Temperature Inversions
- increasing temperature with height
- we’ll look at low-level inversions near surface
Inversion Layer Schematic
Temperature
Increasing with
Height
Inversion Layer Schematic
Temperature
Increasing with
Height
Surface-Based (Radiation)
Inversion (Zbase = 0)
Subsidence Inversion
Temperature
Increasing with
Height
Weather Balloons (Radiosondes)
- Obtain vertical profile of temperature, humidity, wind, and pressure
- Routinely launched in California at Oakland, Vandenberg and San Diego
12Z (4am local) and 00Z (4pm local)
KSAN Radiosonde
July - August
650
00GMT (4PM)
06GMT (10PM)
12GMT (4AM)
18GMT (10AM)
700
Pressure
750
800
850
900
950
1000
12
14
16
18
20
22
24
26
Temperature ( C)
Mean temperature profile at San Diego
as a function of time of day.
Temperature Inversion Measures
KSAN Radiosonde
July - August
650
Possible Measures:
700
DTINV = TTOP - TBASE
T850 = Temperature at 850 mb
DTDZ = lapse rate within inversion
PBASE = Inversion base pressure
750
Pressure
DT850 = T850 - T2M
00GMT (4PM)
06GMT (10PM)
12GMT (4AM)
18GMT (10AM)
800
850
900
950
1000
12
14
16
18
20
22
24
26
Temperature ( C)
Mean temperature profile at San Diego
as a function of time of day.
Temperature Inversion Measures
KSAN Radiosonde
July - August
650
Possible Measures:
700
DTINV = TTOP - TBASE
DTINV
- actual temp diff across inversion
- hard to obtain from GCM results
- undefined when no inversion present
750
Pressure
DT850 = T850 - T2M
00GMT (4PM)
06GMT (10PM)
12GMT (4AM)
18GMT (10AM)
800
850
900
950
DT850
- somewhat of an approximation
- easily obtained from GCM results
- always yields a defined value
1000
12
14
16
18
20
22
24
26
Temperature ( C)
Mean temperature profile at San Diego
as a function of time of day.
Sample Temperature Profile
DTINV
3500
Original Algorithm:
- looked for “largest” inversion
with no inflection points
==> A - A’
3000
Height (m)
2500
2000
1500
1000
A’
A
500
0
-8
-7
-6
-5
-4
Temperature
-3
-2
-1
Sample Temperature Profile
DTINV
3500
Original Algorithm:
- looked for “largest” inversion
with no inflection points
==> A - A’
3000
Height (m)
2500
2000
B’
1500
1000
New Algorithm:
- use all radiosonde levels and
scan all possible values of DT to
find largest inversion that could
include infection points
==> B - B’
A’
Minor difference in results
using New Algorithm
A
500
B
0
-8
-7
-6
-5
-4
Temperature
-3
-2
-1
Inversions vary seasonally, but are a dominant feature in
California air basins
B) Inversion Frequency KOAK00Z
100
80
80
Frequency (%)
Frequency (%)
A) Inversion Frequency KOAK12Z
100
60
40
20
40
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
DEC
JAN
FEB
MAR
APR
MAY
AUG
SEP
OCT
C) Inversion Frequency KVBG12Z
D) Inversion Frequency KVBG00Z
80
80
60
40
20
NOV
DEC
Subsidence Inversions
All Inversions
60
40
20
0
0
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
DEC
JAN
FEB
MAR
APR
MAY
Month
JUN
JUL
AUG
SEP
OCT
