GAW Report No. 190. Instruments to Measure Solar Ultraviolet Radiation Part 3: Multi-channel filter instruments

GAW Report No. 190 Instruments to Measure Solar Ultravolet Radaton Part 3: Mult-channel flter nstruments For more nformaton, please contact: World Meteorologcal Organzaton Research Department Atmospherc Research and Envronment Branch 7 bs, avenue de la Pax P.O. Box 2300 CH 1211 Geneva 2 Swtzerland Tel.: +41 (0) 22 730 81 11 Fax: +41 (0) 22 730 81 81 E-mal: AREP-MAIL@wmo.nt Webste: http://www.wmo.nt/pages/prog/arep/ndex_en.html WMO/TD - No. 1537

World Meteorologcal Organzaton, 2010 The rght of publcaton n prnt, electronc and any other form and n any language s reserved by WMO. Short extracts from WMO publcatons may be reproduced wthout authorzaton, provded that the complete source s clearly ndcated. Edtoral correspondence and requests to publsh, reproduce or translate these publcaton n part or n whole should be addressed to: Charperson, Publcatons Board World Meteorologcal Organzaton (WMO) 7 bs, avenue de la Pax Tel.: +41 (0) 22 730 84 03 P.O. Box 2300 Fax: +41 (0) 22 730 80 40 CH-1211 Geneva 2, Swtzerland E-mal: Publcatons@wmo.nt NOTE The desgnatons employed n WMO publcatons and the presentaton of materal n ths publcaton do not mply the expresson of any opnon whatsoever on the part of the Secretarat of WMO concernng the legal status of any country, terrtory, cty or area, or of ts authortes, or concernng the delmtaton of ts fronters or boundares. Opnons expressed n WMO publcatons are those of the authors and do not necessarly reflect those of WMO. The menton of specfc companes or products does not mply that they are endorsed or recommended by WMO n preference to others of a smlar nature whch are not mentoned or advertsed. Ths document (or report) s not an offcal publcaton of WMO and has not been subjected to ts standard edtoral procedures. The vews expressed heren do not necessarly have the endorsement of the Organzaton.

WORLD METEOROLOGICAL ORGANIZATION GLOBAL ATMOSPHERE WATCH No. 190 Instruments to Measure Solar Ultravolet Radaton Part 3: Mult-channel flter nstruments Lead Author: G. Seckmeyer Co-authors: A. Bas, G. Bernhard, M. Blumthaler, B. Johnsen, K. Lantz and R. McKenze Contrbutors: S. Daz, P. Dsterhoft, L. Jalkanen, A. Kazantzds, P. Kedron, B. Petkov, C. Snclar and C. Wlson WMO/TD-No. 1537 November 2010

TABLE OF CONTENTS 1. INTRODUCTION...1 2. OBJECTIVES...2 3. SPECIFICATIONS...3 4. CALIBRATION...6 4.1 Calbraton procedures based on Approach 1...7 4.1.1 Comparson wth a spectroradometer (spectral response functons NOT requred)...7 4.1.2 Comparson wth a spectroradometer (spectral response functons requred)...8 4.1.3 Transfer from standard of spectral rradance...9 4.1.4 Emprcal calbraton approaches...11 4.1.5 Comparson wth a reference nstrument...11 4.2 Calbraton procedures based on Approach 2...12 5. CHARACTERIZATION OF MULTI-CHANNEL FILTER INSTRUMENTS...12 5.1 Characterzaton of spectral response functons...12 5.2 Angular response...14 5.3 Stablty tests...14 5.3.1 Comparson wth spectroradometers...14 5.3.2 Comparson wth a reference mult-flter radometer...15 5.3.3 Calbratons wth standards of spectral rradance...15 5.3.4 Repeated spectral response measurements...15 5.3.5 Langley Method...15 5.4 Vsble and nfrared leakage test... 16 5.5 Cosne error correcton...16 6. APPLICATIONS...16 6.1 Bologcally effectve rradance...16 6.1.1 Calculaton of bologcally effectve rradance va regresson analyss...17 6.1.2 Method suggested by Dahlback [1996]...17 6.1.3 Calculaton of bologcally effectve rradance from reconstructed spectra...18 6.2. Calculaton of hgh-resoluton spectra...18 6.2.1 Method suggested by Dahlback [1996] and Booth [1997]...18 6.2.2 Method suggested by Mn and Harrson [1998]...18 6.2.3 Method suggested by Fuenzalda [1998]...19 6.2.4 Method suggested by Thorseth and Kjeldstad [1999], and Thorseth et al., [2000]...19 6.2.5 Spectral reconstructon wth neural networks algorthm...19 6.3 Calculaton of total column ozone...19 6.3.1 Method suggested by Dahlback [1996]...19 6.3.2 Method suggested by Slusser [1999]...20 6.4 Calculatons of cloud optcal depth [Dahlback, 1996]...20 6.5 Qualty control of spectroradometers...20 6.6 Langley Method...20 Glossary...22 References... 25 Annex A - Centre Wavelengths of some Avalable Mult-Flter Instruments...29 Annex B - References of Freely Avalable Radatve Transfer Programmes...30 Annex C - Calculatons n Support of Specfcatons Provded n Secton 3...31 Annex D - Maxmum Irradance at the Earth s Surface...37

1. INTRODUCTION Ultravolet radaton from the sun causes a consderable global dsease burden ncludng acute and chronc health effects on the skn, eye and mmune system. Worldwde up to 60,000 deaths a year are estmated to be caused by ultravolet radaton, most of whch are due to malgnant melanoma (Lucas et.al., 2008). Much of the UV-related llness and death can be avoded through a seres of smple preventon measures. On the other hand, some UV s essental for the producton of vtamn D n people. Emergng evdence suggests an assocaton between vtamn D levels as an ndcator of health rsk [WHO, 2008] relatng to some cancers, cardovascular dsease and multple scleross among others, along wth the establshed lnk wth musculo-skeletal health. Ths gude s part three of a seres of documents dedcated to nstruments for the measurement of solar ultravolet radaton. These documents have been drawn up by the WMO Scentfc Advsory Group on UV Montorng and ts UV Instrumentaton Subgroup. The am of the seres s to defne nstrument specfcatons and gudelnes for nstrument characterzaton that are needed for relable UV measurements. Part 1 of ths seres [Seckmeyer et al., 2001] descrbes scannng spectroradometers that are able to separate the radaton n small wavelength bands wth a typcal resoluton of 1 nm or less. Broadband nstruments to measure erythemally-weghted ( sunburnng ) UV radaton are descrbed n Part 2 [Seckmeyer et al., 2005]. The mult-channel flter radometers (MCFRs) descrbed here make measurements n several dscrete wavelength bands wth bandwdths of typcally 1 to 10 nm fwhm (full wdth at half maxmum). These nstruments can be used to reconstruct spectra of solar global rradance, to derve specfc products such as erythemally weghted rradance, or to determne total column ozone. Compared to spectroradometers and broadband nstruments, nterpretaton of data of these nstruments s more complex and the separaton of nstrument characterstcs and data products s not straghtforward. There s a dverse range of nstruments that fall wthn ths category. Ther specfcatons must therefore be more flexble, whle ther detaled characterzaton becomes more mportant. At a mnmum, the nstruments must be capable of measurng global rradances n at least two channels. The fore-optc generally conssts of a dffuser, the angular response of whch should deally be proportonal to the cosne of the zenth angle. The wavelength selecton s typcally acheved by narrow to moderate band nterference flters, and the sgnal s detected wth a photodode or a phototube (wthout dynodes for multplcaton). Data acquston and loggng are automated, and software s sometmes provded by the manufacturer to produce standardzed products. Typcally, the number of channels s larger than two, and some examples of these nstruments nclude shadowbands whch enable the near-smultaneous measurement of dffuse and drect rradances, n addton to global rradance. Applcatons of such nstruments are qute varable. For comparson wth other nstruments, the data processng usually requres some sort of normalzaton, convoluton, or deconvoluton. Examples of standard data products are dscussed n Secton 6. The ntended audence for ths document ncludes scentsts, nstrument manufacturers, governmental organzatons, and fundng agences dealng wth UV montorng and research. The document should serve as a gude and s based on the current experence and scentfc knowledge about the measurement of UV radaton wth mult-channel flter radometers. Advantages of mult-channel flter nstruments These nstruments allow the determnaton of one or more of the followng, normally by emprcal or modelled relatonshps: Bologcally effectve doses for a varety of acton spectra wthout the requrement of supplementary ozone data. Total column ozone amount. Cloud attenuaton, especally at hgh temporal resoluton. Reconstructon of solar spectra at arbtrary wavelength resoluton. 1

Compared to spectroradometers (Part 1 of ths seres), they typcally have a hgher temporal resoluton, a lower prce, and are much smpler to operate. Compared to sngle-channel broadband nstruments or broadband nstruments measurng erythemally weghted solar radaton (Part 2 of ths seres) they allow separaton of the nfluence of dfferent atmospherc parameters affectng UV rradance (e.g., total column ozone, cloud attenuaton, aerosol effects, etc.). Furthermore, they are not restrcted to only one acton spectrum. If equpped wth a solar tracker or a shadowband they have the capablty to determne drect solar spectral rradance, both at the centre wavelength of the ndvdual channels or at all wavelengths n the UV by means of spectral reconstructon. In ths case, the Langley method may also be used as a calbraton tool, to determne aerosol optcal depths, or the extraterrestral solar rradance [Slusser et al., 2000]. Dsadvantages of mult-channel flter nstruments Compared to spectroradometers, whch delver spectra, the sgnal output of mult-flter nstruments usually requres post processng and the development of algorthms to gan meanngful results or measurement quanttes. Although the stablty of these nstruments may match or exceed the stablty of spectroradometers, the absolute calbraton s usually acheved by calbraton aganst the latter. When calbrated aganst spectroradometers, the uncertanty of mult-channel nstruments s usually hgher than that of spectroradometers due to the addtonal transfer uncertanty. Mult-flter nstruments can also be calbrated wth Standards of Spectral Irradance rather than by comparson wth a spectroradometer under the Sun. However, ths calbraton method requres laboratory characterzaton work, ncludng the accurate characterzaton of the flter s response functons and angular response functons. Ths s partcularly the case for channels n the UV-B, due to the great dfference of lamp and the sun spectra. To obtan a small calbraton uncertanty, accurate determnaton of the spectral responsvty s requred. Ths n turn requres a tunable, small-bandwdth monochromatc lght source of suffcent radatve power. Flters that are used for wavelength separaton may be subject to drfts, both absolute and spectral, whch are dffcult to detect durng operaton. 2. OBJECTIVES Mult-channel flter nstruments can be employed for a varety of scentfc applcatons. In a strct sense these nstruments are capable of measurng only the global rradance n the UV at several wavelength channels weghted wth the response functons of the respectve channel. However, the major advantage and present applcaton for these nstruments are the derved data products. In ths respect these nstruments dffer from the other nstruments descrbed n the seres (Part 1: Spectroradometers; Part 2: Broadband Instruments Measurng Erythemal Irradance; and Part 4: Array Spectroradometers). The objectves for employng these nstruments may be summarzed as follows: 1. To derve data products such as erythemal rradance wth hgh temporal resoluton. 2. To provde nformaton on the varablty of solar UV rradance partcularly due to clouds. 3. To contrbute to determnng geographc dfferences n UV and understandng ther causes. 4. To derve spectral global rradance n the UV at the nstruments nomnal wavelengths. These data can be used to calculate spectral rradance at other wavelengths. 5. To help n qualty control (QC) of spectroradometrc UV measurements. 6. To support ground truthng of satellte estmates of UV. 7. To measure global response-weghted-rradance n the UV, whch s the solar spectral rradance weghted by the spectral response functon of each channel. As wth spectroradometers, multflter-nstruments may also be used to derve total column ozone and, n combnaton wth a solar tracker or a shadowband, aerosol optcal depth at varous wavelengths. 2

3. SPECIFICATIONS Quantty Qualty Cosne error (a) < ±5% for ncdence angles <60 (b) < ±5% to ntegrated sotropc radance (c) < 3% azmuthal error at 60 ncdence angle Mnmum spectral range 305-360 nm Wavelength stablty < 0.03 nm for centre wavelength Wavelength accuracy Not applcable (see remark) Bandwdth (fwhm) < 10 nm Bandwdth stablty < 0.04 * fwhm Stray lght ncludng senstvty to < 1 % contrbuton to the sgnal of wavelengths outsde 2.5 fwhm vsble and IR radaton for SZA less than 70 Stablty n tme on tme scales up to a year Sgnal change Currently n use: better than 5% Desred: 2% Mnmum number of channels At least one channel wth centre wavelength < 310 nm and at least one wth centre wavelength > 330 nm Maxmum rradance Sgnal of the Instruments must not saturate at radaton levels encountered on the Earth s surface. Detecton threshold SNR = 3 for rradance at SZA=80 and total ozone column of 300 DU. Instrument temperature Montored and suffcently stable to mantan overall nstrument stablty Response tme < 1 s Multplexng tme < 1 s Accuracy of tme Better than ±10 s Samplng frequency < 1 mnute Levellng < 0.2 Calbraton uncertanty < 10 % (unless lmted by detecton threshold) Remarks on specfcatons: Cosne error Smaller cosne errors would be desrable. Defntons of cosne and azmuthal error are gven n the Glossary. Mnmum spectral range The mnmum spectral range should be large enough to allow calculaton of bologcally effectve rradance (e.g., erythemal rradance). Instrument channels should deally cover the complete UV range as most bologcal systems respond to wavelengths n both the UV-B and UV-A regons. Wavelength stablty In prncple, wavelength stablty has to be wthn the gven range for all observng condtons. Ths specfcaton s hard to verfy n the feld, but can be approxmately verfed n laboratory experments by characterzng the spectral response functon as descrbed n Secton 5.1.. The specfcaton of 0.03 nm was chosen, because calculatons show that a shft of ths magntude leads to a change of up to 2% of the sgnal for a channel of 10 nm bandwdth centred at 305 nm for solar zenth angles between 0-80 degrees and ozone amount less than 500 DU. Calculatons for smaller bandwdths do not change the concluson apprecably. For more detals see Annex C.1. Wavelength accuracy Ths specfcaton s not applcable for multflter nstruments because for meetng the objectves t s suffcent that the nstrument s flter functons are characterzed accurately. For further detals see gudelnes for nstrument characterzaton n Secton 5. 3

Bandwdth The bandwdth of flters used n currently avalable nstruments ranges between 1 and 10 nm. Instruments wth small bandwdth typcally requre smaller correctons to convert ther raw data to spectral rradance (Secton 4.1). For example, calculatons presented n Annex C.3 show that transfer of calbratons from a lamp standard wll result n an error of 200% n solar measurements at 305 nm for a bandwdth of 10 nm fwhm, SZA less than 80º, and total ozone between 250 and 450 DU, f no correctons are appled. The correspondng number for a bandwdth of 1.0 nm fwhm s 0.5 %. On the other hand a small bandwdth reduces the nstrument s senstvty, whch makes t dffcult to detect low-ntensty radaton n the UV-B. The bandwdth s therefore a compromse between the competng requrements of beng able to calbrate nstruments accurately and to detect solar rradance n the UV-B. Smulatons provded n Annexes C.1 and C.2 show that accurate characterzatons of the nstrument channels spectral response functons (ncludng the accurate determnaton of the channel s centre wavelength) are more mportant than the bandwdth or the shape of response functons. These calculatons ndcate that data products such as total ozone and bologcally effectve rradance can be calculated wth smlar accuracy from raw data of nstruments wth 1-nm or 10-nm wde channels. Bandwdth stablty Calculatons presented n Annex C.2 show that small changes n the bandwdth can have a sgnfcant effect on measurements of mult-flter nstruments n the ozone cut-off regon (wavelengths below 315 nm). For an nstrument wth a bandwdth of nomnally 10 nm fwhm, a 0.2 nm broadenng n bandwdth wll result n up to 3% ncrease n the sgnal at 305 nm for SZA between 0 and 80, and ozone amount between 250 and 450 DU. For narrower flters proportonal changes n bandwdth are less of a concern. Stray lght, ncludng senstvty to vsble and IR radaton Multflter nstruments use nterference flters to realze ther spectral response functons. These flters may have secondary peaks n the vsble and nfrared, whch may be outsde the spectral range of the apparatus for measurng response functons. These secondary peaks can sgnfcantly contrbute to the nstrument s sgnal. Ths senstvty should be checked wth cut-off flters, (e.g., wth WG or GG Schott longpass flters) usng both Sun and calbraton lamp as lght source. A descrpton of ths technque can be found n Secton 5.4. Mnmum number and wavelength of channels By defnton the nstrument must have at least two channels. Normally these nclude one n the UV-B that s senstve to total column ozone, and one n the UV-A. Dependng on the applcaton, e.g., dervng of bologcally effectve rradance or aerosol parameters, addtonal channels are usually necessary. Centre wavelengths of exstng nstruments or other wavelengths relevant for specfc applcatons can be found n Annex A. Maxmum rradance A complaton of the maxmum UV rradance to be expected at the Earth s surface s provded n Annex D. Detecton threshold A low detecton threshold s partcularly necessary at locatons where rradance s low, e.g. at hgh-lattude stes or n wnter. Instrument temperature Instrument temperature should be montored and stablzed. Operatng condtons logged should nclude the nternal nstrument temperature (specfed by the manufacturer) and the effect of heatng by solar radaton, whch may warm the nstrument by a consderable margn above the ambent temperature. Temperature stablzaton s requred for accurate measurements snce both nterference flters and photododes are temperature senstve. If the correlaton between temperature and senstvty s well establshed the nstrument may be used wthout stablzaton by applyng a temperature correcton to the data. Temperature stablzaton s preferable, snce temperature effects on flter-functons are dffcult to correct. 4

