Charles R, Stearns
`The Schwerdffeger Library
`1.225 We:st Dayton. Street
`Madiso,~., WI 53706
`
`PUBLISHED BY AMERICAN GEOPHYSICAL UNION
`
`Lupin Ex. 1056 (Page 1 of 11)
`
`

`

`BYRON BOVR,I,~, 099~, ~996)
`JEF~:Rt’~Y "1", KIEHL (t992-~99~5)
`
`0227) ia published we~kly fbr $492 per year ~f<~r AGU
`
`U~Joa, 2(R~ Ek~da Av¢,. N.W . Washi+~gl~m DC
`
`(202) ,$62-~9+~;} TWX 7!0-822-#300. FAX:
`
`Lupin Ex. 1056 (Page 2 of 11)
`
`

`

`JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 97. NO. Dllh. PAGES 18J85-18,I93, NOVEMBER 2~, 1992
`
`Annual Cycles of Tropospheric Water Vapor
`
`DfAN J, GAFFEN
`
`Air Re,sources LaboratoO,, Nafiomd G~z’eanic and Atrao.*pher£ Administration. Silver Spring. Marfl~md, and Department of
`
`Depgtrrment of Meteorology. University of Mat3,1attd, College Park
`
`WILLIAM P. ELLIOTT
`
`Air Resources LabomtoO’, Nalionat Oecanb: and Atmospheric Administration, Silver Spring, Maryla~d
`
`To understand better the annual cycles of atmos pheric humidi.ty, r~diosonde dale were u~ed to
`c~ate eli~atol.o~es of tem~rature, d~’ poim., relative humidity, and precipitahle wa~er Jn lh~ lower
`~opospheve lbr 56 k,calions around the w~ld for the ~od 1973-1~. On lhe basis of the annual
`mnge~ of re~afive humidity at the surfs(cid:128) ~d at the 850, 7~. and 5~ mbar levels and lhe rntio of the
`~nual ma~imum to minimum surface to 5~mbar p~.ip~ble water, we have defined five h~midity
`regimes: ~’1] middle- and ~ghdafitt~de .continental (2) midd~e~ and hiO~qatitude ~eanic, (3) mid-
`latitud~ monsoon, (4) ~rapic~ oceanic, as~d (5) tropi¢~ monism. For .each re,me we descri~ the
`
`v~y in ~base, Relative h~i:~ty ~nges i~ ~he other three r~g~m~s are m~emte ~:o l~e, ~nd i.n the
`
`controls ~¢.aso~M humidity v~tions. T~y¢ ~sults surest that the assumption of consist re}afire
`humidity made in some climate m~els is not always jostled and that precipitable water is not a strong
`function of temperature in the troDes.
`
`LNTaOOt~CTam
`
`IRe, P~z~bhakara et el. [ 1985] presented seasonal maps of PW
`over the open oceans for 197%.198~ and confirmed the
`As a foundation for understanding tong-term change in general global p:~.~er~s deduced from radiosonde-based stud*
`water vapor and related processes, the annual cycle of
`ies but with ~:.t~d~ more detail in the tropics and southern
`oceans. Later ~t~.;.dies by Liu [1986] and Liu and Niiter
`humidity must be understood. Investigations of the water
`have foct~sed on variations of surt~lce relative humidity and
`vapor-greenhouse effect l~edback [e.g., Rind et el., 1991;
`PW at oceani~ ~ites with an eye towa~ improved p~me-
`Ravel andRamanathan, 1989] used seasonal ~d geoga~hic
`variations as su~ogales for long-term change. These studies
`terization of occ::~a-atmosp~ere fluxes t~ sing ~motely sensed
`PW. Recently, t!~e seasonal cycMs of clear sky up~r tropo~
`empIoyed satellite water va~r and radiation dma, respec-
`t:ively and ~e sho~ness of the satellite record ~d ant. allow
`sphmc relative humidity have been dete~ia~ from
`testing the assumption that seasona~ and spatial changes
`et eL, 1991] and by the St~tosphefic Aerosol ~d Gas
`reve~ patterns og long-term change.
`Surprisingly, there are few studies in the literature of the Experiment (SAGE ID [e,g., China et el., 1992]. Because
`observed annual cycle of h~midity above the surface. A few
`these satellite obse~wations are possibIe in eiear skies only,
`the annual cycle may not be fully resolved in regions with a
`early investigators [Reimn 1960a, b; Bannon and ~teele~
`distinct, cloudy season.
`1960; Taller. 1.%81 used radiosonde observations to ch~rac-
`terize seasona~ varialions in precipitable water (PC) b~t did
`While earlier studies have ft’mused o~ particular aspects of
`not consider other measures of water vapor or their vertic;d
`tropospheric l~umidity, none has explicitly explored the
`annual cycle of both relative, and specific humidity at a
`profiles. Later work showed seasonal variations in specific
`humidity’ IRao~mt~sson, 1972; Peixoto et el., 1981: Peixoto
`representative sample of stations globally. Using radiosonde
`and Oort, 1983; Oort, 1983] but was bas~ on radiosonde
`data ..........................................
`humidity variables, which leads us to the identification, of
`data that were (1~ t~en ~f~ I973. wMch ~e known to ~e
`five distinct water vapor regimes. Then seasonal variations
`of poor quality relative to more modern measurements, (2)
`gridded or zonally aver~ed, which is l~ely to mask local
`in relative humidity are explained and interpreted in greater
`variability M humidity, or (3) compiled into monthly means,
`detail.
`wkic~ cannot be ase~ to de~ve supplement~ h~midity
`variables without introducing bias [Elliott and G~ffen, I~1 ~,
`Using microwave observations from the Nimbus 7 s~tel-
`
`Copyright 1992 by lhe American. Geophysical Union.
`
`Paper number 92JD01999.
`0148~0227/92/92JD-01999505
`
`DATA AND METHOD
`
`Daily radiosonde reports, provided by National Climatic
`Data Center (tape deck 6103L from J,muary 1973 t.hrough
`December 1990, formed the basic data set. The 56 stations
`selected t Figure 1 and Table 1) are a subset of the 63-station
`
`This i~,1~teriai n:ay be protected by Copyrigm ~a,,, T;tie "~ iJ.S. Code
`...............................................................................................
`
`18, !85
`
`Lupin Ex. 1056 (Page 3 of 11)
`
`