NOV
DEC
NOV
DEC
Month
E) Inversion Frequency KSAN12Z
F) Inversion Frequency KSAN00Z
100
100
80
80
Frequency (%)
Frequency (%)
JUL
Month
100
60
40
20
60
40
20
0
JAN
JUN
Month
100
JAN
Radiation Inversions
0
FEB
Frequency (%)
Frequency (%)
60
20
0
JAN
Based on DTINV
1960-2007
0
FEB
MAR
APR
MAY
JUN
JUL
Month
AUG
SEP
OCT
NOV
DEC
JAN
FEB
MAR
APR
MAY
JUN
JUL
Month
AUG
SEP
OCT
Inversions vary seasonally, but are a dominant feature in
California air basins
B) Inversion Frequency KOAK00Z
100
80
80
Frequency (%)
Frequency (%)
A) Inversion Frequency KOAK12Z
100
60
40
20
40
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
DEC
JAN
FEB
MAR
APR
MAY
AUG
SEP
OCT
C) Inversion Frequency KVBG12Z
D) Inversion Frequency KVBG00Z
80
80
60
40
20
NOV
DEC
Subsidence Inversions
All Inversions
60
40
20
0
0
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
DEC
JAN
FEB
MAR
APR
MAY
Month
JUN
JUL
AUG
SEP
OCT
NOV
DEC
NOV
DEC
Month
E) Inversion Frequency KSAN12Z
F) Inversion Frequency KSAN00Z
100
100
80
80
Frequency (%)
Frequency (%)
JUL
Month
100
60
40
20
60
40
20
0
JAN
JUN
Month
100
JAN
Radiation Inversions
0
FEB
Frequency (%)
Frequency (%)
60
20
0
JAN
Based on DTINV
1960-2007
0
FEB
MAR
APR
MAY
JUN
JUL
Month
AUG
SEP
OCT
NOV
DEC
JAN
FEB
MAR
APR
MAY
JUN
JUL
Month
AUG
SEP
OCT
Inversions vary seasonally, but are a dominant feature in
California air basins
B) Inversion Frequency KOAK00Z
100
80
80
Frequency (%)
Frequency (%)
A) Inversion Frequency KOAK12Z
100
60
40
20
40
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
DEC
JAN
FEB
MAR
APR
MAY
AUG
SEP
OCT
C) Inversion Frequency KVBG12Z
D) Inversion Frequency KVBG00Z
80
80
60
40
20
NOV
DEC
Subsidence Inversions
All Inversions
60
40
20
0
0
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
DEC
JAN
FEB
MAR
APR
MAY
Month
JUN
JUL
AUG
SEP
OCT
NOV
DEC
NOV
DEC
Month
E) Inversion Frequency KSAN12Z
F) Inversion Frequency KSAN00Z
100
100
80
80
Frequency (%)
Frequency (%)
JUL
Month
100
60
40
20
60
40
20
0
JAN
JUN
Month
100
JAN
Radiation Inversions
0
FEB
Frequency (%)
Frequency (%)
60
20
0
JAN
Based on DTINV
1960-2007
0
FEB
MAR
APR
MAY
JUN
JUL
Month
AUG
SEP
OCT
NOV
DEC
JAN
FEB
MAR
APR
MAY
JUN
JUL
Month
AUG
SEP
OCT
Distribution of Daily Values
B) San Diego / SC
Dθ850
DTINV
ZBASE
DZINV
Red = 00Z (local afternoon)
Blue = 12Z (local morning)
Distribution of Daily Values
B) San Diego / SC
Dθ850
DTINV
ZBASE
DZINV
Red = 00Z (local afternoon)
Blue = 12Z (local morning)
Oakland / SJV
San Diego / SoCal
Dθ850 =>
DTINV =>
Seasonal variations differ between the 2 temperature measures
DTINV vs Dθ850 (Daily Anomaly)
DTINV vs DTH850 KOAK12Z
DAILY ANOMALIES JAN-DEC
DTINV vs DTH850 KSAN12Z
DAILY ANOMALIES JAN-DEC
20
15
15
DTH850 Anomaly (°C)
Better agreement at 12Z (local morning)
R = 0.7 to 0.8
DTH850 Anomaly (°C)
Correlations > 0.60
10
10
5
0
-5
5
0
-5
-10
-10
-15
-15
y = -0.090241 + 1.1487x R= 0.73203
y = -0.63132 + 1.1409x R= 0.8018
