ALTIKA : A MICRO-SATELLITE Ka-BAND ALTIMETRY MISSION

IAF – 01 – B.2.10ALTIKA : A MICRO-SATELLITE Ka-BAND ALTIMETRY MISSIONJ. VERRON 1 , P. BAHUREL 2 , E. CAUBET 3 , B. CHAPRON 4 , J.F. CRETAUX 5 ,L. EYMARD 6 , C. LE PROVOST 5 , P.Y. LE TRAON 7 , L. PHALIPPOU 3 ,F. REMY 5 , E. THOUVENOT 8 , J. TOURNADRE 4 , P. VINCENT 81 , LEGI, BP 53X, 38041 Grenoble Cedex, France2 , SHOM/CMO, 18, Av. E. Belin, 31401, Toulouse Cedex, France3 , Alcatel Space Industries, 26 Av. Champollion, BP 1187, 31037 Toulouse, France4 , IFREMER/DOS, BP 70, 29280 Plouzane, France5 , LEGOS, 18, Av. E. Belin, 31401, Toulouse Cedex, France6 , CETP, 10-12, Avenue de l'Europe, 78140 Vélizy, France7 , CLS, 18 Av. E. Belin, 31401 Toulouse Cedex, France8 , CNES, 18 Av. E. Belin, 31401 Toulouse Cedex, France52 th International Astronautical Congress1-5 Oct 2001 / Toulouse, FranceFOR PERMISSION TO COPY OR REPUBLISH, CONTACT THE INTERNATIONALASTRONAUTICAL FEDERATION3-5 Rue Mario Nikis, 75015 Paris, France

Abstract.Ka-band altimetry has been proposed by Verron etal. (2001) to complement the altimetry referencemissions of the Jason class. The goal of this paper isto provide information on the so-called “AltiKa”proposal in terms of science requirements,responsive technical specifications, and a prioriperformances. Besides the fact that the feasibility ofembarking an AltiKa payload on a microsatellitehas already been assessed, there is also theinformation that a small launcher may have thecapability of launching several microsatellitessimultaneously (up to three) in a number ofconfigurations that may be adapted to the spacetime requirements of high resolution altimetry.IntroductionThe development of altimetry has been steady andthe benefits outstanding. There has been a quasicontinuousseries of missions, starting withGEOSAT (1985), and then ERS-1/2 (1991 and1995), and TOPEX-POSEIDON (1992). These willcontinue from Year 2000 with Jason-1 andENVISAT, that are respectively scheduled forlaunch in 2001.More generally, the heritage of past and currentoceanographic space missions has paved the way forthe incremental implementation of the spacecomponent of the Global Ocean Observing System.Because they have already demonstrated their “preoperational”maturity, in terms of end-to-endperformances, value to users and affordability,altimeter and scatterometer missions are the firstimmediate candidates for operational transition.Short-term decisions are therefore required to securecontinuity and enhancement of the current altimeterand scatterometer systems in the GODAE (GlobalOcean Data Assimilation Experiment) timeframe andbeyond.In the short term, the future looks secure with a twosatellitealtimeter system being maintained for thenext five years with Jason-1 and ENVISAT. It isforeseen that the combined use of their data withmeasurements from independent gravity missionssuch as CHAMP (2000), GRACE (2001) andpossibly GOCE (2004) will be another breakthrough.It should enable absolute estimates of ocean currents,to be obtained by subtracting an independentestimate of the geoid from individual altimeterheight measurements.A first firm decision is however needed in the veryshort term to secure continuity through Jason-2measurements in the GODAE window. Furtherdecisions will be needed to maintain a multi-satellitesystem providing the required accuracy, coverageand sampling beyond the GODAE window, after theend of life of ENVISAT. Such a system shouldprovide an affordable service equivalent to oneJason-class mission and one or two medium accuracymissions of the ENVISAT class. Several optionsneed to be evaluated, taking into account results of R& D and demonstration missions aimed at validatingnew candidate observing techniques.The goal of this paper is then to provide the readerwith the status of Ka-band altimetry, considering theembarkment of the system on a microsatelliteplatform.1. The AltiKa proposal : about scienceobjectivesThe science objectives of the AltiKa mission aredescribed in Verron et al. (2001) and in Remy et al.(1999) in very details.Hereafter, we only recall those of the objectives thatmost impact the design of the altimeter instrumentand the payload.1.1 Ocean Mesoscale variabilityBased on the need to resolve the mesoscalevariations and to study the effect of thesefluctuations on the energetics of the meancirculation, the spatial and temporal resolutionrequired is set by the need to directly observe themesoscale variability to a degree that allows, inconjunction with assimilative numerical models, aproper description of the oceanic eddy field and theinteractions of this field with the mean field.In 1999, G. Mitchum et al. summarized some of therequirements for a well adapted sea-surface heightsystem: The section below partly derives fromMitchum et al.’ statements