`

`between PW and other vaHaNes, Becau~ the stations
`
`Organization Identification Numbers, ~at~ons. and
`
`Fig. 1.
`
`Map of the humidity regimes of each vadiosonde sta6on "
`
`[demifieafioa
`
`Name
`
`rude
`
`rude
`
`Elevation
`
`midd[e-and b~ghqalimd~ ~’eani~ (MO), mi/~-Ia~itud~ modem
`(MM}, tropical oce~c ~TOL a~d tropical monsoon (TML
`
`02836
`113953
`
`$odankla
`Valetitia
`
`9
`
`20
`
`91
`257
`
`- 8,7
`70,9
`26,7
`6Z3
`51.9
`-10,3
`48.2
`IL7
`80.2
`73.5
`70.7
`!6L3
`732
`54,9
`108.2
`57.8
`30,5
`50,3
`55.2
`5~.8
`39,2
`21,7
`22.5
`88.3
`72.8
`19,2
`77,0
`8.5
`114.2
`22d
`45.:f
`141.7
`1~14.O
`1.4
`2.2
`13.5
`- 17.5
`1.4,7
`77.5
`-37.8
`~3,9
`5.3
`47,5
`-18.8
`25.7
`-2£L2
`-40.3
`--9.8
`37,8
`-46,8
`"It .3
`-I56.8
`572 ~/702
`55.0
`-- 131.5
`762
`-119.3
`82,5
`-62_3
`48.5
`--58.5
`~80.7
`51,3
`-97_5
`25.8
`32.7
`-117,2.
`47.5 ~ 111,3
`---66.0
`18,5
`4.7
`-742
`4.8
`-52.3
`~22,8
`~3,2
`-’70.5
`--23.5
`-,I 1.5
`---7L2
`-70.3
`~-.2.3
`-66,3
`110.7
`13.5
`t44.8
`t66.7
`!93
`~9.7
`-~ t 55.0
`7.0
`1713 3
`55
`-9.3
`160,0
`-~172
`t77.5
`~9,8
`--139,0
`-/49.7
`-- I?.5
`-43N
`-176.5
`/~.g
`- t9(3
`118.7
`-20.3
`-34,9
`138_5
`
`64
`62
`3
`3
`233
`24
`29
`7
`~276
`1344
`54
`
`9
`34
`12
`63
`26
`1o
`6
`9
`1115
`3
`2541
`9
`5
`137
`
`52
`9
`11
`4
`
`52
`
`6
`
`Dik~n
`C~etyrekbstol~ovy
`Omsk
`KJrensk
`Kiev
`Orenburg
`Jeddah
`
`Bombay
`Trivandrum
`Hong Kung
`Wakk~ai
`Sh~a~
`Ni~ey
`Dakar
`New Amsterdam
`Abidjnn
`Antan~naf~vO
`Bu~awayo
`Gou@
`
`Barrow
`SL Paul
`
`Mould Buy
`Alert
`Stephenville
`
`Browns~ille
`San Diego
`Great Falls
`S~n Juan
`Bogota
`Cayenne
`Rio de J;meiro
`A~mf~as~a
`Pue~o Montt
`S.A.N.AiE.
`Casey
`Guam
`Wak~
`Hik3
`MNuro
`t!onJara
`N~d~
`A~uona
`rahRi
`Chatham
`Townsv~lle
`Port Hedland
`Addaide
`
`network described and u sod by A ngelt and Korshover l 1983]
`for analysis of t.ropos~heric and stratospheric temperature.
`However. Ettiott et aL [15N~] identified some problems with
`the original 63-st, a(ioa set for the analysis of humidity data,
`Thirty-nine of the 56 stations used here have very few
`missing data, at ieast for one observation time per day. The
`other 17 were indudedto We reasonable global represen~
`tadon of climatic zones, even though they could not be ttsed
`l~r the Rill 18 years.
`The radiosoude repo~ includes temperature ( T~ and dew
`poim depression at aft levels, geopotential heigN of manda-
`tory levels, and surface pressure and pressure of other
`significant levels, These were converted to dew Imin(. {Td)
`and relative humidity (~) al each levd and precipitab]e
`wa~er (PW) inthe su~hce to 850 mbar (PW,,-8) a.nd su~hce to
`(PW,¢.~) layers. Preci.pitab~e water is a measure of
`5~*mbar
`68~67~
`column water va~r coment and is the integral be~’~n two 68~4
`pressure leveN (p~ and p~) of specific humidity (q I: 7~E6
`
`2067~
`21965
`28~8
`302~
`333~5
`3512~
`40477
`
`43~3
`~3371
`~5{~4
`47~1
`~8698
`61052
`
`65578
`67083
`
`where ~¢ is the gravimtion~d acceleration. Because of the
`known poor performance of radiosonde hygristors in cold,
`d~ e~vm)~menIs, and ~cause almosl a~l ot’~hewater vapor
`in the atmosphere is in the lower troposphere, ~o data above
`5N) mb~ were used. Details on data quality control are given
`by ()’t~,n [19921.
`For each station, observation time (~0 and 12~ UT},
`~d variaMe, monthly means were calculated from the daffy
`datm To characterize the mean ~nnuN cycles, the 18 years of
`monlNy means were ~veraged to obtain kmg-te~
`
`To measure the amp!itude of the annual c cries, tfiera~ges
`of monthly mean RH were computed ~s the difference
`between the maximum and minimum long-term monthly
`mean values. A PW ratio was defined as the raao of the
`maximum to minimum long-term monthly mean PW,-~.
`additiom the surface tempm~mre {T;) nmge and mean
`annu:N T.,. and PWs> were computed. To anNyze p~ase
`relationships among vocables, lhe months
`mimma and maxima occurred were recorded. Two varNbles
`were con stalered to be in phase when the maximum and
`miNmum of one occurred within one month of the maximum
`and minimum of the other.
`
`30398
`71072
`71082
`71815
`71836
`72250
`72290
`72775
`78526
`~0~2
`8~5
`837~
`85~2
`857~
`89~N1
`896H
`91217
`91245
`91285
`9~376
`9i5~7
`91~0
`91925
`91~8
`93~
`91294
`94312
`94672
`
`Positive lalitude and km .gitude are degrees rmrth and ea.~t, respe-c-
`fivdy.
`
`Lupin Ex. 1056 (Page 4 of 11)
`
`