-20
-20
-15
-10
-5
0
5
10
15
20
-10
-5
DTINV Anomaly (°C)
During afternoon, better comparison
if limit to lower inversions ZBASE < 1000m
0
5
10
15
DTINV Anomaly (°C)
DTINV vs DTH850 KOAK00Z
DAILY ANOMALIES JAN-DEC
DTINV vs DTH850 KSAN00Z
DAILY ANOMALIES JAN-DEC
20
20
15
15
DTH850 Anomaly (°C)
DTH850 Anomaly (°C)
10
5
0
-5
10
5
0
-5
-10
-10
-15
y = -1.3967 + 0.93754x R= 0.64775
y = -1.1489 + 0.86642x R= 0.61116
-20
-15
-10
-5
0
5
10
15
20
-10
-5
DTINV Anomaly (°C)
DTINV vs DTH850 KOAK00Z
DAILY ANOMALIES JAN-DEC
ZBASE < 1000m
20
0
5
10
15
20
DTINV Anomaly (°C)
DTINV vs DTH850 KSAN00Z
DAILY ANOMALIES JAN-DEC
ZBASE < 1000m
20
15
15
DTH850 Anomaly (°C)
DTH850 Anomaly (°C)
10
5
0
-5
10
5
0
-5
-10
-15
-10
y = -0.30943 + 1.0388x R= 0.82244
y = -0.36939 + 0.97538x R= 0.7865
-20
-10
-5
0
5
10
DTINV Anomaly (°C)
15
20
-15
-10
-5
0
5
10
DTINV Anomaly (°C)
15
20
Oakland
San Joaquin
Air Basin
Vandenberg
South Coast
Air Basin
San Diego
Question:
How representative are radiosondes
at San Diego and Oakland to South
Coast and San Joaquin Air Basins?
Oakland
San Joaquin
Air Basin
Vandenberg
South Coast
Air Basin
San Diego
Note on Dθ850:
Dθ850 = T850 - T2M
==> Sounding locations not in air basins of interest
T2M obtained from available coop stations in each
air basin ==> should be more representative
Compared surface temperatures from several coop stations within
each air basin
- temps are well correlated throughout each air basin
Table 1. Maximum/Minimum inter-station θ2M correlations for both daily values and monthly
means. Surface temperatures are from available cooperative observer stations within Southern
California and the San Joaquin Valley.
Daily
Monthly
NDJF
JJAS
NDJF
JJAS
SC θ2M 00Z
89/68
88/54
91/77
84/56
SC θ2M 12Z
82/69
79/58
83/76
80/64
SJV θ2M 00Z
90/73
94/83
93/85
90/80
SJV θ2M 12Z
87/72
83/64
92/83
83/50
Compared 850 hPa temperatures from available radiosondes.
- some limited radiosonde records at Los Angeles (Long Beach and
Santa Monica) and Merced
- temps are again very well correlated throughout
Table 2. Correlations of T850 between radiosonde locations for both daily values and monthly
means. No overlap (NO) indicates there were less than 30 days in common between a particular
station pair.
Daily Values
Los Angeles
Vandenb erg
Oakland
Merced
NDJF
JJAS
NDJF
JJAS
NDJF
JJAS
NDJF
JJAS
San Diego
94
92
88
84
75
65
83
N.O.
Los Angeles
N.O.
N.O.
82
73
87
N.O.
Vandenb erg
87
85
N.O.
N.O.
Oakland
94
N.O.
High correlations also for Monthly Values
Compared 850 hPa temperatures from available radiosondes.
- some limited radiosonde records at Los Angeles (Long Beach and
Santa Monica) and Merced
- temps are again very well correlated throughout
Table 2. Correlations of T850 between radiosonde locations for both daily values and monthly
means. No overlap (NO) indicates there were less than 30 days in common between a particular
station pair.
