Examensarbete vid Institutionen för geovetenskaper ISSN 1650-6553 Nr 220
Observed Ice Supersaturated Layers over Sweden and Implications for Aviation Induced Contrails over the Baltic Sea
Elin Björklund
Handledare: Tomas Mårtensson och Hans Bergström
Abstract
In the atmosphere vertical extended layers that are in the state of super saturation with respect to ice can appear; these layers have been termed ice‐
supersaturated layers (ISSL). If an aircraft passes through an ISSL, persistent condensation trails can form.
These contrails absorb the long wave radiation from the earth and reflect the incoming short wave radiation from space. The absorbing effect of the long wave radiation is although greater than the reflecting effect of the short wave
radiation and therefore these contrails increase the greenhouse effect.
This study contain statistics of when ice‐supersaturated layers occur in the Swedish airspace, based on data from balloon soundings that take place each day at four locations in Sweden. The soundings that are used in this report were carried out from January 2006 to December of 2010. The results show that ISSL are more common in the Swedish airspace than expected. The layers are by average 42 hPa thick, located at the height of 339 hPa and are present in 44 % of the soundings.
Referat
I atmosfären kan det uppkomma vertikalt utsträckta skikt som är övermättade med avseende på is; dessa lager har blivit benämnda som ice‐supersaturated layers (ISSL). Om ett flygplan flyger igenom dessa ISSL‐luftmassor kan
permanenta kondensationsstrimmor bildas.
Dessa kondensationsstrimmor absorberar den långvågiga strålningen från jorden och reflekterar den inkommande kortvågiga strålningen från rymden.
Den absorberande effekten är större än den reflekterande effekten, vilket bidrar till en förstärkning av växthuseffekten.
Den här studien innehåller statistik för när dessa ISSL uppkommer i det Svenska luftrummet, baserat på data från ballongsonderingar som var utförda från Januari 2006 till December 2010.
Resultatet från denna statistik visar att ISSL är mer vanliga i det Svenska luftrummet än förväntat. Lagren är i medeltal 42 hPa tjocka, placerade på 339 hPa höjd och förekommer i 44 % av sonderingarna.
Table of Contents
Abstract ... 2
Referat ... 3
1. Introduction ... 5
1.1 Contrail formation ... 6
1.2 Aviation and radiative forcing ... 7
1.3 Climate impact of Contrails ... 7
2 Method ... 7
2.1 Sounding stations ... 7
2.2 Geographic placement of the stations ... 8
2.3 Collection of the data ... 9
2.4 Correction of data ... 9
2.5 Calculation of the relative humidity over ice ... 10
2.6 Further data manipulation ... 10
2.6.1 Clouds with super cooled water ... 10
2.6.2 Excluding measurements in the tropopause ... 11
2.6.3 Interpolation of data to simplify calculations ... 11
3 Results ... 11
3.1 ISSL yearly variation ... 11
3.2 Case study of the 24th September 2010 at Visby ... 15
3.3 Implications for contrail formation ... 17
4 Discussion ... 21
4.1 The use of interpolated low resolution data ... 21
4.2 Calibration Constants ... 21
5 Conclusions ... 21
6 Acknowledgements ... 23
7 References ... 23
1. Introduction
Condensation trails behind aircraft is a phenomenon that almost everyone have observed, but are they all beautiful to look at and completely harmless for the atmospheric radiation balance?
The clouds that appear at the same height as these contrails (short for condensation trails) are cirrus clouds. Cirrus clouds are clouds that exist at altitudes above 8000 meters and up to the tropopause, which is located at 10‐12 km at the Polar Regions. A cirrus cloud consists of ice crystals and the most common form is a feather shaped cloud. Clouds at high altitudes have two main properties when it comes to radiation, these are to reduce the incoming
radiation from space, which cools the earth and to absorb the outgoing long wave radiation, which warms the earth. The absorption of the long wave
radiation gives warming and if contrails that have similar properties for radiance are created at this height it will result in an increased green house effect.
