Ann. Geophys., 36, 213–229, 2018
h ps://doi.o g/10.5194/angeo-36-213-2018
© Au ho (s) 2018. This wo k is dis ibu ed unde
he C ea i e Commons A ibu ion 4.0 License.
Rela i e humidi y e ical p o iling using lida -based syne gis ic
me hods in he amewo k o he Hyg a-CD campaign
Le D. Labzo skii1, Alexand os Papayannis2, Ioannis Binie oglou3,4, Robe F. Banks5,6, Jose M. Baldasano5,7,
Flo ica Toanca3, Ch is G. Tzanis8, and John Ch is odoulakis8
1School o En i onmen al Science and Enginee ing, Sou h Uni e si y o Science and Technology o China,
Shenzhen 518055, China
2Lase Remo e Sensing Labo a o y, Physics Depa men , Na ional Technical Uni e si y o A hens,
A hens, Zog a ou, 15780, G eece
3Lase Remo e Sensing Labo a o y, Na ional Ins i u e o R&D o Op oelec onics, Magu ele (Il o ), 07712, Romania
4Raym e ics S.A., Spa is 32, Me amo osi A ikis, 14452, G eece
5Ea h Sciences Depa men , Ba celona Supe compu ing Cen e -Cen o Nacional de Supe comu ación (BSC-CNS),
Ba celona, 08034, Spain
6Depa men o Geoscience and Remo e Sensing, Facul y o Ci il Enginee ing and Geosciences, Del Uni e si y o
Technology, Del , 2628 CN, he Ne he lands
7En i onmen al Modelling Labo a o y, Technical Uni e si y o Ca alonia (UPC), Ba celona, 08028, Spain
8Sec ion o En i onmen al Physics and Me eo ology, Depa men o Physics, Na ional and Kapodis ian Uni e si y o
A hens, A hens, 15784, G eece
Co espondence: Le D. Labzo skii (le[email p o ec ed])
Recei ed: 25 Oc obe 2017 – Accep ed: 14 Decembe 2017 – Published: 14 Feb ua y 2018
Abs ac . Accu a e con inuous measu emen s o ela i e hu-
midi y (RH) e ical p o iles in he lowe oposphe e ha e
become a signi ican scien i ic challenge. In ecen yea s
a syne gy o a ious g ound-based emo e sensing ins u-
men s ha e been success ully used o RH e ical p o il-
ing, which has esul ed in he imp o emen o spa ial eso-
lu ion and, in some cases, o he accu acy o he measu e-
men . Some s udies ha e also sugges ed he use o high-
esolu ion model simula ions as inpu da ase s in o RH e -
ical p o iling echniques. In his pape we apply wo syn-
e ge ic me hods o RH p o iling, including he syne gy o
lida wi h a mic owa e adiome e and high- esolu ion a -
mosphe ic modeling. The wo me hods a e employed o RH
e ie al be ween 100 and 6000 m wi h inc eased spa ial es-
olu ion, based on da ase s om he Hyg A-CD (Hyg oscopic
Ae osols o Cloud D ople s) campaign conduc ed in A hens,
G eece om May o June 2014. RH p o iles om syne ge ic
me hods a e hen compa ed wi h hose e ie ed using single
ins umen s o as simula ed by high- esolu ion models. Ou
p oposed echnique o RH p o iling p o ides imp o ed s a-
is ical ag eemen wi h e e ence o adiosoundings by 27 %
when he lida – adiome e (in compa ison wi h adiome e
measu emen s) app oach is used and by 15 % when a lida
model is used (in compa ison wi h WRF-model simula ions).
Mean unce ain y o RH due o empe a u e bias in RH p o-
iling was ∼4.34% o he lida – adiome e and ∼1.22 %
o he lida –model me hods. Howe e , maximum unce -
ain y in RH e ie als due o empe a u e bias showed ha
lida -model me hod is mo e eliable a heigh s g ea e han
2000 m. O e all, ou esul s ha e demons a ed he capabil-
i y o bo h combined me hods o day ime measu emen s in
heigh s be ween 100 and 6000 m when lida – adiome e o
lida –WRF combined da ase s a e a ailable.
Keywo ds. A mosphe ic composi ion and s uc u e (ins u-
men s and echniques)
1 In oduc ion
Rela i e humidi y (RH) is a c ucial pa ame e o a mo-
sphe ic esea ch, as i ep esen s he cu en s a e o wa e
apo and ambien ai ela ed o sa u a ion. Changes in RH
Published by Cope nicus Publica ions on behal o he Eu opean Geosciences Union.
214 L. D. Labzo skii e al.: Rela i e humidi y e ical p o iling using lida -based syne gis ic me hods
may in luence a mosphe ic op ical p ope ies such as isibil-
i y, which is o en educed due o RH a ia ions in he a mo-
sphe e (Tang e al., 1981). Mo eo e , inc eased RH in he a -
mosphe e may in luence he physical p ope ies o ae osols,
causing condensa ion on o hei su ace, which subsequen ly
igge s hei hyg oscopic g ow h. No only does his g ow h
a ec he di ec sca e ing o adia ion (Hanel and Zankl,
1979; Hegg e al., 1996; Ziege e al., 2013), bu also he p o-
cess o cloud condensa ion nuclei (CCN) o ma ion (Cha l-
son e al., 1992; Pe e s and K eidenweis, 2007; Wex e al.,
2008; Mochida, 2014). P e ious s udies ha e ound ha e en
mino changes in RH may a ec p ocesses such as cloud
o ma ion o p ecipi a ion (Kulmala e al., 1993; Tomkins,
2003; She wood e al., 2010; Al a a z e al., 2013). The in-
e es in he ole o RH in he modi ica ion o ae osol, p e-
cipi a ion and cloud mic ophysics, including CCN o ma-
ion, has ecen ly inc eased mainly due o he c ucial ole
o ae osol–cloud in e ac ions in clima e change (Fan e al.,
2007; Veselo skii e al., 2009; Ziege e al., 2013; G anadoz-
Munoz e al., 2015; López and Á ila, 2016).
In addi ion, RH measu emen s a e equen ly used o
e alua ion s udies aiming o p edic he o ma ion o clouds
(Hee waa den and A ellano, 2008) and ai c a con ails
(Radel and Shine, 2007). No less impo an a e he signi i-
can unce ain ies in he es ima ion o global clima e change
pa ame e s using clima e modeling (Schneide e al., 2010).