Response tme A tme constant of one second s suffcent for most applcatons and s easly achevable wth the exstng nstrumentaton. In some specalzed applcatons, e.g., statstcs for clear sky determnaton or the nvestgaton of transent cloud effects, shorter tme constants may be desrable. Accuracy n tme Tme errors of 10 s can lead to measurable dfferences as SZA and cloud condtons change. Tme-keepng of better than 10 s s requred f an nstrument s to be compared to other nstruments, n partcular durng cloudy stuatons. Uncertantes of less than one second are readly achevable wth current technology (e.g., Internet tme server, GPS). Levellng Levellng to better than ±0.2 can be acheved wth a smple bubble level. Care should be taken that the reference plane used for levellng s parallel to the nstrument s collector. Samplng frequency Less than 1 mnute for most applcatons. However, for specfc cloud studes, a much hgher frequency (e.g., 1 s) may be necessary. Calbraton uncertanty The calbraton uncertanty ncludes all uncertantes assocated wth the rradance calbraton procedures. Qualty Assurance and Qualty Control (QA/QC) To obtan data of hgh qualty t s not suffcent that nstruments meet the basc specfcatons dscussed above. In addton, measurements of ancllary data for nterpretng the measurements should be avalable, nstruments have to be well mantaned, and a Qualty Assurance and Qualty Control (QA/QC) plan has to be followed [Webb et al., 1998; 2003]. Recommended ancllary data and QA/QC procedures are compled below. Recommended ancllary data Total ozone column, from ndependent nstruments or satelltes to establsh correcton factors (Secton 4) or to check for bas n ozone retrevals of mult-channel nstrument data (Secton 6.3). Data from ndependent radometers such as pyranometers, broadband UV sensors or spectroradometers to help to valdate the nstrument s stablty n tme (Secton 5.3). Meteorologcal data. Mantenance and QA/QC 1. Daly: Checkng of nput optcs (rradance collector), and cleanng f necessary. Determnaton of offset (Most nstruments provde an automated offsetdetermnaton durng the dark hours. Offset checks may have to be done manually n polar regons durng perods wth 24 hours of sunlght). 2. Weekly: Checkng of the effectveness of temperature stablzaton, tme-keepng, levellng, and data loggng. 3. At least once per year (every sx months s desrable): Checkng of nstrument s stablty by comparson to a reference nstrument, lamp, or spectroradometer. Checkng of the correct operaton and calbraton of electronc supportng devces (data loggers, A/D boards, sgnal amplfers, cables, computers etc.). 5

Checkng of dark stablty durng the year. Instablty may suggest temperature dependence of the electroncs or other problems. 4. At deployment and f qualty checks above ndcate a problem: Verfcaton of the spectral and angular response. Checkng of the accuracy of the nstrument s level ndcator. (The optcal plane of an nstrument s sometmes not consstent wth the reference plane that s used for checkng whether the nstrument s level). 4. CALIBRATION There are several fundamentally dfferent approaches to calbratng mult-channel flter nstruments, some resultng n dfferent radometrc quanttes. Each approach can also be mplemented n dfferent ways. Several methods and ther advantages and dsadvantages are dscussed below. Some mplementatons requre that the spectral response functons of all channels are accurately known. Other methods are based on a comparson wth a spectroradometer under varyng atmospherc condtons. In ths case, measurements of spectral response functons may not be necessary. Approach 3 s applcable only to nstruments equpped wth shadowbands. Approach 1 Spectral Irradance The objectve of Approach 1 s to establsh a calbraton functon for each channel of the nstrument whch, when appled to the raw sgnal, returns spectral rradance at the nomnal wavelength of the channel. For example, a calbraton functon may be defned such that the calbrated measurement of each channel approxmates spectral rradance, measured by a spectroradometer wth a 1-nm bandpass at the channel s nomnal wavelength. (Ths example assumes that spectroradometer and flter nstruments are exposed to the same radaton feld.) The advantage of Approach 1 s that the measurement spectral rradance s a common radometrc quantty. The dsadvantage s that calbraton functons usually depend on the radaton source measured. Ths s partcularly a problem for measurements n the UV-B part of the solar spectrum due to the rapd change of the Sun s spectrum wth wavelength n ths regon. In ths case, calbraton functons wll depend on solar zenth angle and other atmospherc parameters affectng the shape of the solar spectrum, such as total column ozone. Approach 2 Response-weghted rradance The objectve of Approach 2 s to apply a calbraton factor to each channel of the nstrument such that the calbrated measurement of a gven channel s response-weghted rradance. Ths quantty s defned as the wavelength ntegral of the product of spectral rradance and spectral response functon of the channel (see Glossary). For example, f measurements from a spectroradometer are weghted wth the spectral response functons of a collocated flter nstrument, the resultng response-weghted-rradance wll be dentcal wth the calbrated measurement of the flter nstrument. The advantage of ths calbraton approach s that the calbraton functon smplfes to a factor, whch s ndependent of the radaton source. When measurng solar radaton, ths factor does not depend on solar zenth angle and other atmospherc parameters. If the calbraton factor was establshed wth a standard lamp, t can be appled to solar measurements wthout correctons. The dsadvantage of the approach s that the quantty measured response-weghted rradance s nstrument dependent. Comparng the results of dfferent nstruments s therefore dffcult. However, standardzed data products such as erythemal rradance can be calculated wth good accuracy from calbrated values wthout the need of consderng the atmospherc condtons durng the tme of the measurement (Secton 6.1). 6

Approach 3 Langley Method Ths approach s based on the Langley method and requres that nstruments are equpped wth shadowbands (Secton 6.6). From consecutve measurements of global and dffuse rradance (the latter determned by blockng the Sun wth the shadowband), drect rradance s calculated. Measurements of drect rradance are performed at dfferent armasses and extrapolated to armass zero to derve the nstrument s sgnal that would be expected outsde the Earth s atmosphere. The sgnals of the dfferent channels at armass zero are then compared wth a reference extraterrestral spectrum, whch s weghted wth the response functon of each channel to establsh calbraton factors as n approach 2. These factors are fnally appled to measurements of global rradance. The method has been descrbed by Slusser et al. [2000], and s not dscussed n more detal here. Implementatons of Approach 1 and 2 are descrbed n the followng sectons. 4.1 Calbraton procedures based on Approach 1 4.1.1 Comparson wth a spectroradometer (spectral response functons NOT requred) For ths method, the mult-channel nstrument s operated next to a hgh-resoluton (1) spectroradometer, both of whch are exposed to sunlght. A calbraton factor C s establshed by dvdng the net sgnal measured by channel of the mult-channel radometer, wth spectral solar rradance E S ( λ ), measured by the spectroradometer at nomnal wavelength λ : C (1) = E V S S, ( λ ) ( V = S,, G E S V ( λ ) S,,0 ). Here V S,, G s the lght sgnal (e.g., measured n volts) of channel when exposed to the radaton of the Sun, V S,, 0 s the dark sgnal, obtaned by coverng the collector, and V S, s the net sgnal, calculated as the dfference of V S,, G and V S,, 0. Measurements of the spectroradometer have to be corrected for all systematc errors, such as the (1) cosne error, pror to the comparson. Once C has been derved, solar spectral rradance at the nomnal wavelength λ of the mult-channel radometer s calculated wth: V E S ( λ ) = C S, (1) Due to the msmatch of the spectral response functons and the slt functon of the (1) spectroradometer, C wll usually depend on solar zenth angle (SZA), total ozone (O 3 ), and (1) other atmospherc parameters x. For ths reason, C s a functon dependng on these parameters: C (1) = C (1) ( SZA, O3, x) The argument ( SZA, O 3, x) s omtted n the followng for better readablty. The comparson of the mult-channel nstrument and the spectroradometer has to be performed 7

over a perod suffcent to nclude a large set of envronmental condtons. Snce the calbraton functon can be establshed only for condtons that occurred durng the comparson perod, deployment of the flter nstrument at a locaton wth dfferent condtons may be problematc. (1) In practce, the dependence of C on SZA, O 3, and other factors can be descrbed wth a lookup table or a numerc parameterzaton. For example, Díaz et al. [2005] used a mult-regressve model to parameterze the relatonshp between V S,, E S ( λ ), SZA and O 3 : ln( ES ( λ )) = a1 ln( VS, ) + a2o3 + a3 f (90 SZA) + b, o o where f ( 90 SZA) s a functon of SZA, and a 1, a 2, a 3, and b are coeffcents determned by regresson. Although parameterzatons such as the one suggested by Díaz et al. [2005] may delver suffcently accurate results, great care must be appled when extrapolatng parameterzatons to condtons for whch they were not desgned. For example, a parameterzaton derved from measurements at a hgh-lattude locaton may lead to systematc errors f the nstrument s deployed at a low-lattude ste. 4.1.2 Comparson wth a spectroradometer (spectral response functons requred) As an alternatve to the method descrbed n Secton 4.1.1, a calbraton functon may be establshed from a sngle reference solar spectrum that s measured by a mult-channel nstrument and a spectroradometer for well known SZA and atmospherc condtons. The (2) (2) alternatve calbraton functon s denoted C and ncludes a correcton term K, whch depends on SZA and O 3, and possbly other atmospherc parameters x. (2) VR, (2) C = K = C E ( λ ) R ( R) K (2). Here E R ( λ ) s solar spectral rradance of the reference spectrum, V R, s the net sgnal of (R) channel when measurng ths spectrum, and C s the calbraton factor for the reference spectrum, calculated as rato of V R, and E R ( λ ). Solar spectral rradance E S ( λ ) for arbtrary condtons s then calculated wth: V V E S ( λ ) = = C C S, ( R) (2) K S, (2) Note that ths formula does not account for the cosne error of the nstrument or any other systematc errors. In practce, t s usually necessary to correct for these errors (Secton 5.5), leadng to a further modfcaton of the measurement equaton: VS, E S ( λ ) = X ( 1, 2,..., ) (2) p p pn, C where X ( p1, p2,..., pn) s a correcton term dependng on n parameters such as SZA, O 3, and cloud condton. The correcton functon (2) K s defned by: 8

K (2) = E E S R ( λ' ) R ( λ' ) dλ' E ( λ' ) R ( λ' ) dλ' E R S ( λ ) ( λ ) where R (λ) s the spectral response functon of channel. Systems for measurng the spectral (2) response of flter radometers are descrbed n Secton 5.1. Determnaton of K requres the knowledge of the solar spectrum E S ( λ ), the quantty to be measured by the nstrument. An exact (2) (2) determnaton of K s therefore not possble. However, K can be estmated from model calculaton usng SZA, O 3, and x as nput parameters. For most applcatons knowledge of SZA and O 3 s suffcent. O 3 can be taken from satellte measurements. A complaton of radatve (2) transfer models that can be used for the calculaton of K s provded n Annex B. 4.1.3 Transfer from standard of spectral rradance In ths mplementaton, the radometer s set up n front of a standard lamp. The spectral rradance produced by the lamp at the place of the nstrument s collector s denoted E L ( λ ) ; the (L) assocated net sgnal measured by channel of the radometer s V L,. The calbraton factor C s defned as the rato of V L, and spectral rradance at the nomnal wavelength λ of channel : ( L) VL, C = EL ( λ ) Wth ths defnton, solar spectral rradance at wavelength λ may be approxmated wth: E S S, λ ), ( L) C ( V where V S, s agan the net sgnal of channel when measurng the Sun. Solar spectral rradance calculated wth ths approach wll have a large systematc error n the ozone cut-off regon of the solar spectrum. The error ncreases wth decreasng wavelength and ncreasng bandwdth of the radometer s channel. It s caused by the dfference of lamp and solar spectra at short wavelengths, where the second dervatves of both sources devate consderably (Fgure 1). For a hypothetcal nstrument wth a centre wavelength of 305 nm, the error can be as large as 200% for an nstrument wth a bandwdth of 10-nm for SZA 0 80 and total ozone 250 450 DU. It s less than 1% for an nstrument wth a bandwdth of 1-nm. 9

1000 1 Spectral rradance n mw/(m² nm) 100 10 1 0.1 0.01 0.001 Solar spectral rradance at SZA=30 and total ozone of 300 DU Solar spectral rradance at SZA=60 and total ozone of 300 DU Spectral rradance of 1000W FEL standard lamp n 50 cm dstance Typcal spectral response functons of moderate-bandwdth radometer 280 300 320 340 360 380 400 Wavelength n nm 0.1 0.01 0.001 Spectral response functon Fgure 1 - Comparson of solar spectral rradance at SZA=30 and 60, the spectrum of a 1000-Watt FEL calbraton standard at 50 cm dstance, and typcal response functons of a moderate-bandwdth mult-channel flter radometer wth nomnal wavelengths of 305, 320, 340, and 380 nm and a bandwdth of approxmately 10 nm To correct for the error, a correcton functon a modfed calbraton functon (3) VL, (3) C = K = C E ( λ ) L ( L) K (3) C : (3) The correct expresson for calculatng solar spectral rradance s then: V V E S ( λ ) = = C C S, ( L) (3) K S, (3) (3) K has to be appled for each channel, resultng n Smlar to the procedure descrbed n Secton 4.1.2, t s usually necessary to correct for addtonal systematc errors, such as the cosne error, leadng to the followng modfcaton of the measurement equaton: VS, E S ( λ ) = X ( 1, 2,..., ) (3) p p pn, C where X ( p1, p2,..., pn) s agan a correcton term dependng on n parameters such as SZA, O 3, and cloud condton. (3) The correcton functon K of channel s defned as: K (3) = E E S L ( λ' ) R ( λ' ) dλ' E ( λ' ) R ( λ' ) dλ' E L S ( λ ) ( λ ) 10

(3) Here R ( λ ) s the spectral response of channel. K s a functon of SZA, O 3, and other parameters affectng the transfer or radaton through the atmosphere, and can be estmated from (2) modelled spectra n a smlar manner as K. For most applcatons, knowledge of SZA and O 3 s suffcent. For lamp-based calbratons, R ( λ ) must be known very accurately. Accordng to Annex C.1, the wavelength applcable to a gven spectral responsvty needs to be known to an accuracy of 0.03 nm to gve an error n the solar rradance of less than 2% for a flter wth centre wavelength at 305 nm. The error n the sgnal s essentally ndependent of the fwhm of the radometer for bandwdths from 1 10 nm and the error s smaller wth ncreasng centre wavelength. 4.1.4 Emprcal calbraton approaches The dependence of the calbraton functon on SZA and O 3 can partly be accounted for by ncludng measurements of all channels n the calbraton [Díaz et al., 2005]. For example, solar spectral rradance at wavelength λ may be expressed as a lnear combnaton of net sgnals measured by all channels: E S (λ ) = cj VS, j j The coeffcents c j are determned va mult-lnear regresson of solar spectral rradance E S ( λ ), measured wth a spectroradometer under varyng condtons, aganst net sgnalsv S, j, measured wth the mult-channel radometer at channels j. To further mnmze errors caused by the SZAdependence of the calbraton, Díaz et al. [2005] suggested the followng modfcatons to the regresson equaton: o For SZA < 40 : ES ( λ ) = [ cj VS, j ] + c f (90 SZA) j For SZA > 40 : o ln( ES ( λ )) = [ cj ln( VS, j )] + c f (90 SZA) + d j o where f ( 90 SZA) s a polynomal ft-functon, and c and d are ft-coeffcents. Emprcal approaches should be valdated over a large range of condtons wth dfferng SZA, O 3, and other parameters x, and should not be appled to condtons outsde ths range. 4.1.5 Comparson wth a reference nstrument In ths case, the nstrument to be calbrated should operate alongsde a reference multchannel nstrument ( R ) wth well-establshed calbraton factors c (R), whch are generally dependent on SZA and O 3. If the spectral response functons of the two nstruments are vrtually (T) dentcal for all channels, the calbraton factor c of the nstrument under test ( T ) s: (T) (T) V (R) c = c (R) V By ths comparson, also c (T) wll also become dependent on SZA and O 3. The nstruments should run sde-by-sde for several days coverng a full range of SZA, and deally a wde range of ozone columns. If the spectral response functons of the two nstruments are dfferent, a correcton 11

factor must be appled, whch can be calculated n a smlar way as descrbed n Secton 4.1.2. 4.2 Calbraton procedures based on Approach 2 * In ths approach, the calbraton factorc s the rato of the net sgnal of channel, V, to rradance E (λ), weghted wth the relatve spectral response R (λ) of channel : C * = V E( λ) R ( λ) dλ The radaton source producng E (λ) can ether be a standard lamp or the Sun. In the former case, E (λ) s known from the standard s certfcate; n the latter case, E(λ) s typcally measured * by a spectroradometer deployed next to the flter nstrument to be calbrated. In theory, C does not depend on the lght source beng measured, whch can be the Sun or any artfcal lght source. * * In practce, C s subject to uncertanty f R (λ) s not accurately known. The uncertanty of C as a functon of varous characterstcs of R (λ) are dscussed n Annex C. The radatve quantty beng measured by flter nstruments calbrated wth ths approach s response-weghted rradance E wth V E = C * Ths quantty s dfferent for every nstrument. However, standardzed data products such as erythemal rradance can be calculated from E wth hgh accuracy. Ths s dscussed n Secton 6. 5. CHARACTERIZATION OF MULTI-CHANNEL FILTER INSTRUMENTS Proper characterzatons of angular and spectral response of mult-channel flter nstruments are crucal for obtanng accurate measurements. Informaton from the characterzaton s used to convert raw sgnals of the nstruments to physcal quanttes such as spectral rradance. For approprate qualty control and assurance of flter nstruments, characterzaton of the spectral response and angular response should be undertaken at regular ntervals. Characterzng nstruments requres well-desgned systems. It s suggested that qualfed laboratores carry out spectral response and cosne response characterzatons. In addton, the stablty and calbraton of any nstrument needs to be montored over tme. The followng gves a general descrpton of typcal characterzaton systems and procedures. 5.1 Characterzaton of spectral response functons The spectral senstvty of each channel should be measured wth a dynamc range and wavelength range large enough to detect small flter leakages outsde the man flter bandpass. Generc response functons should not be used because t has been shown that the spectral transmsson of flters may vary sgnfcantly, even for flters of the same batch [Bernhard et.al., 2005]. The centre (nomnal) wavelength of each channel should be calculated from the measurement response functons, e.g. (1) Determne centrod wavelength of the response functon, (2) Multply the response functon wth a reference spectrum and determne ts centrod wavelength. A system for characterzng the spectral response of mult-channel radometers requres a spectral lght source. Ths can be provded ether by a tunable laser or by an optcally dspersng nstrument, such as a monochromator. The followng descrpton wll concentrate on the latter. Typcally, the output from a hgh-ntensty lght source such as a xenon arc lamp s maged onto the 12