To determine if contrails can be created one should look at details of how an air mass at this height acts. For example an air mass with clouds can have the same level of saturation as an air mass without clouds. To transform a cloud free region into a region that contains clouds, a process where humidity condensate on condensation nucleus need to be performed.
A condensation nuclei is a small particle with a diameter of approximately 0.2 μm and can be, for example aerosols. The process when a cirrus cloud is formed, usually is an adiabatic cooling which starts when cloud droplets condensate on condensation nucleus.
To determine if an air mass can create persistent contrails a classification has been developed and there are two main terms, ice‐supersaturated regions (ISSR) and ice‐supersaturated layers (ISSL).
An ISSR is an air mass that has the characteristics of being in the state of super saturation with respect to ice, has temperature below zero degrees Celsius and holds no clouds.
An ISSL is an air mass that has the characteristics of being in the state of super saturation with respect to ice, a temperature below zero degrees Celsius including clouds.
According to Schumann and Ström (2001); aircraft fly through ISSR 15 % of their flight time and as aircraft fly through ISSR contrails may be created; this gives an enhanced chance of persistent contrails.
To determine if an air mass is an ISSL or ISSR one can use data that are measured by a radiosonde.
Observations made by radiosondes contain measurements of the temperature, humidity, wind speed, wind direction and pressure in the surrounding air mass.
The radiosonde can only detect ISSL since it cannot automatically tell if a measure point is inside a cloud. To tell if the atmosphere at that sounding
contained clouds, a synoptic observation or a satellite image should be consulted.
Previous studies have been made to determine ISSR; Spichtinger et al (2003) with the help of satellite data, equipment placed on aircraft by S. J. Moss (1999) and with model data by P. Spichtinger and K. Gierens (2003).
The method to use data from balloon soundings, as used in this study, has previously been carried out over Lindenberg by Spichtinger et al. (2003). They have shown that the average frequency of occurrence over Lindenberg is about 28 % and occur within a 200 hPa layer down from the tropopause. A publication for the British Isles including Gibraltar, the Falkland Islands and St Helena written by Dickson et al. (2003) show that the frequency of occurrence over the British Isles, Gibraltar and The Falklands Island range from 15‐25 % at
wintertime and 5‐15 % at summertime. At St Helena there is a 2,1 % occurrence with thickness of 600‐800 m.
Early, studies carried out with the help of Vaisala RS80 radiosonde had issues with humidity measurements at low temperatures. RS80 need careful calibration and have the disadvantage of instability issues below ‐40 degrees. This causes problems when contrail occurs below ‐35 degrees Celsius.
Since the release of the Vaisala RS92 the problem with humidity measurement below ‐40 degrees has improved. A part of the problem is solved for the Vaisala RS92 by the help of two measuring points instead of one placed under a cap, they alternately are measuring while the other is heated to keep the cap clear from ice at low temperatures.
1.1 Contrail formation
Contrails were first observed in 1915 and have been observed along with air traffic since. Studies have been done to explain why these contrails form and there are two main ways to reach condensation.
Contrails that form as an aero dynamical phenomenon. Contrails of this type are for example created at the wingtips of the airplane when the air mass that the airplane passes through is cooled adiabatically.
Contrails that form when air from the aircraft's exhaust mix with the surrounding air. These contrails form by the isobaric mixing between the heated warm air that exits the aircrafts engine and the surrounding colder air.
In this report the focus is set at contrails that form by isobaric mixing.
Contrails are formed in the atmosphere when the temperature is below 233 K (‐
40 degrees Celsius). This formation takes place when the hot air exit the engines exhaust and it cools down rapidly and a mixing with surrounding air occurs and a contrail is formed.
A commonly used method to forecast if contrails can form is the Appleman‐
Schmidth criteria derived by Appleman and Schmidt (Appleman, 1953).
One way of expressing the Appleman‐Schmidth criteria is equation (1) derived by SJ. Moss (1999), here ∆T is the difference between the exhaust and the surrounding air. The factor A (the contrail factor) contains engine specific parameters. Contrails form below temperature Tc.