concerning spacealtimetry. The constraint concerning the spatial andtemporal resolution of a sea-surface height systemmeans that we must be able to resolve signals thatoccur at spatial scales on the order of several oceanicRossby radii, and on time scales of days to weeks,which cannot presently be done with a singlealtimeter. The important question is how manyaltimeters must be available at any given time, andwhat orbits should these satellites occupy? Jacobs etal. (1999) have estimated that three altimeters areneeded to directly observe the mesoscale variationswithout additional statistical or dynamical modelinput. If assimilative models are used, however, twosatellites may be adequate. In a different set ofsimulations, Le Traon and Dibarboure (1999)

conclude that most of the improvement inreproducing the mesoscale variability is obtained ingoing from a single altimeter to two flyingsimultaneously. Adding further altimeters improvesthe results, but to a lesser degree.It is clear that at least two altimeters are required,and that additional altimeter would still improve thescience return, but the optimal number is still understudy.Other ocean topics may also add spatial andtemporal requirements that will impact the number ofaltimeters needed to get high space-time resolution.1.2 Contribution in coastal altimetry andcontinental watersThe initial single satellite AltiKa proposal did nothave the goal to answer the space-time requirementscorresponding to a fully adapted sampling of thecoastal dynamical features. We’ll come to thepossibility of flying a multisatellite AltiKaconstellation in the conclusion.However, coastal altimetry has been a driver toderive new instrumental spcecifications to answernew performance requirements both in terms ofnoise and along-track space resolution.At the same time, we also considered how a newaltimetry system may answer requirements dealingwith the observation of continental waters (lakes,reservoirs, rivers) such as the requirements listed inthe HYDRASAT proposal. Once again, besides thetime sampling issue, we looked at the orbit patternthat may be useful to have in complement of Jason tosample the major continental water areas; we alsotried to focus on space resolution requirementsavoiding to design an instrument that would bespecific to the altimetry of lakes and rivers keepingin mind that the main driver is ocean altimetry.Some main technical features of the Ka-bandaltimeter that is part of the AltiKa payload aredirectly derived from the coastal and continentalwater requirements : they are described in moredetails in sections 2 and 3.1.3. Ice sheet ice monitoringAs the AltiKa payload has been foreseen to alsocontinue the ERS/ENVISAT series of altimetermeasurement of ice sheets, we also considered somebasic features of ice altimetry that may impact thetechnical specifications of AltiKa.The monitoring of the ice-sheet height overAntarctica and Greenland is of primary importancefor climate studies. However, one of the majorlimitations of the ice sheet mass balance study, ispresently the poor knowledge of the radar wavepenetration within the snowpack, in Ku-band for theERS1/2 altimeters. The induced volume echo is ofthe same magnitude as the surface echo, yielding to acritical situation. The induced effect is twofold: itdirectly acts on the surface elevation accuracy and,due to the temporal change in the snowcharacteristics from daily to yearly scales, itproduces a long term analysis error that can hide thelong term trend.Then, the objective was to try specifying thealtimeter so that penetration effects would beminimized with respect to Ku-band altimetry.2. Ka-band payload characteristics2.1. General requirements and proposed payloadThe above objectives were translated into some mainfeatures of the Ka-band payload :• get an altimeter instrument whose range noiseperformance may be so that the recovery of theocean short wavelength features is improved,• get an altimeter instrument with an improvedspace resolution along-track and a betterperformance when approaching or leavingcoastal boundaries,• get an altimeter instrument that will minimizethe penetration effects over media such ascontinental ice,• embark an orbitography system that will ensurea high level of accuracy in terms of orbitographyand that will ease the connection of historicalaltimetry series within a common well surveyedgeodetic reference frame,• embark a microwave radiometer that will helpcorrecting altimeter measurements for wettroposphere effects.To answer the previous requirements, it is proposedto compose an AltiKa payload with :• A single frequency Ka-band (35 GHz) altimeterinstrument (see Phalippou et al. (2000) fordetails)• A two-frequency radiometer,• A DORIS receiver,• A passive laser retroreflector array.