entrance slt of a monochromator. The monochromator scans across the desred wavelength range (e.g., 270 420 nm) n wavelength ncrements small enough to resolve the spectral response. The output of the monochromator s maged onto one of two separate detecton systems, the multchannel flter radometer (MCFR) under test and a reference detector wth known spectral response. A measurement of the MCFR and the reference detector output sgnals are taken at each wavelength step. The spectral response R ( λ ) of the MCFR s calculated from measurements of the MCFR and the reference detector: ( V R ( λ ) = DUT,, L ( V R, L ( λ ) V ( λ ) V DUT R,0,,0( λ )) S ( λ )) R ( λ ) Here VDUT,, L( λ ) s the lght sgnal of channel of the MCFR, VDUT,,0 ( λ ) s the correspondng dark sgnal, VR, L( λ ) and VR, 0 ( λ ) are the lght and dark sgnal of the reference detector, respectvely, and S R ( λ ) s the spectral response of the reference detector. Instrument specfc spectral response functons may be avalable from the manufacturer, but can change wth tme. The reference detector SRF should deally be confrmed by a standards laboratory, and must be checked perodcally to assure stablty. Measurng R ( λ ) accurately n absolute terms s dffcult and t s therefore conventonal to normalze R ( λ ) to one at ts maxmum value. The monochromator s bandwdth should deally be more than an order of magntude smaller than the bandwdth of the MCFR. Its wavelength accuracy should be better than 0.03 nm for flters centred at 305 nm to acheve solar rradance errors less than 2% (see Annex C.1), n partcular f the MCFR s calbrated aganst a lamp (Secton 4.1.3) Ths crteron s less strct when measurng spectral response at longer wavelengths. The stray-lght rejecton of the monochromator should be suffcent to ensure spectral purty at each measurement step. Ths typcally requres the use of a double-monochromator. The lght output from the monochromator should be suffcent to gve a dynamc range of at least three orders of magntude n the measured spectral response of the MCFR. The spectral response measurement system should have an optmum balance between acceptable stray-lght rejecton, band-pass sze, wavelength step sze and adequate sgnal throughput to obtan the spectral response curve of the MCFR wth the desred dynamc range. It s often not possble to set the monochromator to a bandwdth that s one order of magntude smaller than that of the MCFR s channels, and stll have suffcent sgnal over the desred dynamc range. It may therefore be necessary to deconvolve the spectral response functon wth the monochromator s slt functon. A sutable technque has been suggested by Bernhard et al. [2005]. As an alternatve, the core part of the response functons may be scanned wth a small bandwdth and the wngs wth a large bandwdth to have suffcent sgnal. Measurements wth the two bandwdth settngs can then be sttched together. Ths technque has been successfully used by Johnsen et al. [2008b]. Spectral response functons should deally be characterzed once per year. Ths s often not possble n practce. At a mnmum, nstruments should be retested f comparsons wth other nstruments ndcate potental changes n the detector s spectral response. The followng publcatons provde descrptons of systems for the characterzaton of spectral response functons: Bernhard et al. [2005]; Bolsée et al. [2000]; d Sarra et al. [2002]; Hülsen and Gröbner [2007]; Johnsen et al. [2008b]; and Lantz et al. [2005]. These papers should be consulted f a user of mult-channel nstruments chooses to perform such characterzatons hmself. As these measurements are rather demandng, t s advsed these characterzatons are performed by establshed laboratores such as the WMO/GAW regonal calbraton centers. 13

5.2 Angular response As wth any other nstrument measurng solar UV rradance the angular response of a mult-channel flter nstrument should be characterzed for zenth angles between -90 and 90, and several azmuth angles n ncrements suffcent to obtan any structure that may be present (e.g., 1 zenth angle ncrement, 45 azmuth angle ncrement). The system for measurng the angular response typcally conssts of a lght source, a computer-controlled rotary table, and algnment fxtures. The MCFR s mounted to the rotary table such that ts axs of rotaton touches the reference plane of the MCFR s dffuser. For performng the angular response measurement, the rotary table s turned from -90 to 90 whle the sgnals of the MCFR detectors are recorded. The measurement s repeated after turnng the nstrument n ts holder to a dfferent azmuth angle. Cosne and azmuthal errors can be calculated from these measurements usng the defntons gven n the Glossary. The lght source may ether be a hgh-ntensty ncandescent lamp, such as a 1000-W FEL lamp, or a dscharge lamp, such as a Xenon lamp. If a convex mrror or a lens s used to collmate the lamp s output, t has to be ensured that the beam s homogeneous and overflls the dffuser of the MCFR. The measurement system should have optcally flat black surroundngs to lmt scattered lght. A large baffle should be nstalled between lamp and MCFR such that scattered off-axs radaton from the lamp cannot reach the MCFR. For accurate measurements t s crtcal that the MCFR s algned correctly. Ths can be acheved by means of an algnment laser mounted behnd the lght source and drected toward the rotary stage. Frst, lamp and MCFR are algned such that the laser goes through the centre of the lamp and the centre of MCFR s dffuser. Second, a mrror s placed n front of the MCFR such that t s parallel to the MCFR s reference surface used for levellng the nstrument n the feld. The algnment of the MCFR s then adjusted such that the laser beam s back-reflected to the laser. Wth ths method the 0 poston of the rotary stage can also be determned accurately. If the measurements ndcate that the nstrument has a sgnfcant azmuthal error, t s mportant that the orentaton of the nstrument can be marked and transferred to the feld such that azmuthal asymmetres n solar measurements can be nterpreted and corrected. The followng publcatons provde descrptons of systems for the angular response characterzatons: Harrson et al. [1994a]; Hülsen and Gröbner [2007]; Johnsen et al. [2008b]. 5.3 Stablty tests Several methods can be used to check the temporal stablty of the calbraton of mult-flter nstruments. These nclude: Comparson wth solar measurements performed by well-mantaned spectroradometers. Comparson wth other mult-flter nstruments. Va the Langley Method of analyss and extrapolaton to extraterrestral solar rradance. A combnaton of the methods lsted above. 5.3.1 Comparson wth spectroradometers For the mplementaton of ths method, the mult-flter radometer to be tested and the spectroradometer measure sde by sde for a perod of deally one week or longer. The measurements of the spectroradometer are weghted wth the response functons of the mult-flter radometer as descrbed n Secton 4.2, and compare wth the net sgnal of the mult-flter radometer. The rato of the two measurements should deally not depend on tme. If the drft s beyond an acceptable lmt, the mult-flter radometer should be recalbrated. The method s routnely appled to measurements performed by the NSF UV Montorng Network [Bernhard et al., 2008]. 14

5.3.2 Comparson wth a reference mult-flter radometer As n the prevous secton, the mult-flter radometer to be tested and the reference radometer measure sde by sde for a suffcently long perod. Measurements of the two nstruments are compared as descrbed n Secton 4.1.5. Both nstruments should deally have very smlar spectral response functons. If ths s not the case, the tme seres analyss should be restrcted to a subset of measurements performed at smlar SZA and total ozone column. 5.3.3 Calbratons wth standards of spectral rradance The mult-flter radometer to be tested s regularly (e.g., annually) placed n front of a Standard of Spectral Irradance and the net-sgnal s measured. Assumng that the lamp s stable, the radometer should measure the same net-sgnal at every event. If dfferent lamps are used, measurements of the radometer should be converted to spectral rradance and compared wth the values provded n the lamp s certfcates. Addtonal correctons smlar to those descrbed n Secton 4.1.3 may be necessary f the colour temperature of the varous lamps s not dentcal. Calbraton standards, ether reference standards or workng standards, should be recalbrated (or replaced) after 20 hours of use, unless otherwse stated n the lamp s certfcate. A recalbraton of reference standards should be performed by standard laboratores (see also Webb et.al., 1998) 5.3.4 Repeated spectral response measurements Repeated spectral response measurements can help to determne reasons for changes n nstrument senstvty uncovered by one of the three methods descrbed above. For example, these measurements may detect changes n the response functons centre wavelengths, bandwdth, peak response, and wngs [Bgelow and Slusser, 2000]. Fgure 2 presents a tme seres of centrod wavelengths and bandwdths determned from measurements of an nstrument that s used by the USDA UV-B Montorng and Research Programme. Fgure 2 - Results of repeated spectral response measurements of a nstrument used by the USDA UV-B Montorng and Research Programme. Left panel: devaton of centrod wavelength from ntal measurement. Rght panel: Bandwdth expressed as fwhm 5.3.5 Langley Method The Langley method allows determnaton of the solar spectrum outsde the Earth s atmosphere from measurements at varous armasses [Slusser et al., 2000]. Applcaton of the technque s possble only for nstruments that are equpped wth a shadowband. If the radometer s stable, repeated estmates of the extraterrestral spectrum should deally be dentcal. The technque s mostly suted to detect long-term changes of an nstrument s stablty as the Sun s a 15

very stable lght source. The method s not affected by changes n reference nstruments or lamps that mght affect the methods descrbed n Sectons 5.3.1-5.3.3. However, the Langley method gves robust results only f the atmosphere s stable durng the Langley analyss, for example ths requres that the sky s cloud-free and atmospherc ozone and aerosol concentraton are constant over the tme requred for a Langley analyss (typcally a few hours are needed). Apart from hgh alttude stes, these condtons are rarely met. More nformaton on the Langley method s provded n Secton 6.6. Bgelow and Slusser [2000] have compared the methods descrbed n Sectons 5.3.3, 5.3.4, and 5.3.5 and dscussed ther advantages and dsadvantages. More nformaton on long-term stablty of mult-channel nstruments can also be found n Johnsen et al. [2002], Janson and Slusser [2003], and Janson et al. [2004]. 5.4 Vsble and nfrared leakage test The senstvty to vsble and nfrared radaton can be tested wth cut-off flters that transmt vsble and nfrared radaton but block UV radaton (e.g., GG 400 produced by Schott). The measurement should be performed outdoors wth the Sun as the lght source. The flter should be placed on top of the radometer s dffuser. It s mportant that there s a good lght-tght seal between flter and radometer to prevent unfltered radaton from reachng the dffuser. Wth the flter n place the sgnal should be less than 1% compared to the sgnal wthout the flter for SZA smaller than 70. If the nstrument s calbrated wth a Standard of Spectral Irradance (Secton 4.1.3), then the lght leakage should also be tested usng the same calbraton lamp. Ths s mportant as ncandescent lamps have a larger contrbuton from the nfrared than the Sun [Lantz et al., 2005]. 5.5 Cosne error correcton The effect of the cosne error on solar data should be corrected. Correcton methods must take nto consderaton: (1) the devaton of the drectonal response of the radometer from the deal cosne response and (2) the dstrbuton of the radaton feld,.e., the dstrbuton of radance, when measurng solar radaton. Because the radaton feld s generally not known n detal, approxmatons have to be made. The most common approxmatons and smplfcatons are: The global spectral rradance s defned as the sum of drect horzontal spectral rradance and dffuse spectral rradance. For clear-sky condtons, the proporton of both can be ether measured drectly or calculated by a model. For overcast condtons, the drect spectral rradance s set to zero. For partly cloudy condtons, the accuracy of cosne error correcton methods s generally lmted. The drectonal dstrbuton of sky radance s regarded as sotropc. Ths assumpton has proved to be approxmately vald n the UV-B regon (Blumthaler et al., 1996). Methods of cosne error correctons should provde estmates of ther uncertanty. Descrpton of mplementatons and valdatons of cosne error correcton algorthms can be found n (Bernhard and Seckmeyer, 1997), (Seckmeyer and Bernhard, 1993), (Gröbner et al., 1996), (McKenze et al., 1992), (Fester et al., 1997), (Bas et al., 1998) and (Cordero et.al.,2008). 6. APPLICATIONS Several data products can be derved from mult-channel flter nstruments ncludng bologcally effectve rradance (such as erythemal rradance), reconstructed hgh-resoluton solar spectra, total column ozone, and aerosol and cloud optcal depth. The followng secton gves an overvew of methods for calculatng these data products. 6.1 Bologcally effectve rradance The followng methods are sutable for dervng values of bologcally weghted rradance (such as erythemally weghted rradance) from measurements of mult-channel flter nstruments. 16

6.1.1 Calculaton of bologcally effectve rradance va regresson analyss Bologcally effectve rradance D (see Glossary) s estmated va the lnear equaton D = a V, where V are the net sgnals of the nstrument s channels. The coeffcents a are determned va multple lnear regresson aganst D determned from hgh-resoluton spectra measured wth a collocated spectroradometer. For most accurate results all systematc errors of the spectroradometer, such as the cosne error, have to be corrected. Results obtaned wth ths regresson method are usually affected by systematc errors, whch depend on SZA and total ozone O 3. Results can be mproved by multplyng the results of the regresson, D regress, wth a correcton functon ε ( SZA, O3 ). In practce, t s suffcent to consder the effect of the SZA only. In ths case, the correcton functon smplfes to ε (SZA) : D corrected = ε SZA) av = ε ( SZA) ( D regress The functon s determned wth a two-step process: frst, D obtaned from the spectroradometrc measurements s ratoed aganst D regress and plotted versus SZA. In a second step, ε (SZA) s determned by fttng a polynomal to the data of ths scatter plot. The method has been descrbed n detal by Johnsen et al. [2008a; 2008b]. 6.1.2 Method suggested by Dahlback [1996] Ths method s based on calbraton Approach 2 (Secton 4.2) and requres a set of lnear equatons to be solved. The soluton of the set of equatons gves coeffcents a, whch allow the calculaton of bologcally effectve rradance D va D = a V, smlar to the method descrbed n Secton 6.1.1. Method n detal: The net sgnal V of channel s: * * * V = C R ( λ) E( λ) dλ C R λ= λ EλΔλ 0 0 Ths equaton follows mmedately from the defnton of C ntroduced n Secton 4.2. * C can be derved from one sngle solar spectrum measured wth a hgh-resoluton spectroradometer next to the mult-channel nstrument. The exact bologcally effectve rradance Dexact s defned by (see Glossary): D exact = W ( λ) E( λ) dλ W Δ 0 λ Eλ λ, λ = 0 W s the bologcal weghtng functon. where (λ) m D can be approxmated by Dapprox = av, where m s the number of channels of the = 1 mult-channel nstrument. exact 17

By settng D approx = D exact and replacng V by the frst equaton leads to m * a = C Rλ Eλ Wλ Eλ = 1 λ = 0 λ= 0 Ths set of m equatons can be solved wth m dfferent solar spectra, modelled for dfferent SZA and total column ozone. Dahlback [1996] uses 4 spectra wth SZA of 40 and 60 and total ozone columns of 320 and 340 DU. The accuracy of the dose-rate estmate D approx thus derved depends slghtly on SZA and total ozone column, requrng a correcton term ε ( SZA, O3 ), smlar to the method descrbed n Secton 6.1.1. Wth ths correcton appled, the accuracy of the method for 0 < SZA < 80, cloud optcal depth between 0 and 60, and ozone between 200 and 500 DU, s estmated to be better than 5% for the nstrument chosen by Dahlback [1996]. The method suggested by Dahlback has been mplemented by the Norwegan UV-montorng programme [Johnsen et al., 2002] and the U.S. Natonal Scence Foundatons UV Montorng Network for Polar Regons [Bernhard et al., 2005]. In the latter reference, data products for more than 15 dfferent acton spectra are ntroduced, n addton to erythemally weghted rradance. 6.1.3 Calculaton of bologcally effectve rradance from reconstructed spectra Bologcally effectve rradance D can be drectly derved from reconstructed spectra (Secton 6.2). D = W λ) E 0 reconstruc ted ( λ) ( dλ 6.2 Calculaton of hgh-resoluton spectra Solar hgh-resoluton (e.g., 1-nm) spectra may be calculated from mult-channel flter radometer data based on the followng methods. The major characterstcs of these methods are lsted below. For a detaled analyss, please refer to the orgnal papers. 6.2.1 Method suggested by Dahlback [1996] and Booth [1997] In a frst step, total column ozone and cloud optcal depth are determned as descrbed n Sectons 6.3. and 6.4. Informaton on albedo and aerosol optcal depth s obtaned from measurements of other nstruments or from a clmatology. In a second step, a hgh-resoluton solar spectrum s calculated wth a radatve transfer model and the prevously derved nput parameters. References to sutable radatve transfer models are gven n Annex B. 6.2.2 Method suggested by Mn and Harrson [1998] For ths method, the spectrum to be determned s wrtten as a product of extraterrestral spectrum and atmospherc transmsson. The atmospherc transmsson s constructed as an analytcal functon wth several coeffcents X j. In the second step, these coeffcents are determned by non-lnear least-square ft calculaton, resultng n the spectrum sought. Method n detal: The least square-ft s based on mnmzng the sum 1 m m = 1 ( E F ) 2 18

where E s the measured rradance of channel of the mult-flter nstrument (1<= <= Total number of flters) and F s the ntegral F = R ( λ) synthetc_spectrum( λ,x) dλ = R ( λ) S( λ) T( λ, X) dλ 0 In ths equaton, R (λ) s the spectral responsvty of the flter. The synthetc spectrum s constructed as the product of an extraterrestral spectrum S(λ) and an atmospherc transmsson functon T( λ, X ), whch depends on a set of parameters X. (See Mn and Harrson [1998] for dervaton of the functon T( λ, X ).) Before the least squares ft can be carred out, the mult-channel nstrument has to be calbrated based on Approach 2 (Secton 4.2). A comparson of synthetc spectra retreved wth ths method wth spectra measured by spectroradometers has been presented by Gao et al. [2002]. The accuracy of the method has been further mproved by Davs and Slusser [2005]. 6.2.3 Method suggested by Fuenzalda [1998] Ths method s based on a constraned nverson algorthm. See paper for detals. 6.2.4 Method suggested by Thorseth and Kjeldstad [1999], and Thorseth et al., [2000] The method utlzes both a scannng spectroradometer and a mult-channel radometer. The spectroradometer provdes hgh spectral resoluton whle the mult-channel radometer gves hgh temporal resoluton. Combnng both systems provdes solar spectra n hgh temporal resoluton (e.g., 0.5 Hz), whch may be useful for studyng cloud effects. Advantages of the method No radatve transfer modellng requred; error sources such as unknown model nputparameters (e.g., albedo, aerosols, and broken-cloud effects) are nherently excluded. Hgh temporal resoluton. No absolute calbraton of the mult-flter nstrument s requred. The only requrements are a good short-term stablty (stable to wthn 10 mnutes), good lnearty and spectral response at a well defned wavelength. The overall accuracy s manly based on the calbraton and the characterstcs of the spectroradometer. Dsadvantage of the method Addtonal spectroradometer requred. See paper by Thorseth and Kjeldstad [1999] for mplementaton of method. 6.2.5 Spectral reconstructon wth neural networks algorthm A method for reconstructng hgh-resoluton spectra from measurements of mult-flter radometers based on a neural networks algorthm has been suggested by Fester et al. [2005]. Addtonal detals of the algorthm are provded by Schwander et al., [2001]. 6.3 Calculaton of total column ozone 6.3.1 Method suggested by Dahlback [1996] Ths method s based on a modfed algorthm suggested by Stamnes et al. [1991], whch was developed for dervng total column ozone from global rradance spectra. The mplementaton suggested by Dahlback [1996] s based on the followng equaton: 0 19