Daily Values
Los Angeles
Vandenb erg
Oakland
Merced
NDJF
JJAS
NDJF
JJAS
NDJF
JJAS
NDJF
JJAS
San Diego
94
92
88
84
75
65
83
N.O.
Los Angeles
N.O.
N.O.
82
73
87
N.O.
Vandenb erg
87
85
N.O.
N.O.
Oakland
94
N.O.
High correlations also for Monthly Values
California Reanalysis Downscaling at 10-km (CaRD10)
Cross-Correlation (Spatial Coherence) of daily mean values of Dθ850
California Reanalysis Downscaling at 10-km (CaRD10)
Cross-Correlation (Spatial Coherence) of daily mean values of Dθ850
Inversion Magnitude vs. Pollutant Concentration
Pollution measurements from California Air Resources Board (CARB)
website at: www.arb.ca.gov/aqmis2/aqinfo.php
Must view results with caution, as some pollutants like ozone are
temperature dependent.
Dθ850 vs Pollution Concentration
Temperature inversions and pollution
0.8
Correlation of Daily Means
Inversion Measure vs Pollutant
Month ==JUNE
June
MONTH
Ozone
0.7
PM-SO4
Correlation
0.6
0.5
0.4
0.3
0.2
0.1
0
DTINV
DT850
T850
PBASE
Inversion Measure
DTDZ
Relationship of inversion strength to large-scale circulation
Model data from:
• NCEP Reanalysis 2 (2.5° x 2.5°)
- similar resolution to most climate models
- hindcast
- incorporates available observations
- represents best estimate of atmospheric state 1979-present
Composite Daily Atmospheric Patterns During Strong/Weak Inversion Events
- examine weather balloon data at Oakland (Jun-Aug 1979-2001)
- find the 30 events with largest/smallest inversion magnitudes
- examine mean large-scale circulation for these 30 events
- consider anomalies (departure from long-term average)
500mb Height and Wind Anomalies
Strong Inversions
at Oakland
Weak Inversions
at Oakland
- Strong inversions associated with above normal 500mb heights (large-scale high pressure systems)
- Weak inversions associated with below normal 500mb heights (large-scale low pressure systems)
===> Inversions in California associated with large-scale circulation
Relationship of inversion strength to large-scale and
regional-scale circulation
Model data from:
• NCEP Reanalysis 2 (2.5° x 2.5°)
- similar resolution to most climate models
- hindcast
- incorporates available observations
- represents best estimate of atmospheric state 1979-present
• California Reanalysis Downscaling at 10km (CaRD10)
- dynamical downscaling
DOWNSCALED COMPOSITE MEANS JUN-AUG
SURFACE WIND AND INVERSION MAGNITUDE
ACTUAL VALUES (NOT ANOMALIES)
STRONG INVERSIONS AT OAKLAND
WEAK INVERSIONS AT OAKLAND
Large-Scale 500mb Height Difference
2
2
1
Define DH500 = H500,reg1 - H500,reg2 using historical analysis data
How does this large-scale variable relate to local inversion measures
throughout California? On daily timescales? Monthly timescales?
HOW DO LOCAL INVERSION MAGNITUDES COMPARE
TO LARGE-SCALE FEATURES?
CORRELATION OF DAILY MEANS
Downscaled Inversion Magnitude vs. Large-Scale 500mb Height Difference
• KOAK R = 0.55
• KOAK R = 0.57
• KSAN R = 0.51
Correlation
• KSAN R = 0.34
Correlation
California Inversion Index GFDL A2
500hPa height diff, Elko minus Churchill
25
20
n
15
10
5
0
1900
1950
2000
2050
2100
yr
Figure 9. Frequency (5-year running total) of positive ² h500 anomalies exceeding 1.0 standard deviations from the
SRES A2 runs of the GFDL CM2.1 model. Here ²h 500 is defined as the difference in 500 mb height between 42°N,
115°W (Elko) and 60°N, 95°W (Churchill). The anomalies are referenced to the 1961-1990 climatology.