2.2. Summary of some characterisitics of the KabandaltimeterFrom the above requirements, we derived altimeterinstrument specifications that are compared with thePOSEIDON-2 specifications in the table below :AltiKa POSEIDON2Altitude 500 to 800 1340 kmkmInclination ~98° 66°Lifetime 2 years 5 years(objective :3 years)Frequency 35.75 GHz 13.65 GHzBandwidth 480 MHz 320 MHzPRF 4000 Hz ~2000 HzSignal to noise > 10dB > 10dBratioSpectral analysis 128 gates 128 gatesIntegration 25 to 50 ms 50 msdurationMass

2.4. An integrated Altimeter + RadiometerinstrumentOne of the initial requirements to design the AltiKapayload was that it should be possible to embark iton a microsatellite and that it could also be providedas a whole to become a passenger on an apportunityplatform.This has lead us to define an integrated instrumentthat allows for interface optimization and reductionof the number of units (boxes) . Indeed, theintegrated instrument is composed of :• One microwave unit that gathers all microwavefunctions of the altimeter and the radiometer,including the calibration functions of theradiometer and the sources of the antenna.• One processing unit that gathers all functionsdedicated to the altimeter and radiometerprocessing, as well as a global management unit.3. Performances and impact on somescience objectives3.1. Minimizing the range noise of the altimetermeasurementThe Ka-band (35 GHz) is more interesting than theKu-band for the altimeter since it improves the linkbudget and allows larger bandwidth (up to 500MHz) and pulse repetition frequency (4 kHz).The selected 480 MHz bandwidth provides a 0.3 mvertical resolution.Due to the smaller antenna beamwidth, the Brownecho has a sharper shape in Ka-band than that whatis obtained with conventional altimeters in Ku/Cband(e.g. POSEIDON 2).The proposed architecture for the Ka-band altimeteris based on the classical deramp technique for pulsecompression and it takes benefits of Frenchexperiences from the realisations of POSEIDON 1 &2. A wide band chirp (~480 MHz) of about 107 µsduration is emitted. The digital processing includesthe real time FFT, echo integration and radiometerprocessing (power averaging). The main limitationof Ka-band is the non-operationality for rain ratehigher than several mm/h (5 to 10% of time forAltiKa depending on the geographic area).Range noise comparison with POSEIDON 2altimeter shows that AltiKa will provide excellentperformances (see Figure 1 where the Ku/C and Kucurves correspond to the dual and Ku-band-onlyfrequency estimates of POSEIDON 2 range noise).Figure 1: Comparison of AltiKa (at 800 km) andPOSEIDON 2 (at 1340 km) performanceSince no ionospheric correction is required forAltiKa (DORIS might indeed provide a correction inextreme, rare ionospheric conditions), the expectedimprovement is close to a factor of 3 (for example,1-s range noise at 2-m SWH would be less than 0.8cm for AltiKa vs 2.2 cm for POSEIDON 2).3.2. Space resolution improvement wrt classicalKu-band altimetryThis very good range performance will be associatedwith reduced footprints and better coastalperformance as well.Performance for coastal applications depends :(i) on waveform / footprint relationship(ii) on antenna diagram / footprint relationship(to obtain attenuation of land contributionto echo(iii) on tracking performance for complex(coastal or in-land) echoes.AltiKa characteristics have been optimized for these3 issues :- AltiKa is close to a beam limited altimeter :there is no ‘plateau’ in the echo, since it stronglyattenuates shortly after leading edge due to thesmall antenna aperture. This will greatly reducethe pollution of ‘land gates’ into ‘ocean gates’.- Tracking loops were designed at the same timeas SIRAL/CRYOSAT altimeter for whichcontinuous tracking in ocean/land/ice conditionsis required. Thus, AltiKa is supposed to providedata in most coastal zones and in-land waterareas. Moreover, in case of loss of track,acquisition/locking phase is required to beshorter than 500 ms, allowing a quick return to