. ∆ ∆ .
. (1)
If the contrail forms in an air mass that has relative humidity below 100 % the lifetime of the contrail are minutes before it dissolves because it dries out. For contrails that are formed in an air mass that has relative humidity with respect to ice over 100% the lifetime of the contrail persist from hours to days. These persistent contrails can transform into cirrus clouds and therefore has an impact on the radiation balance.
1.2 Con the rad 1.3
Figu
Con albe the et a 0.01 con rad
2 M
2.1 To d per ball Serv
Aviation a ntrails are o earth. Rele iate forcing Climate im
ure 1 Aviation
ntrails have edo so that
outgoing r al. (2009) sh
118 Wm‐2 ( ntrails is low
iate forcing
Method
Sounding detect ice‐s formed at f loon sound vice.
and radiati only one pa ease of CO2 g. For furth mpact of C
Radiative For
e an impact t the incom radiation fr how in Figu (IPCC AR4 v w. The scie g is very low
stations supersatur four station dings each d
ive forcing art of aviati 2, NOx, H2O her reading
ontrails
rcing Compon
t on the clim ing radiatio rom the ear
ure 1 that t values). Th ntific unde w and has a
ated layers ns in Swede day is prov
g
ion’s contri O, SOx and s g, see Lee et
nents in 2005
mate. At da on from th rth, which e
the radiativ he scientific erstanding f an estimat
s, data from en are anal vided by SM
ibution to t soot also ac t al. (2009)
5. Figure from
aytime they e sun is low enhance th ve forcing f c understan
for induced ed value of
m balloon so lyzed. The MHI and the
the radiatio ct in a way
.
Lee et al. (20
y reduce the wered. They e greenhou for linear co nding abou d cirrus clo f 0.03 Wm‐2
ounding th service to c e Armed Fo
on balance to alter the
009)
e earths y also abso use effect. L ontrails is ut linear oudiness an
2.
hat are carry out orce Weath
of e
orb Lee
nd
er
The Göt RS9 2.2
Figu
Tabl
Lule Sun Göt 025 Visb Figu Lule the Göt
e four place teborg/Lan 92 radioson
Geograph
ure 2 Placeme le 1 Informati
eå/Kallax 0 ndsvall 023
teborg/Lan 527
by 02591 ure 2 show eå/Kallax i Gulf of Bot teborg/Lan
es used her ndvetter. St
nde.
hic placeme
nt of the soun
ion of the rad
S la 02185 6
65 6
ndvetter 5 5 ws the geogr s located a tnia in the m ndvetter by
re are Luleå arting Janu
ent of the
nding stations
iosonde statio
Station atitude 65.55 62.53 57.66 57.66
raphical loc t the north midst of Sw
the west c
å/Kallax, Su uary of 200
stations
s.
ons
Station longitu 22.13 17.45 12.50 18.35 cations of t hwest shore weden, Visb coast by Kat
undsvall, V 06 these loc
n
ude As
Ye Ye Ye Ye the stations e of the Bay by at the isl ttegat. In T
isby and cations all u
scent 00z es
es es es
s in Sweden y of Bothnia land of Got Table 1 long
use the Vais
Ascent 1 Yes No No Yes n;
a, Sundsval tland and gitude and
sala
2z
ll by
latitude of the stations are shown. All four of the stations have balloon ascents at 00z each day, but only Visby and Luleå have ascents that take place at 12z.
Table 2 Number of ascents for soundings carried out at 00z
00z First
ascent 2006 2007 2008 2009 2010 Total Luleå/Kallax 02185 20060101 326 334 357 350 332 1699 Sundsvall 02365 20060101 329 353 344 324 339 1638
Göteborg/Landvetter 02527
20060101 313 328 339 342 353 1689
Visby 02591 20060127 319 321 352 338 353 1683
6709
In Table 2 the number of ascents that took place for each year and station are listed for the 00z soundings.