`

`TABLE 2, Amplitudes m~d Phase Rela~mnships of the Annual Cycle~ of Tre, p~spheric HumidRy ~’or the F~ve }tlJ.midi~y Regimes
`
`Humidity Regime
`
`PW Ratio
`
`Rtt Range*
`
`Variables i~ Pha~ Wi~h PW
`
`Middle- and highdatimde contmentM
`Middle- and high4atitude oceanic
`Mid-la~itud~ mousc~n
`
`Tropical
`
`1 an~ t~ at el! leve~s
`
`m}dtro~ospheric RH
`Midtropospherie RH and 7)
`
`PW ratio i~ the ratio of the maximum to rnia}mt~m m(mthty mean v;dues of ~m~)ce to 5Ni-mbar precipitabl~ water. RH
`difference ~etween maxhn~m and minimum moathiy me.a~ rela,iv~ humid~ y, evaJua~ed at l~e suff~ and a~ ~he 830, 700, and 5tY~ mbar
`]eve~s,
`~’Ihe ranges are categorized as small. ~15%; moderale, 15~30%; and large,
`~Daytime su~ace relative humidity does nol a~ways fulh~,~, these patterns and can exhiN~ a moderate ammal range.
`
`each regime tended to have geographic similartties, we
`naraed ~hem a¢cordi,~giy.
`Poleward of about 20= latitude where the annual cycies ot
`T and PW are in phase, most stations ha~e small RH. tanges.
`A subset of these with hig~ PW ~-atios we designate the
`midd~e- and high-latitude continental regime. The subse~
`with less annum variation of PW. found along windward
`coasts or o~ islands, is called the midale- and bighdatimde
`i~eanic regime. The third mid-lmiiude ~gime atso shows
`m~est annual changes in PW but sizable annual ranges of
`RH. These stations have a distinct rainy sea,on at~d so were
`grouped as the midqatitude monsoon regime.
`Stations between 20~N at~d 20’~S have distinct annual
`cycles in PW buI not in T~ The tlopicN oceanic
`characterized by a small PW ratio and stuN1 RH ranges in the
`planela~, boundary layer IPBL) but substantial variations
`RH in the midtroposphere. (I~ Nis analysis, the s~rface and
`850-mb~r data are considered representative ol" the PB L, and
`the 7(gg and 5(~)-mbar levels are considered the midtroN>
`sphere.} The tropic~d monsoon regime stations have larger
`PW ratios and a la.rger RH range throI~ghout ihe lower
`tm~sphere. The stations are idemified by their regimes i~
`FNure 1, and the quantitative tim.its oF the RH ranges and
`PW ratios a~ given in qMbte 2.
`In the remainder of this section we describe additiom~!
`%aiu~gs of these humidity regimes and show an example of
`each. Where availaN.e, we show nighttime data because
`the middle- and high-ia~mde, continental and mid-Iatimde
`monsoon reNmes, daytime surface RH is often more v~i-
`able than RH Mott or t~an su~tce RH at night.
`influence o}? the la~d ~lfface as a source of moistare, a~d the
`possibility of evaporation being less than its 0otential value.
`probabt~ accom~ts R~r tNs dagtime variabiti:ty in surNc¢ RH
`in these two reNmes.) Therefore we ~lied more on
`than on daytime surface RH variabiiiiy to determine
`station’s water vaN~r regime, although some siations, par-