nominal tracking conditions (about 3 km alongtrack).These values will be associated with small footprintsfor the radiometer as well since altimeter andradiometer will share the same antenna. In additionto the altitude reduction, which is the maincontributor to the radiometerfootprint reduction, it may also be noticed that theabsence of an additional 19-GHz channel to theradiometer is also an advantage since this lowfrequency would give the largest footprint.POSEI--DON2(Ku-band)POSEI--DON2 (Cband)AltiKa (800km)AltiKa (500km)Leadingedge(2-mSWH)128-gateecho(2-mSWH)Footprintdiameter at3-dBattenuation5.3 19.1 19.55.3 19.1 49.54.1 12.7 7.23.4 10.3 4.6Table : Footprint diameter for different criteria(values given for a 75-cm antenna for AltiKa)This is why it is expected that AltiKa will give usefuldata as close as 5 km from a majority of coastalareas. Associated with a shorter data cycle (probably500 ms instead of 1 s), it will probably represent amajor improvement in mesoscale and coastalapplications of radar altimetry.3.3. Minimizing penetration effects overcontinental iceAn empirical analysis of the temporal variability ofthe satellite radar altimetric observationdemonstrates that the Ku-band radar penetrationabove the dry snowpack of Antarctica is between 5m in the interior to 14 m at a lower altitude, beforedecreasing due to wetness near the coast. Theabsorption coefficient of dry snow being 0.05 m -1 , itinduces a scattering coefficient comprised between0.05 and 0.16 m -1 , depending on ice the grain sizewhich varies from 0.2 to 3 mm.In Ka-band, Mie scattering can be assumed, then thescattering coefficient is inversely proportional to theradar wavelength at a power 4. From Ku to Kabands, the scattering coefficient will increase by afactor 55: the volume scattering will then be clearlydominant over the surface scattering. Inversely, theradar wave extinction will also increase to valuescomprised between 2.75 and 8.75 m -1 , leading to apenetration depth over snow surface between 0.3 and0.1 m. The altimetric observation and heightrestitution will thus correspond to a thin subsurfacelayer. The accuracy will then be considerablyimproved.Moreover, ice grain size, that is one of the pertinentclimatological snow parameters, could be directlyderived from Ka-band scattering coefficient, whichis not possible from present day Ku measurementsthat also highly depend on surface roughness.However, the major benefit will lie in the reductionof the induced long term bias due to the temporalchange in the snowpack surface. Indeed, a strongtemporal variability of the altimetric observationsover Antarctica ice sheet has been recently exhibitedwith the help of the ERS-1 three day orbit mission.These temporal variations are linked tometeorological events and consequently play a roleon a large band of the temporal spectrum, from fewhours to few decades. They are due to changes insurface roughness induced by change in wind, whichmodifies the echo surface part and then, both thewaveform shape and the elevation recovery. Evenwith a dedicated correction, the residual correctionyields to a raw noise estimate of 10 cm.3.4. Propagation losses3.4.1. Attenuation effectsIt is known that propagation of EHF waves throughthe troposphere may suffer from severe attenuationdue to the interaction of the electromagnetic wavewith the atmospheric contents. As far as the Ka-bandaltimeter is concerned, current requirement is thatthe instrument provides nominal performance 90%of time, with an objective as high as 95%. Anextensive study has been made to analyze thecontributions from :(i) gazes : attenuation is calculated for oxygen andwater vapour. Even if the allocated frequency rangeis nearly ideally chosen with respect to the waterpropagation spectral window, strong attenuation mayresult from moderate to high humidity contents.Values may vary from 0.4 to 2.1 dB when humidityvaries from 7.5 to 50 g/m3.(ii) clouds : 7 types of clouds are taken into accountfor the propagation models. Attenuation is estimatedfrom their liquid vapour contents and depths. Asexpected, cumulus and cumulonimbus are generally