V / V j * = C R E λ λ / C λ= 0 λ= 0 * j R jλ E λ N The rght sde of ths equaton s calculated wth a radatve transfer model for a wde range of SZAs and total column ozone O 3 usng spectral response functons ( R λ and R jλ ) of two channels ( and j) of the flter radometer. The nomnal wavelength of one channel must be n a spectral regon that s senstve to ozone absorpton, the other channel must be at a wavelength that s less affected by ozone absorpton. A look-up table s constructed from these calculatons, specfyng N for dfferent values of SZA and O 3. In a second step, the ratos of the measured sgnals of channel and j, V and V j, are compared wth the N-ratos n the look-up table gven at the SZA of the measurement, resultng n an estmate for O 3. Both Dahlback [1996] and Stamnes et al. [1991] base the calculaton on the wavelength par 305 nm (senstve to ozone) and 340 nm (nsenstve to ozone). The accuracy was determned by comparson wth Dobson and Brewer measurements. Devatons were found to be smaller than a few percent. 6.3.2 Method suggested by Slusser [1999] Ths method s smlar the approach by Dahlback [1996], and s also based on a modfed Stamnes -method. The wavelength par s 300 (senstve to ozone) and 338 nm (nsenstve to ozone). The paper gves detaled nformaton on nfluence of atmospherc and other parameters on the accuracy of the ozone estmate. These nclude: atmospherc pressure, aerosols, cloud cover, asymmetry factor, sngle-scatterng albedo, ground albedo, ozone absorpton coeffcents used by the model, atmospherc temperature profles, change n spectral response of flters, sgnal-to-nose rato, and cosne-error. 6.4 Calculatons of cloud optcal depth [Dahlback, 1996] Response-weghted rradance for a radometer channel n the UV-A s calculated wth a radatve transfer model for dfferent cloud optcal depths, and a look-up table s created. In a second step, solar measurements of ths channel are compared wth the values n the look-up table to derve cloud optcal depth. As one-dmensonal radatve models can descrbe only stratform clouds, the applcaton of the method s problematc for general cloud condtons. 6.5 Qualty control of spectroradometers Spectroradometers can be more prone to short-term nstabltes than mult-channel flter nstruments due to ther more complex desgn. Comparsons of measurements from spectroradometers and mult-channel nstruments on an operatonal bass can help to uncover problems, track the stablty of the spectroradometer and ntate remedal acton. For a most effcent comparson, the net sgnals V of the mult-channel nstrument should be compared wth response-weghted rradance calculated from spectroradometrc measurements. An applcaton example s provded by Bernhard et al. [2008]. 6.6 Langley Method The Langley Method s based on the work of S. P. Langley n the early 1900s to determne the solar constant. If the nstrument s equpped wth a movable shadowband or f t s mounted on a solar tracker and equpped wth a feld-of-vew lmtng baffled tube, drect measurements of solar rradance can be performed. When a shadowband s used, drect rradance s derved by subtractng dffuse rradance wth the shadowband n place from global rradance measurements wth the shadowband removed from the lght path. Based on drect measurements and the armass, so-called Langley plots can be performed [Slusser et al., 2000]. Ths technque allows several evaluatons: By extrapolaton of the drect measurements at dfferent solar zenth angles (or dfferent armasses) spectral rradance at the top of the atmosphere (.e., the extraterrestral solar spectrum) can be derved. These measurements can be compared to an extraterrestral solar spectrum from the lterature to regularly check the calbraton 20

coeffcents of the mult-flter nstrument. Wth ths technque t s possble to use the Sun as the calbraton source for UV measurements. From measurements at dfferent wavelengths, values of total column ozone, aerosol optcal depth (ncludng wavelength dependence of aerosol parameters) can be derved. Wth sophstcated analyss algorthms and f measurements at approprate wavelengths exst, aerosol sze dstrbutons, and columnar amounts of water vapour and ntrogen doxde may also be extracted. A Langley plot s possble only f the atmosphere s stable for a suffcent amount of tme,.e. the optcal depths of Raylegh and aerosol scatterng as well as ozone absorpton should be constant over a range of at least 3 armasses. Best condtons for Langley plots are usually found on mountans wth alttudes of 2500 meters or hgher. Varyng condtons near ctes may only sporadcally be sutable. Good ntroductons to the Langley Method usng flter nstruments can be found n the works by Schmd and Wehrl [1995]; Schmd et al. [1998]; Harrson and Mchalsky [1994a]; Harrson and Mchalsky [1994b]; and Slusser et al., [2000]. Note that these lterature examples are not a complete lst of all the valuable work that has been done concernng the Langley Method. 21

GLOSSARY Acton spectrum See bologcal weghtng functon. Armass Armass (or more precsely relatve optcal armass ) s defned as the rato of the actual (slant) optcal path length taken by the drect solar beam to the analogous vertcal path when the Sun s overhead from the surface to the top of the atmosphere. Azmuthal Error The azmuthal error f a descrbes the varaton of the angular response of a radometer at a fxed ncdence angle ε as a functon of the azmuthal angleϕ. It s defned by f a Y ( ε, ϕ) = < Y readng readng ( ε, ϕ) ( ε) > 1 100% where Y readng (ε,ϕ) s the readng of the radometer at angles ε and ϕ < Y ( ε)> s the average response at ncdence angle ε defned by: readng < Y readng ( ε ) >= n = 1 Y readng n ( ε, ϕ ) Yreadng ( ε, ϕ) s measured at n dscrete azmuth anglesϕ wth 1<=<=n at ncdence angle ε. The angular response should be measured at least at four dfferent azmuth angles (e.g., 0,90,180, and 270 ). Bologcally effectve rradance; bologcal weghtng functon A bologcal weghtng functon descrbes the wavelength dependence of effects ntroduced by electromagnetc radaton on bologcal matter. Dependng on the effect and the nvolved organsm dfferent bologcal weghtng functons are used. The bologcally effectve rradance s calculated by multplyng global spectral rradance E G (λ) wth the acton spectrum and ntegratng over wavelength λ: An mportant weghtng functon s the acton spectrum for erythema proposed by CIE [McKnlay and Dffey, 1987], whch descrbes the wavelength dependence of the reddenng of human skn by UV radaton (see also below erythemally weghted rradance E CIE ). Centrod wavelength The centrod wavelength λ C of a response functon R(λ) s defned as defned as follows: λ C = λr( λ) dλ R( λ) dλ 22

Cosne error The devaton of the angular response of a radometer from the deal cosne response s specfed wth two parameters n ths document. The frst of these (a) s defned accordng to CIE [1982] and s expressed by the quantty f 2a ( ε, ϕ) : f2 a ( ε, ϕ) = (, ) Yreadng ε ϕ 1 readng ( 0, ) cos( ) Y ε = ϕ ε 100% where ε s the ncdence angle of the radaton, ϕ s the azmuth angle, Y ( ε, ϕ) s the readng of the radometer at ε and ϕ, readng Y ( ε = 0, ϕ) cos( ε ) s the deal response. readng The second specfcaton (b) refers to sotropc radaton and s defned as follows: 2π π / 2 Yreadng ( ε, ϕ) / Yreadng ( ε = 0, ϕ) sn( ε ) dε dϕ 0 0 2 ( ε, ϕ) = 1 2π π / 2 cos( ) sn( ) ε ε dε dϕ 0 0 f b 100% Detecton threshold Mnmum rradance that s detectable. Erythemally weghted rradance E CIE Global spectral rradance E G (λ) multpled wth the acton spectrum for erythema, C (λ), proposed by CIE [McKnlay and Dffey, 1987], and ntegrated over wavelength λ: E CIE = 400nm E G 250nm ( λ) C( λ) dλ where C (λ) = 1 for 250< λ 298 nm = 10 (0.094(298-λ)) for 298< λ 328 nm = 10 (0.015(139-λ)) for 328< λ 400 nm Global spectral rradance E G (λ) Radant energy dq arrvng per tme nterval dt, per wavelength nterval dλ, and per area da on a horzontal surface from all parts of the sky above the horzontal, ncludng the dsc of the sun tself: dq E G( λ) = = ED( λ) cos( ψ ) + ES ( λ) ; (unts: W m -2 nm -1 s -1 ) dtdadλ where ψ s the solar zenth angle, E D (λ) s drect normal spectral rradance,.e. radant energy dq arrvng from the dsk of the sun per tme nterval dt, per wavelength nterval dλ, and per area da on a surface normal to the solar beam and E S (λ) s dffuse spectral rradance,.e. radant energy dq arrvng per tme nterval dt, per wavelength nterval dλ, and per area da on a horzontally orented surface from all parts of the sky above the horzontal, excludng the dsc of the sun. 23

Lnearty The lnearty of a radometer s the degree to whch the output quantty of the radometer (e.g., a voltage) s proportonal to the nput quantty (e.g., global spectral rradance). Response-weghted-rradance Response-weghted rradance E W s defned n ths document as the ntegral of the product of spectral rradance E (λ) and the spectral response functon R (λ) of channel from a mult-flter radometer: E W = 0 R ( λ) E( λ) dλ Spectral senstvty or spectral response functon R (λ) Rato of the sgnal from a specfc channel of a flter radometer output, dv (λ), to the spectral rradance de (λ) at the place of the radometer s collector, as a functon of wavelength λ: dv ( λ) R ( λ) = de ( λ) Remarks: a) For measurng R (λ), a tunable radaton source s requred. A system for measurng R(λ) s descrbed n Secton 5.1. b) In ths document R(λ) s regarded a relatve functon and can be multpled wth a wavelength ndependent factor. Spectroradometer Instrument for the spectrally resolved measurement of electromagnetc radaton. Instruments of ths type are descrbed n Part 1 of ths document seres [Seckmeyer et al., 2001]. Total ozone column Heght of a hypothetcal layer whch would result f all ozone molecules n a vertcal column above the Earth s surface were brought to standard pressure (1013.25 hpa) and temperature (273.15 K). The total ozone column s usually reported n mll-atmosphere-centmeters (m-atm-cm), commonly called Dobson unts (DU). ). The global average ozone amount s close to 300 DU, whch corresponds to a layer of thckness 3 mm at STP. One DU Defnes the amount of ozone n a vertcal column whch, when reduced to standard pressure and temperature, wll occupy a depth of 0.01 mm. Corresponds to molecules/cm 2. UV-A radaton Electromagnetc radaton between 315 and 400 nm. UV-B radaton Electromagnetc radaton between 280 and 315 nm. 24

REFERENCES Bas A.F., S. Kazadzs, D. Bals, C. S. Zerefos, and M. Blumthaler. (1998). Correctng global solar UV spectra recorded by a Brewer spectroradometer for ts angular response error, Appl. Opt., 37(27). 6339-6344. Bernhard G. and G. Seckmeyer. (1997). New entrance optcs for solar spectral UV measurements, Photochem. and Photobol., 65(6), 923-930. Bernhard, G., C. R. Booth, and J. C. Ehramjan. (2005). Real-tme ultravolet and column ozone from multchannel ultravolet radometers deployed n the Natonal Scence Foundaton s ultravolet montorng network. Optcal Engneerng, 44(4), 041011-1 - 041011-12. Bernhard, G., C. R. Booth, and J. C. Ehramjan. (2008). Comparson of UV rradance measurements at Summt, Greenland; Barrow, Alaska; and South Pole, Antarctca, Atmos. Chem. Phys., 8(16), 4799 4810. Bgelow, D. S. and J. R. Slusser. (2000). Establshng the stablty of mult-flter UV rotatng shadow-band radometers, J. Geophys. Res.,105(D4), 4833-4840. Blumthaler M., J. Gröbner, M. Huber, and W. Ambach. (1996). Measurng spectral and spatal varatons of UVA and UVB sky radance, Geophys. Res. Lett., 23(5), 547-550. Bolsée, D., A. R. Webb, D. Gllotay, B. Dörschel, P. Knuschke, A. Krns, and I. Terenetskaya. (2000). Laboratory Facltes and Recommendatons for the characterzaton of Bologcal Ultravolet Dosmeters, Appl. Opt. 39(16), 2813-2822. Booth, C. R., Synthetc UV spectroradometry. (1997). n: IRS 96 Current Problems n Atmospherc Radaton, Eds. W.L. Smth and K. Stamnes, Deepak Publ. Hampton, Vrgna, U.S.A., pp. 849-852. Commsson Internatonal de l Éclarage (CIE) (Eds.). (1982). Methods of charactersng the performance of radometers and photometers, CIE publcaton, No 53 (TC-2.2), Pars, France. Cordero R. R., G. Seckmeyer G., and F. Labbe. (2008). Cosne error nfluence on ground-based spectral UV rradance measurements, Metrologa, 45, 406-414. Dahlback, A. (1996). Measurements of bologcally effectve UV doses, total ozone abundances, and cloud effects wth multchannel, moderate bandwdth flter nstruments, Appl. Opt., 35(33), 6514-6521. Davs, J. M. and J. Slusser (2005). New USDA UVB synthetc spectra algorthm, n: Ultravolet Ground- and Space-based Measurements, Models, and Effects V, edted by G. Bernhard, J. R. Slusser, J.R. Herman, and W. Gao, Proceedngs of SPIE, Vol. 5886, page 0B-1-0B- 7. Daz, S. C. R. Booth, R. Armstrong, C. Brunat, S. Cabrera, C. Camlon, C. Cascca, G. Deferrar, H. Fuenzalda, C. Lovengreen, A. Paladn, J. Pedron, A. Rosales, H. Zagarese, and M. Vernet. (2005). Multchannel radometer calbraton: a new approach, Appl. Opt., 44(26), 5374-5380. d Sarra, A., P. Dsterhoft, and J. J. De Lus. (2002). On the mportance of spectra responsvty of Robertson-Berger-type Ultravolet Radometers for Long-term Observatons, Photochem, Photobol, 76(1), 64-72. Fester U., R. Grewe and K. Gercke. (1997). A method for correcton of cosne errors n measurements of spectral UV rradance, Solar Energy, 60(6), 313-332. Fester, U., A. Kafel, R.-D. Grewe, J. Kaptur, O. Reutter, M. Wohlfart, and K. Gercke (2005). Fast measurements of solar spectral UV rradance frst performance results of two novel spectroradometers, Optcal Engneerng, Volume 44(4), 041007-1 - 041007-9. Fuenzalda, H. A. (1998). Global ultravolet spectra derved drectly from observatons wth multchannel radometers, Appl. Opt., 37(33), 7912-7919. 25

