Design, calibration, and performance of ... - Wiley Online Library

JOURNAL OF GEOPHYSICAL

RESEARCH, VOL. 106, NO. D13, PAGES 14,573-14,582, JULY 16, 2001

Design, calibration, and performance of MICROTOPS handheld ozone monitor and Sun photometer

II

Marian Morys1 Solar Light Co., Inc., Philadelphia,Pennsylvania

Forrest M. Mims III and Scott Hagerup Geronimo Creek Observatory,Seguin,Texas

StanleyE. Anderson,Aaron Baker, JesseKia, and Travis Walkup Westmont College, Santa Barbara, California

Abstract. MICROTOPS II is a five-channel,handheld Sun photometer that can be configuredto measuretotal ozone,total water vapor, or aerosoloptical thicknessat variouswavelengths.The instrumentmeasures10 x 20 x 4.3 cm and weighs600 g. A principal designgoal was the measurementof total ozone to within 1% of ozone measurementsmade by much larger, heavier, and more expensiveDobson and Brewer spectrophotometers. This goal has been met for a maximum air massof up to -2.5, as demonstratedby comparisonsof MICROTOPS II and its immediatepredecessor, Supertops,with Dobson and Brewer instrumentsat various locations.Conventional interferencefilters are subjectto gradualand unpredictabledegradation.MICROTOPS II avoidsthese problemsby usinghighly stableultraviolet filters manufacturedwith an ion depositionprocess.The 2.4 nm (FWHM) band passof the UV filterswas selectedto balancenoise and ozone measurementperformance.The optical collimatorsand electronicsof the instrumentwere carefullydesignedto optimize pointing accuracy,stray light rejection,thermal and long-termstability,signal-to-noiseratio, and data analysis.An internal microcomputerautomaticallycalculatesthe total ozone column based on measurementsat three UV wavelengths,the site'sgeographiccoordinates,and universal time, altitude, and pressure.The coordinatescan be entered manuallyor by a Global PositioningSystem(GPS) receiver.A built-in pressuretransducerautomaticallymeasures pressure.MICROTOPS II savesin nonvolatilememory up to 800 scansof the raw and calculateddata. Measurementscan be read from a liquid crystaldisplayor transferredto an external computer. 1.

Introduction

Aerosols and ozone within the atmospheremodulate the intensityof ultraviolet radiation at the surfaceof the Earth. Sinceozoneabsorbsshorterwavelengthsmore effectivelythan longerwavelengths,the ratio of the intensityof direct sunlight at two wavelengthswithin the range of 300-320 nm is related to the total abundanceof ozone in a column through the atmosphere.This forms the basic operating principle for a variety of instrumentsthat measurethe ozone layer. The bestknown ground-basedozone-monitoringinstruments are the Dobson and Brewer spectrophotometers.Both these instrumentsdividesunlightinto its constituentwavelengthsby means of a spectrometer.The dispersingelementis a quartz prism in the Dobson and a diffraction grating in the Brewer. While the Dobson and the Brewer are universallyaccepted instrumentsfor measuring column ozone, these instruments are expensive,heavy,and large.There haslong been a need for

a compactinstrumentto measurecolumn ozonewith an accuracy that approachesthat of the Dobson and Brewer. This paper describesan advancedfilter ozonometer,MICROTOPS II, which providesthis capability.

2.

Paper number 2001JD900103. 0148-0227/01/2001JD900103509.00

Ozonometers

Many attempts have been made to use optical filters in ozone measuringinstruments.Optical filters of the glassabsorption type are used in the Russian M-83 spectrometer [Gushchin,1963; Khrgian, 1973]. This instrument has been comparedwith a Dobson spectrometerby Bojkov[1969],who found

differences

between

the measured

ozone values of the

two instrumentsof up to 30%. Osherovichet al. [1969] found even greater discrepanciesin his analysis.The discrepancies have been attributed to the fairly wide filter bandwidth of --•25-30 nm, which significantlyaffectsthe applicabilityof the Beer-Lambert law used to compute the ozone amount. Correctionsbasedon the accurateozone spectrumcannot easily be applied becauseof atmosphericaerosolscatteringeffects.

•Nowat Numar,a division of Halliburton, Malvern,Pennsylvania. Other Copyright2001 by the American GeophysicalUnion.