`ficulady in the tropics, have oniy daytim~ observalions.
`
`Middle- and High~Lz~dtude COntinental Hamidi~.~ Regime
`
`~,e middle- and high-latitude continental (MC) Immidity
`re, me is chim~ctefized by a pronounced a.nm~al cycle in PW
`and small variations in RH. Munich, Germany, Is a go,~
`e~ample of the MC regime I Figure 2). Both PW.~ and PWs-.
`reac~ a maximum in late st~mmer, as do -bolh T and Td, A~
`MC ~t.ations. summer PWs-5 can be 3 to l(} times greater tha~
`
`winter PW.~.._s, while boundary layer and midtroposphcric
`RH ranges are generally less than 15%.
`Ah:hough not used ~o classiC’ slations, ~he annual mca~
`and annual range of T, and the annum mean PW~.,~ provide ~::
`additionM distinguishing c~agacteristics of each regime "’he
`MC sites exhibit h~.rge annual ranges in T.,. more than 20~’C,
`mid low annual mean T,, less than iff’C. As one would ~:.
`exact, .on the ~asis of ~he Clausius-.Clapeyro~ rdaliorL ::~::~
`annuat ;nea~ PW~,..~ is also low. 0.2 ~o 1.5
`
`Mi~Mle- a~d High-Latitude Oceanic HumidiO’ Regime
`
`Qualitativdy resembling t.he MC regime, the middle- and
`high-}ati~lIde oceanic ~MO) regime shows the moderating and
`moistening influences of the ocean, island and coastal sta-
`tions polewa~ of 40~ tend ~o show MO characteristics
`(Figure IL and Valentia, Ireiand, is a typical exam#e
`(Figure 3). At MO sites, RH ranges a~ small, paaictflarly
`above the PBL where ~he}, tend to be less than 10%. This,
`combined wi~h small annual r., ranges {less than 15~CL
`resuits in smaller .PW,.~ ratios (1.5 to 3) than at MC s~tes..As
`at MC sites, 1", Ta. and PW are in phase. ~a~iag iu summer.
`AnmtM mean T, is generally less than 15"C. and anm~a] mean
`PW,.-~ is (L5 to 2.0 ¢m. Jan Mayen IMaad. at 70.6°N. and
`Casey, An:tarctica, at 66.2~S~ are. respectively, the most
`noO~herly and southerly MO stations.
`"[’~e smaller T., range and PW ratio ai MO stalions
`compared with MC stations a~ consistent with the findings
`of Reitan l1960a, bl, who noted that the PW ratio is a good
`indicator of comi~wntali~y, commonly expressed as the
`anntml range of tempe.rotate. Peixot.o e? aL [19811 also tkmud
`seasonal changes in suH~ce to 300-tuber PW more marked
`over [and lha.o over sea.
`
`Mid-Latitt¢de Mot~aoon Humidi~= Regime
`
`The mid-latitude monsooo (MM} stations are coastal sites
`between about 20~ and 40° latitude and show more anm)al
`variabiiily in RH ~h;m either MC or MO stations, At Rio de
`J~eiro, Brazil, tk~r example, surface RH is maximum m
`winter, but aloft RH is maximum in summer, it~ phase wi.t~
`T, T,~. and PW (Figure 4L This aummertmxe maximum in
`midtroposp~eric hnmidity corresponds with a summertime
`maximum in precipitation [Eischefd er el, i~I I. The sea-
`sonal humidity variations at MM stations are related to
`
`Lupin Ex. 1056 (Page 5 of 11)
`
`