This dataset provides sonde data carried out for 92 % of the whole period from January 2006 to December of 2010.
2.3 Collection of the data
Data from these soundings are stored in a high resolution dataset, but before this information is released publicly, the information is stored into a low resolution file that only save significant points for the measurements and standard pressure levels. In this study it has not been possible to retrieve high‐resolution datasets, therefore low‐resolution files have been used.
To get access to the data from the radiosounding a script written in the programming language Python was used. In this script several options can be selected, choice of station number, year, month and day. After these options have been determined the data is downloaded from University of Wyoming’s data server.
Observations made by radiosondes contain measurements of the temperature (T in Kelvin), humidity (RHwobs measured in %), wind speed (measured in m/s) and wind direction (measured in degrees 0‐360) and pressure (p measured in hPa).
2.4 Correction of data
It is well known that most radiosondes have problems with humidity
measurements at low temperatures and are also affected by radiation from the sun. There are some calibration algorithms developed by the scientific
community, which are used to improve the quality of the data.
In this study, the observed relative humidity is corrected with the help of the calibration constants Ccal and Crad derived by Vömel et al. (2007). Relative humidity with respect to water is then calculated:
∗ (3)
The calibration constant Ccal is derived from nighttime soundings made by Vömel et al. (2007), by comparison of the CFH (Cryogenic Frostpoint
Hygrometer) reference measurement and the Vaisala RS92. Ccal correct the data
so that the values of the relative humidity are closer to the values given by the CFH. In Table 3 the temperature dependence of Ccal is shown.
Table 3 Table of Ccal. T given in Kelvin.
208 1.13
208 218 1.04
219 233 0.94
233 0.98
The Vaisala RS92 does not have a radiation shield over the sensors to prevent effects from solar radiation, therefore a calibration constant Crad is used to reduce these effects. The constant is expressed in equation (2).
0.12158 log 4.7855 (2)
Crad is derived from measurements in the tropics, when applied to Swedish 12z soundings an unrealistic compensation for relative humidity was found. Based on this compensation only 00z soundings were analyzed when there is no direct sunlight on the equipment.
2.5 Calculation of the relative humidity over ice
Relative humidity with respect to ice can be calculated for all temperatures, but it is not physically relevant for this to appear in the atmosphere above 273 K.
To calculate the relative humidity with respect to ice, formulas for partial pressures over ice that are used. These formulas are derived by Murphy and Koop (2005).
(4)
Where ew and ei are the partial pressures over water and ice respectively.
Following Murphy and Koop (2005). we use equation (6) for the vapor pressure over water and equation (5) for the vapor pressure over .
exp 9.550426 . 3.53068 0.00728332 T>110K (5)
ln
54.842763 . 4.210 ln tanh 0.0415 218.8 0.000367 53.878
9.44523 . 0.014025 (6)
2.6 Further data manipulation
In order to reduce errors in the observation data the data is processed further.
2.6.1 Clouds with super cooled water
When the radiosonde passes through a cloud that have supercooled water drops in it, it can be mistaken for an ISSL since the sensor will register RHi~100 %.
Therefore the ISSL and the waterclouds need to be divided. A way of excluding clouds that contain supercooled water is to remove measure points where the temperature is above 243 degrees Kelvin (‐30 degrees Celsius). Below 243 degrees Kelvin super cooled water rarely exists in clouds. At this temperature the cloud mainly consists of ice crystals.
2.6.
The the a fa trop To f that calc 2.6.