Usually hese unce ain ies a e associa ed wi h RH a ia-
ions, since wa e apo ac s as a global cons ain in he
clima e sys em (She wood e al., 2010). Despi e ad ance-
men s in sa elli e emo e sensing o wa e apo , con inu-
ous quan i ica ion o RH in he low oposphe e emains
challenging. Ve ical esolu ion o spacebo ne measu emen s
o wa e apo and empe a u e cons ains he accu acy o
RH e ie al close o he g ound (wi h spa ial esolu ion o
∼1–2 km) (Wul meye e al., 2015). RH obse a ions a e
based on wa e apo and empe a u e measu emen s which
a e oge he equen ly e e ed o as he modynamic a mo-
sphe ic p o iling. A comp ehensi e desc ip ion o he mod-
e n echniques o he modynamic p o iling by di e en in-
s umen s is gi en by Wul meye e al. (2015), who ha e ou -
lined he ad an ages, disad an ages and unce ain ies o each
ins umen . He e, we p esen a b ie desc ip ion o di e en
g ound-based echniques o RH e ical p o iling. One o
he mos equen ly used ins umen s o RH e ical p o il-
ing a e adiosondes. Radiosoundings p o ide e ical p o iles
o RH wi h spa ial esolu ion o a ew me e s, and ela i ely
high accu acy (±4–5%, depending on he ime o he day)
(Miloshe ich e al., 2009). Mo e au oma ed RH e ical p o-
iling is usually pe o med using passi e and ac i e emo e
sensing senso s which a e able o measu e bo h wa e a-
po and empe a u e e ical dis ibu ion. The A mosphe ic
Emi ed Radiance In e e ome e (AERI) is an example o
a passi e senso ha can be used o wa e apo and em-
pe a u e quan i ica ion, based on a mosphe ic adiance mea-
su emen s a he 15 µm CO2band. P o iles o RH a e p o-
ided using a combina ion o AERI wa e apo and em-
pe a u e da ase s om he g ound up o 3000m, wi h a em-
po al esolu ion o 10min (Fel z e al., 1998; Knu eson e
al., 2004). The AERI sys em is, howe e , limi ed by coa se
spa ial esolu ion and o en AERI canno de ec sha p and
s ong in e sion laye s (Ma is e al., 2002). Mic owa e a-
diome ic measu emen s o empe a u e and humidi y can
p o ide wa e apo and empe a u e e ical p o iles as well.
A mic owa e adiome e pe o ms mul i equency measu e-
men s o b igh ness empe a u es a high empo al esolu ion
(∼1 s) and high accu acy om he su ace (0.6K) up o he
middle oposphe e (1.5 o 2K) (Hogg e al., 1983; Wa e
e al., 2003). Ne e heless, adiome e measu emen s su e
om coa se spa ial esolu ion and subs an ial unce ain ies in
he e ie al o humidi y and empe a u e a heigh s g ea e
han 4000 m, whe e only 5 % o he independen in o ma-
ion o igina es om adiome e measu emen s hemsel es
(Rose e al., 2005). Ac i e emo e sensing ins umen s such
as a mosphe ic lida ins umen s ha e he abili y o ob ain
high- esolu ion measu emen s o RH. To his end, wo majo
echniques can be employed: he di e en ial abso p ion li-
da (DIAL) and he Raman lida echniques. Each echnique
has i s unique ad an ages and limi a ions. The DIAL wa e
apo p o iling is based on he a io o wo elas ic backsca -
e signals a wo adjacen wa eleng hs and is a ec ed by
he empe a u e dependence o he wa e apo molecula
abso p ion, which is g ea e han 1–2% (Wul meye e al.,
2015). Addi ionally, he p esence o s ong ae osol g adi-
en s may esul in high sys ema ic unce ain ies ha exceed
he equi emen s o mos desi ed applica ions (Theopold and
Bosenbe g, 1993). The Raman lida echnique o RH e -
ical p o iling is based on he ib a ional Raman sca e ing,
which can be combined wi h he o a ional Raman sca e ing
o p o ide also he empe a u e e ical p o iles (A shino e
al., 1983). When a lase beam is emi ed o he a mosphe e a
355 nm, he use o a Raman lida a 387 nm (Raman shi ing
by a mosphe ic N2) and a 407 nm (Raman shi ing by wa e
apo ) enables he humidi y-dependen pa ame e o be de-
i ed, which is subsequen ly no malized o he mixing a io
o wa e apo . The mos impo an cons ain s o cu en
lida ins umen s make mos o hem no applicable o wa-
e apo measu emen s du ing day ime (due o high a mo-
sphe ic backg ound le els) and in he lowes se e al hund ed
me e s o he oposphe e (due o geome ical op ics limi a-
ions).
Despi e he ela i ely high pe o mance o emo e sens-
ing ins umen s wi h ega d o RH e ical p o iling, we s ill
su e om a lack o a consis en obus me hod o con in-
uous RH e ical p o iling. One s ep o wa d in RH e ical
p o iling wi hou he echnological imp o emen o senso s
is o use syne gis ic app oaches, as p oposed by Tu ne e
al. (2000). This s udy p esen ed he syne gis ic e ie al o
RH based on a Raman lida - e ie ed wa e apo mixing a-
io and empe a u e p o iles om he AERI ins umen . Such
syne gy allowed he p o iling o RH wi h high empo al es-
Ann. Geophys., 36, 213–229, 2018 www.ann-geophys.ne /36/213/2018/
L. D. Labzo skii e al.: Rela i e humidi y e ical p o iling using lida -based syne gis ic me hods 215
olu ion o be pe o med. Ano he example o a syne ge ic ap-
p oach owa ds RH e ical p o iling has been p esen ed by
Nagel e al. (2001), whe e he au ho s app oached RH p o-
iling by combining lida -de i ed humidi y and empe a u e
measu emen s om adiosoundings (launched e e y 10min).
Mo eo e , Wang e al. (2011) demons a ed ha densi y and
he wa e apo mixing a io can be combined wi h empe -
a u e obse a ions om a colloca ed o a ional Raman lida
o p o ide RH e ical p o iles. They showed ela i ely good
ag eemen be ween lida - e ie ed and adiosonde obse a-
ions, wi h a bias o up o 10% in he lowes 2000 m o
he oposphe e. A s ep o wa d in syne gis ic app oaches o-
wa ds accu a e RH e ical p o iling has been pe o med by
Na as-Guzman e al. (2014). They demons a ed he me hod
o RH e ical p o iling based on he combina ion o Raman
lida humidi y and empe a u e measu emen s om a collo-
ca ed mic owa e adiome e , oge he wi h ai densi y p o-
iles aken om a s anda d a mosphe ic model scaled o nea -
g ound densi y measu emen s (COESA, 1976). This com-
bined e ie al me hod esul ed in an inc eased accu acy o
con inuous lida -de i ed RH measu emen s in compa ison
wi h o he emo e sensing echniques. The esul an mean ab-
solu e de ia ion in RH compa ed o adiosonde da a based
on lida – adiome e e ie als a ied om 6 o 7% om
1000 o 5000m, espec i ely. Recen ly, Ba e a-Ve dejo e
al. (2016) once again es ed lida – adiome e combina ion
pe spec i es o wa e apo s udies. They de eloped a new
app oach o lida – adiome e syne gy o absolu e humidi y
(AH) e ical p o iling using an op imal es ima ion me hod.
They combined mul i equency b igh ness empe a u e ob-
se a ions om a mic owa e adiome e and mixing a io
obse a ions om a Raman lida o e ie e high- esolu ion
p o iles o AH. Thei esul s p o ed ha he combina ion o
lida and adiome e da a can educe he heo e ical e o by
a ac o o 2 in he lowe oposphe e when wa e apo in-
o ma ion is e ie ed. All o hese a o emen ioned me hods
and app oaches ha e shown he pe spec i es o syne gis ic
app oaches o RH e ical p o iling using colloca ed emo e
sensing ins umen s. Mo e ecen ly, Schu gens e al. (2017)
p esen ed p omising esul s om he combina ion o spa-
ially colloca ed obse a ions and model simula ions, poin -
ing ou ha high- esolu ion model simula ions can se e as a
obus da a sou ce.
Based on he ecommenda ions o Wul meye e al. (2015)
and Schu gens e al. (2017), in ou s udy we used wo syn-
e gis ic app oaches o RH e ical p o iling. The i s ap-
p oach is based on he syne gy o lida and adiome e ins u-
men s, while he second me hod is based on lida and nume -
ical simula ions om he Wea he Resea ch and Fo ecas ing
(WRF) model ou pu . Ou app oaches use a combina ion o
da ase s, including wa e apo mixing a io om a Raman
lida , empe a u e p o iles om adiome e , high- esolu ion
simula ions om he WRF model and ai densi y p o iles
om he US S anda d A mosphe e (1976) (COESA, 1976).
