman has received NASAresearch grants during almost all of his 37 yearsat Hofstra. Invariably, these have involved closecollaborations with engineers and scientists atresearch centers or other universities. During asabbatical in 1990, he was a visiting scientist atthe Woods Hole Oceanographic Institution(WHOI), where he began an additional long-term, ocean remote sensing research programunder the auspices of the U.S. Navy, Office ofNaval Research. There, he collaborated with sci-entists at WHOI and the University ofWashington. In 1999 he began receiving supportand collaborating with the New York City officeof the National Weather Service, which led tosome of the work discussed in this article. He hasauthored or co-authored more than 40 articles inarchival journals and edited proceedings, andhas presented more than 100 conference papers.In 1990 he was elected a Fellow of the Institute ofElectrical and Electronic Engineers (IEEE). In1977 he received the Best Journal Paper Awardfrom the IEEE Transactions on Antennas andPropagation. Additionally, he has received otherawards from NASA and the IEEE. He currentlyserves on the Executive Committees of the IEEEOceanic Engineering Society (since 1975), andthe Geoscience and Remote Sensing Society(since 1999).

The ultimate solution for the rain prob-lem will be found when NASA, in col-laboration with the Japanese AerospaceExploration Agency (JAXA) and ourNational Oceanic and AtmosphericAdministration (NOAA), launches anew satellite carrying both a scatterome-ter and passive microwave radiometerwhose main function is to measureatmospheric properties, including rain-rate [Yueh et al., 2003]. This radiometerscans the same atmospheric region asthe Scatterometer. Its complete designa-tion is the Advanced MicrowaveScanning Radiometer (AMSR). Theknowledge base and techniques devel-oped for the NEXRAD precipitation meas-urements, in this Hofstra project, will beneeded to evaluate the AMSR’s ability tocorrect the Scatterometer. In effect, thismethodology will be the calibrationinstrument that will define the limits ofrain corrections for global wind studies.

Acknowledgments:

The principal support for theseendeavors has been provided by thePhysical Oceanography Program of theNational Aeronautics and SpaceAdministration in Washington, D.C.However, vital contributions (both col-laborative and financial) have also beenprovided by the New York City office ofthe National Weather Service and itsOffice of Science and Technology inSilver Spring, Maryland. In addition, weappreciate the financial resources pro-vided by the University Corporation forAtmospheric Research (Boulder,Colorado) through the CooperativeProgram for Operational Meteorology,Education and Training (COMET).

References:

Buderi, R. (1996). The Invention ThatChanged the World. New York: Simon andSchuster.

Weissman, D.E., M.A. Bourassa and J.Tongue. (2002, May). “Effects of rain-rateand wind magnitude on SeaWinds scat-terometer wind speed errors,” J. Atmos.Oceanic Technol., Vol. 19, No. 5, pp.738-746.

Weissman, D.E., M.A. Bourassa, J.J. O’Brienand J. Tongue. (2003, December).“Calibrating the QuikSCAT/SeaWindsradar for measuring rain-rate over theoceans,” IEEE Trans. Geosci. Remote Sens.,Vol. 41, No.12, pp. 2814-2820.

Weissman, D.E., G. Apgar, J. Tongue andM.A. Bourassa. (2005, May). “CorrectingScatterometer winds by removing raineffects,” Bulletin of the AmericanMeteorological Society, Vol. 86, No. 5, pp.621-622.

Yueh, S.H., B.W. Stiles and W. Timothy Liu.(2003, November). “QuikSCAT windretrievals for tropical cyclones,” IEEETrans. Geosci. Remote Sens., Vol. 41, No.11, pp. 2616-2628.