`

`Fig. 2. The a~nua! cycles of temperature, dew ~mt, relative humidity (at lhe ,surface and at 850. 7~, and 5~)
`
`UT, Data are load-term monthly me~s ~or the period 1973-t986. MuMch is im example of the middle- and hi~-latit~de
`
`Tropical Oceanic Humidity Regime
`
`seasonal circulation ~lnd convection patterns, so the term with the rainy season at ~ ake [National Oceanic and
`"’monsoon,+ is appropriate for the regime
`Atmospheric AdrninistraUo:i, 1982]. As ,~ result, roost Of the
`At MM slations, RH ~’anges can approach 50%. both in the
`a~nual change in PW is above 850 tabor, and PW,:5 is in
`PBL and the told’troposphere, Ra~ios of PW~.~ are compara-
`plaase with midtropospheric Ta and
`ble to the MO stations, 1.3 to 3, At MM stations, annual
`At TO stalions the ~atio &maximum to minimum PW,.~ is
`mea~ PW,¢.~. is beiween t.4 and 4.0 c.m, a~d a~ual mean T,
`less than :,’ ,6. and monlhly :mean PBL RH is high, always
`is typicaIIy greater than t.5~C, with annual ra~es of 5<’ to
`a~ve ~t%, with a smMI annum ra~e. Midtroposphedc RH
`I3~C.
`is lower than in the PBL (in fact:, a monolonic decrease of
`RH with decreasing pressure is noted in most re#mes),
`it is more v~fiable; the annnal range can approach 5~)L For
`TO stations, ammal meat~ Ts is generally ~eater than 20~C,
`In marked conlrast t~ H~e MC. MO. and MM regimes, the
`and annu~ ~nge of T, is less than 5~C, warm all year.
`~ropicM ~ceanic (TO) ~:me is ch~acterized by a PBL with
`Annual mean PW:: at TO sites is ~igh. 3 to 5 cm, All TO
`a relatively constant n~¢,~sture content and a more. vadable
`stations ~e within 20~ of the equator and are either island or
`midtroposphere. W~e l~]and is a typical TO station ¢Figure
`coastal sites (Figure
`5). Bo~nda~ layer T and T,~ are in phase, wi~h a s~i~t
`The PW ratios we compute are ¢om~ble to
`maximum in the summer. The~ combine to give a uniformly
`m~ist PBL as seen in the RH data. resulting in high PW in
`infer from Prabhakara et aL [1985] and Liu [1986] for the
`tropical Pacific~. Our restdts are also consistent with those of
`the surface to 850-tabor layer, greater lhan 2 era. "Ehe annual
`range of T in the midtroposphere is negligible, but Tu varies Liu and Niiter { 199. 0j~ who found tr~picaI oceanic surface
`by more than 10°C> leading to RH ra.nges of abot~t 30%. The RH to vary by tess than 1~:$ over the course of the year and
`high midtropospheric RH in summer and fall is associated
`to be generally between 70 and 80%. The RH variability
`
`Lupin Ex. 1056 (Page 6 of 11)
`
`

`

`GAFF’EN ET AL.: ANNUAL CYCLES OF TROPOSPKERIC WATER VAPOR
`
`18 A89
`
`30-
`
`30-
`
`~, 850
`"1" "1" ". 700.
`
`-4O
`
`-40-’
`
`1
`
`2
`
`3
`
`4-
`
`S
`
`6
`
`7
`
`8
`
`9
`
`i0 1t
`
`~2
`
`2
`
`3
`
`4
`
`5
`
`6
`
`7
`
`8
`
`~’
`
`10
`
`It
`
`!2
`
`~ sfc-SDO
`
`80,
`
`~ ~0~
`
`7= 40-~
`
`20-
`
`2 3 4 5 5 7 8 9 10 !1 12 2 3 (cid:128) 5 6 7 ~ 9 ~0 1~ 12
`Month Month
`
`Fig. 3. Same as Figure 2 but for Valentim Ireland. Data are long-term 0000 UT monthly means for the period
`1973-1990. Valemia is an example of the middle- and high-latitude oceanic humidity regime
`
`above the PBL can be compared with the results of Rasmus-
`effects of cloud cover and precipitation. As in the MM and
`son 1i972], who noted that zonally averaged RH showed TO regimes, seasonM variations in convection and in mois-
`maximum January to July changes, in excess of 20%~ in the
`tare advection, related to Mrge-scale atmospheric circulation
`subtropical midtroposphere. Our RH range is larger proba- patterns, control seasonal humidity variations at TM sites.
`bly because we are not taking a zonal mean. The large-seaM
`At TM stations, PW~-.~ ratios are between 2 and 3. In both
`the PBL and midtroposphere, RH ranges are large, generally
`cause of these seasonal changes is probably seasonal
`changes in convective activity and precipitation associated greater than 30%. Annual mean PW<5 at these sites is i .5 to
`with the north-south movement of the intertropical conver-
`5 cm. Annual mean T, is high. i5° to 30°C, with annual
`gence zone, as suggested by Prabhakara et al. [1979].
`ranges between 4° and 12°C.
`
`Tropical Monsoon Humidity Regime
`
`SEASONAL CHANGES IN RELATIVE HUMIDITY
`
`Much more seasonal variability is evident, in tropical
`The assumption that RH is approximately constant, both
`over the course of the year and over longer periods, is often
`monsoon (TM) locations. There are two subtypes of the TM
`regime. The first is the humid Coastal climate, for example made and formed the basis of the early modeling study of the
`effects of increasing atmospheric carbon dioxide by Manabe
`Bombay, India, where the boundary layer remains humid all
`year but the midtroposphere has a distinct dry season. The
`and Wetherald [1967]. They argued that holding RH constant
`second is an inland regime, as at Niamey, Niger (Figure 6),
`in a radiative-convective climate model is more reasonable
`where the dry season is evident at all levels. In both
`than holding absolute humidity constant, on the basis of
`subtypes, midtropospheric T shows little variation, but PBL
`observations of both humidity variables~ They showed fig-
`T has a minimum in winter and a secondary summer mini-
`ures from Telegadas and London [1954] as evidence that
`mum when T,~ is maximum. The secondary minimum during
`relative humidity changes little over the course of the year.
`the moist season is probably due to the surface cooling
`Close examination of those figures, and of figures pro-
`
`Lupin Ex. 1056 (Page 7 of 11)
`
`