Gao, W., J. R. Slusser, L. C. Harrson, P. Dsterhoft, Q. Mn, B. Olson, K. Lantz and B. Davs. (2002). Comparsons of UV Synthetc Spectra Retreved from the USDA UV Mult-flter Rotatng Shadow-band Radometer wth Collocated USDA Reference UV Spectroradometer and NIWA UV Spectroradometer, n: Ultravolet Ground- and Spacebased Measurements, Models, and Effects, edted by J.R. Slusser, J.R. Herman, and W. Gao, Proceedngs of SPIE, Vol. 4482, 408-414. Grenfell, T. C., S. G. Warren, and P. C. Mullen. (1994). Reflecton of solar radaton by the Antarctc snow surface at ultravolet, vsble, and near nfrared wavelengths, J. Geophys. Res., 99(D9), 18,669 18,684. Gröbner J., M. Blumthaler, W. Ambach. (1996). Expermental nvestgaton of spectral global rradance measurement errors due to non deal cosne response, Geophys. Res. Lett., 23(18), 2493-2496. Harrson, L., J. J. Mchalsky, and J. Berndt. (1994a). Automated multflter rotatng shadow-band radometer: an nstrument for optcal depth and radaton measurements, Appl. Opt., 33(22), 5118-5121. Harrson L. and J. Mchalsky (1994b). Objectve algorthms for the retreval of optcal depths from ground based measurements, Appl. Opt., 33, 5126-5132. Hofmann, D. J., S. J. Oltmans, G. L. Koeng, B. A. Bodhane, J. M. Harrs, J. A. Lathrop, R. C. Schnell, J. Barnes, J. Chn, D. Kunyuk, S. Ryan, R. Uchda, A. Yoshnaga, P. J. Neale, D. R. Hayes, J., V. R. Goodrch, W. D. Komhyr, R. D. Evans, B. J. Johnson, D. M.Quncy, and M. Clark. (1996). Record low ozone at Mauna Loa Observatory durng wnter 1994-1995: a consequence of chemcal and dynamcal synergsm?, Geophys. Res. Lett., 23 (12), 1533-1536. Hülsen G. and J. Gröbner. (2007). Characterzaton and calbraton of ultravolet broadband radometers measurng erythemally weghted rradance, Appl. Opt. 46(23), 5877-5886. Janson, G. T. and J. R. Slusser. (2003). Long-term stablty of UV multflter rotatng shadowband radometers, n: Ultravolet Ground- and Space-based Measurements, Models, and Effects III, edted by J. R. Slusser, J. R. Herman, and W. Gao, Proceedngs of SPIE, Vol. 5156, 94-100. Janson, G., J. Slusser, G. Scott, P. Dsterhoft, K. Lantz. (2004). Long-term stablty of UV mult-flter rotatng shadowband radometers, Part 2: lamp calbratons versus the Langley method, n: Ultravolet Ground and Space-based Measurements Models, and Effects IV, edted by J. R. Slusser, J. R. Herman, W. Gao, and G. Bernhard, Proceedngs of SPIE, Vol. 5545, 43-48. Johnsen, B. J., O. Mkkelborg, M. Hannevk, L. T. Nlsen, G. Saxebøl, and K. G. Blaasaas. (2002). The Norwegan UV montorng programme. Perod 1995/96 to 2001, Strålevern Rapport 2002:4, Norwegan Radaton Protecton Authorty. Johnsen, B., B. Kjeldstad, T. N. Aalerud, L. T. Nlsen, J. Schreder, M. Blumthaler, G. Bernhard, C. Topaloglou, O. Menander, A. Bagher, J. R. Slusser, and J. Davs. (2008a). Intercomparson and harmonzaton of UV Index measurements from multband flter radometers, J. Geophys. Res., 113, D15206, do:10.1029/2007jd009731. Johnsen, B., Bert Kjeldstad, T. N. Aalerud, L. T. Nlsen, J. Schreder, M. Blumthaler, G. Bernhard, C. Topaloglou, O. Menander, A. Bagher, J. R. Slusser and J. Davs. (2008b). Intercomparson of global UV ndex from Multband flter radometers: Harmonzaton of global UVI and spectral rradance, WMO GAW Report No. 179, WMO TD No. 1454. Avalable at: http://www.wmo.nt/pages/prog/arep/gaw/documents/gaw179-web.pdf Lantz, K., P. Dsterhoft, C. Wlson, G. Janson, J. Slusser, S. Bloms, and J. Mchalsky. (2005). Out of Band Rejecton Studes of the UV mult-flter rotatng shadow-band radometers, n: Remote Sensng of Clouds and the Atmosphere X, edted by K. Schäfer, A. T. Comerón, J. R. Slusser, R. H. Pcard, and M. R. Carleer; Proceedngs of SPIE, Vol. 5979, 59791N-1. Lley, J.B., and R.L. McKenze. (2006). Where on Earth has the hghest UV?, n: UV Radaton and ts Effects: an update, pp. 26-37. Avalable at: http://www.nwascence.co.nz/rc/atmos/uvconference 26

Lucas RM, McMchael AJ, Armstrong BK, Smth WT. (2008) Estmatng the global dsease burden due to ultravolet radaton exposure. Internatonal Journal of Epdemology; 37 (3): 654-667. Madronch, S., and S. Flocke. (1995). Theoretcal estmaton of bologcally effectve UV radaton at the earth s surface, n Solar Ultravolet Radaton. NATO, Seres I: Advanced Study Insttute, edted by C.S. Zerefos, and A.F. Bas, pp. 23-48, Sprnger, Berln. Mayer B. and A. Kyllng. (2005). Techncal Note: The lbradtran software package for radatve transfer calculatons: descrpton and examples of use. Atmos. Chem. Phys., 5(7), 1855-1877. McKenze R. L., P. V. Johnston, M. Kotkamp, A. Bttar, J. D. Hamln. (1992). Solar ultravolet spectroradometry n New Zealand: nstrumentaton and sample results from 1990, Appl. Opt., 31(30), 6501-6509. McKnlay, A. F., and B. L. Dffey (Eds.) (1987). A reference acton spectrum for ultravolet nduced erythema n human skn, CIE J., 6(1), 17 22. Mn, Q. and L. Harrson. (1998). Synthetc spectra for terrestral ultravolet measurements, J. Geophys. Res., 103(D14), 17033-17039. Schmd, B. and C. Wehrl. (1995). Comparson of Sun photometer calbratons by use of the Langley technque and the standard lamp, Appl. Opt., 34(21), 4500-4512. Schmd B., P. R. Spyak, S. F. Bggar, C. Wehrl, J. Sekler, T. Ingold, C. Maetzler, and N. Kaempfer. (1998). Evaluaton of the applcablty of solar and lamp radometrc calbratons of a precson Sun photometer operatng between 300 and 1025 nm, Appl. Opt., 37(18), 3923-3941. Schwander H., Mayer B., Ruggaber A., Albold A., Seckmeyer G., Koepke P. (1999). Method to determne snow albedo values n the UV for radatve transfer modellng, Appl. Opt., 38(18), pp 3869-3875. Schwander, H., A. Kafel, A. Ruggaber, and P. Koepke. (2001). Spectral radatve-transfer modellng wth mnmzed computaton tme by use of a neural-network technque, Appl. Opt., 40(3), 331-335. Seckmeyer G. and G. Bernhard. (1993). Cosne error correcton of spectral UV-rradances, n: Atmospherc Radaton, edted by Knut H. Stamnes, Proceedngs, SPIE, 2049, 140-151. Seckmeyer, G., A. Bas, G. Bernhard, M. Blumthaler, C. R. Booth, P. Dsterhoft, P. Erksen, R. L. McKenze, M. Myauch, and C. Roy. (2001). Instruments to measure solar ultravolet radaton. Part 1: Spectral nstruments, World Meteorologcal Organzaton, Global Atmosphere Watch Publcaton No. 125, WMO TD No. 1066, Geneva, Swtzerland. Seckmeyer, G., A. Bas, G. Bernhard, M. Blumthaler, C. R. Booth, K. Lantz, and R. L. McKenze. (2005). Instruments to measure solar ultravolet radaton. Part 2: Broadband nstrument measurng erythemally weghted solar rradance, World Meteorologcal Organzaton, Global Atmosphere Watch Publcaton No. 164, WMO TD No. 1289, Geneva, Swtzerland. Avalable at: http://www.wmo.nt/pages/prog/arep/gaw/gaw-reports.html Slusser, J. R., J. H. Gbson, D. S. Bgelow, D. Kolnsk, W. Mou, G. Koeng, and A. Beauben. (1999). Comparson of column ozone retrevals employng a UV mult-flter rotatng shadowband radometer wth those from Brewer and Dobson spectrophotometers, Appl. Opt., 38(9), 1543-1551. Slusser, J. R., J. H. Gbson, D. Kolnsk, P. Dsterhoft, K. Lantz and A. F. Beauben. (2000). Langley Method of Calbratng UV Flter Radometers, J. Geophys. Res., 105(D4), 4841-4849. Stamnes K., J. R. Slusser, and M. Bowen. (1991). Dervaton of total ozone abundance and cloud effects from spectral rradance measurements. Appl. Opt., 30(30), 4418-4426. Thorseth, T. M. and B. Kjeldstad. (1999). All-weather ultravolet solar spectra retreved at a 0.5 Hz samplng rate, Appl. Opt., 38(30), 6247-6252. Thorseth T. M., B. Kjeldstad, and B. Johnsen. (2000). Comparson of solar ultravolet measurements performed wth spectroradometers and moderate bandwdth multchannel radometers for dfferent cloud condtons, J. Geophys, Res., 105 (D4), 4809-4820. 27

Webb, A. R., B. G. Gardner, K. Leszczynsk, V. Mohnen, P. Johnston, N. Harrson, and D. Bgelow. (1998). Qualty assurance n montorng solar ultravolet radaton: the state of the art, World Meteorologcal Organzaton, Global Atmosphere Watch Publcaton No. 146, WMO TD No. 1180, Geneva, Swtzerland. Webb, A. R., B. G. Gardner, T. J. Martn, K. Leszczynsk, J. Metzdorf, and V. A. Mohnen. (2003). Gudelnes for ste qualty control of UV montorng, World Meteorologcal Organzaton, Global Atmosphere Watch Publcaton No. 126, WMO TD No. 884, Geneva, Swtzerland. Wehs P., J. Lenoble, M. Blumthaler, T. Martn, G. Seckmeyer, R. Phlpona, de la Casnere, C. Sergent, J. Gröbner, T. Cabot, D. Masserot, T. Pchler, E. Pougatch, G. Rengarajan, D. Schmuck, and S. Smc. (2001). Modellng the effect on an nhomogeneous surface albedo on ncdent UV radaton n mountanous terran: determnaton of an effectve surface albedo, Geophys. Res. Lett., 28(16), 3111-3114. WHO (2008) World Health Organzaton, Internatonal Agency for Research on Cancer. Vtamn D and cancer. IARC Workng Group Reports, WHO Press, 5, 148. WMO (1986). Recent progress n sun photometry; Determnaton of the aerosol optcal depth, Envronmental Polluton Montorng and Research Programme 43, WMO TD No.143, Geneva, Swtzerland. WMO (1993). Report of the WMO workshop on the measurement of atmospherc optcal depth and turbdty. Slver Sprng, Md., 6-10 December, (B. Hcks, ed.) Global Atmosphere Watch Publcaton No. 101, WMO TD No. 659, Geneva, Swtzerland. Wuttke S., G. Seckmeyer, and G. Koeng-Langlo. (2006). Measurements of spectral snow albedo at Neumayer, Antarctca, Annales Geophyscae, 24, 7-21. 28

ANNEX A Centre Wavelengths of some Avalable Mult-Flter Instruments Instruments to measure global rradance Bosphercal Instruments GUV: 305, 313, 320, 340, 380, 395 (all 10 nm fwhm), PAR. NILU-UV: 300, 312, 320, 340, 380, (all 10 nm fwhm), PAR. UV-SPRAFIMO: 303.5, 309, 314.5, 327, 387, (all ~2nm). Instruments equpped wth shadowbands measurng global and dffuse rradance Yankee-SPUV: 300, 311, 317, 325, 332, 368, 500, 673, 778, 870. Yankee UV-MFRSR: 300, 305.5, 311.5, 317.5, 325, 332.5, 368 (all 2 nm fwhm, partly Dobson wavelengths). Yankee MFRSR: 415, 500, 615, 673, 870, 940 (all 10 nm fwhm). Instruments mounted on a solar tracker for measurng drect rradance UV-PFR: 305, 311, 318, 332 (all 1,5 nm fwhm), optonal 368, 412, 500, 862. CIMEL sunphotometer: 440, 670, 870, 936,1020 (10 nm fwhm). SPM-2000 Sun photometer 300, 313, 305, 310, 320, 340, 368, 412, 450, 500, 610, 675, 719, 778, 817, 862, 946, 1024 [Schmdt et al., 1998]. Wavelength for total ozone dervaton Dobson wavelengths: A (305.5, 325.4), B (308.8, 329.1), C (311.45, 332.4), D (317.66, 339.8). Brewer wavelengths: 306.3, 310.1, 313.5, 316.8, 320. Ozone dervaton from global spectra suggested by Stamnes et al. [1991]: 305, 340. Wavelengths frequently used for specfc applcatons Column waver vapour retreval: 940 nm. Atmospherc aerosol optcal depth measurements: 368, 412, 450, 500, 610, 675, 719, 778, 817, 862, 946, 1024 nm. Wavelength recommended by [WMO, 1986; WMO, 1993]. Wavelength affected by NO 2 and SO 2 absorpton. 29

ANNEX B References of Freely Avalable Radatve Transfer Programmes lbradtran - Lbrary for Radatve Transfer http://www.lbradtran.org TUV - Tropospherc Ultravolet and Vsble Radaton Model http://www.acd.ucar.edu/tuv/ STAR - System for Transfer of Atmospherc Radaton http://www.meteo.physk.un-muenchen.de/strahlung/uvrad/star/starnfo.htm FASTRT - Fast radaton transfer modellng (Onlne radatve transfer model based on Look-Up Tables that were calculated wth lbradtran.) http://zardoz.nlu.no/~olaeng/fastrt/fastrt.html SBDART - Santa Barbara DISORT Atmospherc Radatve Transfer Model http://www.ncga.ucsb.edu/projects/metadata/standard/uses/sbdart.htm Addtonal codes may be found here: http://en.wkpeda.org/wk/lst_of_atmospherc_radatve_transfer_codes 30

ANNEX C Calculatons n Support of Specfcatons Provded n Secton 3 Calculatons presented n ths Annex explore the magntude of measurement errors resultng from: Wavelength-shfts of spectral response functons. Changes n the bandwdth of spectral response functons. Systematc errors that may occur when mult-flter nstruments are calbrated wth lamps. Lght leakage (.e. sgnfcant response outsde the flter bandwdth). Calculatons are based on modelled solar clear sky UV spectra, E s, generated by the lbradtran radatve transfer software package [Mayer and Kyllng, 2005]. Input parameters are total ozone O 3, SZA θ, surface albedo (set to 5%), and default aerosol amount. A set of dealzed, rectangular spectral response functons are used. These functons have unty response wthn the fwhm, and a base level of 10-10 ( zero ) outsde ths range. C.1 Effects from wavelength-shfts of spectral response functons Calculatons of relatve errors n the sgnal resultng from wavelength-shfts of spectral response functons are based on the followng equaton. ΔU ( λ, δλ, FWHM ) = R ( λ + λ, fwhm, λ' ) ES ( λ', θ, O3 ) dλ' 1 100% (,, ') ( ',, 3) ' R λ fwhm λ ES λ θ O dλ Here ΔU ( λ, δλ, FWHM) s the relatve change n sgnal at nomnal wavelength λ, for a small wavelength error δλ, and a gven bandwdth expressed n fwhm, and R s the spectral response functon of channel. Results correspondng to a wavelength shft of δλ = 0.03 nm at the nomnal centre wavelength of λ = 305 nm and fwhm = 10.0 nm are shown n Fgure C.1. Errors are smaller than 2% for SZA less than 80º and ozone amounts between 250 DU and 450 DU. Errors n the sgnal dmnsh wth longer centre wavelengths but are almost ndependent of fwhm for bandwdths between 1.0 nm and 10 nm fwhm. 31

Errors n lamp based calbratons [%] from 0.03nm wavelength errors n srf Center wavelength 305nm, FWHM 10.0nm 600 2 550 1 1.2 1.4 1.6 1.8 2.4 2 2.2 2.2 500 Total ozone, DU 450 400 350 0.8 1 0.8 1.2 1.4 1.6 1.8 2 1.8 1.6 1.4 300 0.6 11 1.2 1.6 250 1.4 1.2 200 0.6 10 20 30 40 50 60 70 80 90 SZA 0.8 1 Fgure C.1 - Contour dagram of % errors n the sgnal from a 0.03 nm shft for nomnal centre wavelength at 305 nm, a bandwdth of 10.0 nm fwhm, as functon of SZA and total ozone amount. Contour numbers are gven n percent C.2 Effects from changes n bandwdth Errors from shfts n bandwdth are based on the equaton below, where the bandwdth s expressed as fwhm and modfed wth a constant δ. ΔU ( λ,fwhm, δ) = R ( λ, λ',fwhm (1 δ)) ES( λ', θ,o3 )dλ' 1 100% R ( λ, λ',fwhm) ES( λ', θ,o3 )dλ' Results are normalzed to SZA = 40º and total ozone of 300 DU n order to focus on the relatve changes as functon of SZA and O 3. Errors resultng from a 2% change n fwhm at centre wavelength of λ = 305 nm and a nomnal bandwdth of 10.0 nm fwhm can be as hgh as 3% for typcal ozone amounts and SZA less than 80º (Fgure C.2). Smlar calculatons for a nomnal bandwdth of 2.0 nm result n very small errors (Fgure C.3). Errors are wavelength dependent. At UV-B wavelengths n the ozone cut-off regon, the errors become ncreasngly larger wth shorter centre wavelengths, whereas the effect s nsgnfcant outsde the ozone cut-off regon. 32

Dfference [%] nduced by 0.02 [%] varaton n FWHM Center wavelength 305.0nm, FWHM 10.0nm, total Ozone 250 to 450 DU 0.5 0-0.5 Dfference [%] -1-1.5-2 -2.5-3 0 10 20 30 40 50 60 70 80 90 SZA Fgure C.2 - Dfference n sgnal at 305 nm for 2% varatons n bandwdth, calculated for a bandwdth of 10.0 nm fwhm. Dfferences are relatve to SZA 40º and total ozone 300 DU. Dfference [%] nduced by 0.02 [%] varaton n FWHM Center wavelength 305.0nm, FWHM 2.0nm, total Ozone 250 to 450 DU 0.01 0-0.01-0.02 Dfference [%] -0.03-0.04-0.05-0.06-0.07-0.08 0 10 20 30 40 50 60 70 80 90 SZA Fgure C.3 - Same as Fgure C.2 but calculated for a bandwdth of 2.0 nm fwhm 33