Da ase s we e acqui ed du ing he Hyg A-CD (Hyg oscopic
Ae osols o Cloud D ople s) campaign conduc ed in A hens,
G eece, om May o June 2014 (Papayannis e al., 2017).
The main scope o his pape is o show he e ec i eness o
he wo syne gis ic app oaches in compa ison wi h single-
ins umen obse a ions o RH om mic owa e adiome e
and RH single simula ions om WRF. As a second objec i e,
we de e mine he e ec i eness o hese app oaches acco d-
ing o c ucial equi emen s applied o he modynamic p o-
iling echniques o mula ed and gene alized by Wul meye
e al. (2015). These equi emen s can be applied o RH e i-
cal p o iling and include se e al poin s, among which he ac-
cu acy, spa ial esolu ion and he minimum–maximum ange
o measu emen s.
2 Ins umen s and models
2.1 Expe imen al si e
The da ase s used in his wo k we e collec ed du ing he
Hyg A-CD expe imen al campaign o ganized in he g ea e
A hens a ea in he pe iod 15 May–22 June 2014. The aim
o he campaign was o b ing oge he a ious ins umen s
o a mosphe ic measu emen s in o de o imp o e ou cu -
en unde s anding o he impac o ae osols on clouds nea
he op o he plane a y bounda y laye (PBL). Du ing he
campaign pe iod a a ie y o emo e sensing and in si u in-
s umen s p o ided an impo an eco d o da a on ae osols,
clouds and local me eo ology condi ions. Among he i e
measu emen si es in ol ed in he Hyg A-CD campaign,
mos o he da a we e ob ained a he Na ional Technical Uni-
e si y o A hens (NTUA) (37.97◦N, 23.79◦E, 212 m a.s.l.)
and he Na ional Cen e o Scien i ic Resea ch Demok i os
(DEM) (37.99◦N, 23.82◦E, 275 m a.s.l.) (Papayannis e al.,
2017).
2.2 Mul iwa eleng h lida
The ae osol and ozone lida sys em (EOLE) mul iwa eleng h
Raman lida sys em loca ed a he campus o NTUA emi -
ed pulses a h ee wa eleng hs: 355, 532 and 1064nm, wi h
ene gies pe pulse o 240, 260 and 300mJ, espec i ely,
wi h a 10 Hz epe i ion a e. A ecei ing Casseg ainian ele-
scope (p ima y mi o o 300 mm diame e and 600 mm o-
cal leng h) was used o simul aneously ecei e he elas ic
backsca e ed lida signals and he Raman ones (387, 407,
607 nm). The ull o e lap o he sys em is achie ed a ∼
300 m om he lida sys em (Kokkalis e al., 2012). Since he
Raman signals a e ela i ely weak, he Raman lida measu e-
men s we e pe o med only a nigh ime unde clea -sky con-
di ions. The Raman-de i ed e ical p o iles o he wa e a-
po mixing a io we e calcula ed o 26 days o he campaign
wi h di e en empo al scales (2min, 1 h and 2 h), which
we e selec ed depending on he a ious ins umen s’ in e -
compa ison: 2 min a e aged da a we e used o compa isons
wi h in eg a ed adiome ic alues, and 2 h a e aged da a
www.ann-geophys.ne /36/213/2018/ Ann. Geophys., 36, 213–229, 2018
216 L. D. Labzo skii e al.: Rela i e humidi y e ical p o iling using lida -based syne gis ic me hods
o he in e compa ison wi h he adiosonde da a, depend-
ing on he signal- o-noise a io (SNR) o he lida signals,
as ou lined as he minimum equi ed esolu ion o e ec i e
he modynamic p o iling in he e iew om Wul meye e
al. (2015). The signal de ec ion a he ib a ion Raman chan-
nels o 387 and 407nm gi es he possibili y o e ie e he
wa e apo mixing a io p o iles de ined as he a io µo he
mass o wa e apo o he mass o d y ai (gkg−1) (Gold-
smi h e al., 1998) as ex ensi ely used in he las 2 decades
o bo h day ime and nigh ime measu emen s (Whi eman e
al., 2006, 2010; Adam e al., 2007; Leblanc e al., 2011).
µ=CPWV(R)
PN2(R)
exp"−
R
R0
αN2(d )d #
exp"−
R
R0
αWV(d )d #,(1)
whe e PWV is he de ec ed lida signal a he wa e apo
channel, PN2is he de ec ed Raman signal a he ni ogen
channel and Cis he calib a ion cons an (see Sec . 3.1).
The exponen ial pa o Eq. (1) akes in o accoun he a-
io o he a mosphe ic ansmission a 387 (αN2)and 407 nm
(αWV)(Wei kamp, 2005). This di e ence in ansmission is
mainly con ibu ed by Rayleigh sca e ing and can be calcu-
la ed using empe a u e and p essu e p o iles aken om he
US S anda d A mosphe e 1976 (COESA, 1976) and ange-
independen Rayleigh sca e ing c oss sec ions a app op i-
a e wa eleng hs (Buchol z, 1995). The s anda d p o iles o
p essu e a e assumed o be accu a e o ou pu poses since
he unce ain ies in oduced by hei use a e lowe han 5%.
The di e ence be ween p o iles o p essu e ob ained om
adiosoundings and s anda d p essu e p o iles did no ex-
ceed 0.055 kgm−3(4.5%). A mo e de ailed desc ip ion o
he calib a ion p ocedu e and he analysis o wa e apo ob-
se a ions o he EOLE sys em can be ound in Landul o e
al. (2009) and Mamou i e al. (2008), espec i ely, as well in
Sec . 3.1 below.
2.3 Mic owa e adiome e
The HATPRO-G2 mic owa e adiome e consis s o se -
e al componen s: wo ecei e uni s (22.24–31.4 and 51.3–
59 GHz) wi h he ele an ecei ing op ics, he ambien load,
he in e nal scanning mechanism, he elec onics and he da a
acquisi ion sys em (Rose e al., 2005). The mic owa e a-
diome e used in his s udy is manu ac u ed by Radiome e
Physics GmbH and belongs o he Na ional Ins i u e o R&D
in Op oelec onics (Bucha es , Romania). The HATPRO-
G2, ins alled a NTUA, was calib a ed be o e he obse a-
ion campaign acco ding o he p ocedu e o adiome e ab-
solu e calib a ion using liquid ni ogen (Liljeg en, 2002).
The a mosphe ic adia ion is measu ed a se en channels
loca ed in he K band, along he wing o he wa e apo
abso p ion line (22.35 GHz), and se en channels loca ed in
he V band (Wes wa e , 1965), along he oxygen abso p-
ion complex (cen e is a ound 60GHz) (Wes wa e e al.,
2005). The e ical p o iles o wa e apo and empe a-
u e a e in e ed om obse ed b igh ness empe a u es by
using s a is ical eg ession algo i hms, based on long- e m
da ase s o colloca ed adiosoundings. The adiome e mea-
su emen s used in his wo k ha e a empo al esolu ion o
15 s and a heigh -dependen e ical esolu ion: 200m om
0 o 2000 m, 400 m om 2000 o 5000m and >500 m o
heigh s abo e 5000m (Löhne e al., 2004; Rose e al., 2005;
Mashwi z e al., 2013). The adiome ic measu emen s ( he
da a p o ided a e in eg a ed wa e apo , IWV and e ical
p o iles o AH, RH and empe a u e) we e con inuously pe -
o med om 15 May o 20 June 2014, excep on 12 June
whe e hey we e only a ailable om 00:00 o 05:55UTC
and om 20:38 o 23:59 UTC, due o a echnical shu down.
