Interannual variability of
tropospheric water vapour
Mark McCarthy (Met Office)
Ralf Toumi (Imperial college, London)
Simon Tett (Met Office)
Thanks to Darren Jackson (etl, NOAA)
1
03/0045b
© Crown copyright
Introduction
2
„
Humidity is a critical component of the climate
system.
„
Variability is poorly documented on annual to
decadal timescales.
„
Principle mode of variability is related to ENSO.
„
Look at regional aspects of humidity response to
ENSO by season. Review interpretations of
trends.
„
Caution: we are discussing relative humidity not
absolute humidity.
03/0045b
© Crown copyright
Data
„
High-resolution InfraRed Sensor (HIRS)
channel 12 measures 6.7µm water vapour
emission.
„
Sensitive to RH and temperature between
approx. 500hPa and 200hPa - FTRH.
„
Major uncertainties stem from
– Inter-satellite bias.
– Conversion algorithm.
– Cloud clearing.
3
03/0045b
© Crown copyright
HIRS Relative Humidity Climatology − DJF 1979−98
30N
0
30S
0
90E
10
20
180
Relative Humidity %
30
40
90W
50
60
0
70
HIRS Relative Humidity Climatology − JJA 1979−98
30N
0
30S
0
90E
10
20
180
Relative Humidity %
30
40
90W
50
60
0
70
Mean Distribution of FTRH (%) for DJF (top) and JJA (bottom)
4
03/0045b
© Crown copyright
EOF analysis
„
EOF analysis of 20-year time-series of each
season. Only physically identifiable mode is related
to ENSO.
„
DJF:
– Variance explained - 25%
– PC time-series - Niño3.4 r = 0.87
„
JJA:
– Mixed modes (15%, 13% of variance)
– PC1 time-series - Niño3.4 r = 0.75
– PC2 time-series - Niño3.4 (lead 2 seasons) r = -0.72
5
03/0045b
© Crown copyright
Composite difference analysis
„
Select strongest +ve and -ve events based on Niño3.4
index of pacific SSTs
„
NCEP re-analysis provides velocity potential, wind vector,
and temperature fields
„
Caution: few strong La Niña in observed period. Two
‘protracted’ El Niño events
El Niño
La Niña
1982/83
1987/88
1991/92
1994/95
6
03/0045b
© Crown copyright
1984/85
1988/89
1995/96
Winter (DJF) Composite
2
0
0
−0
0.5
−1
−3
3
1
0.5
0
.5
0
30N
−0.5
−2
Composite El Nino − La Nina DJF UTRH
5
01.
0
2
0.5
0
180
UTRH (%)
5
−10
−5
−3
−0.5
1
90E
−1
30S
−2
0
−0.5
0
−4
−1
4
0.5
3
1
0
90W
3
5
0
10
Boreal winter (DJF) composite difference map of HIRS UTRH (filled
colour) for significant regions at the 5% level. Over-plotted are 0.2σ
level velocity potential contours, (*106m2/s) solid contours are positive,
dashed contours are negative and vector wind arrows at 300hPa.
7
03/0045b
© Crown copyright
Separating Temperature and
Water Vapour
„
Define RH as:
„
We find:
„
ea′ + ea
RH ′ + RH =
es′ + es
es′
RH t′ = − RH
es
and
RH w′ =
Substituting and rearranging:
RH t′RH w′
RH ′ = RH t′ + RH w′ +
RH
8
03/0045b
© Crown copyright
es
RH − RH
es
Estimated temperature component of FTRH composite
3 5
0
−3
−5
−3
−10
−3
30oS
−10
−5
5 5
−5
−10
−5
−3
−3
3
5
3
o
180
Relative humidity (%)
−3
3
5
© Crown copyright
−−35
10
03/0045b
90 E
10
5
10
5
o
3
10
5
3
3
3
3
−5
−5
5
0
9
10
5
53
5
3
5
−10
3
−5
−5
−3
0
−1
−5
5
30 S
3
5
5
5
−−3
o
3
−−35
−3−5
−3
0
−3
0
−3
(b)
−3
3
30 N
90oW
180
Relative humidity (%)
Estimated water vapour component of FTRH composite
3
53
−3
5
3
90oE
0
o
−5
−1 −5
0
−3
(a)
5
−3
30 N
−3
−5 −3
3
o
3
90 W
5
0
10
Upper tropospheric
temperature
„
Estimate sat. vap. press. from NCEP reanalysis
temperatures.
„
50%-70% of RH anomaly in East Pacific can be
explained through temperature changes.
„
30°N-30°S El Niño-La Niña DJF
– Composite difference -0.62 (RHt=-2.4, Rhw=+1.87)
– Extreme events 82/83 (-1.11), 88/89 (+1.79)
dominated by RHt.