Filter

drawbacks of the M-83

include its wide field of view and

analogmeter readout.While the M-83 hasbeen supplantedby the newerand better-designedM-124, the filter band pass,field of view, and analog output limitations remain. 14,573

14,574

MORYS

ET AL.: MICROTOPS

II OZONE

Ozonometersbasedon narrow-band-pass interferencefilters were first developednearly 30 years ago, the first instrument havingbeen describedby Boleshakova et al. [1961].Osherovich et al. [1969] and Steblova[1975] subsequently reportedon its use.Matthews[1971],Matthewset al. [1974],andBasher[1975] developedan ozonometer for New Zealand's national atmosphericmonitoringlaboratorythat usednarrow-band-pass interference filters. The peak wavelengthsof the filters were closelymatchedto the wavelengthsusedby the Dobsonspec-

MONITOR

AND SUN PHOTOMETER

1020 nm channelspermits the measurementof total column water vapor. S. Hagerup designedthe electroniccircuitry in Microtops, and Advanced Concept Electronicsfabricated 32 instruments.

Several Microtops instrumentswere calibrated using the Langleymethodat Mauna Loa Observatory(MLO), Hawaii, in June 1995. Resultsof the Langley calibrationsand ozone amountsmeasuredby the MLO Dobson(instrument76) were incorporatedinto the algorithm discussedin detail below to trometer. provide calibrationalgorithmsfor each instrument.Labow et al. [1996] and Flynn et al. [1996] of NASA's Goddard Space Flight Center compareda Microtopspreviouslycalibratedat 3. Limitations of Filter Ozonometers MLO with a Brewer spectrophotometer.This independent The key limitations of ozonometersthat use narrow-band- comparisonestablishedthat ozone retrievalsfrom the Micropassinterferencefiltershavebeen reviewedin detailbyBasher topswere typicallywithin 2% of thoseby the Brewer. and Matthews[1977] and Basher[1977]. The most important Supertops,a Microtopswith narrow-bandfilters(FWHM difficultieswith early interferencefilter instrumentswere rel2.2 nm) at 297, 300, 303, 310, and 312 nm, wasthe first instruatively wide bandwidth, temperature dependencies,out-of- ment to detect record low ozone over the southern United bandradiationleakage,and agingof the filters,whichproduces Statesduringthe fall of 1994 [Mimset al., 1995b].Supertops a gradual increase in attenuation and a shift of the center was comparedwith a Brewer spectrophotometer (instrument wavelength. 112) placed at Geronimo Creek Observatoryin South Texas Filter degradationcan be causedby excessiveexposureto (29.6øN,97.9øW)by the Universityof Georgiaand the U.S. directsunlight,whichis knownas solarization,and possiblyby EnvironmentalProtectionAgencyduringthe summerof 1995. changesin dimensionalitydue to reactionwith absorbedmoisAs the resultsof this comparisonled to the decisionby the ture. Solarization can be reduced by placing one or more Solar Light Companyto manufactureMICROTOPS II, it is blockingfilters before the interferencefilter. There is alsothe appropriateto briefly summarizethem here. problemof manufacturingthe filtersto closelyrepeatabletolSimultaneous measurements of total ozone were made with erances,which is related to the difficultyof accuratelydeposBrewer 112 and Supertopsat or near solar noon on 52 days. iting the manythin layersthat comprisean interferencefilter. The mean ozone measuredduring this comparisonby the Many of theseearly problemsin fabricatinginterferencefilters Brewer and Supertopswas 291.3 (standarddeviationof 6.5) have been overcomeby advancesin technology,and filters Dobsonunits(DU) and293.5(standarddeviationof 6.6) DU, havinga bandwidthof 1 nm or lessare now available.Filters respectively.Thus the mean difference in total ozone meahavingmetal oxidelayersare muchmore resistantto agingand suredby the two instrumentswas2.2 DU, or 0.9%. While there solarization.While filterscannotcurrentlybe madewith absowas excellentagreementon mostdays,scatterin the measurelutely identicalperformancespecifications, they canbe manufactured with a tolerance

4.

of 1-2%.

