Validation of Integrated Water Vapor from GOME-2

Validation of Integrated Water Vapor from GOME-2

Validation of water vapor from GOME-2 satellite instrument
against reference GPS data at the Iberian Peninsula
Javier Vaquero-Martı́nez1, Manuel Antón1, Roberto Román2, José P. Ortiz de Galisteo3,4,
Victoria E. Cachorro3, Diego Loyola5
1
5
Departamento de Fı́sica, Universidad de Extremadura, Spain
2
Departamento de Fı́sica Aplicada, Universidad de Granada, Spain
3
Grupo de Óptica Atmosférica (GOA), Universidad de Valladolid, Spain
4
Agencia Estatal de Meteorologı́a (AEMET), Valladolid, Spain
Remote Sensing Technology Institute (IMF), German Aerospace Center (DLR), Oberpfaffenhofen, Germany
This
work
by
was
the
supp
of E
S
orte
cono panish
d
itive
my
ness
and Ministry
CGL
C
2014 through ompet
-562
55-C pro jec
t
2.
Introduction
Water vapor plays a fundamental role in atmospheric processes (radiative transfer, energy transport, greenhouse effect, and so on). However, it is one of the most variable gases in the
atmosphere, both temporally and spatially. Thus, it is difficult to provide good measurements of the integrated water vapor (IWV).
Among the numerous methods for IWV retrieval, GPS has proven to be one of the most reliable. Low costs of GPS receivers allow to have dense networks of these devices. However, GPS
networks spatial resolution is not enough for applications like weather forecast or global climate studies. In those cases, satellite measurements seem to be more suitable. Nevertheless,
satellite measurements have two problems: the temporal resolution (polar orbiting satellites overpass the Iberian Peninsula once or twice a day) and the unreliability of measurements
under cloudy conditions.
Thus, in this work the satellite instrument Global Ozone Monitoring Experiment - 2 (GOME-2) has been validated using GPS ground-based stations as reference. Nine stations were
selected in the interior of the Iberian Peninsula in the period 2007-2012.
GOME-2
• Ozone
profile,
integrated
water vapor, sulfur
dioxide, . . .
Results - Dependence of
with variables
Results - Dependence of IQR[δ]
with variables
δ̄
• GOME
Data
Processor (GDP)
by DLR-IMF
●
28
20
• DOA technique.
0
●
●
●
●
●
●
24
●
●
●
●
●
●
●
●
●
●
20
●
●
●
●
●
●
●
●
−20
• Pixel resolution:
80 × 40km
●
●
IQR[δ(%)]
δ(%)
• Covers Earth surface in 1.5 days
●
de IW
pe V
nd
en
ce
●
●
●
●
●
0
10
20
30
40
0
10
20
IWV(mm)
Subsets
●
All
●
High SZA (>47.5º)
●
N ● 250 ● 500
750
●
Low SZA (<47.5º)
Subsets
● 1000 ● 1250
●
All
●
High SZA (>47.5º)
●
• Instituto Geográfico Nacional ⇒ EUREF
●
●
●
●
●
●
●
●
●
●
●
18
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
Low SZA (<47.5º)
●
●
●
●
20
IQR[δ(%)]
SZA
dependence
0
40
● 1000 ● 1250
●
25
δ(%)
750
●
●
●
• Temperature and pressure from AEMET ⇒
Interpolation
●
N ● 250 ● 500
GPS stations
• Measurement of ZTD ⇒ IWV
30
IWV(mm)
●
●
●
●
●
●
−25
●
●
●
●
16
●
●
●
●
• Cloud free conditions
20
40
●
●
60
80
20
40
60
SZA(º)
Subsets
●
All
●
High IWV (>13.7mm)
N ● 200
45.0
●
400
80
SZA(º)
●
Low IWV (<13.7mm)
Subsets
● 600 ● 800
All
●
High IWV (>13.7mm)
●
N ● 200
40
●
400
●
Low IWV (<13.7mm)
● 600 ● 800
●
●
●
●
●
42.5
leon
●
rioj
vala
sala
vill
cace sons
21
●
●
lat
40.0
●
teru
0
●
●
● ●
●
● ● ●
●
●
●
●
●
●
●
●
●
●
●
●
−20
●
●
●
37.5
●
coba
●
●
●
●
●
●
●
●
●
l
a
n ce
o
s
n
a
e
e
S nd
e
p
de
●
IQR[δ(%)]
●
δ(%)
●
●
●
●
20
●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
18
●
●
●
●
●
●
●
●
●
15
●
−40
●
●
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
Jan
Feb
Mar
Apr
May
Month
N ● 100 ● 200
●
300
Jun
Jul
Aug
Sep
Oct
Nov
Month
● 400 ● 500
N ● 100 ● 200
●
300
● 400 ● 500
35.0
●
−12
−8
−4
0
lon
Subsets
●
All
high IWV (>13.7 mm) &
high SZA (>47.5º)
●
●
δi = 100 ·
wisat − wiGPS
wiGPS
• w = IWV
measured by
satellite
or
GPS.
* Median: δ̄
* Interquartile range: IQR[δ]
●
low IWV (<13.7 mm) &
low SZA (<47.5º)
●
Subsets
●
All
high IWV (>13.7 mm) &
high SZA (>47.5º)
●
●
high IWV (>13.7 mm) &
low SZA (<47.5º)
low IWV (<13.7 mm) &
high SZA (>47.5º)
●
low IWV (<13.7 mm) &
low SZA (<47.5º)
Conclusions
Methodology
• index i = certain location
and time
high IWV (>13.7 mm) &
low SZA (<47.5º)
low IWV (<13.7 mm) &
high SZA (>47.5º)
1. Low IWV data
performs worse
than high: there
is more underestimation (negative
median)
and higher variability (higher
IQR).
2. GOME-2 tends
to
underestimate IWV, specially when it
is small and/or
SZA is high.
The influence of
SZA in this underestimation is
very strong.
3. Variability (IQR)
decreases with
increasing IWV,
but it does not
have a clear
dependence on
SZA.
4. Seasonal
behaviour clearly
influenced
by
SZA and IWV
values troughout the year.
GOME-2 measurements
at
summer
perform
better
than at winter.
LATEX Tik Zposter