C.3 Systematc errors resultng from calbratng mult-flter nstruments wth a standard of spectral rradance If mult-flter nstruments are calbrated wth a Standard of Spectral Irradance (3) (Secton 4.1.3) the correcton functon K must be appled to account for the msmatch of the solar spectrum and the spectrum of the standard lamp. If ths correcton functon s not appled as descrbed n Secton 4, large errors n the measured solar rradance wll result. These errors can be expressed by the error functon Δ U ( λ, FWHM, θ, O3 ) : E,, ( ) (,,, ) S approx λ ΔU λ FWHM θ O3 = 1 100%,, (,,, 3) ES complete λ FWHM θ O ( ) ( ', ) ( ',, 3) ' E λ R λ FWHM E λ θ O dλ L S = 1 100% E ( λ, θ, 3) ( ', ) ( ') ' S O R λ FWHM EL λ dλ Fgures C.4 and Fgure C.5 show Δ U( λ, FWHM, θ, O3 ) as a functon of SZA for a centre wavelength of 305 nm and bandwdths of 10.0 nm and 1.0 nm, respectvely. Data shown n both fgures are normalzed to one for SZA = 40º and total ozone of 300 DU. For a hypothetcal nstrument wth a bandwdth of 10 nm, the error can be as large as 200% for a flter centred at 305 nm (Fgure C.4). However, t s less than 1.2% for an nstrument wth a flter centred at 305 nm wth a bandwdth of 1.0 nm (Fgure C.5). If a correcton functon s appled to the calbraton usng a lamp Standard of Spectral Irradance as descrbed n Secton 4, the % errors n the sgnal wll be sgnfcantly less. Because the correcton functon ncludes the spectral response functon, errors or changes n the wavelength and the bandwdth as descrbed n Sectons C.1 and C.2 wll propagate to the correcton functon. Dfference [%] solar to lamp based calbraton factors at nomnally 305nm FWHM 10.0nm, total ozone 250-450 DU 200 150 100 50 0-50 0 20 40 60 80 100 SZA Fgure C.4 - Error Δ U( λ, FWHM, θ, O3 ) for λ = 305 nm and FWHM = 10 nm. Dfferences are relatve to SZA 40º and total ozone 300 DU. Errors for total ozone of 250 DU (450 DU) are ndcated by the bottom (top) functon. The graph ndcates that measurements of spectral rradance at 305 nm can be n error by as much as 200% f a mult-flter nstrument wth a bandwdth of 10 nm s calbrated wth a standard lamp, and no correctons are appled 34

Dfference [%] solar to lamp based calbraton factors at nomnally 305nm FWHM 1.0nm, total ozone 250-450 DU 1.5 1 0.5 0-0.5 0 20 40 60 80 100 SZA Fgure C.5 - Same as Fgure C.4, but for a bandwdth of 1.0 nm FWHM. The maxmum error s less than ±1.2% C.4 Errors from stray lght (lght leakage) n spectral response functons The senstvty of mult-channel flter radometer to radaton outsde the core wavelength range of ther spectral response functons can lead to consderable errors. These errors are quantfed here wth the functon Δ U ( λ, FWHM, Tal), defned as: ΔU ( λ, FWHM, Tal) = R ( λ, λ', FWHM, Tal)ES( λ', θ,o3 )dλ' 1 100% R ( λ, λ' FWHM, NoTal)ES( λ', θ,o3 )dλ' The parameter NoTal s equal to 10-10, whch for all practcal purposes s dentcal to zero, and the parameter Tal, whch s set to 10-4. (Ths means that the nstrument channel n queston has a wavelength-ndependent senstvty of 10-4 across the UV band relatve to the peak senstvty of the spectral response functon.) Fgures C.6 and Fgure C.7 show Δ U ( λ, FWHM, Tal) as a functon of SZA for a centre wavelength of 305 nm and bandwdths of 10.0 nm and 1.0 nm, respectvely. For a hypothetcal nstrument wth a bandwdth of 10 nm, the error may be as hgh as 80% for large SZA, but s below 5% for SZA smaller than 55. For an nstrument wth a bandwdth of 1.0 nm, the error can be as large as 1500% (Fgure C.7). Ths ndcates that lght leakage can be an mportant error source for flter nstruments wth small bandwdth. 35

Dfference [%] nduced by secondary peak of magntude 1E-004 Center wavelength 305.0nm, FWHM 10.0nm, total Ozone 250 to 450 DU 100 80 Dfference [%] 60 40 20 0-20 0 20 40 60 80 100 SZA Fgure C.6 - Error Δ U ( λ, FWHM, Tal) for λ = 305 nm and FWHM= 10 nm. Dfferences are relatve to SZA 40º and total ozone 300 DU. Errors for total ozone of 250 DU (450 DU) are ndcated by the bottom (top) functon. The graph ndcates that measurements of spectral rradance at 305 nm can be n error by up to 80% f a mult-flter nstrument wth a bandwdth of 10 nm has a wavelength-ndependent senstvty of 10-4 across the UV band relatve to the peak senstvty of the spectral response functon Dfference [%] nduced by secondary peak of magntude 1E-004 Center wavelength 305.0nm, FWHM 1.0nm, total Ozone 250 to 450 DU 2000 1500 Dfference [%] 1000 500 0-500 0 20 40 60 80 100 SZA Fgure C.7 - Same as Fgure C.6, but for a bandwdth of 1.0 nm FWHM. The maxmum error s ±1500% 36

ANNEX D Maxmum Irradance at the Earth s Surface The maxmum UV rradance from 20 years of grdded satellte estmatons from TOMS nstruments corresponds to a UV Index of 24.8, as reported by Lley and McKenze [2006]. That value occurred n Cusco, Peru (13.5 S, 72 W) n February 1998, when the Sun was drectly overhead. The mean alttude of the grd cell (~100 km x 100 km) was 3655 m, though the surroundng terran extended to alttudes of 6500 m. At the tme of ths maxmum, the total column ozone was 235 DU. Spectral rradances were calculated wth the TUV radatve transfer code [Madronch and Flocke, 1995] to match these condtons. No clouds or aerosols were ncluded, and to smulate the UVI value of 24.8 a surface albedo of 0.8 s used. It should be noted that such hgh surface albedo are not realstc for ths locaton and season. Effectve albedos above 0.8 are realstc only for Antarctca [Grenfell et al., 1994; Schwander et.al., 1999; Wehs et.al., 2001; Wuttke et.al., 2006]. To estmate the maxmum spectral rradances (W m -2 nm -1 ) a calculaton was performed for the hghest peak n ths regon (6500 m) wth assumpton of full snow cover (surface albedo 0.99) and assumng a total column ozone amount of 200 DU; such low ozone values have been observed prevously wthn the tropcs. For example the lowest ozone amount measured at Mauna Loa Observatory was 200 DU [Hofmann et al., 1996]. Fnally, a further 20% ncrease was ntroduced to account for possble enhancements due to clouds that do not obscure the sun. These spectral rradances are compared wth the extraterrestral solar spectral rradance n the fgure and table below. Peak Irradances (from Cusco_Peak UV.xls) Irradance (W m -2 nm -1 ) 2.0 1.5 1.0 0.5 Mean Sun ET:UVI>300 Cusco TOMS:UVI=24.8 Cusco Peak: UVI=32.1 Cld Enhance: UVI=38.5 0.0 280 300 320 340 360 380 400 Wavelength (nm) 37

Wvl (nm) Mean Sun ET: UVI>300 Cusco TOMS UVI=24.8 Cusco Peak UVI=32.1 20% Cloud Enhanced UVI=38.5 290.5 6.48E-01 1.10E-04 4.41E-04 5.29E-04 291.5 6.12E-01 3.26E-04 1.07E-03 1.28E-03 292.5 5.26E-01 6.64E-04 1.91E-03 2.29E-03 293.5 5.83E-01 1.65E-03 4.19E-03 5.03E-03 294.5 5.30E-01 2.75E-03 6.38E-03 7.66E-03 295.5 6.10E-01 6.52E-03 1.36E-02 1.63E-02 296.5 5.16E-01 9.44E-03 1.81E-02 2.17E-02 297.5 5.93E-01 1.85E-02 3.26E-02 3.91E-02 298.5 4.44E-01 2.03E-02 3.38E-02 4.06E-02 299.5 5.64E-01 3.85E-02 6.03E-02 7.24E-02 300.5 4.08E-01 3.84E-02 5.73E-02 6.88E-02 301.5 5.36E-01 6.84E-02 9.74E-02 1.17E-01 302.5 5.47E-01 8.81E-02 1.21E-01 1.45E-01 303.5 6.85E-01 1.41E-01 1.87E-01 2.24E-01 304.5 6.33E-01 1.55E-01 2.00E-01 2.40E-01 305.5 6.11E-01 1.85E-01 2.31E-01 2.77E-01 306.5 6.32E-01 2.11E-01 2.60E-01 3.12E-01 307.5 7.00E-01 2.83E-01 3.39E-01 4.07E-01 308.5 6.77E-01 2.97E-01 3.52E-01 4.22E-01 309.5 5.27E-01 2.61E-01 3.04E-01 3.65E-01 310.5 6.94E-01 3.76E-01 4.32E-01 5.18E-01 311.5 7.59E-01 4.33E-01 4.94E-01 5.93E-01 312.5 7.21E-01 4.57E-01 5.14E-01 6.17E-01 313.5 7.62E-01 4.94E-01 5.53E-01 6.64E-01 314.5 6.99E-01 5.00E-01 5.53E-01 6.64E-01 315.5 6.66E-01 4.79E-01 5.29E-01 6.35E-01 316.5 6.69E-01 5.18E-01 5.67E-01 6.80E-01 317.5 8.66E-01 6.79E-01 7.41E-01 8.89E-01 318.5 7.27E-01 6.10E-01 6.60E-01 7.92E-01 319.5 7.30E-01 6.06E-01 6.56E-01 7.87E-01 320.5 8.94E-01 7.74E-01 8.33E-01 1.00E+00 321.5 7.17E-01 6.45E-01 6.91E-01 8.29E-01 322.5 7.12E-01 6.28E-01 6.73E-01 8.08E-01 323.5 7.24E-01 6.81E-01 7.24E-01 8.69E-01 324.5 8.47E-01 7.94E-01 8.44E-01 1.01E+00 325.5 9.61E-01 8.93E-01 9.49E-01 1.14E+00 326.5 1.05E+00 1.02E+00 1.08E+00 1.30E+00 327.5 1.00E+00 9.61E-01 1.02E+00 1.22E+00 328.5 9.61E-01 9.28E-01 9.80E-01 1.18E+00 329.5 1.16E+00 1.16E+00 1.22E+00 1.46E+00 330.5 1.07E+00 1.04E+00 1.10E+00 1.32E+00 331.5 1.04E+00 1.02E+00 1.07E+00 1.28E+00 332.5 9.79E-01 9.85E-01 1.03E+00 1.24E+00 333.5 9.76E-01 9.70E-01 1.02E+00 1.22E+00 334.5 9.98E-01 1.00E+00 1.05E+00 1.26E+00 335.5 1.04E+00 1.05E+00 1.10E+00 1.32E+00 336.5 8.19E-01 8.25E-01 8.63E-01 1.04E+00 337.5 9.22E-01 9.25E-01 9.67E-01 1.16E+00 38

338.5 9.83E-01 9.97E-01 1.04E+00 1.25E+00 339.5 9.99E-01 1.01E+00 1.06E+00 1.27E+00 340.5 1.05E+00 1.07E+00 1.11E+00 1.33E+00 341.5 9.94E-01 1.01E+00 1.05E+00 1.26E+00 342.5 1.06E+00 1.08E+00 1.12E+00 1.34E+00 343.5 1.05E+00 1.07E+00 1.11E+00 1.33E+00 344.5 7.65E-01 7.76E-01 8.07E-01 9.68E-01 345.5 1.03E+00 1.04E+00 1.08E+00 1.30E+00 346.5 9.77E-01 9.96E-01 1.03E+00 1.24E+00 347.5 9.53E-01 9.72E-01 1.01E+00 1.21E+00 348.5 1.00E+00 1.02E+00 1.06E+00 1.27E+00 349.5 9.08E-01 9.25E-01 9.59E-01 1.15E+00 350.5 1.17E+00 1.19E+00 1.24E+00 1.49E+00 351.5 1.04E+00 1.06E+00 1.10E+00 1.32E+00 352.5 9.16E-01 9.34E-01 9.66E-01 1.16E+00 353.5 1.15E+00 1.17E+00 1.21E+00 1.45E+00 354.5 1.18E+00 1.20E+00 1.24E+00 1.49E+00 355.5 1.09E+00 1.11E+00 1.15E+00 1.38E+00 356.5 9.62E-01 9.81E-01 1.01E+00 1.21E+00 357.5 9.09E-01 9.27E-01 9.57E-01 1.15E+00 358.5 6.42E-01 6.55E-01 6.75E-01 8.10E-01 359.5 1.17E+00 1.19E+00 1.23E+00 1.48E+00 360.5 9.89E-01 1.01E+00 1.04E+00 1.25E+00 361.5 9.27E-01 9.45E-01 9.74E-01 1.17E+00 362.5 1.20E+00 1.23E+00 1.26E+00 1.51E+00 363.5 9.90E-01 1.01E+00 1.04E+00 1.25E+00 364.5 1.05E+00 1.07E+00 1.10E+00 1.32E+00 365.5 1.32E+00 1.34E+00 1.38E+00 1.66E+00 366.5 1.29E+00 1.32E+00 1.35E+00 1.62E+00 367.5 1.27E+00 1.30E+00 1.33E+00 1.60E+00 368.5 1.13E+00 1.15E+00 1.18E+00 1.42E+00 369.5 1.37E+00 1.39E+00 1.43E+00 1.72E+00 370.5 1.12E+00 1.14E+00 1.17E+00 1.40E+00 371.5 1.35E+00 1.38E+00 1.42E+00 1.70E+00 372.5 1.11E+00 1.13E+00 1.16E+00 1.39E+00 373.5 8.69E-01 8.86E-01 9.08E-01 1.09E+00 374.5 9.25E-01 9.43E-01 9.66E-01 1.16E+00 375.5 1.19E+00 1.21E+00 1.24E+00 1.49E+00 376.5 1.15E+00 1.17E+00 1.20E+00 1.44E+00 377.5 1.35E+00 1.38E+00 1.41E+00 1.69E+00 378.5 1.40E+00 1.42E+00 1.45E+00 1.74E+00 379.5 1.05E+00 1.07E+00 1.10E+00 1.32E+00 380.5 1.35E+00 1.38E+00 1.41E+00 1.69E+00 381.5 1.15E+00 1.17E+00 1.20E+00 1.44E+00 382.5 7.86E-01 8.00E-01 8.18E-01 9.82E-01 383.5 7.28E-01 7.41E-01 7.58E-01 9.10E-01 384.5 1.11E+00 1.13E+00 1.15E+00 1.38E+00 385.5 1.03E+00 1.05E+00 1.07E+00 1.28E+00 386.5 1.15E+00 1.17E+00 1.19E+00 1.43E+00 387.5 1.03E+00 1.05E+00 1.07E+00 1.28E+00 388.5 9.89E-01 1.01E+00 1.03E+00 1.24E+00 39

389.5 1.33E+00 1.35E+00 1.38E+00 1.66E+00 390.5 1.31E+00 1.33E+00 1.36E+00 1.63E+00 391.5 1.47E+00 1.49E+00 1.52E+00 1.82E+00 392.5 1.03E+00 1.04E+00 1.06E+00 1.27E+00 393.5 5.24E-01 5.33E-01 5.43E-01 6.52E-01 394.5 1.19E+00 1.21E+00 1.23E+00 1.48E+00 395.5 1.47E+00 1.49E+00 1.52E+00 1.82E+00 396.5 6.92E-01 7.03E-01 7.17E-01 8.60E-01 397.5 1.10E+00 1.11E+00 1.13E+00 1.36E+00 398.5 1.63E+00 1.65E+00 1.68E+00 2.02E+00 399.5 1.74E+00 1.77E+00 1.80E+00 2.16E+00 400.5 1.76E+00 1.79E+00 1.82E+00 2.18E+00 401.5 1.88E+00 1.91E+00 1.94E+00 2.33E+00 402.5 1.92E+00 1.95E+00 1.98E+00 2.38E+00 403.5 1.77E+00 1.79E+00 1.83E+00 2.20E+00 404.5 1.72E+00 1.74E+00 1.77E+00 2.12E+00 405.5 1.79E+00 1.82E+00 1.85E+00 2.22E+00 406.5 1.74E+00 1.77E+00 1.80E+00 2.16E+00 407.5 1.66E+00 1.69E+00 1.72E+00 2.06E+00 408.5 1.94E+00 1.97E+00 2.00E+00 2.40E+00 409.5 1.82E+00 1.84E+00 1.87E+00 2.24E+00 410.5 1.61E+00 1.64E+00 1.66E+00 1.99E+00 411.5 1.94E+00 1.97E+00 2.00E+00 2.40E+00 412.5 1.93E+00 1.95E+00 1.98E+00 2.38E+00 413.5 1.89E+00 1.92E+00 1.95E+00 2.34E+00 414.5 1.87E+00 1.89E+00 1.92E+00 2.30E+00 415.5 1.86E+00 1.89E+00 1.92E+00 2.30E+00 416.5 1.99E+00 2.02E+00 2.05E+00 2.46E+00 417.5 1.79E+00 1.81E+00 1.84E+00 2.21E+00 418.5 1.81E+00 1.83E+00 1.86E+00 2.23E+00 419.5 1.83E+00 1.85E+00 1.88E+00 2.26E+00 420.5 1.89E+00 1.91E+00 1.94E+00 2.33E+00 421.5 1.94E+00 1.96E+00 1.99E+00 2.39E+00 422.5 1.72E+00 1.75E+00 1.77E+00 2.12E+00 423.5 1.84E+00 1.87E+00 1.89E+00 2.27E+00 424.5 1.89E+00 1.91E+00 1.94E+00 2.33E+00 425.5 1.83E+00 1.85E+00 1.88E+00 2.26E+00 426.5 1.85E+00 1.87E+00 1.90E+00 2.28E+00 427.5 1.69E+00 1.71E+00 1.73E+00 2.08E+00 428.5 1.71E+00 1.73E+00 1.76E+00 2.11E+00 429.5 1.58E+00 1.60E+00 1.62E+00 1.94E+00 430.5 1.22E+00 1.23E+00 1.25E+00 1.50E+00 431.5 1.81E+00 1.83E+00 1.85E+00 2.22E+00 432.5 1.76E+00 1.78E+00 1.81E+00 2.17E+00 433.5 1.85E+00 1.87E+00 1.89E+00 2.27E+00 434.5 1.78E+00 1.80E+00 1.83E+00 2.20E+00 435.5 1.84E+00 1.86E+00 1.88E+00 2.26E+00 436.5 2.06E+00 2.08E+00 2.11E+00 2.53E+00 437.5 1.92E+00 1.94E+00 1.96E+00 2.35E+00 438.5 1.69E+00 1.71E+00 1.73E+00 2.08E+00 439.5 1.95E+00 1.97E+00 1.99E+00 2.39E+00 40