`

`18,190
`
`GAFFEN ET AL.: ANNUAL CYCLES OF TROPOSPHERIC WATER VAPOg
`
`Month
`
`Manlh
`
`6.~v~v~v~ sfc--500
`~ sfc-850
`
`-40-
`
`-5O
`
`IO0
`
`80’
`
`2O
`
`.~.." .’. ".’. 700
`
`Month
`
`Month
`
`Fig. 4. Same as Figure 2 but for Rio de Janeiro, Brazil. Data are long-term 1200 UT monthly means for the period
`1973-1990. Rio de laneiro is an example of the mid-latitude monsoon humidity regime.
`
`sented by Rasmusson [1972], Rind et al. [1991], and Chiou et
`al. [1992], reveals that summer to winter RH changes as
`large as 20% are typical in some regions of the troposphere
`for zonally averaged RH for seasons or representative
`months. Furthermore, from the above analysis of station
`data, it is clear that lower tropospheric RH can range over
`50% locally over the course of a year. To understand better
`the nature of the annual cycle of RH, this section presents an
`expression for the time rate of change of RH and discusses it
`in terms of the water vapor regimes described above.
`Relative humidity does not vary independently but is a
`result of the combined effects of T and Td:
`
`es(Td)
`RH = ~
`e,(T)
`where e, is the saturation vapor pressure. Using the Clau-
`sius-Clapeyron equation, we can express es as
`
`(2)
`
`eo = es(To), and To = 273 K. The time rate of change of
`RH is then given by
`
`ORH L
`
`Ot Rv
`
`RH
`
`1 OTd 1
`
`T~ dt
`
`T2 ~t
`
`(4)
`
`Of course, here we are not considering the day-to-day
`variations in T, Ta, and RFI; rather, we are taking monthly
`mean values of each variable as representative of the daily
`values and considering month-to-month changes in the
`means. To gain some insight into the processes that can lead
`to seasonal changes in RH, we consider the two cases of
`approximately constant and of seasonally varying RH.
`
`Approximately Constant Relative Humidity
`
`If
`
`es(T) = e0 exp
`
`-
`
`(3)
`
`where L is the latent heat of vaporization, taken to be
`constant, Rv is the gas constant for water vapor,
`
`then
`
`ORH
`--~ 0 (~)
`ot
`
`Lupin Ex. 1056 (Page 8 of 11)
`
`