To i an i whe in th Thi stud end
3 R
3.1
Figu
In F mon A sl the min who Fro for
.2 Excluding e radiosond tropopaus actor of 2 (V
popause we find the tro t maintain culated heig .3 Interpola increase th interpolatio en analyzin he vertical.
s interpola dy the inter d of the set,
Results
ISSL yearly
ure 3 Frequen
Figure 3 the nth for the light season
ISSL. An av nimum valu ole period a
m the indiv Visby in No
g measurem de measure se. In the tro
Vömel et. 2 ere exclude opopause a
a negative ght which i ation of dat he number o
on was mad ng the resu
.
ation was d rpolation g by stretch
y variation
cy of occurren
e average fr soundings nal variatio verage max ue of 28 %
and all stat vidual aver ovember 2
ments in the es until the
opopause t 007) by the ed in order
calculation lapse rate is in the tro
a to simplif of data poin de. This als lts. This int one linearl gave a point
ing the dat
n
nce of ISSL 20
requency d are shown on is found ximum valu for Visby in tions is 44 % rage values
007 and th
e tropopau balloon br the values f e Vaisala R r to simplify
n is made w are include opopause a
fy calculatio nts withou so automat terpolation ly between
t for each 1 ta to 100 da
006‐2010 00z
distribution n.
with lower ue of 59 % n August ar
%.
s for each m he minimum
se
reaks, whic for humidit RS92. There y the calcul where data
ed. The dat are exclude
ons
ut manipula tically bins n gives the d
the points 10 hPa from
ata points w
soundings fo
n of all ISSL r frequenci for Sundsv re found. A month, the m
m value of 1
h usually h ty are unde efore values
lations.
points abo ta points ab
d.
ating the da the data, w data a com
given in th m ground le with a 10 h
r various mon
L for each st ies in summ vall in Febru n average v maximum 7 11 % is fou
happens ab erestimated
s from the ove 400 hP bove this
ata too muc which will h mmon basel
he data. In t evel up to t hPa spacing
nths
tation and mertime for
uary and a value for th 73 % is fou und in July
ove d by
a
ch help
ine this the g.
r he und
200 latit
% f
Figu
In F axis with A ge mon the nor for high In Ja also sho For sum diff a he unu
06 for Luleå tudinal var for Luleå, 5
ure 4 ISSL 200
Figure 4, th s and mont h a clear se eographic v nths Januar stations in rthern locat the layer h hest layer h anuary 201 o be seen fo own higher r Sundsvall mmer the av
ference of a eight scale usually cold
å. When cal riation is fo 0 % Sundsv
6‐2010. Avera
e ISSL heig ths on the x easonal shif variation fo ry to March n January an tion where height, corr height for t 10 the ISSL or the Visby
in the trop in August 2 verage ISSL approximat
of 67 hPa.
d. Luleå had
lculating th ound. The a
vall, 42 % f
age ISSL for th
ght in the tr x‐axis. The a
ft.
or the heigh h. The aver nd 405 to 3
the winter esponding these mont L started as y and Land posphere th
2006 the IS L height are tely 5 hPa i
The winter d, for exam
he annual fr average freq for Landve
he four statio
roposphere average va hts of the la rage height 370 for Feb
r season is to Landvet ths.
low as 530 dvetter stat
han any of t SSL ended a
e almost pl n July. For r months Ja mple an aver
requencies quencies fo tter and 47
ons
e is shown w alue shows
ayers can b varies betw bruary. Lule
the longest tter and Vis 0 hPa. The tions, their the other m at a height laced at the
January th anuary to M
rage value
s for each st or the each 7 % for Visb
with pressu a spatial di be seen at fo
ween 350 t eå that has t shows the sby that sh
seasonal va average va months.
of 242 hPa e same leve e average v March of 20 of 530 hPa
tation, no station are by
ure on the istribution for winter
to 410 hPa the most e lowest va hows the
ariation ca alues are a. For the el with a
values diffe 010 was a for the low
e 48
y‐
for alue
n
er at west
poin abo The
Figu
The Figu foun A m 60 h (cor
Figu
Mos app to 7 betw dist Figu cirr
nt of the IS ove 400 hPa
e average h
ure 5 ISSL thic
e annual va ure 5 using nd for the w maximum va
hPa (corres rrespondin
ure 6 Distribut
st cirrus clo proximately 70 hPa are I ween 60 an tribution of ure 6 also c rus clouds.