Total Ozone Portable Spectrometer (TOPS)

F. M. Mims applied the work of R. E. Basher and W. A. Matthewsand high-qualityinterferencefiltersto the development of a miniature, handheld filter ozonometer known as

Total Ozone Portable Spectrometer(TOPS) [Mims, 1992a]. TOPS detects direct sunlight at 300 and 305 nm with filters havinga full-width,half-maximum(FWHM) band passof 5 nm. Two TOPS instrumentswere calibratedagainstthe Total Ozone MappingSpectrometer(TOMS) aboardNASA's Nimbus 7 satellite in the fall of 1990. This empiricalcalibration method providesa simple yet sensitivemethod for tracking slight differences in ozone measured from satellite and ground-basedinstruments [Mims, 1993]. The original two TOPS instrumentshave provided an 11 year, ongoingtime seriesof total ozoneoversouthTexasat localsolarnoon.They havealsomeasuredfluctuationsin the ozonelayer duringthe total solar eclipseof 1991 [Mims and Mims, 1993] and the reduced ozone following the eruption of Mount Pinatubo in 1991 [Mimset al., 1995a]. The success of TOPS resultedin a 1993RolexAward [Reed, 1993]that fundedthe developmentof Microtops,a microprocessor-controlledversion of TOPS with five channels:297, 303, 310, 940, and 1020 nm. The three UV channels measure total column ozone and direct UVB.

The ratio of the 940 nm and

mentslimitedthecorrelation (r2 = 0.47).

On August 5, 1995, 18 simultaneousobservationsof total ozone by Brewer 112 and Supertopswere conductedfrom morning(0815localstandardtime (LST); air massm = 2.04) to afternoon(1642 LST; m = 1.82). Cumuluscloudswere present,the columnwatervaporwasmoderatelyhigh(3.9 cm), there was typical summer haze (aerosol optical thickness (AOT) at 500 nm equals0.35), and the ratio of diffuseto full-skyUVB at 308nm washigherthanusual(0.64).The total ozonemeasuredat localsolarnoonby the TIROS Operational Vertical Sounder (TOVS) was 282 DU. (A working Total OzoneMappingSpectrometer (TOMS) wasnot in orbit.)The mean ozone measuredby the Brewer and Supertopsduring this8.5 hourcomparison was289.8(standarddeviationof 3.1) DU and 289.3 (standarddeviationof 2.7) DU, respectively. Thusthe mean differencein total ozonemeasuredby the two instrumentswas0.5 DU, or only 0.16%. This very closeagreement is especiallysignificantin view of the independentcalibrationof the two instruments prior to the comparison. An importantoperationaladvantageof Supertopsquickly became apparent during the comparison.The Brewer measuresozoneat preprogrammedintervalsby scanninga rangeof UV wavelengths.Thus cloudsor cloud fragmentsthat pass beforethe Sunduringa scancanadversely influencethe ozone measurement.SinceSupertopsmeasuresozone in the time it takesto pressa button,it can make observations duringeven brief openingsbetweenpassingclouds.

MORYS ET AL.' MICROTOPS

channel I

II OZONE MONITOR

AND SUN PHOTOMETER

14,575

•

•lc0h20an lifte-••:• Amp OPTICAL BLOCK

SIGNAL-PROCESSING BLOCK

Figure 1. MICROTOPS II operationalblock diagram.

other data can be viewed on a digital readout or sent to an external computer. The success of Microtopsand Supertopsled to the decision MICROTOPS II stores in nonvolatile memory up to 800 to use similar filter wavelengthsin MICROTOPS II, the sigscans of the raw and calculated data. Each scan includes the nificantlyadvancedversionof Microtopsdesignedby M. Morys while he was at the Solar Light Company,which is the subject time and date, the temperature inside the instrument, the of the remainder of this paper (more information about barometricpressure,and the geographiccoordinates.While the ozoneandcolumnwatervaporversion MICROTOPS II is availablefrom the Solar Light Companyat thispaperdescribes of the instrument, an alternativeversionis configuredas a Sun http://www.solar.com/mtops.htm). Original versions of MICROTOPS II have five channels at 300, 305, 312, 940, and photometer that measuresup to five wavelengthsof sunlight 1020nm. Becausethe signalat 300 nm is quiteweak, especially and automatically computes the aerosol optical thickness during winter and at high latitudes,the 300 nm channel has (AOT). been deleted. The UV wavelengthsof current versions of The overalldesignof MICROTOPS II is shownin Figure 1. MICROTOPS II are 305.5, 312.5, and 320.0 nm. The optical block determinesthe field of view of the instruThe filters used for the three UV channels have a FWHM ment, filtersthe incomingradiation,and detectsand facilitates bandpassof 2.4 nm, and the precisionof the peak wavelength targetingat the Sun. The electricalsignalsfrom the five phois +0.3 nm. The filters are fabricatedusingan ion deposition todetectorsare amplified, converted to a digital form, and processthat enhancestheir stabilityover time. The near-IR numericallyprocessedin the signal-processing block. filters have a FWHM band passof 10 nm and a precisionof