TRENDS OVER 48 YEARS (1960-2007)
( ) = CHANGE OVER 48 YEARS / STANDARD DEV OF ANNUAL MEANS
dtinv
dzinv
ztop
zbase
Site
00Z
12Z
00Z
12Z
00Z
12Z
00Z
12Z
brownsville -0.3(149 %) -0.4(145 %) -118.9(279 %) -97.0(253 %) -79.5(133 %) -97.7(109 %) 39.4( 74 %) -0.7( %)
tucson
-0.3( 84%) 0.9(153 %) -142.9(214 %) -132.0(189 %) -158.2( 84%) -456.6(190 %) -15.3( 9%) -324.6(168 %)
amarillo
-0.1( 24%) 0.2( 43%) -142.0(254 %) -114.4(258 %) -165.7(145 %) -156.1(232 %) -23.7( 27 %) -41.7(100 %)
albuquerque -0.1( 20%) 0.2( 68%) -94.5(191 %) -87.4(188 %) -104.2(131 %) -169.4(211 %) -9.7( 18 %) -82.0(145%)
dodge city -0.3( 85%) -0.4( 85%) -109.7(245 %) -121.3(235 %) -99.8(176 %) -118.1(204 %) 9.9( 18%) 3.2( 7%)
denver
-0.3( 63%) 0.2( 41%) -95.6(182 %) -76.4(177 %) -116.2(142 %) -78.4(114 %) -20.6( 37%) -2.3( 4%)
grand junct -0.1( 30%) -0.2( 57%) -154.9(195 %) -116.5(146 %) -155.6(150 %) -182.9(117 %) -25.9( 30 %) -67.6( 64 %)
north platt -0.2( 71%) -0.1( 29%) -96.9(239 %) -128.3(246 %) -70.8(128 %) -171.4(221 %) 26.0( 51 %) -43.1( 87%)
saltlake c 0.2( 19%) 0.4( 36%) -155.7(198 %) -97.2(162 %) -143.7(130 %) -111.4(121 %) 12.0( 12 %) -14.2( 25 %)
medford
-0.6(187 %) -1.2(235 %) -140.3(242 %) -109.4(214 %) -262.8(227 %) -339.5(227 %) -123.8(138 %) -230.0(198 %)
rapid city -0.2( 40%) -0.1( 29%) -119.4(249 %) -107.8(244 %) -228.6(246 %) -154.6(236 %) -111.3(163 %) -46.8(117 %)
boise
-0.3( 48%) 0.7(136 %) -160.5(173 %) -190.1(224 %) -228.9(175 %) -339.4(191 %) -68.7( 63 %) -150.5(117 %)
salem
0.0( 1%) -0.2( 42%) -206.8(204 %) -198.2(199 %) -156.4(174 %) -384.3(272 %) 50.4( 68 %) -186.3(169 %)
glasgow
-0.9(177 %) -0.8(163 %) -150.8(253 %) -109.5(207 %) -186.8(188 %) -162.1(184 %) -36.0( 40%) -52.6( 89 %)
great falls -1.2(132 %) -0.2( 21%) -75.0(164 %) -49.7(105 %) -123.1(115 %) -152.0(143 %) -48.1( 52 %) -102.3(114 %)
spokane
-0.1( 50%) -0.1( 20%) -91.9(221 %) -72.0(200 %) -134.6(147 %) -93.8(140 %) -42.8( 50%) -21.8( 44%)
desert rock 0.5(103 %) -0.3( 46%) -75.9(159 %) -119.5(186 %) 24.2( 18 %) -358.2(215 %) 92.2( 74 %) -238.7(194 %)
miramar
-0.6(136 %) -0.6(105 %) -117.1(242 %) -78.1(205 %) -23.8( 41 %) -64.8(101 %) 93.3(149%) 13.3( 19%)
oakland
-0.6(148 %) -0.5( 96%) -78.7(210 %) -50.4(182 %) -22.3( 36 %) -64.7(103%) 56.7( 89% ) -14.9( 24%)
greensboro