10
03/0045b
© Crown copyright
Summer (JJA) composite
−0
−1
2
90E
0
180
UTRH (%)
2.5
−10
5
−1
1
0
−1
0
0.
0.5
30S
−5
−3
00.
5
−1 −0.5
2
0
0
11
.5
1
0.5
−1
30N
−0.5
Composite El Nino − La Nina JJA (after peak) UTRH
90W
3
5
0
10
„
Delayed response over Indian and Atlantic
oceans
„
Regions of strong horizontal humidity gradients
03/0045b
© Crown copyright
Composite El Nino − La Nina JJA (after peak) SST
.5
30N
5
−0.
−0
0.75
0.5
0.175
0.5
5
0.
0.5
0.75
0
0.5
.75
01
−0.5
0
90E
−1
0.5
0
0.75
1
3
53
−5
30S
0
90E
−10
5
−3
−3
35
−5 −10
3
5
35
3
5
5
−5
10
−5
−3
−5
−10
3
53
35
3
−3
−3
−3
−5
−3
5
180
OLR (W/m^2)
−5
−5
−3
3
−5
−3
5
−3
5
−3
−10
−3
−10
5
3
53
−3
3
−3
−5
3
05
−3
© Crown copyright
−0.5
90W
−3
−3
−0.75
3
30N
180
Temperature (Deg C)
Composite El Nino − La Nina JJA (after peak) OLR
−3
03/0045b
.5
0.5
30S
12
−0
0.5 0
.5
−0.5
90W
5
−5
−3
0
10
Ocean-atmosphere interaction
Correlations
SST (JJA)
Nino3.4
(DJF)
UTH (JJA)
0.60
0.49
Nino3.4 (DJF) 0.72
13
„
Remote ocean warming. ref: Klein et al (April 1999), Xie
et al (April 2002) both J.Clim.
„
Anomalous convection in the Indian ocean.
„
No significant correlation with Indian rainfall data.
03/0045b
© Crown copyright
Long-term
Trends
Global UTRH Trends 03/79−08/97
60oN
30oN
(a)
0
30oS
60oS
0
o
−0.2
60oN
−0.1
0
0.1
0.2
Zonal UTRH trends 03/79−08/98
40oN
Latitude
20oN
(b)
0
20oS
40oS
60oS
−0.15 −0.10 −0.05 −0.00 0.05
Trend (%yr−1)
14
03/0045b
© Crown copyright
0
90 W
180
%yr−1
90 E
−0.3
o
0.10
0.15
0.3
„ No significant global
trend
„ Cancellation of
significant regional
trends
„ Stable to removal of
ENSO
Zonal trends
Monthly mean anomalies 10degN to 10degS
UTRH anomaly (%RH)
4
2
0
−2
−4
−6
1975
1980
1985
1990
1995
2000
1995
2000
Year
Monthly mean anomalies 25degS to 45degS
UTRH anomaly (%RH)
6
4
2
0
−2
−4
−6
1975
1980
1985
1990
Year
„
15
03/0045b
Intensified Hadley circulation not adequate to describe
time period of zonal mean trends.
© Crown copyright
Model Evaluation
16
„
Comparison with set of HadAM3 simulations
forced with SSTs and greenhouse gases.
„
Overactive hydrological cycle in HadAM3
„
ENSO represents equivalent fraction of model
inter-annual variance.
03/0045b
© Crown copyright
Model Evaluation
60oN
HIRS and HadAM3 FTRH trends 03/79−08/98
40oN
„
Model does not
reproduce HIRS12
trends.
„
Small negative
trends at southern
midlatitudes.
Latitude
20oN
0
20oS
40oS
60oS
−0.150
17
03/0045b
−0.075
© Crown copyright
0.000
Trend (%yr−1)
0.075
0.150
Conclusions
„
Simple changes in ENSO and hydrological
cycle are not adequate to describe trends in
tropical UTRH.
„
Issues surrounding ‘trends’:
– Further analysis of inter-satellite calibration
issues. Lessons learned from MSU.
– Uncertainties in temperature trends.
– Potential decadal scale ocean-atmosphere
variations projected onto 20 year time series.
18
03/0045b
© Crown copyright
Conclusions
„
„
„
19
HIRS12 useful for diagnosing seasonal to interannual scale variability of FTRH both from
observations and in GCMs.
Further research into temperature-humidity
relationships in the tropical atmosphere on long
time-scales required.
Current and future HIRS instruments place channel
12 higher in the troposphere - compromising the
stability of HIRS12 datasets for use in climate
research.
03/0045b
© Crown copyright