Radiome e - ela ed andom e o s a e analyzed based on p e-
ious s udies which had de e mined ha he e o ela ed o
he sys ema ic bias o AH e ie als using he eg ession
me hod below 4000m equals 0.8 gm−3. The andom e o
o 0.5 K is aken in o accoun in he PBL (up o 1000m) and
1.7 K be ween 1000 and 6000 m (Güldne , 2013; C ewell e
al., 2001; Liljeg en e al., 2005; Löhne and Maie , 2012).
2.4 Sun pho ome e
The sun pho ome e is a passi e emo e sensing ins umen
ha e ie es columna a mosphe ic ae osol p ope ies du -
ing day ime while poin ing a he sun. Fo his wo k, we
used ae osol columna op ical p ope ies e ie ed om a
CIMEL CE-318-NEDPS9 sun pho ome e (Holben e al.,
1998), which is a membe si e o he NASA AERONET
(Ae osol Robo ic Ne wo k) and was loca ed on a nea by si e
a ound 5 km om NTUA (Papayannis e al., 2017). The sun
pho ome e pe o ms di ec sun and sky measu emen s o so-
la adiances a eigh wa eleng hs (340, 380, 440, 500, 675,
870, 1020 and 1640 nm). Ae osol op ical dep h is e ie ed
om di ec sun measu emen s, while di use sun measu e-
men s a e used by he in e sion algo i hms o columna mi-
c ophysical p ope ies (Dubo ik and King, 2000; Dubo ik e
al., 2006). Finally, measu emen s a 940nm a e used o es-
ima e he in eg a ed amoun o p ecipi able wa e which is
used in ou wo k o he pu poses o mic owa e adiome e
alida ion and calib a ion.
2.5 Radiosoundings
Du ing he campaign 17 high- esolu ion adiosondes
(Vaisala RS92-SGP) we e launched by he Hellenic Na ional
Me eo ological Se ice (HNMS) a he Hellinikon ai po
(37.88◦N, 23.73◦E). Ve ical p o iles o empe a u e we e
measu ed wi h an unce ain y o 0.3–0.4◦C, ela i e humid-
i y wi h an unce ain y a ound 4% and heigh unce ain y
a ound 20 m (Nash e al., 2011). He e we use only he a-
diosoundings launched a 00:00UTC, since only nigh ime
Ann. Geophys., 36, 213–229, 2018 www.ann-geophys.ne /36/213/2018/
L. D. Labzo skii e al.: Rela i e humidi y e ical p o iling using lida -based syne gis ic me hods 217
lida measu emen s we e conside ed o wa e apo p o il-
ing. When high e ical esolu ion adiosonde da a we e no
a ailable, we used he adiosounding da a (p essu e, em-
pe a u e, humidi y) wi h spa se e ical esolu ion o abou
50–700 m (low- esolu ion adiosoundings) ob ained om he
Uni e si y o Wyoming websi e (h p://wea he .uwyo.edu/
uppe ai /uamap.sh ml).
2.6 WRF model con igu a ion
The WRF model (Skama ock e al., 2005) is a nume ical
wea he p edic ion sys em designed o bo h a mosphe ic
esea ch and ope a ional o ecas ing needs. He e, we use
WRF model e sion 3.4.1 wi h he Ad anced Resea ch WRF
(ARW) dynamical co e. Th ee one-way nes ed domains a e
con igu ed, wi h he ines domain on a 1km ×1 km g id
o e he g ea e A hens a ea. This high esolu ion is deemed
su icien o make simula ions compa able wi h obse a-
ional measu emen s in he complex a ea o A hens. Mois-
u e pa ame e s we e simula ed wi h 39 e ical le els a 50–
100 m e ical spacing and wi h a 1h w i ing pe iod o he
esul s. Daily simula ions we e un wi h a 36 h o ecas cy-
cle, including an allowance o 12 h o model spin-up. Mo e
de ails abou he con igu a ion o he WRF model du ing he
Hyg A-CD campaign a e desc ibed in Banks e al. (2016).
3 E alua ion o indi idual e ie als
RH calcula ion equi es p o iling o wa e apo , empe a u e
and p essu e. In his pape we used Raman lida , a mic owa e
adiome e and he WRF model o calcula e he RH p o iles.
The calib a ion and e alua ion o wa e apo and empe a-
u e measu emen s a e u he p esen ed.
3.1 EOLE calib a ion o wa e apo measu emen s –
alida ion o lida calib a ion esul s
The lida calib a ion is equi ed since he measu ed quan-
i y canno be di ec ly e e ed o as he wa e apo mix-
ing a io. The e a e se e al me hods o calib a e a Raman
lida sys em o wa e apo p o iling: by compa ing wi h
adiosonde obse a ions, wi h colloca ed ins umen obse -
a ions (Whi eman e al., 1992; Fo h e al., 2015) o by us-
ing hyb id no maliza ion by calib a ing lamps (Leblanc and
McDe mid, 2008). He e we apply he mos commonly used
me hod o he calib a ion using colloca ed adiosondes da a
wi hin speci ic al i ude anges depending on he lida signal-
o-noise a io o wo main easons. Fi s ly, calib a ion using
passi e emo e senso s may esul in incomple e sampling o
he wa e apo column obse ed by lida . Secondly, he use
o calib a ed lamps can be challenging and may lead o he
inc eased amoun o unexpec ed e o s ha a e di icul o
obse e and compensa e o (Whi eman e al., 2011). To de-
i e he wa e apo mixing a io (µ), a calib a ion cons an
(C) has o be es ima ed (see Eq. 2) aking, as well, he a io o
he ib a ional Raman lida signals a 407 (PWV) and 387 nm
(PN2):
µ=CPWV(R)
PN2(R) .(2)
The mean alue o Cis calcula ed om he bo om o he op
laye s o six high- esolu ion adiosoundings, aking in o ac-
coun he al i udes whe e i s s anda d de ia ion is less han
10 % (Table 1). In ou case Cwas es ima ed o be equal
o 23.65, wi h a weigh ed s anda d de ia ion o 9.5% (see
Fig. 1, le panel). The alues o he calib a ion cons an wi h
app op ia e de ia ions o ce ain al i udes whe e a e aging
was pe o med (calcula ed om a se ies o adiosonde da a),
a e p esen ed in Table 1.
Fu he mo e, we compa ed he Raman lida da a wi h col-
loca ed low- esolu ion adiosoundings o check he sani y o
he es ima ed C alue ( he one shown in Table 1). The low-
esolu ion calib a ion cons an is 24.36 wi h a weigh ed s an-
da d de ia ion o 8.8% (Fig. 1., igh panel) and he di -
e ence be ween he mean calib a ion cons an s is negligi-
ble. Rela i ely low di e ences wi h e e enced adiosound-
ing di ec ly app o es he alidi y o ou calib a ion o mee
he equi emen s (unce ain y should no exceed 10%) om
Leblanc and McDe mid (2008). The e o e in his s udy we
used calib a ion cons an om high esolu ion adiosounding
(C=23.65 ±2.28).