440.5 1.82E+00 1.84E+00 1.86E+00 2.23E+00 441.5 2.06E+00 2.09E+00 2.11E+00 2.53E+00 442.5 2.12E+00 2.14E+00 2.17E+00 2.60E+00 443.5 2.06E+00 2.08E+00 2.10E+00 2.52E+00 444.5 2.11E+00 2.13E+00 2.15E+00 2.58E+00 445.5 1.94E+00 1.96E+00 1.98E+00 2.38E+00 446.5 2.00E+00 2.02E+00 2.04E+00 2.45E+00 447.5 2.20E+00 2.23E+00 2.25E+00 2.70E+00 448.5 2.10E+00 2.12E+00 2.14E+00 2.57E+00 449.5 2.17E+00 2.19E+00 2.21E+00 2.65E+00 41

GLOBAL ATMOSPHERE WATCH REPORT SERIES 1. Fnal Report of the Expert Meetng on the Operaton of Integrated Montorng Programmes, Geneva, 2-5 September 1980. 2. Report of the Thrd Sesson of the GESAMP Workng Group on the Interchange of Pollutants Between the Atmosphere and the Oceans (INTERPOLL-III), Mam, USA, 27-31 October 1980. 3. Report of the Expert Meetng on the Assessment of the Meteorologcal Aspects of the Frst Phase of EMEP, Shnfeld Park, U.K., 30 March - 2 Aprl 1981. 4. Summary Report on the Status of the WMO Background Ar Polluton Montorng Network as at Aprl 1981. 5. Report of the WMO/UNEP/ICSU Meetng on Instruments, Standardzaton and Measurements Technques for Atmospherc CO2, Geneva, 8-11; September 1981. 6. Report of the Meetng of Experts on BAPMoN Staton Operaton, Geneva, 23 26 November 1981. 7. Fourth Analyss on Reference Precptaton Samples by the Partcpatng World Meteorologcal Organzaton Laboratores by Robert L. Lampe and John C. Puzak, December 1981. 8. Revew of the Chemcal Composton of Precptaton as Measured by the WMO BAPMoN by Prof. Dr. Hans-Walter Georg, February 1982. 9. An Assessment of BAPMoN Data Currently Avalable on the Concentraton of CO2 n the Atmosphere by M.R. Mannng, February 1982. 10. Report of the Meetng of Experts on Meteorologcal Aspects of Long-range Transport of Pollutants, Toronto, Canada, 30 November - 4 December 1981. 11. Summary Report on the Status of the WMO Background Ar Polluton Montorng Network as at May 1982. 12. Report on the Mount Kenya Baselne Staton Feasblty Study edted by Dr. Russell C. Schnell. 13. Report of the Executve Commttee Panel of Experts on Envronmental Polluton, Fourth Sesson, Geneva, 27 September - 1 October 1982. 14. Effects of Sulphur Compounds and Other Pollutants on Vsblty by Dr. R.F. Pueschel, Aprl 1983. 15. Provsonal Daly Atmospherc Carbon Doxde Concentratons as Measured at BAPMoN Stes for the Year 1981, May 1983. 16. Report of the Expert Meetng on Qualty Assurance n BAPMoN, Research Trangle Park, North Carolna, USA, 17-21 January 1983. 17. General Consderaton and Examples of Data Evaluaton and Qualty Assurance Procedures Applcable to BAPMoN Precptaton Chemstry Observatons by Dr. Charles Hakkarnen, July 1983. 18. Summary Report on the Status of the WMO Background Ar Polluton Montorng Network as at May 1983. 19. Forecastng of Ar Polluton wth Emphass on Research n the USSR by M.E. Berlyand, August 1983. 20. Extended Abstracts of Papers to be Presented at the WMO Techncal Conference on Observaton and Measurement of Atmospherc Contamnants (TECOMAC), Venna, 17-21 October 1983. 21. Ffth Analyss on Reference Precptaton Samples by the Partcpatng World Meteorologcal Organzaton Laboratores by Robert L. Lampe and Wllam J. Mtchell, November 1983. 22. Report of the Ffth Sesson of the WMO Executve Councl Panel of Experts on Envronmental Polluton, Garmsch- Partenkrchen, Federal Republc of Germany, 30 Aprl - 4 May 1984 (WMO TD No. 10). 23. Provsonal Daly Atmospherc Carbon Doxde Concentratons as Measured at BAPMoN Stes for the Year 1982. November 1984 (WMO TD No. 12). 43

24. Fnal Report of the Expert Meetng on the Assessment of the Meteorologcal Aspects of the Second Phase of EMEP, Fredrchshafen, Federal Republc of Germany, 7-10 December 1983. October 1984 (WMO TD No. 11). 25. Summary Report on the Status of the WMO Background Ar Polluton Montorng Network as at May 1984. November 1984 (WMO TD No. 13). 26. Sulphur and Ntrogen n Precptaton: An Attempt to Use BAPMoN and Other Data to Show Regonal and Global Dstrbuton by Dr. C.C. Wallén. Aprl 1986 (WMO TD No. 103). 27. Report on a Study of the Transport of Sahelan Partculate Matter Usng Sunphotometer Observatons by Dr. Gullaume A. d'almeda. July 1985 (WMO TD No. 45). 28. Report of the Meetng of Experts on the Eastern Atlantc and Medterranean Transport Experment ("EAMTEX"), Madrd and Salamanca, Span, 6-8 November 1984. 29. Recommendatons on Sunphotometer Measurements n BAPMoN Based on the Experence of a Dust Transport Study n Afrca by Dr. Gullaume A. d'almeda. September 1985 (WMO TD No. 67). 30. Report of the Ad-hoc Consultaton on Qualty Assurance Procedures for Incluson n the BAPMoN Manual, Geneva, 29-31 May 1985. 31. Implcatons of Vsblty Reducton by Man-Made Aerosols (Annex to No. 14) by R.M. Hoff and L.A. Barre. October 1985 (WMO TD No. 59). 32. Manual for BAPMoN Staton Operators by E. Meszaros and D.M. Whelpdale. October 1985 (WMO TD No. 66). 33. Man and the Composton of the Atmosphere: BAPMoN - An nternatonal programme of natonal needs, responsblty and benefts by R.F. Pueschel, 1986. 34. Practcal Gude for Estmatng Atmospherc Polluton Potental by Dr. L.E. Nemeyer. August 1986 (WMO TD No. 134). 35. Provsonal Daly Atmospherc CO2 Concentratons as Measured at BAPMoN Stes for the Year 1983. December 1985 (WMO TD No. 77). 36. Global Atmospherc Background Montorng for Selected Envronmental Parameters. BAPMoN Data for 1984. Volume I: Atmospherc Aerosol Optcal Depth. October 1985 (WMO TD No. 96). 37. Ar-Sea Interchange of Pollutants by R.A. Duce. September 1986 (WMO TD No. 126). 38. Summary Report on the Status of the WMO Background Ar Polluton Montorng Network as at 31 December 1985. September 1986 (WMO TD No. 136). 39. Report of the Thrd WMO Expert Meetng on Atmospherc Carbon Doxde Measurement Technques, Lake Arrowhead, Calforna, USA, 4-8 November 1985. October 1986. 40. Report of the Fourth Sesson of the CAS Workng Group on Atmospherc Chemstry and Ar Polluton, Helsnk, Fnland, 18-22 November 1985. January 1987. 41. Global Atmospherc Background Montorng for Selected Envronmental Parameters. BAPMoN Data for 1982, Volume II: Precptaton chemstry, contnuous atmospherc carbon doxde and suspended partculate matter. June 1986 (WMO TD No. 116). 42. Scrpps reference gas calbraton system for carbon doxde-n-ar standards: revson of 1985 by C.D. Keelng, P.R. Guenther and D.J. Moss. September 1986 (WMO TD No. 125). 43. Recent progress n sunphotometry (determnaton of the aerosol optcal depth). November 1986. 44. Report of the Sxth Sesson of the WMO Executve Councl Panel of Experts on Envronmental Polluton, Geneva, 5-9 May 1986. March 1987. 45. Proceedngs of the Internatonal Symposum on Integrated Global Montorng of the State of the Bosphere (Volumes I-IV), Tashkent, USSR, 14-19 October 1985. December 1986 (WMO TD No. 151). 44

46. Provsonal Daly Atmospherc Carbon Doxde Concentratons as Measured at BAPMoN Stes for the Year 1984. December 1986 (WMO TD No. 158). 47. Procedures and Methods for Integrated Global Background Montorng of Envronmental Polluton by F.Ya. Rovnsky, USSR and G.B. Wersma, USA. August 1987 (WMO TD No. 178). 48. Meetng on the Assessment of the Meteorologcal Aspects of the Thrd Phase of EMEP IIASA, Laxenburg, Austra, 30 March - 2 Aprl 1987. February 1988. 49. Proceedngs of the WMO Conference on Ar Polluton Modellng and ts Applcaton (Volumes I-III), Lenngrad, USSR, 19-24 May 1986. November 1987 (WMO TD No. 187). 50. Provsonal Daly Atmospherc Carbon Doxde Concentratons as Measured at BAPMoN Stes for the Year 1985. December 1987 (WMO TD No. 198). 51. Report of the NBS/WMO Expert Meetng on Atmospherc CO2 Measurement Technques, Gathersburg, USA, 15-17 June 1987. December 1987. 52. Global Atmospherc Background Montorng for Selected Envronmental Parameters. BAPMoN Data for 1985. Volume I: Atmospherc Aerosol Optcal Depth. September 1987. 53. WMO Meetng of Experts on Strategy for the Montorng of Suspended Partculate Matter n BAPMoN - Reports and papers presented at the meetng, Xamen, Chna, 13-17 October 1986. October 1988. 54. Global Atmospherc Background Montorng for Selected Envronmental Parameters. BAPMoN Data for 1983, Volume II: Precptaton chemstry, contnuous atmospherc carbon doxde and suspended partculate matter (WMO TD No. 283). 55. Summary Report on the Status of the WMO Background Ar Polluton Montorng Network as at 31 December 1987 (WMO TD No. 284). 56. Report of the Frst Sesson of the Executve Councl Panel of Experts/CAS Workng Group on Envronmental Polluton and Atmospherc Chemstry, Hlo, Hawa, 27-31 March 1988. June 1988. 57. Global Atmospherc Background Montorng for Selected Envronmental Parameters. BAPMoN Data for 1986, Volume I: Atmospherc Aerosol Optcal Depth. July 1988. 58. Provsonal Daly Atmospherc Carbon Doxde Concentratons as measured at BAPMoN stes for the years 1986 and 1987 (WMO TD No. 306). 59. Extended Abstracts of Papers Presented at the Thrd Internatonal Conference on Analyss and Evaluaton of Atmospherc CO2 Data - Present and Past, Hnterzarten, Federal Republc of Germany, 16-20 October 1989 (WMO TD No. 340). 60. Global Atmospherc Background Montorng for Selected Envronmental Parameters. BAPMoN Data for 1984 and 1985, Volume II: Precptaton chemstry, contnuous atmospherc carbon doxde and suspended partculate matter. 61. Global Atmospherc Background Montorng for Selected Envronmental Parameters. BAPMoN Data for 1987 and 1988, Volume I: Atmospherc Aerosol Optcal Depth. 62. Provsonal Daly Atmospherc Carbon Doxde Concentratons as measured at BAPMoN stes for the year 1988 (WMO TD No. 355). 63. Report of the Informal Sesson of the Executve Councl Panel of Experts/CAS Workng Group on Envronmental Polluton and Atmospherc Chemstry, Sofa, Bulgara, 26 and 28 October 1989. 64. Report of the consultaton to consder desrable locatons and observatonal practces for BAPMoN statons of global mportance, Bermuda Research Staton, 27-30 November 1989. 65. Report of the Meetng on the Assessment of the Meteorologcal Aspects of the Fourth Phase of EMEP, Sofa, Bulgara, 27 and 31 October 1989. 66. Summary Report on the Status of the WMO Global Atmosphere Watch Statons as at 31 December 1990 (WMO TD No. 419). 45

67. Report of the Meetng of Experts on Modellng of Contnental, Hemspherc and Global Range Transport, Transformaton and Exchange Processes, Geneva, 5-7 November 1990. 68. Global Atmospherc Background Montorng for Selected Envronmental Parameters. BAPMoN Data For 1989, Volume I: Atmospherc Aerosol Optcal Depth. 69. Provsonal Daly Atmospherc Carbon Doxde Concentratons as measured at Global Atmosphere Watch (GAW)-BAPMoN stes for the year 1989 (WMO TD No. 400). 70. Report of the Second Sesson of EC Panel of Experts/CAS Workng Group on Envronmental Polluton and Atmospherc Chemstry, Santago, Chle, 9-15 January 1991 (WMO TD No. 633). 71. Report of the Consultaton of Experts to Consder Desrable Observatonal Practces and Dstrbuton of GAW Regonal Statons, Halkdk, Greece, 9-13 Aprl 1991 (WMO TD No. 433). 72. Integrated Background Montorng of Envronmental Polluton n Md-Lattude Eurasa by Yu.A. Izrael and F.Ya. Rovnsky, USSR (WMO TD No. 434). 73. Report of the Experts Meetng on Global Aerosol Data System (GADS), Hampton, Vrgna, 11 to 12 September 1990 (WMO TD No. 438). 74. Report of the Experts Meetng on Aerosol Physcs and Chemstry, Hampton, Vrgna, 30 to 31 May 1991 (WMO TD No. 439). 75. Provsonal Daly Atmospherc Carbon Doxde Concentratons as measured at Global Atmosphere Watch (GAW)-BAPMoN stes for the year 1990 (WMO TD No. 447). 76. The Internatonal Global Aerosol Programme (IGAP) Plan: Overvew (WMO TD No. 445). 77. Report of the WMO Meetng of Experts on Carbon Doxde Concentraton and Isotopc Measurement Technques, Lake Arrowhead, Calforna, 14-19 October 1990. 78. Global Atmospherc Background Montorng for Selected Envronmental Parameters BAPMoN Data for 1990, Volume I: Atmospherc Aerosol Optcal Depth (WMO TD No. 446). 79. Report of the Meetng of Experts to Consder the Aerosol Component of GAW, Boulder, 16 to 19 December 1991 (WMO TD No. 485). 80. Report of the WMO Meetng of Experts on the Qualty Assurance Plan for the GAW, Garmsch-Partenkrchen, Germany, 26-30 March 1992 (WMO TD No. 513). 81. Report of the Second Meetng of Experts to Assess the Response to and Atmospherc Effects of the Kuwat Ol Fres, Geneva, Swtzerland, 25-29 May 1992 (WMO TD No. 512). 82. Global Atmospherc Background Montorng for Selected Envronmental Parameters BAPMoN Data for 1991, Volume I: Atmospherc Aerosol Optcal Depth (WMO TD No. 518). 83. Report on the Global Precptaton Chemstry Programme of BAPMoN (WMO TD No. 526). 84. Provsonal Daly Atmospherc Carbon Doxde Concentratons as measured at GAW-BAPMoN stes for the year 1991 (WMO TD No. 543). 85. Chemcal Analyss of Precptaton for GAW: Laboratory Analytcal Methods and Sample Collecton Standards by Dr Jaroslav Santroch (WMO TD No. 550). 86. The Global Atmosphere Watch Gude, 1993 (WMO TD No. 553). 87. Report of the Thrd Sesson of EC Panel/CAS Workng Group on Envronmental Polluton and Atmospherc Chemstry, Geneva, 8-11 March 1993 (WMO TD No. 555). 88. Report of the Seventh WMO Meetng of Experts on Carbon Doxde Concentraton and Isotopc Measurement Technques, Rome, Italy, 7-10 September 1993, (edted by Graeme I. Pearman and James T. Peterson) (WMO TD No. 669). 89. 4th Internatonal Conference on CO2 (Carqueranne, France, 13-17 September 1993) (WMO TD No. 561). 46

90. Global Atmospherc Background Montorng for Selected Envronmental Parameters GAW Data for 1992, Volume I: Atmospherc Aerosol Optcal Depth (WMO TD No. 562). 91. Extended Abstracts of Papers Presented at the WMO Regon VI Conference on the Measurement and Modellng of Atmospherc Composton Changes Includng Polluton Transport, Sofa, 4 to 8 October 1993 (WMO TD No. 563). 92. Report of the Second WMO Meetng of Experts on the Qualty Assurance/Scence Actvty Centres of the Global Atmosphere Watch, Garmsch-Partenkrchen, 7-11 December 1992 (WMO TD No. 580). 93. Report of the Thrd WMO Meetng of Experts on the Qualty Assurance/Scence Actvty Centres of the Global Atmosphere Watch, Garmsch-Partenkrchen, 5-9 July 1993 (WMO TD No. 581). 94. Report on the Measurements of Atmospherc Turbdty n BAPMoN (WMO TD No. 603). 95. Report of the WMO Meetng of Experts on UV-B Measurements, Data Qualty and Standardzaton of UV Indces, Les Dablerets, Swtzerland, 25-28 July 1994 (WMO TD No. 625). 96. Global Atmospherc Background Montorng for Selected Envronmental Parameters WMO GAW Data for 1993, Volume I: Atmospherc Aerosol Optcal Depth. 97. Qualty Assurance Project Plan (QAPjP) for Contnuous Ground Based Ozone Measurements (WMO TD No. 634). 98. Report of the WMO Meetng of Experts on Global Carbon Monoxde Measurements, Boulder, USA, 7-11 February 1994 (WMO TD No. 645). 99. Status of the WMO Global Atmosphere Watch Programme as at 31 December 1993 (WMO TD No. 636). 100. Report of the Workshop on UV-B for the Amercas, Buenos Ares, Argentna, 22-26 August 1994. 101. Report of the WMO Workshop on the Measurement of Atmospherc Optcal Depth and Turbdty, Slver Sprng, USA, 6-10 December 1993, (edted by Bruce Hcks) (WMO TD No. 659). 102. Report of the Workshop on Precptaton Chemstry Laboratory Technques, Hradec Kralove, Czech Republc, 17-21 October 1994 (WMO TD No. 658). 103. Report of the Meetng of Experts on the WMO World Data Centres, Toronto, Canada, 17-18 February 1995, (prepared by Edward Hare) (WMO TD No. 679). 104. Report of the Fourth WMO Meetng of Experts on the Qualty Assurance/Scence Actvty Centres (QA/SACs) of the Global Atmosphere Watch, jontly held wth the Frst Meetng of the Coordnatng Commttees of IGAC-GLONET and IGAC-ACE, Garmsch-Partenkrchen, Germany, 13 to 17 March 1995 (WMO TD No. 689). 105. Report of the Fourth Sesson of the EC Panel of Experts/CAS Workng Group on Envronmental Polluton and Atmospherc Chemstry (Garmsch, Germany, 6-11 March 1995) (WMO TD No. 718). 106. Report of the Global Acd Deposton Assessment (edted by D.M. Whelpdale and M-S. Kaser) (WMO TD No. 777). 107. Extended Abstracts of Papers Presented at the WMO-IGAC Conference on the Measurement and Assessment of Atmospherc Composton Change (Bejng, Chna, 9-14 October 1995) (WMO TD No. 710). 108. Report of the Tenth WMO Internatonal Comparson of Dobson Spectrophotometers (Arosa, Swtzerland, 24 July - 4 August 1995). 109. Report of an Expert Consultaton on 85Kr and 222Rn: Measurements, Effects and Applcatons (Freburg, Germany, 28-31 March 1995) (WMO TD No. 733). 110. Report of the WMO-NOAA Expert Meetng on GAW Data Acquston and Archvng (Ashevlle, NC, USA, 4-8 November 1995) (WMO TD No. 755). 111. Report of the WMO-BMBF Workshop on VOC Establshment of a World Calbraton/Instrument Intercomparson Faclty for VOC to Serve the WMO Global Atmosphere Watch (GAW) Programme (Garmsch-Partenkrchen, Germany, 17-21 December 1995) (WMO TD No. 756). 47