the main contributor (0.2 to 0.7 dB) to the total loss,and sometimes stratocumulus (0.05 to 0.2 dB).(iii) rain : attenuation by rain, and possibly snow, iscomputed from rain rate and rainfall thickness.Other possible contributors (ice, fog) have beenshown to have no significant influence (order of 0.1dB for the fog).As far as rain is concerned, maximum precipitationrates are known as less than 1.5 mm/hr everywhere(95% through time) or even 1.0 mm/hr (90% throughtime) over ocean and polar areas, except in a regionoff the Philippines islands for which previous figureshave to be doubled. Following accepted statisticaldistributions, one can associate some classicallayering distributions with the previous rates. Then,total attenuation due to rain may be computed.Results show that attenuation is lower than 1.8 dB(95%) or even 1.2 dB (90%). Given these figures,we can conclude that :- Case 1 : no rain or very low rain rates, then anominal functioning of AltiKa is ensured,- Case 2 : rain rates greater than 1.5 mm/h, thenaltimeter measurements will be degraded oreven not acquired for the largest rain rates,- Case3 : moderate rain rates less than 1.5 mm/h,then the echoes will have to be corrected using aspecific algorithm modeling the perturbation ofthe echoes by rain cells.To refine the previous statements, 7 years of TMRdata from TOPEX/POSEIDON have been used tocheck some seasonal effects and to analyze the effectof local time which may be an important parameterto select in the case of a sun-synchronous orbit forAltiKa. Translation of TMR measurements into rainattenuation has been performed using dedicatedalgorithms and results have been mapped versuslocation, local time, month in the year, and on amulti-year time scale. On the one hand, the studyhas confirmed that Ka-band data availability will begreater than 90% over the world ocean and evengreater than 95% for some areas such as the NorthAtlantic. On the other hand, it appeared that localtimes from 6 am to 12 am, and from 6 pm to 12 pmare the worst cases for rain, which may provide a« light » constraint for the optimization of the linkbudget of the instrument and select the mostadequate local time for a sun-synchronous orbit.3.4.2. Impact of rain cells on waveformanalysisFollowing considerations are derived fromTournadre (1999).Approximations made in previous studies to estimatethe effect of rain cells on Ku-band altimeterwaveforms are no longer valid ; a second newmethod has thus been developed based on the trueintegration of rain attenuation.It may be shown that the difference between the twomethods is significant even for low rain rates (fewtenths of mm/hr). The method can be used tocompute the altimeter waveform for any rain field. Itshows that a precise retrieval of altimetricparameters (significant wave height, wind speed andsea surface topography) will require to take intoaccount corrections for rain cell perturbations evenfor low rain rates.The sensitivity of Ka band measurements to rain mayalso lead to a way of estimating very light rainfallover the oceans, for which we dramatically lack ofinformation and which could lead to a greatimprovement of our knowledge of the oceanic rainclimatology.3.5. Anticipated radiometer performancesThe following figures present the major elements ofthe radiometer performance budget :-resolution : about 0.3 K for the 23.8 GHzfrequency, and 0.4 K for the 37 GHz frequency-sensitivity : better than 1K-accuracy : better than 3K (worst case estimate).4. Conclusion : Status of the Ka-bandaltimetry studiesThe Ka-band altimetry concept that is proposedexhibits enhanced performances in terms of verticalresolution, time decorrelation of echoes, spaceresolution, range noise. In addition, an adaptedtracker algorithm has already been designed toperform near-continuous altimetric tracking aboveall surfaces, which is especially important whenapproaching or leaving coasts. Also, it appears that,even if Ka-band measurements are more sensitive torain than Ku-band measurements, rain does notprevent from acquiring a fairly high percentage ofmeasurements (objective is up to 95%) except for