As ou lined by Leblanc e al. (2011), he calib a ion s a-
bili y o a wa e apo Raman lida has o be insu ed by
colloca ed wa e - apo measu emen s. The e o e, an in e -
compa ison o he IWV be ween he Raman lida , sun pho-
ome e and mic owa e adiome e was pe o med. Fi s , we
compa ed he IWV om colloca ed adiome e and sun pho-
ome e measu emen s. Secondly, he lida -de i ed IWV al-
ues (2min esolu ion) we e compa ed wi h he IWV alues
e ie ed om he adiome e da a a he same ime wi h a
empo al di e ence o less han 30s. The in e compa ison
was done in wo s eps because he Raman lida p o ided he
wa e apo mixing a io only du ing nigh ime, while he
sun pho ome e measu ed only du ing day ime. In he i s
s ep, he sun pho ome e and adiome e da ase s we e in e -
compa ed, while in he second s ep he adiome e and li-
da da ase s we e in e compa ed. The sun pho ome e unce -
ain y in IWV calcula ion was conside ed o be 1.2kgm−2
(Wang, 2008), while ha o he mic owa e adiome e o be
0.8 kgm−2(Rose e al., 2005). In ou analysis we used in o-
al 36 measu emen pai s. The de e mina ion coe icien (R2)
was es ima ed o be equal o 0.89 and only in one case was
no ed as inconsis en (Fig. 2, le panel). The a e age bias (b)
o he adiome e da a ega ding he sun pho ome e was es i-
ma ed o be equal o 1.02. The in e compa ison be ween lida
and he mic owa e adiome e was pe o med wi h some ad-
di ional assump ions as in ou s udy EOLE was no able o
p o ide eliable da a below 750 m, due o geome ic op ical
limi a ions o he sys em (see Sec . 3.2). An uppe limi o
he in eg a ion ange o 9000m has been chosen, such as o
www.ann-geophys.ne /36/213/2018/ Ann. Geophys., 36, 213–229, 2018
218 L. D. Labzo skii e al.: Rela i e humidi y e ical p o iling using lida -based syne gis ic me hods
Table 1. Calib a ion cons an mean alues o each case o high- esolu ion adiosounding.
Da e Lida ime Radiosonde Bo om Top Cons an mean σ
(UTC) launcha(km) (km) (±σ) (%)
15 May 2014 21:00–22:00 00:00 1.00 2.04 22.52±2.23 9.94
17 May 2014 19:00–20:00 00:00 1.00 4.00 24.51±2.47 10.00
18 May 2014 22:00–23:00 00:00 1.00 2.57 23.41±2.27 9.78
20 May 2014 22:00–23:00 00:00 2.40 4.24 24.00±2.29 9.70
21 May 2014 21:00–22:00 00:00 2.26 3.34 21.58±2.03 9.50
1 Jun 2014 23:00–00:00 00:00 1.00 4.04 25.89±2.27 8.84
aAll adiosondes a e launched on he ollowing day a e he Raman lida measu emen s.
Figu e 1. Va ia ions o he wa e apo calib a ion cons an calcula ed based on low- esolu ion (LR) and high- esolu ion (HR) adiosound-
ings. G een ba s: calib a ion cons an a ia ions, ed solid line: mean calib a ion cons an , dashed lines: maximum and minimum calib a ion
cons an s.
Figu e 2. IWV in e compa ison be ween di e en ins umen s:
adiome e –sun pho ome e (a) and adiome e –lida (b) wi h he
R2coe icien s p o ided.
ha e Raman lida signals wi h SNR >3 in o de o de i e
accu a e in eg a ed wa e apo alues. The ag eemen be-
ween lida -de i ed IWV adiome e -de i ed IWV was e y
high (R2=0.98) as shown in Fig. 2 ( igh panel). Also he
Raman lida e ie als we e no signi ican ly biased om he
mic owa e adiome e esul s. In mos o he cases he abso-
lu e di e ence be ween he IWV alues de i ed om hese
wo ins umen s was no highe han 1. We he e o e in e ed
ha he Raman lida calib a ion can be conside ed eliable
and he sys em is sui able o pe o m accu a e wa e apo
mixing a io measu emen s.
We pe o med ano he in e compa ison o alida e bo h
WRF simula ions and ins umen al measu emen s o wa-
e apo . We analyze he absolu e di e ence (%) be ween
high- esolu ion adiosonde obse a ions and each measu e-
men and simula ion echnique (lida , mic owa e adiome-
e , WRF simula ions) sepa a ely. F om Fig. 3 we can see
ha in he lowes oposphe ic laye , lida demons a es he
poo es ag eemen wi h adiosounding measu emen s due o
incomple e geome ical o e lap o he signal. In he lowes
500 m, he median di e ence be ween adiosounding and li-
da wa e apo measu emen s is 151.6 %. P e ious s udies
ha e shown ha he o e lap egion o he EOLE sys em is
a ound 300m (Kokkalis e al., 2012). Howe e , when wa e
apo pa ame e s a e e ie ed, he o e lap issue may a ec
lida measu emen s be ween 500 and 1000m due o di e -
en o e lap cha ac e is ics o he Raman channels a 387 and
407 nm. So, we in es iga ed he bo om laye o lida wa-
e apo measu emen s o ou s udy by a e aging he a-
diosounding mixing a ios om 7 days and calcula ed he
s anda d de ia ion om he mean wa e apo mixing a io
alue. The esul an s anda d de ia ion (0.58) is used o ap-
ply he da a quali y es . We se he al i ude h eshold based
on 3 s anda d de ia ion so ha heigh s whe e he mean di -
e ence be ween lida and adiosounding da a exceeded 1.74
a e no used. Based on his analysis we de e mined he bo -
om heigh o 750m, abo e which he lida measu emen s
Ann. Geophys., 36, 213–229, 2018 www.ann-geophys.ne /36/213/2018/
L. D. Labzo skii e al.: Rela i e humidi y e ical p o iling using lida -based syne gis ic me hods 219
Figu e 3. Absolu e di e ence (%) be ween he mixing a io o wa e
apo om adiosoundings and lida (g een), mic owa e adiome-
e ( ed) and WRF simula ions (blue). All six high- esolu ion a-
diosounding cases om he Hyg A-CD campaign a e used (15, 17,
18, 20, 21 May and 1 June).
o he mixing a io can be accu a e enough o be used o
RH e ie als. Then, when we analyze he ag eemen wi h
adiosounding measu emen s be ween 750 and 6000 m, li-
da shows he bes esul s. Lida – adiosounding median di -
e ence is 11.8 %, while WRF– adiosounding median di -
e ence equals 28.6 % and adiome e – adiosounding me-
dian di e ence eaches 86.6% (Fig. 3). As expec ed, he a-
diome e demons a es easonable ag eemen wi h e e ence
adiosoundings only in he lowes 4000m since abo e his
laye , only 5 % o e ie ed obse a ions o igina e om he
adiome e i sel .