5.

MICROTOPS II Design

_+1.5 nm.

5.1.

MICROTOPS II measures10 x 20 x 4.3 cm and weighs 600 g. Four AA 1.5 V cellsprovidepower. A quartz window providesaccessto the collimatortubesfor the fivephotodiodes and a Sun alignmenttarget.A spring-loadeddoor protectsthe windowwhen the instrumentis not in use. In operationthe windowdoor is left closedwhile poweris appliedto the instrument.The instrumentadjuststhe offsetof eachchannelto zero while the photodiodesview the inner black surface of the window door. Unless they have been previouslyentered, the date, universaltime, and geographiccoordinatesmustnext be entered manually into the instrument'skeypad or automaticallyby a Global PositioningSystem(GPS) receiver. The operator then opensthe window door and points the instrumenttowardthe Sun until a bright point of light is centered over a crosshair in a Sun target window. A scanbutton is then pressedto initiate a programmablenumber of rapid

The success of the instrumentdependson its ability to measure the ozone columnwith long-term stabilityunder a broad range of air massesand atmosphericconditions.The performancefor the entire instrumentand eachof its subsystems was analyzedin a seriesof computersimulations.The initial design goal of an overallprecisionof better than 3%, for an air mass of -5 years) model a relationshipbetween r a• and ra2 is found for a stanstabilityof its ion-depositedUV filters, and the resultsof an dard atmosphere,and becauseof a closeproximityof the two

MORYS ET AL.: MICROTOPS bands

it is assumed

constant

for other

conditions.

II OZONE For

the

filters usedin MICROTOPS II the relationshipis

(15)

The verticalwater vapor columnis calculatedas

9.

km b l.lS%2m) (ln(l/o•) - ln (l/•)•/b (16)

Column Water Vapor Results

are far more

stable than instruments

that use con-

ventionalinterferencefilters [Mirns,1999].The measureddrift of the extraterrestrial

constant

derived

from such instruments

14,581

Conclusion

II is a low-cost, handheld instrument that

provides quick and accurate measurementsof total ozone, water vapor, and aerosol optical thickness.The instrument, which can be operated manually or under computer control, storesin nonvolatile memory up to 800 measurementscans. The filter degradationproblemsthat have adverselyaffected previousfilter ozonometershavebeen largelyovercomeby the useof ion-depositedinterferencefilters.Testsindicatethat the instrument gives reproducible results over a wide range of altitudes

Column water vapor is measuredeach time MICROTOPS II measures column ozone. As the emphasis with MICROTOPS II and its predecessors has been the measurement of total ozone,the columnwater vapor measurementsby MICROTOPS II haveyet to be fully analyzed.Here we report preliminary resultsfrom daily observationsat solar noon at Geronimo Creek Observatory. Column water vapor has been regularly measured at Geronimo Creek ObservatorysinceFebruary 4, 1990, using a novel near-infrared hygrometerdevelopedby Mirns [1992b]. This instrumentusesa pair of near-infraredlight-emittingdiodes(LEDs) as spectrallyselectivephotodiodes.Annual Langley calibrationsconductedat Mauna Loa Observatorysince 1992 have demonstratedthat this and similar LED Sun photometers

10.

AND SUN PHOTOMETER

MICROTOPS

1.16%2.

u-

MONITOR

and under

various

weather

and climatic

conditions.

Comparisonsby F. M. Mims of well-calibratedMICROTOPS II instruments

with Dobson

and Brewer

instruments

at Mauna

Loa Observatoryand elsewhereyield agreementon the order of