`

`GAFFEN I~T AL.: ANNUAL CYCLES OF TROPOSPHERIC WATER VAPOR
`
`lg, lgl
`
`40
`
`-2,0
`
`-40-
`
`-50
`
`~fc
`
`700
`
`700 soo
`
`Monih
`
`Mot’tth
`
`4O
`
`30-
`
`-40~
`
`-50
`
`1 O0
`
`20.
`
`Month
`
`CCCCO sfc-850
`
`~onth
`
`12
`
`Fig.
`
`Same as Figure 2 but for Wake Island. Data are long-term 0000 UT monthly means for the period 1973-1990.
`Wake Island is an example of the tropical oceanic humidity regime.
`
`and when RH decreases with time,
`
`(6)
`
`Since T,i can never exceed T, the time rate of change of Td
`must be slightly less than that of T, and this discrepancy
`grows as RH decreases.
`The water vapor regimes that exhibit minimal RH changes
`over the course of the year are the MC and MO regimes and
`the TO regime in the PBL. In the latter tropical case, PBL T
`and Td are almost constant, so (6) is satisfied because both
`sides are approximately zero. In the MC and MO regimes, T
`and Ta vary in phase, therefore their time rates of change
`always have the same sign. At Valentia (Figure 3), seasonal
`changes in T and Td are almost identical, but at Munich
`(Figure 2) the time rate of change of T exceeds that of Ta, as
`required by (6).
`
`Relative Humidity Varies Seasonally
`
`From (4) it follows that when RH increases with time,
`
`1 aTd 10T
`-- > T2
`T~ Ot Ot
`
`(7)
`
`I aTd 1 aT
`<
`T~ at T2 at
`
`(8)
`
`There are two possible scenarios.
`Dew point variations greater than temperature variations.
`If T is approximately constant, as at TO stations above the
`PBL, then (4) requires that Ta and RH be in phase, as
`observed (e.g., at Wake Island, Figure 5). Even if T varia-
`tions are not negligible, when Ta variations are much larger,
`RH cannot be constant. In the TM regime the midtropo-
`sphere experiences a period of increasing Td and RH as the
`humid season approaches and a period of decreasing Td and
`RH as the dry season approaches. At Niamey, for example
`(Figure 6), 700- and 500-mbar T changes are minimal com-
`pared with Td changes. The changes are consistent with (7)
`and (8), with Ta increasing faster than T as the moist summer
`season approaches and decreasing faster than T as the dry
`season approaches.
`Large temperature variations. If there is a pronounced
`seasonal cycle in T, as at MM stations, T and Ta vary in
`phase with RH, but the time rate of change of Ta is greater
`
`Lupin Ex. 1056 (Page 9 of 11)
`
`

`

`18,192
`
`GAFFF~ E’r AL.: ANNUAL CYCLF_.~ OF TROPOSPHERIC WATER VAPOR
`
`40
`
`30-
`
`20-
`
`lO.
`
`-40-
`
`850
`?00 ~oo
`
`Month
`
`40-
`
`30-
`
`*0-
`
`O-
`
`-5O
`
`10o
`
`80
`
`700
`
`700
`soo
`
`Month
`
`Month
`
`Fig. 6. Same as Figure 2 but for Niamey, Niger. Data are long-term 1200 UT monthly means for the period ! 973-1990.
`Niamey is an example of the tropical monsoon humidity regime.
`
`(in absolute magnitude) than that of T (e.g., at Rio de
`Janeim, Figure 4). Thus the advection of moisture (as
`measured by Td) by the monsoon circulation is more signif-
`icant than thertrml (T) advection for determining P,H varia-
`tions at both tropical and mid-latitude monsoon stations.
`
`DISCUSSION
`
`Examining the annual march of RH at stations around the
`globe reveals different patterns at different locations. In
`middle and high latitudes, RH is approximately constant
`throughout the year, except at locations influenced by mon-
`soon circulations. But in low-latitude regions, horizontal and
`vertical moisture advection, combined with low-amplitude
`changes in T over the course of the year, allows for substan-
`tial seasonal variation in RH. This is especially true above
`the PBL, which highlights the need for analysis of humidity
`measurements above the surface.
`The annual cycles of PW are consistent with the RH
`patterns. In middle and high latitudes, PW follows T. In the
`tropics, however, the annual cycle of PW is more closely
`related to midtropospheric RH variations than to T variations
`and reaches maximum values during the local rainy season
`when deep convection is strong. Our results for the trope-
`
`sphere below 500 mbar are qualitatively consistent
`with RH variations at higher levels measured by sat*Rites [van
`de Berg et ai., 1991; Chiou et el., 1992], that is, larger
`variations in the tropics than at mid-latitudes. While these five
`regimes are quite clearly delineated in our 56-station data set,
`future work will establish whether they are sufficient to char-
`acterize the annual cycles of tropospheric humidity globally.
`The existence of distinct humidity regimes suggests that a
`single set of assumptions above the vertical structure of
`humidity, and its seasonal variations, should not be applied
`globally. One-dimensional climate models need to disfin-
`gnish between tropical and mid-latitude and between conti-
`nental and oceanic atmospheres not just in terms of temper-
`ature profiles but also in terms of humidity profiles. Our
`results can also help to evaluate three-dimensional climate
`model simulations of the global climatology of tropospheric
`water vapor.
`If one wishes to draw analogies between long-term
`changes in climate and seasonal variations in the present
`climate, tl~e assumption that PW will increase as T increases
`and RH remains constant might be valid at some middle- and
`high-latitude locations. However, it cannot be supported in
`tropical regions at least at individual stations.
`
`Lupin Ex. 1056 (Page 10 of 11)
`
`