3.2 E alua ion o empe a u e p o iles
We analyzed he ag eemen and sys ema ic bias o he a ail-
able a mosphe ic empe a u e p o iles compa ed o he 6
high- esolu ion adiosondes da a, men ioned in he p e i-
ous sec ion. Only MWR (mic owa e adiome e ) and WRF
model simula ions we e applied, since he EOLE lida sys-
em is no capable o empe a u e measu emen s. Jus like
humidi y, physical empe a u e in he a mosphe e is ela ed
o he b igh ness empe a u e o he objec . The e o e, he
basic p inciple o empe a u e e ie al by he adiome e is
close o an analogical p inciple as applied o humidi y p o-
iling. Tempe a u e a ies e ically a he same a e o bo h
adiome e measu emen s and model simula ions o he con-
side ed da es, and he e is no s ong in e sion by any ins u-
men o simula ion. Bo h he WRF model simula ions and
MWR measu emen s show high ag eemen wi h adiosound-
ings (R2=0.99 and 0.98, espec i ely). Howe e , some mi-
no empe a u e in e sions measu ed by he adiosoundings
a e no e ie ed by he WRF model o he adiome e mea-
su emen s (Fig. 4). Fo mos o he cases, he ag eemen
be ween he WRF model simula ions and adiosoundings
is be e han he ag eemen be ween mic owa e adiome-
e and adiosoundings. In pa icula , he di e ence o em-
pe a u es be ween WRF model simula ions and adiosound-
ings does no exceed 2◦C in he lowes egion o he o-
posphe e. WRF– adiosounding mean absolu e bias (0.45 ◦C)
is lowe han adiome e – adiosounding mean absolu e bias
(1.84 ◦C). Radiome e esul s mee ou expec a ions, acco d-
ing o Sec . 2.3, as he andom e o o empe a u e equals
1.7 ◦C. Conside ing he e y high s a is ical ag eemen be-
ween adiome e and WRF simula ions wi h adiosound-
ings, he low sys ema ic bias and he ai ag eemen wi h
li e a u e alues o he adiome e , we u he used hese
empe a u e da ase s as inpu obse a ions o RH calcula-
ion. The ole o he empe a u e unce ain y e ec s in e-
sul ing RH e ical p o iling echniques a e conside ed in he
Sec . 4.2.3.
4 Combined algo i hms o ela i e humidi y e ical
p o iling: lida – adiome e and lida –WRF me hods
We used wo combined algo i hms o RH e ical p o iling:
lida – adiome e and lida –WRF me hods. The basics o RH
calcula ions a e explained below and p esen ed along wi h
desc ip ion o bo h me hods.
4.1 Lida – adiome e combina ion o ela i e
humidi y e ical p o iling
The syne gis ic lida – adiome e me hod (LD-MWR he e-
a e ) is based on wo inpu da ase s. The i s inpu da ase is
based on lida measu emen s o he wa e apo mixing a-
io. The second inpu da ase is aken om wa e apo and
empe a u e measu emen s om he adiome e . AH p o iles
o lida a e calcula ed om a simple con e sion o mula in
which mixing a io p o iles a e mul iplied wi h ai densi y
p o iles ( aken om he US S anda d A mosphe e). We sep-
a a ed he a mosphe ic column in o h ee di e en egions.
The ange o heigh s in lowes egion (0–750m) is dic a ed
by lida o e lap limi a ions and only adiome e measu e-
men s a e applied below 750m. In he middle egion (750–
2000 m) we combined mixing a io measu emen s om he
lida wi h empe a u e and humidi y measu emen s om he
mic owa e adiome e in o de o e ie e combined mea-
su emen s o RH. The adiome e - e ie ed RH p o iles a e
in e pola ed o lida measu emen s and hen a e aged wi h
app op ia e lida measu emen s, which esul ed in combined
RH p o iles. The combina ion ac s o smoo h he ansi ion
om he egion whe e lida is “blind” o he egion whe e
only lida is used o wa e apo measu emen s (>2000 m).
In o de o e ie e he RH p o ile, we calcula e he sa u a-
ion densi y o wa e apo (es) (Eq. 3a) which can be de ined
www.ann-geophys.ne /36/213/2018/ Ann. Geophys., 36, 213–229, 2018
220 L. D. Labzo skii e al.: Rela i e humidi y e ical p o iling using lida -based syne gis ic me hods
Figu e 4. Tempe a u e p o iles in e compa ison o all heigh s. All six high- esolu ion adiosounding cases om he Hyg A-CD campaign
a e used (15, 17, 18, 20, 21 May and 1 June): adiosoundings e sus adiome e (a, ed) and adiosoundings e sus WRF model simula ions
(b, blue).
as he a io be ween he molecula mass o wa e (M=18 g)
and gaseous cons an o wa e apo (R=0.0623) mul iplied
by he a mosphe ic empe a u e (T) aken om adiome e
obse a ions a he app op ia e heigh s. Then, he a io be-
ween (M) and (RT ) is mul iplied by he empi ical alue o
he wa e apo densi y (ps), he calcula ion o which implies
he use o empe a u e (T) as well (Eq. 3b).
es=M
RT ·ps(T ) (3a)
ps=(0.61078 ·7.501)e17.2694·T
238.3+T(3b)
RH calcula ion includes he implemen a ion o a simple con-
e sion o mula as he a io be ween AH om lida and sa -
u a ion apo densi y calcula ed om Eq. (3a).
4.2 Lida –WRF combina ion o ela i e humidi y
e ical p o iling
Using he lida –WRF (LD-WRF) combined me hod o RH
p o iling we apply he same p inciples and o mulas o he
RH calcula ion as o he LD-MWR me hod. One di e ence
is ha ins ead o mic owa e adiome e measu emen s, we
use WRF simula ions in he same h ee laye s o he opo-
sphe e chosen o analysis. Bo h syne ge ic me hods a e used
simul aneously and compa ed wi h colloca ed adiosound-
ings.
4.3 E alua ion o he ela i e humidi y p o iles
4.3.1 Mic owa e adiome e and WRF simula ions
e sus high- esolu ion adiosounding
We de e mine he e ec i eness o each conside ed RH e i-
cal p o iling me hod and examine adiome e measu emen s
(MWR), WRF model simula ions, LD-MWR and LD-WRF
me hods. A i s , only MWR and WRF model simula ions
a e analyzed e sus RH e ical p o iles calcula ed om a-
diosoundings (Fig. 5). In some cases, MWR measu emen s
do no depic some ine -scale humidi y ea u es in he mid-
dle oposphe e (18 and 21 May, 1 June). The esul s a e
somewha ambiguous in he i s 1000m whe e he ag ee-
men be ween adiosoundings wi h MWR o WRF simula-
ions depend on e e y case. No clea pa e n is iden i ied o
see whe he MWR o WRF simula ions ag ee wi h he a-
diosoundings o no . Fo ins ance, he ag eemen wi h a-
diosoundings is e y simila o MWR and he WRF model
in he case o 15 and 20 May, whe e he di e ence o
bo h echniques does no exceed 10% be ween 100 and
1000 m. Mo eo e , some highe humidi y laye s de ec ed by
adiosondes a e no seen by bo h MWR measu emen s and
WRF simula ions (20 May, a 4500 m, 21 May a 5000 m,
1 June be ween 4000 and 5000 m). These de iciencies ha
a e indi ec ly seen om he bias calcula ion can be alle ia ed
by add essing he combined algo i hms.
4.3.2 Syne ge ic me hods e sus high- esolu ion
adiosounding
The LD-MWR and LD-WRF me hods a e in e compa ed
o he RH e ical p o iling p o ided by high esolu ion a-
diosoundings and he esul s a e shown o 6 days (15 May
o 1 June). In gene al, bo h me hods ag ee qui e well wi h
he adiosonde da a om 1000 o 6000 m heigh ; howe e ,
signi ican di e ences a e e idenced. Fo example on 15 and
17 May he di e ences o bo h me hods o he adiosonde
da a emain qui e la ge be ween 500 and 3000 m. On 20 and
21 May he ag eemen o bo h me hods wi h he RH a-
diosonde da a is e y good in he heigh egion om 1000
o 6000 m. Finally, on 1 June bo h me hods show he bes
ag eemen wi h he adiosonde da a. In Tables 2 and 3 he e -
Ann. Geophys., 36, 213–229, 2018 www.ann-geophys.ne /36/213/2018/
L. D. Labzo skii e al.: Rela i e humidi y e ical p o iling using lida -based syne gis ic me hods 221
Figu e 5. Ve ical p o iles o ela i e humidi y om WRF model simula ions (blue), mic owa e adiome e ( ed) and adiosoundings (g ey).