SL, compar a.
height of the
kness 2006‐2
ariation of a g data from
winter mon alue is show sponding to ng to 600 m
tion of ISSL th
ouds have a y 70 hPa at
ISSR and th nd 80 hPa m
f how thick contain line
red to the w e ISSL for a
2010
average thi the years 2 nths and du
wn for Lule o 1200 met meter) for S
hickness 2006
a thickness t 8000m, th hicknesses my data als k ISSL are.
ear regress
winters of 2 all months a
ckness of t 2006 to 20 uring summ eå in Febru
ter) and a m Sundsvall in
6‐2010 for all
s above 100 his suggest
above 80 h so suggests sions for es
2006‐2009 and station
the ISSL for 010. The lar mer the ISS uary with a minimum o n August.
stations.
00 meters, that thickn hPa are ma s that there timation fo
where ISS ns is 339 hP
r all station rgest thickn
L are typic value of ap of approxim
correspond nesses in th
inly clouds is a clear s or dividing
SL where Pa.
ns is shown nesses are cally thinne pproximate mately 30 h
ding to he interval
s. Somewhe shift in the
ISSR from
in er.
ely hPa
10 ere
The relation between Cirrus and ISSR is 1 to 4 with the chosen "cut" at 70 hPa.
This implies that the real frequencies of ISSR are 75 % of the calculated frequencies for ISSL (Figure 3).
Figure 7 The distribution of relative humidity over ice for Visby 2006‐2010.
The distribution of RHi for temperatures below 243 K for Visby from January 2006 to October 2010 is shown in Figure 7 The maximum supersaturation was found to be 138 % in this data sample and the bars in the plot are for 5 % bins.
The rise occurring above 85 % has been found in other investigations Dickson et al. (2010) and Spichtinger and Gierens (2003). The reason for the values that are larger in relation to values at 70 and 80 % are unknown.
0 20 40 60 80 100 120 140
0 500 1000 1500 2000
RHi%
Number of observations
3.2
Figu
Case stud
ure 8 Satellite
y of the 24
image by Mod
4th Septem
dis (cropped)
mber 2010
) for 25 Septem
at Visby
mber 2010
Figure 9 Sounding carried out over Visby from the 00z soundings. From the left 20100924,
20100925, 20100926, 20100927. Pressure layers between 400 hPa up to 180 hPa are shown.
At the 24 September 2010 a lot of persistent contrails over the Baltic Sea was seen by satellite (Figure 8). Looking at the sonde data for 00z from Visby (Figure 9 ) showed that there were a ISSL present. Figure 8 shows that the ISSR first appeared on the 24th September and stayed until the 27th of September. The persistent contrail was visible until the 25th of September; the 26th and on the 27th clouds appeared over the area.
0 50 100
200
250
300
350
400
RH [%]
P [hPa]
RH(w) RH(i)
0 50 100
200
250
300
350
400
RH [%]
P [hPa]
RH(w) RH(i)
0 50 100
200
250
300
350
400
RH [%]
P [hPa]
0 50 100
200
250
300
350
400
RH [%]
P [hPa]
RH(w) RH(i)
RH(w) RH(i)
Table 4 Statistics for Visby 20100924‐25
Start
height (hPa)
End height (hPa)
Average height (hPa)
Thickness
(hPa) Maximum
value (%)
24 September 350 290 320 60 118.7
25 September 210 200 205 10 100.2
Average 280 245 262.50 35
Statistics for these the first two days are presented in Table 4. It shows the average values for the start, end and average heights of the layers, the thicknesses for the layers for each day and maximum values for the relative humidity for these days.
3.3 Implications for contrail formation
After studying the case in section 3.2 one might ask how common are days with a clear sky and a persistent visible contrail cover? To study this, one have to know how much air traffic there is over the Baltic Sea. The horizontal extent of ISSR is 150 km according to Gierens et al. (1999). Making the assumption that the sounding from Visby is valid approximately in an area of 75 km around the sounding station in Visby we only need to find out the number of aircraft that fly through that area.