0.0( 10%) -0.6(203 %) -102.5(256 %) -84.2(210 %) -190.0(220 %) -206.9(208 %) -87.5(116 %) -122.7(156 %)
charleston -0.3( 33%) -0.2( 91%) -73.9(107 %) -81.2(215 %) -189.3(160 %) -218.0(208 %) -115.4( 79 %) -137.2(162 %)
nashville -0.3(131 %) -0.3(100 %) -116.9(264 %) -116.7(245 %) -110.4(148 %) -204.7(194 %) 6.5( 9%) -88.1(118 %)
caribou
-0.3(159 %) -0.1( 60%) -111.7(250 %) -91.5(225 %) -164.2(224 %) -160.7(202 %) -53.2(106 %) -69.1(117 %)
buffalo
-0.4(128 %) -0.3(106 %) -99.1(208 %) -92.1(189 %) -162.1(193 %) -177.8(194 %) -63.0(111 %) -85.9(147%)
green bay
-0.3(136 %) -0.4(155 %) -96.9(228 %) -114.2(254 %) -74.6(134 %) -134.6(174 %) 22.3( 45 %) -20.1( 33 %)
intnlfalls -0.2(109 %) -0.5(152 %) -153.2(284 %) -156.5(275 %) -135.7(236 %) -142.1(181 %) 17.5( 42 %) 14.4( 26 %)
bismarck
-0.6(164 %) -0.5(151 %) -112.7(227 %) -125.9(241 %) -122.7(213 %) -121.5(186 %) -11.6( 23%) 4.0( 9 %)
dulles a.p. -0.4(123 %) -0.2(107 %) -125.6(251 %) -124.6(262 %) -137.2(176 %) -205.7(182 %) -12.3( 18%) -81.1( 86 %)
pittsburgh -0.3(166 %) -0.2( 62%) -116.4(270 %) -116.6(242 %) -154.4(227 %) -195.1(193 %) -38.0( 70 %) -78.4( 92 %)
lakecharles -0.5(222 %) -0.6(153 %) -140.0(307 %) -126.1(266 %) -139.2(152 %) -231.3(210 %) 0.8( 1%) -105.2(116 %)
topeka
-0.2( 87%) -0.2( 57%) -115.0(254 %) -115.8(244 %) -94.0(188 %) -132.7(203 %) 21.0( 44 %) -16.9( 42%)
midland
0.2( 59%) -0.2( 49%) -80.8(208 %) -60.6(191 %) -184.1(185 %) -107.4(194 %) -103.7(115 %) -46.8(118 %)
jackson
-0.2( 63%) -0.2( 53%) -83.1(110 %) -70.5(188 %) -14.0( 12 %) -95.3(148%) 69.1( 43% ) -24.7( 44%)
little rock -0.5(192 %) -0.6(176 %) -112.9(242 %) -140.2(264 %) -162.0(195 %) -250.1(214 %) -49.1( 88 %) -109.8(139 %)
quillayute -0.6(221 %) -0.5(173 %) -125.7(267 %) -84.6(222 %) -307.6(300 %) -331.2(284 %) -181.9(223 %) -246.7(253 %)
albany
-0.3(136 %) 0.4( 55%) -122.5(253 %) -108.4(222 %) -235.2(239 %) -222.6(232 %) -112.8(163 %) -114.7(175 %)
Examination of Radiosonde Metadata (1960-1983)
C = Change Computer
S = Change Sonde Model
L = Station Location Change
R = Change to R H
G = Change Gravity
D = Change Data Cutoffs
E = Change Ground Equip
I = Change Station ID
M = Miscellaneous
1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983
1
E
R
C
2
L L
SR
S
R
C
SR
3
R
R
C
R
4
E
R
R
C
SR
5
E
L
R
R
C
R
6
E
R
R
C
R
7
S R
L L
R
C
SR
8
S R
L L
R
C
SR
9
E
R
R