Table 2. De e mina ion coe icien s (R2) be ween a ious measu emen s echniques o simula ions o ela i e humidi y p o iling and high-
esolu ion adiosounding o di e en heigh egions. LD: lida , MWR: mic owa e adiome e , WRF: WRF model.
R2MWR WRF LD-MWR LD-WRF
Heigh egions Low Mid. Full Low Mid. Full Low Mid. Full Low Mid. Full
15 May 0.79 0.79 0.82 0.55 0.82 0.86 0.79 0.91 0.91 0.55 0.92 0.82
17 May 0.88 0.78 0.78 0.92 0.81 0.77 0.88 0.95 0.94 0.92 0.96 0.91
18 May 0.86 0.20 0.21 0.94 0.82 0.77 0.86 0.92 0.91 0.94 0.92 0.85
20 May 0.52 0.58 0.32 0.40 0.21 0.21 0.52 0.91 0.88 0.40 0.90 0.88
21 May 0.96 0.91 0.88 0.89 0.84 0.87 0.96 0.91 0.87 0.89 0.91 0.93
1 June 0.27 0.94 0.87 0.05 0.90 0.90 0.27 0.96 0.96 0.05 0.90 0.90
Mean 0.71 0.70 0.65 0.63 0.73 0.73 0.71 0.93 0.92 0.63 0.92 0.88
iciency o hese app oaches is u he analyzed by add ess-
ing a s a is ical analysis o he da ase s ob ained.
We calcula ed he R2and he esul ing mean bias o RH
e ical p o iling ob ained by each echnique and he a-
diosonde da a (Tables 2 and 3). The R2 alues o he com-
bined me hods a e ob iously simila o he associa ed sin-
gle me hods (e.g., LD-MWR wi h adiome e and LD-WRF
wi h WRF model simula ions) in he lowes egion. In e es -
ingly, analysis o he middle egion shows ha he use o
lida -de i ed humidi y da a in he combined me hods d as i-
cally inc eases he ag eemen wi h adiosoundings in com-
pa ison o he MWR and WRF model simula ions alone. Fo
ins ance, he mean R2 alue be ween MWR and adiosound-
ings in his egion is 0.70, while he eplacemen o low-
esolu ion humidi y da a om he MWR o lida humidi y
da a imp o es R2(R2=0.93). Howe e , he WRF model-
simula ed RH in he same egion shows easonable ag ee-
men wi h adiosounding da a (R2=0.73). The ag eemen
inc eases when we combine WRF model-simula ed empe -
a u e and lida mixing a io in his egion (R2=0.92). Bo h
examples demons a e ha he ole o he lida da a be-
comes mo e e ec i e in heigh s abo e 1000m in combina-
ion wi h measu emen da a (modeling da a) ob ained (simu-
la ed) wi hin he PBL. The highes R2 alues (R2=0.92) in
he ull conside ed oposphe ic column a e obse ed when
he LD-MWR me hod is used.
The adiome e shows he lowes mean bias (7.09) in he
low egion, while he WRF model shows a poo e ag eemen
wi h adiosoundings (12.32). We obse e a signi ican im-
p o emen in he middle and ull egion when we add ess he
combined algo i hms. Fo example, he mic owa e adiome-
e when used alone shows a mean bias o 13.63 and 11.61
www.ann-geophys.ne /36/213/2018/ Ann. Geophys., 36, 213–229, 2018
228 L. D. Labzo skii e al.: Rela i e humidi y e ical p o iling using lida -based syne gis ic me hods
B.: Rela i e-humidi y p o iling in he oposphe e wi h a Raman
lida , Appl. Op ics, 41, 6451–6462, 2002.
Miloshe ich, L. M., Vomel, H., Whi eman, D. N., and Leblanc, T.:
Accu acy assessmen and co ec ion o Vaisala RS92 adiosonde
wa e apo measu emen s, J. Geophys. Res., 114, D11305,
h ps://doi.o g/10.1029/2008JD011565, 2009.
Mochida, M.: Simul aneous Measu emen s o Hyg oscopic P op-
e y and Cloud Condensa ion Nucleus Ac i i y o Ae osol Pa i-
cles o Ma ine Biogenic O igin, Wes e n Paci ic Ai -Sea In e ac-
ion S udy, 71–81, h ps://doi.o g/10.5047/w-pass.a01.008, 2014.
Nagel, D., Lei e e , U., Die , H., Ka s, A., Reicha d , J.,
and Beh end , A.: High accu acy humidi y measu emen s
using he s anda dized equency me hod wi h a esea ch
uppe -ai sounding sys em, Me eo ol. Z., 10, 395–405,
h ps://doi.o g/10.1127/0941-2948/2001/0010-0395, 2001.
Nash, J., Oakley, T., Vömel, H., and Wei, L.: WMO in e compa ison
o high quali y adiosonde sys ems, Yangjiang, China, WMO,
TD, 1580, 2011.
Na as-Guzmán, F., Fe nández-Gál ez, J., G anados-Muñoz, M.
J., Gue e o-Rascado, J. L., B a o-A anda, J. A., and Alados-
A boledas, L.: T oposphe ic wa e apou and ela i e humid-
i y p o iles om lida and mic owa e adiome y, A mos. Meas.
Tech., 7, 1201–1211, h ps://doi.o g/10.5194/am -7-1201-2014,
2014.
Papayannis, A., A gy ouli, A., Bougia io i, A., Remoundaki, E.,
V a olis, S., Nenes, A., Solomos, S., Komppula, M., Giannakaki,
E., Kalogi os, J., Banks, R., Ele he iadis, K., Man as, E., Di-
apouli, E., Tzanis, C. G., Kazadzis, S., Binie oglou, I., Lab-
zo skii, L., Vande Hey, J., and Ze e os, C. S.: An o e iew om
hyg oscopic ae osols o cloud d ople s: The Hyg A-CD cam-
paign in he A hens basin, Sci. To al En i on., 574, 216–233,
h ps://doi.o g/10.1016/j.sci o en .2016.09.054, 2017.
Pe e s, M. D. and K eidenweis, S. M.: A single pa ame e
ep esen a ion o hyg oscopic g ow h and cloud condensa-
ion nucleus ac i i y, A mos. Chem. Phys., 7, 1961–1971,
h ps://doi.o g/10.5194/acp-7-1961-2007, 2007.
Radel, G. and Shine, K. P.: E alua ion o he use o a-
diosonde humidi y da a o p edic he occu ence o pe sis-
en con ails, Q. J. Roy. Me eo ol. Soc., 133, 1413–1423,
h ps://doi.o g/10.1002/qj.128, 2007.
Rose, T., C ewell, S., Löhne , U., and Simme , C.: A ne -
wo k sui able mic owa e adiome e o ope a ional moni o -
ing o he cloudy a mosphe e, A mos. Res., 75, 183–200,
h ps://doi.o g/10.1016/j.a mos es.2004.12.005, 2005.
Schneide , T., O’Go man, P. A., and Le ine, X. J.: Wa e apo and
he dynamics o clima e changes, Re . Geophys., 48, RG3001,
h ps://doi.o g/10.1029/2009RG000302, 2010.
Schu gens, N., Tsy o, S., G yspee d , E., Go o, D., Weigum, N.,
Schulz, M., and S ie , P.: On he spa io- empo al ep esen a-
i eness o obse a ions, A mos. Chem. Phys., 17, 9761–9780,
h ps://doi.o g/10.5194/acp-17-9761-2017, 2017.