`

`GAFFEN ET AL.: ANNUAL CYCLES OF TROPOSPHERIC WATER VAPOR
`
`18,193
`
`REFERENCES
`
`Angell, J. K., and J. Korshover, Global temperature variations in
`the troposphere and stratosphere, i958-1982, Man. Weather
`Rev., 11I, 901-921, I983.
`Bannon, J. K., and L. P. Steele, Average water-vapour content of
`the air, Geophys. Mere. 102, U.K. Meteorol. Office, London,
`1960.
`Chiou, E. W., J. C. Larsen, M. P. McCormick, W. P. Chu, and D.
`Rind, Global distributions of upper troposphere relative humidity
`derived from SAGE II observations, paper presented at 72n~1
`American Meteorological Society, Annual Meeting, Am. Meteo-
`rol. Sac., Boston, Mass., Jan 5-10, 1992.
`Eiseheid, J. K., H. F. Diaz, R. S. Bradley, and P. D, Jones, A
`comprehensive precipitation data set for global land areas, DOEt
`ER-69.017T-HI, 81 pp., U.S. Dep. of Energy, Washington, D, C.,
`1991.
`Elliott, W. P., and D. J. Gaffen, On the utility of radiosonde
`humidity ~a~e, hives for elimat.e studies, Bull. Am. Meteorol. Sac.,
`72, 1507-1520, 199I.
`Elliott, W. P., M. E. Smith, and J. K. Angell, On monitoring
`tropospheric water vapor changes using radiosond.e data, in
`Greenhouse-Gas-Induced Climatic Cha~ge: A Critical Appraisal
`of Simulaiions and Observations, edited b~ M. Schlesinger, pp.
`311-328, Elsevier, l~ew York, 1991.
`Gaffen, D. J., Observed annt~al~ and interannual variations in tropo-
`sphere water vapor, Ph.D. dissertation, Univ. of Md., College
`Park, 1992.
`Liu, W. T., Statistical relation between monthly mean precipitable
`water and surface-level humidity over gIobal oceans, Man.
`Weather. Rev., 114, 1591-1602, 1986.
`Liu, W. T., and P. P. Niiler, The sensitivity of latent heat flux to the
`air humidity approximations used in ocean circulation models, J.
`Geophys, Res., 95, 9745-9753, 1990.
`Manabe, S., and R. T. Wetherald, Thermal equilibrium of the
`atmosphere with a given distribution of relative humidity, J.
`Atmos. S¢i., 24,241-259, 1967.
`National Oceanic and Atmospheric Administration, Local Climato-
`logical Data: Annual Summades for 1982, vol. 1 and 2, National
`Climatic Data Center, Asheville, N. C., 1982.
`tort, A. H., Global atmospheric circulation statistics, 1958-1973,
`NOAA Prof. Pap. 14, 180 pp., .Natl. Oceanic and Atmos. Admin.,
`U.S. Dep. of Commer., Washington, D. C., t983.
`Peixoto, J. P., and A. H. tort, The atmospheric branch of the
`hydrological cycle and. climate, in Variations in the Global Water
`Budget, edited by A. Street-Perrott, M. Beran, and R. Ratcliffe,
`pp. 5-65, D. Reide/.~ Hingham, Mass., 1983.
`
`Peixoto, $..P., D. A. Salstein, and R. D. Rosen, Intraannual
`variation in large-scale moisture fields, J, Geophys. Res., 86,
`1255-1264, 1981.
`Prabhakara, C’., G. Dalu, R. C. Lo, and N, R, Nath, Remote sensing
`of sea~onal distribution of precipitable water vapor over the
`oceans and th.e interference of boundary-layer structure, Man.
`Weather R.ev., 107, 1388-1401, I979.
`Prabhakara, C., D. A. Short, and B. E. Volmer, El Nifio and
`atmospheric water vapor: Observations from Nimbus 7 SMMR, J.
`Clim. Appl. Meteorol., ~24, I311-1324, 1985.
`Rasmusson, E. M., Seasonal variation of tropical humidity param-
`eters, in The General Circulation of the Tropical Atmosphere, vol.
`1, edited by R. E. Newell et al., pp. I93-237, MIT Press,
`Cambridge, Mass., 1972.
`Raval, A., and V. Ramanathan, Observational determination of the
`greenhouse efl’eet, Nature, 342,758-761, 1989.
`Reitan, C. H., Mean monthly values of preci~itable water over the
`United States, 1946-56, Man. Weather Rev, 88, 25-35, 1960a.
`Re!tan, C. H., Distribution of preeipitable water vapor over the
`continental United’States, Bull. Am. Meteorol. Sac., 41, 79--87,
`1960b.
`Rind, D,, E.-W, Chip, u, W. Chu, 3. Larseo, S, Oltmaos, J. Lerner,
`M. P, !vJcCormie~, and L.’" McMaster, Positive water vapour
`feedback in climate models confirmed by satellite d..ata, Nature,
`349, 500L503, 1991.
`Telegad~.s, K., and J. London, A physical model of northern
`hemis~her.e troposphere for winter and summer, S’cL Rep. 1,
`contract AF19(122)-165, 55 pp., Res. Div., College of Eng., N. Y.
`Univ., 1954.
`Tuller, S. E., World distribution of mean monthly and annual
`precipitable water, Man. Weather Rev., 96, 785-797, 1968.
`van de Berg, L., A. Pyomjamsri, and J. Schmetz, Monthly mean
`upper tropospheric hu