C
SR
10
RS
R
C
SR
11
LE
R
R
C
R
12
R
L
L
R
C
SR
13
E
R
R
C
R
14
E
R
R
15
E
R
L LR
R
SR
16
L R
L L
R
C
SR
17
L
18
R
L
S
R
C
M R S
19
R
R
C
M
SR
20
E
R
L
R
C
LR
21
R
R
C
SR
22
E
L E SR
R
C
SR
23
SR
S
R
C
R
24
LE
R
R
C
M R
25
E
L L R
R
C
R
26
E
R
L
R
SR
27
R
LS L
R
C
SR
28
L
E
R
R
C
R S
29
LE L RS
L
R
C
SR
30
E L
R
R
R
31
E
R
L L
R
C
R
32
EL
R
R L
C
SR
33
L
R
L
R
C
L
R
34
L LE
E L R
R
LC
R
35
ERL
R
SR M
36
R
R
C
R
Examination of Radiosonde Metadata (1984-2007)
C = Change Computer
S = Change Sonde Model
L = Station Location Change
R = Change to R H
G = Change Gravity
D = Change Data Cutoffs
E = Change Ground Equip
I = Change Station ID
M = Miscellaneous
1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007
1
C
S C
GDR
S
C
2
LC
S S C
S
3
C
S S C
S
4
C
S S C
S
5
C
S C
DGR
L S
C
6
C
S
C
DRG
S
7
C
S S C
S
8
C
S
C
RGD
9
C
S
CS
S
10
L
C
S
C
GDR
S
C
11
C
S C
RGD
12
C
S C S
S
13
C
S C
RGD
14
C
S C
DRG
15
C C
S
C
RGD L
S
C
16
C
S
C
RGD
17
C
C
RGD
18
C
S
C
RGD
19
C
S CLI
RGD
S
C
S
20
C
S C
DRG
21
C
S
C
RDG
22
C
S
C
GRD
23
C
C
S
C
RGD
24
C
S
C
GRD
25
C
S C
GDR
26
C
S
C
RGD
S
27
C
S
SC
S
S
C
28
C
S C
RGD
S
C
L
S
29
C
S C
RGD
S
C
30
C
S
C
RGD
31
C
C
S C
RGD
32
C
S S C
S S
33
C
S
C
RGD
34
S C
DGR
35
C
S C
RGD
36
C
S C
GDR
post 1996 updates only at a few sites
Annual Mean Values of 12Z sounding
from all stations used in EOF analysis (36 total)
No of Levels (Srf to 700mb)
18
Number of Radiosonde Levels (Surface to 700mb) (with Temperature reading)
16
14
12
10
8
1960
1970
1980
1990
Year
2000
Annual Mean Values of 12Z sounding
from all stations used in EOF analysis (36 total)
No of Levels (Srf to 700mb)
18
Number of Radiosonde Levels (Surface to 700mb) (with Temperature reading)
16
14
12
10
8
1960
1970
1980
1990
2000
Year
No longer mandatory levels at
950, 900, 800, and 750
Temperature no longer included at
significant levels that were due to
inflection points in wind speed or
direction.
Annual Mean Values of 12Z sounding
from all stations used in EOF analysis (36 total)
No of Levels (Srf to 700mb)
18
Number of Radiosonde Levels (Surface to 700mb)
16
14
12
10
8
1960
1970
1980
1990
2000
Year
Inversion Frequency
Inversion Frequency (%)
98
96
94
92
90
88
1960
1970
1980
1990
2000
Year
Inversion Thickness
Inversion Thickness (m)
400
380
360
340
320
300
1960
1970
1980
1990
Year
2000