She wood, S. C., Ing am, W., Tsushima, Y., Sa oh, M., Robe s,
M., Vidale, P. L., and O’Go man, P.: Rela i e humidi y
changes in a wa me clima e, J. Geophys. Res., 115, 1–11,
h ps://doi.o g/10.1029/2009JD012585, 2010.
Skama ock, W. C., Klemp, J. B., Dudhia, J., Gill, D. O., Ba ke ,
D., Wang, W., and Powe s, J. G.: A Desc ip ion o he Ad anced
Resea ch WRF Ve sion 2, echnical epo NCAR/TN 468+STR,
h ps://doi.o g/10.5065/D68S4MVH, a ailable a : h p://www2.
mmm.uca .edu/w /use s/docs/a w_ 2_070111.pd (las access:
2 Feb ua y 2018), 2005.
Tang, I. N., Wong, W. T., and Munkelwi z, H. R.: The ela i e im-
po ance o a mosphe ic sulpha es and ni a es o isibili y educ-
ion, A mos. En i on., 15, 2763–2471, 1981.
Theopold, F. A. and Bosenbe g, J.: i e en ial abso p-
ion lida measu emen s o a mosphe ic empe a u e
p o iles: Theo y and expe imen , J. A mos. Oceanic
Technol., 10, 165–179, h ps://doi.o g/10.1175/1520-
0426(1993)010<0165:DALMOA>2.0.CO;2, 1993.
Tomkins, A. M.: Impac o empe a u e and humidi y a iabili y on
cloud co e assessmen using ai c a da a, Q. J. Roy. Me eo ol.
Soc., 129, 2151–2170, 2003.
Tu ne , D. D., Fel z, W. F., and Fe a e, R. A.: Con inuous Wa e Va-
po P o iles om Ope a ional G ound-Based Ac i e and Passi e
Remo e Senso s, B. Am. Me eo ol. Soc., 8, 1301–1317, 2000.
Van Hee waa den, C. C. and de A ellano, V.-G.: Rela i e hu-
midi y as an Indica o o Cloud Fo ma ion o e He e oge-
neous Land Su aces, J. Am. Me eo ol. Soc., 65, 3263–3277,
h ps://doi.o g/10.1175/2008JAS2591.1, 2008.
Veselo skii, I., Whi eman, D. N., Kolgo in, A., And ews, E., and
Ko enskii, M.: Demons a ion o ae osol p ope y p o iling by
mul i-wa eleng h lida unde a ying ela i e humidi y condi-
ions, J. A mos. Ocean. Technol., 26, 1543–1557, 2009.
Wang, L. Z.: Sys ema ic e o s in global adiosonde p e-
cipi able wa e da a om compa isons wi h g ound-
based GPS measu emen s, J. Clima e, 21, 2218–2238,
h ps://doi.o g/10.1175/2007JCLI1944.1, 2008.
Wang, Y., Hua, D., Mao, J., Wang, L., and Xue, Y.: A de ec ion
o a mosphe ic ela i e humidi y p o ile by UV Raman Lida , J.
Quan . Spec . Ra., 112, 214–219, 2011.
Wa e, R., Ca pen e , R., Guldne , J., Liljeg en, J., Neh ko n, T., Sol-
heim, F., and Vandenbe ghe, F. A.: Mul i-Channel Radiome ic
P o iles o Tempe a u e, Humidi y and Cloud Liquid, Radio Sci.,
38, 8079–8032, 2003.
Wei kamp, C.: Lida . Range-Resol ed Op ical Remo e Sensing
o he A mosphe e, ISSSN 0342-4111, 102 Ausg., New Yo k:
Sp inge Science, 2005.
Wes wa e , E.: G ound-based passi e p obing using he mic owa e
spec um o oxygen, T oposphe e kine ic empe a u e s uc u e
om g ound-based measu emen s o oxygen emission spec a
in mic owa e egion, using in e s ion echnique and i e a i e
me hod, J. Res. Na l. Ins . S and. Technol., 69D, 1201–1211,
1965.
Wes wa e , E., C ewell, S., Ma zle , C., and Cimini, D.: P inciples
o Su ace-based Mic owa e and Milime e wa e Radiome e ic
Remo e Sensing o he T oposphe e, Quade ni Della Socie a I al-
iana Di Ele omagne ismo, 1, 52, 2005.
Wex, H., S a mann, F., Hennig, T., Ha mann, S., Niede meie , D.,
Nilsson, E., Ocskay, R., Rose, D., Salma, I., and Ziese, M.: Con-
nec ing hyg oscopic g ow h a high humidi ies o cloud ac i a-
ion o di e en pa icle ypes, En i on. Res. Le ., 3, 035004,
1–10, 2008.
Whi eman, D., Mel i, S., and Fe a e, R.: Raman lida sys em o
he measu emen o wa e apo and ae osols in he Ea h’s a -
mosphe e, Appl. Op ics, 31, 3068–3078, 1992.
Whi eman, D. N., Demoz, B., Di Gi olamo, P., Come , J.,
Veselo skii, I., E ans, K., Wang, Z., Cadi ola, M., Rush, K.,
Schwemme , G., Gen y, B., Mel i, S. H., Mielke, B., Ven-
Ann. Geophys., 36, 213–229, 2018 www.ann-geophys.ne /36/213/2018/
L. D. Labzo skii e al.: Rela i e humidi y e ical p o iling using lida -based syne gis ic me hods 229
able D., and an Ho e, T.: Raman Lida Measu emen s du -
ing he In e na ional H2O P ojec . Pa I: Ins umen a ion and
Analysis Techniques, J. A mos. Ocean. Tech., 23, 157–169,
h ps://doi.o g/10.1175/JTECH1838.1, 2006.
Whi eman, D. N., Rush, K., Rabenho s , S., Welch, W., Cadi ola,
M., McIn i e, G., Russo, F., Adam, M., Venable, D., Con-
nell, R., Veselo ski, I., Fo no, R., Mielke, B., S ein, B.,
Leblanc, T., McDe mid, S., and Vomel, H.: Ai bo ne and
G ound-Based Measu emen s Using a High-Pe o mance Ra-
man Lida , J. A mos. Ocean. Technol., 27, 1781–1800,
h ps://doi.o g/10.1175/2010JTECHA1391.1, 2010.
Whi eman, D. N., Venable, D. D., and Landul o, E.: Commen s on
accu acy o Raman lida wa e apo calib a ion and i s applica-
bili y o long- e m measu emen s, Appl. Op ics, 50, 2170–2176,
h ps://doi.o g/10.1364/AO.47.005592, 2011.
Wul meye , V., Ha des y, R. M., Tu ne , D. D., Beh end , A.,
Cadeddu, M. P., Di Gi olamo, P., Schlussel, P., Van Baelen, J.,
and Zus, F.: A e iew o he emo e sensing o lowe oposphe ic
he modynamic p o iles and i s indispensable ole o he unde -
s anding and he simula ion o wa e and ene gy cycles, Re .
Geophys., 53, 819–895, h ps://doi.o g/10.1002/2014RG000476,
2015.
Ziege , P., Fie z-Schmidhause , R., Weinga ne , E., and Bal-
enspe ge , U.: E ec s o ela i e humidi y on ae osol ligh
sca e ing: esul s om di e en Eu opean si es, A mos. Chem.
Phys., 13, 10609–10631, h ps://doi.o g/10.5194/acp-13-10609-
2013, 2013.
www.ann-geophys.ne /36/213/2018/ Ann. Geophys., 36, 213–229, 2018