strong rain rates : nevertheless, the development ofnew algorithms taking into account rainperturbations of the altimeter echoes is necessary andpresently under study.Of course, even if not presented in this note, thepresence of the DORIS system and of a laserretroreflector array in the AltiKa package ensuresthat the classically required orbitographyperformances will be reached and that the AltiKaaltimeter series may be linked toTOPEX/POSEIDON, Jason and ENVISAT series ina well monitored geodetic reference frame.one being to have the 3 satellites flying on differentorbit planes that would be separated up to 0.24degree corresponding to one third of the typical 35day ground track spacing (formation flying). These 2configurations may of course answer differentlyspace time requirements of high resolution oceanaltimetry and maybe inputs for ocean simulation toillustrate their relative interest in terms of sciencereturn.Preliminary studies have shown that a microsatelliteplatform such as the MYRIADE series underdevelopment at CNES is well suited to embark theAltiKa payload (Figure 2). A recent study has beenperformed considering a classical sun-synchronous35 day ENVISAT-type orbit (6 am ascending node)at an altitude of 800 km and a 98 degree inclination,also requiring a +/- 1km repeatability for the groundtrack. The main issues relating with theaccommodation of the AltiKa payload on themicrosatellite platform have been analyzed : mass,power, pointing, fields of view of the variousinstruments, TM/TC rate, thermal andelectromagnetic environment, etc. A detailed reportis still under writing at the present time, theconclusion of it being that AltiKa is a validcandidate for the embarkment on such amicrosatellite.Figure 3 : Example of accommodation of 3 AltiKamicrosatellites inside a DNEPR fairingFigure 2 : Artist view of AltiKa satelliteDuring the same study, a very preliminary analysisof possibly launching several (up to 3) AltiKamicrosatellites with the same small launcher hasbeen performed (Figure 3). The feasibility of such anapproach has been assessed for 2 configurations :one is to have 3 satellites flying in the same orbitplane with a 120 degree lag in terms of position onthe orbit (classical constellation type), the secondIt shall be noticed that the technical studies that arereferred to are preliminary ; they need to beconsolidated by more detailed analysis including thepossibility of considering other microsatelliteplatforms and launchers. However, the results thatare presented demonstrate the potential of Ka-bandaltimetry both in technical and performances aspects,validating the AltiKa concept for a high performancecontinuation of ENVISAT altimetry and topotentially contribute in high resolution altimetry.Finally, it may be stressed that the Ka-band altimeterpayload defined above may be embarked as apassenger of opportunity missions as a complementto existing altimeters such as the POSEIDON ones :

we would then get multi-frequency measurementsthat may be of much use for all applicationsmentioned above.5. ReferencesJacobs et al. (1999), Navy altimeter datarequirements, Technical Report, Naval ResearchLaboratory, Stennis Space Center, MS 39529-5004,USA.Le Traon, P.Y. and Dibarboure G., 1999, Mesoscalemapping capabilities from multiple altimetermissions. J. Atm. Ocean. Tech., 16, 1208-1223Mitchum G. et al., 1999, The future of sea-surfaceheight observations, OCEANOBS99, Saint-Raphael,Oct. 1999.Phalippou L., E. Caubet and E. Thouvenot, 2000, AKa-band altimeter for future altimetry missions,IGARSS’2000,Rémy F., B. Legresy and P.Vincent, 1999, ‘Newscientific opportunities from Ka-band altimetry’,IGARSS’99, Hamburg, Germany.Tournadre J.,1999, ‘Estimation of rain fall from Kabandaltimeter data’, IGARSS’99, Hamburg,GermanyVerron J., P. Bahurel and P. Vincent, 2001,« AltiKa: Etude de la circulation océaniquemésoéchelle par altimétrie en bande Ka surmicrosatellite », Research proposal to CNES.