112. Report of the WMO/STUK Intercomparson of Erythemally-Weghted Solar UV Radometers, Sprng/Summer 1995, Helsnk, Fnland (WMO TD No. 781). 112A. Report of the WMO/STUK 95 Intercomparson of broadband UV radometers: a small-scale follow-up study n 1999, Helsnk, 2001, Addendum to GAW Report No. 112. 113. The Strategc Plan of the Global Atmosphere Watch (GAW) (WMO TD No. 802). 114. Report of the Ffth WMO Meetng of Experts on the Qualty Assurance/Scence Actvty Centres (QA/SACs) of the Global Atmosphere Watch, jontly held wth the Second Meetng of the Coordnatng Commttees of IGAC-GLONET and IGAC- ACE Ed, Garmsch-Partenkrchen, Germany, 15-19 July 1996 (WMO TD No. 787). 115. Report of the Meetng of Experts on Atmospherc Urban Polluton and the Role of NMSs (Geneva, 7-11 October 1996) (WMO TD No. 801). 116. Expert Meetng on Chemstry of Aerosols, Clouds and Atmospherc Precptaton n the Former USSR (Sant Petersburg, Russan Federaton, 13-15 November 1995). 117. Report and Proceedngs of the Workshop on the Assessment of EMEP Actvtes Concernng Heavy Metals and Persstent Organc Pollutants and ther Further Development (Moscow, Russan Federaton, 24-26 September 1996) (Volumes I and II) (WMO TD No. 806). 118. Report of the Internatonal Workshops on Ozone Observaton n Asa and the Pacfc Regon (IWOAP, IWOAP-II), (IWOAP, 27 February-26 March 1996 and IWOAP-II, 20 August-18 September 1996) (WMO TD No. 827). 119. Report on BoM/NOAA/WMO Internatonal Comparson of the Dobson Spectrophotometers (Perth Arport, Perth, Australa, 3-14 February 1997), (prepared by Robert Evans and James Easson) (WMO TD No. 828). 120. WMO-UMAP Workshop on Broad-Band UV Radometers (Garmsch-Partenkrchen, Germany, 22 to 23 Aprl 1996) (WMO TD No. 894). 121. Report of the Eghth WMO Meetng of Experts on Carbon Doxde Concentraton and Isotopc Measurement Technques (prepared by Thomas Conway) (Boulder, CO, 6-11 July 1995) (WMO TD No. 821). 122. Report of Passve Samplers for Atmospherc Chemstry Measurements and ther Role n GAW (prepared by Greg Carmchael) (WMO TD No. 829). 123. Report of WMO Meetng of Experts on GAW Regonal Network n RA VI, Budapest, Hungary, 5 to 9 May 1997. 124. Ffth Sesson of the EC Panel of Experts/CAS Workng Group on Envronmental Polluton and Atmospherc Chemstry, (Geneva, Swtzerland, 7-10 Aprl 1997) (WMO TD No. 898). 125. Instruments to Measure Solar Ultravolet Radaton, Part 1: Spectral Instruments (lead author G. Seckmeyer) (WMO TD No. 1066), 2001. 126. Gudelnes for Ste Qualty Control of UV Montorng (lead author A.R. Webb) (WMO TD No. 884), 1998. 127. Report of the WMO-WHO Meetng of Experts on Standardzaton of UV Indces and ther Dssemnaton to the Publc (Les Dablerets, Swtzerland, 21-25 July 1997) (WMO TD No. 921). 128. The Fourth Bennal WMO Consultaton on Brewer Ozone and UV Spectrophotometer Operaton, Calbraton and Data Reportng, (Rome, Italy, 22-25 September 1996) (WMO TD No. 918). 129. Gudelnes for Atmospherc Trace Gas Data Management (Ken Masare and Peter Tans), 1998 (WMO TD No. 907). 130. Jülch Ozone Sonde Intercomparson Experment (JOSIE, 5 February to 8 March 1996), (H.G.J. Smt and D. Kley) (WMO TD No. 926). 131. WMO Workshop on Regonal Transboundary Smoke and Haze n Southeast Asa (Sngapore, 2 to 5 June 1998) (Gregory R. Carmchael). Two volumes. 132. Report of the Nnth WMO Meetng of Experts on Carbon Doxde Concentraton and Related Tracer Measurement Technques (Edted by Roger Francey), (Aspendale, Vc., Australa). 48

133. Workshop on Advanced Statstcal Methods and ther Applcaton to Ar Qualty Data Sets (Helsnk, 14-18 September 1998) (WMO TD No. 956). 134. Gude on Samplng and Analyss Technques for Chemcal Consttuents and Physcal Propertes n Ar and Precptaton as Appled at Statons of the Global Atmosphere Watch. Carbon Doxde (WMO TD No. 980). 135. Sxth Sesson of the EC Panel of Experts/CAS Workng Group on Envronmental Polluton and Atmospherc Chemstry (Zurch, Swtzerland, 8-11 March 1999) (WMO TD No.1002). 136. WMO/EMEP/UNEP Workshop on Modellng of Atmospherc Transport and Deposton of Persstent Organc Pollutants and Heavy Metals (Geneva, Swtzerland, 16-19 November 1999) (Volumes I and II) (WMO TD No. 1008). 137. Report and Proceedngs of the WMO RA II/RA V GAW Workshop on Urban Envronment (Bejng, Chna, 1-4 November 1999) (WMO-TD. 1014) (Prepared by Greg Carmchael). 138. Reports on WMO Internatonal Comparsons of Dobson Spectrophotometers, Parts I Arosa, Swtzerland, 19-31 July 1999, Part II Buenos Ares, Argentna (29 Nov. 12 Dec. 1999 and Part III Pretora, South Afrca (18 March 10 Aprl 2000) (WMO TD No. 1016). 139. The Ffth Bennal WMO Consultaton on Brewer Ozone and UV Spectrophotometer Operaton, Calbraton and Data Reportng (Halkdk, Greece, September 1998)(WMO TD No. 1019). 140. WMO/CEOS Report on a Strategy for Integratng Satellte and Ground-based Observatons of Ozone (WMO TD No. 1046). 141. Report of the LAP/COST/WMO Intercomparson of Erythemal Radometers Thessalonk, Greece, 13-23 September 1999) (WMO TD No. 1051). 142. Strategy for the Implementaton of the Global Atmosphere Watch Programme (2001-2007), A Contrbuton to the Implementaton of the Long-Term Plan (WMO TD No.1077). 143. Global Atmosphere Watch Measurements Gude (WMO TD No. 1073). 144. Report of the Seventh Sesson of the EC Panel of Experts/CAS Workng Group on Envronmental Polluton and Atmospherc Chemstry and the GAW 2001 Workshop (Geneva, Swtzerland, 2 to 5 Aprl 2001) (WMO TD No. 1104). 145. WMO GAW Internatonal Comparsons of Dobson Spectrophotometers at the Meteorologcal Observatory Hohenpessenberg, Germany (21 May 10 June 2000, MOHp2000-1), 23 July 5 August 2000, MOHp2000-2), (10 23 June 2001, MOHp2001-1) and (8 to 21 July 2001, MOHp2001-2). Prepared by Ulf Köhler (WMO TD No. 1114). 146. Qualty Assurance n montorng solar ultravolet radaton: the state of the art. (WMO TD No. 1180), 2003. 147. Workshop on GAW n RA VI (Europe), Rga, Latva, 27-30 May 2002. (WMO TD No. 1206). 148. Report of the Eleventh WMO/IAEA Meetng of Experts on Carbon Doxde Concentraton and Related Tracer Measurement Technques (Tokyo, Japan, 25-28 September 2001) (WMO TD No 1138). 149. Comparson of Total Ozone Measurements of Dobson and Brewer Spectrophotometers and Recommended Transfer Functons (prepared by J. Staeheln, J. Kerr, R. Evans and K. Vancek) (WMO TD No. 1147). 150. Updated Gudelnes for Atmospherc Trace Gas Data Management (Prepared by Ken Masere and Peter Tans (WMO TD No. 1149). 151. Report of the Frst CAS Workng Group on Envronmental Polluton and Atmospherc Chemstry (Geneva, Swtzerland, 18-19 March 2003) (WMO TD No. 1181). 152. Current Actvtes of the Global Atmosphere Watch Programme (as presented at the 14 th World Meteorologcal Congress, May 2003). (WMO TD No. 1168). 153. WMO/GAW Aerosol Measurement Procedures: Gudelnes and Recommendatons. (WMO TD No. 1178). 154. WMO/IMEP-15 Trace Elements n Water Laboratory Intercomparson. (WMO TD No. 1195). 49

155. 1 st Internatonal Expert Meetng on Sources and Measurements of Natural Radonucldes Appled to Clmate and Ar Qualty Studes (Gf sur Yvette, France, 3-5 June 2003) (WMO TD No. 1201). 156. Addendum for the Perod 2005-2007 to the Strategy for the Implementaton of the Global Atmosphere Watch Programme (2001-2007), GAW Report No. 142 (WMO TD No. 1209). 157. JOSIE-1998 Performance of EEC Ozone Sondes of SPC-6A and ENSCI-Z Type (Prepared by Herman G.J. Smt and Wolfgang Straeter) (WMO TD No. 1218). 158. JOSIE-2000 Jülch Ozone Sonde Intercomparson Experment 2000. The 2000 WMO nternatonal ntercomparson of operatng procedures for ECC-ozone sondes at the envronmental smulaton faclty at Jülch (Prepared by Herman G.J. Smt and Wolfgang Straeter) (WMO TD No. 1225). 159. IGOS-IGACO Report - September 2004 (WMO TD No. 1235), 68 pp, September 2004. 160. Manual for the GAW Precptaton Chemstry Programme (Gudelnes, Data Qualty Objectves and Standard Operatng Procedures) (WMO TD No. 1251), 186 pp, November 2004. 161 12 th WMO/IAEA Meetng of Experts on Carbon Doxde Concentraton and Related Tracers Measurement Technques (Toronto, Canada, 15-18 September 2003), 274 pp, May 2005. 162. WMO/GAW Experts Workshop on a Global Surface-Based Network for Long Term Observatons of Column Aerosol Optcal Propertes, Davos, Swtzerland, 8-10 March 2004 (edted by U. Baltensperger, L. Barre and C. Wehrl) (WMO TD No. 1287), 153 pp, November 2005. 163. World Meteorologcal Organzaton Actvtes n Support of the Venna Conventon on Protecton of the Ozone Layer (WMO No. 974), 4 pp, September 2005. 164. Instruments to Measure Solar Ultravolet Radaton: Part 2: Broadband Instruments Measurng Erythemally Weghted Solar Irradance (WMO TD No. 1289), 55 pp, July 2008, electronc verson 2006. 165. Report of the CAS Workng Group on Envronmental Polluton and Atmospherc Chemstry and the GAW 2005 Workshop, 14-18 March 2005, Geneva, Swtzerland (WMO TD No. 1302), 189 pp, March 2005. 166. Jont WMO-GAW/ACCENT Workshop on The Global Tropospherc Carbon Monoxde Observatons System, Qualty Assurance and Applcatons (EMPA, Dübendorf, Swtzerland, 24 26 October 2005) (edted by J. Klausen) (WMO TD No. 1335), 36 pp, September 2006. 167. The German Contrbuton to the WMO Global Atmosphere Watch Programme upon the 225 th Annversary of GAW Hohenpessenberg Observatory (edted by L.A. Barre, W. Frcke and R. Schleyer (WMO TD No. 1336), 124 pp, December 2006. 168. 13 th WMO/IAEA Meetng of Experts on Carbon Doxde Concentraton and Related Tracers Measurement Technques (Boulder, Colorado, USA, 19-22 September 2005) (edted by J.B. Mller) (WMO TD No. 1359), 40 pp, December 2006. 169. Chemcal Data Assmlaton for the Observaton of the Earth s Atmosphere ACCENT/WMO Expert Workshop n support of IGACO (edted by L.A. Barre, J.P. Burrows, P. Monks and P. Borrell) (WMO TD No. 1360), 196 pp, December 2006. 170. WMO/GAW Expert Workshop on the Qualty and Applcatons of European GAW Measurements (Tutzng, Germany, 2-5 November 2004) (WMO TD No. 1367). 171. A WMO/GAW Expert Workshop on Global Long-Term Measurements of Volatle Organc Compounds (VOCs) (Geneva, Swtzerland, 30 January 1 February 2006) (WMO TD No. 1373), 36 pp, February 2007. 172. WMO Global Atmosphere Watch (GAW) Strategc Plan: 2008 2015 (WMO TD No. 1384), 108 pp, August 2008. 173. Report of the CAS Jont Scentfc Steerng Commttee on Envronmental Polluton and Atmospherc Chemstry (Geneva, Swtzerland, 11-12 Aprl 2007) (WMO TD No.1410), 33 pp, June 2008. 174. World Data Centre for Greenhouse Gases Data Submsson and Dssemnaton Gude (WMO TD No. 1416), 50 pp, January 2008. 50

175. The Nnth Bennal WMO Consultaton on Brewer Ozone and UV Spectrophotometer Operaton, Calbraton and Data Reportng (Delft, Netherlands, 31-May 3 June 2005) (WMO TD No. 1419), 69 pp, March 2008. 176. The Tenth Bennal WMO Consultaton on Brewer Ozone and UV Spectrophotometer Operaton, Calbraton and Data Reportng (Northwch, Unted Kngdom, 4-8 June 2007) (WMO TD No. 1420), 61 pp, March 2008. 177. Jont Report of COST Acton 728 and GURME Overvew of Exstng Integrated (off-lne and on-lne) Mesoscale Meteorologcal and Chemcal Transport Modellng n Europe (ISBN 978-1-905313-56-3) (WMO TD No. 1427), 106 pp, May 2008. 178. Plan for the mplementaton of the GAW Aerosol Ldar Observaton Network GALION, (Hamburg, Germany, 27-29 March 2007) (WMO TD No. 1443), 52 pp, November 2008. 179. Intercomparson of Global UV Index from Multband Radometers: Harmonzaton of Global UVI and Spectral Irradance (WMO TD No. 1454), 61 pp, March 2009. 180. Towards a Better Knowledge of Umkehr Measurements: A Detaled Study of Data from Thrteen Dobson Intercomparsons (WMO TD No. 1456), 50 pp, December 2008. 181. Jont Report of COST Acton 728 and GURME Overvew of Tools and Methods for Meteorologcal and Ar Polluton Mesoscale Model Evaluaton and User Tranng (WMO TD No. 1457), 121 pp, November 2008. 182. IGACO-Ozone and UV Radaton Implementaton Plan (WMO TD No. 1465), 49 pp, Aprl 2009. 183. Operatons Handbook Ozone Observatons wth a Dobson Spectrophotometer (WMO TD No. 1469), 91 pp, March 2009. 184. Techncal Report of Global Analyss Method for Major Greenhouse Gases by the World Data Center for Greenhouse Gases (WMO TD No. 1473), 29 pp, June 2009. 185. Gudelnes for the Measurement of Methane and Ntrous Oxde and ther Qualty Assurance (WMO TD No. 1478), 49 pp, September 2009. 186. 14 th WMO/IAEA Meetng of Experts on Carbon Doxde Concentraton and Related Tracers Measurement Technques (Helsnk, Fnland, 10-13 September 2007) (WMO TD No. 1487), 31 pp, Aprl 2009. 187. Jont Report of COST Acton 728 and GURME Revew of the Capabltes of Meteorologcal and Chemstry-Transport Models for Descrbng and Predctng Ar Polluton Epsodes (ISBN 978-1-905313-77-8) (WMO TD No. 1502), 69 pp, December 2009, electronc verson - July 2009. 188. Revson of the World Data Centre for Greenhouse Gases Data Submsson and Dssemnaton Gude (WMO TD No.1507), 55 pp, November 2009. 189. Report of the MACC/GAW Sesson on the Near-Real-Tme Delvery of the GAW Observatons of Reactve Gases, Garmsch- Partenkrchen, Germany, 6-8 October 2009, (WMO TD No. 1527), 31 pp. August 2010. 51