Figu
Figu traf each pilo UNL con that airc The betw sho and app dep 100 370
ure 10 ATC cha
ure 10 show ffic controll
h square of ots are alloc Limited, GN ntroller at L
t is marked craft pass th ey are equa
ween 260 h ould be add d take off cy proximately parting and 0 km from V 0 hPa.
art for the air
ws with pu ler in Malm f these pur cated by th ND stands G LFV in Malm d UNL FL28
hrough eve ally distribu hPa and 18
ed. Here w ycles a year y 26 every d landing at Visby they
rspace over th
urple marki mö ATCC is ple marked he Air Traffi GrouNDlev mö has prov 85 close to ery day.
uted betwe 80 hPa. Air t we have use
r is done w day. Those t Visby. But are at their
he southern B
ings the are responsibl d areas rep fic Controlle vel and FL s
vided this s Gotland. In een FL330 a traffic depa ed LFV stati with comme e 26 aircraft
t when they r cruising a
altic Sea
ea in the Ba le for. The c presents be
er. The acro stands for F
study with n that secto and FL420;
arting and istics show ercial jet or fts are at low
y are appro altitude ran
altic Sea tha code that is
tween whi onyms UNL FlightLevel.
the data fo or approxim
; this corres arriving fro wing that 95
turboprop w altitudes oximately a nging from
at the air s printed in
ch flight lev L stands fo
. The air tra or the secto mately 500
sponds om Visby 500 landing p aircraft,
s when at a distanc
300 hPa to
n vels r affic or
g
e of o
Figu
Figu ISSL heig dep
Tabl
Janu Feb Mar Apr May Jun July Aug Sep Oct Nov Dec In T data
ure 11 Low, To
ure 11 how L and also t ghts at whi partures are
le 5 ISSL depe
IS fr
uary 0
bruary 0
rch 0
ril 0
y 0
e 0
y 0
gust 0
ptember 0 ober 0 vember 0 cember 0 Table 5 we a together
op, Medium an
w the averag the tropopa ich the plan
e marked w
endence of clo
SSR
req. N
C
0.33 0
0.32 0
0.33 0
0.26 0
0.32 0
0.28 0
0.24 0
0.21 0
0.27 0
0.33 0
0.39 0
0.35 0
list data fro with air tra
nd Tropopaus
ge values fo ause level f nes fly are m with a light
oud informati
N=8 &
Ch=/
0.53 0.47 0.34 0.25 0.18 0.14 0.14 0.15 0.20 0.32 0.43 0.49
om the inve affic data. T
se level over V
for the start for Visby fo marked in g t grey color
on
Ch=0
0.61 0.55 0.44 0.36 0.32 0.32 0.33 0.36 0.43 0.47 0.61 0.61 estigated s The monthl
Visby 2006‐2
t, end and a or the years grey and h r.
N<3
0.13 0.16 0.26 0.31 0.40 0.41 0.34 0.34 0.30 0.23 0.15 0.3
oundings w ly average
010
average hei s 2006‐201 eights for a
ISSL/
Tropopau (hPa) 133 134 125 99 86 81 52 51 76 103 89 112 with ground
ISSL freque
ight of the 10. The arrival and
use ISSL‐
to tra (hPa) 95 86 110 71 59 50 18 18 23 52 57 69 d observat encies from
top affic
)
ion m
section 3.1 multiplied with the average ratio between clouds and ISSR found in the thickness distribution in figure 7 are multiplied by a factor of 0.75 due to that from Figure 6 it has shown that 3/4 of the measured ISSL are ISSR. This is shown in column one.
Column two, three and four contains SYNOP data from Visby. The SYNOP data is from the 30‐year period 1960‐1990. For this period Visby always made manual observations of clouds. These monthly averages are calculated as the average of the observations made at 00, 06, 12 and 18 UTC. The data is from Swedish Armed Forces Weather Service (Meteorological and oceanographic Center) and are provided to this study by Peter Löfwenberg at METOCC.
Column two (N=8 and Ch = /) show the frequency when an observation of high clouds has not been possible because of overcast clouds at lower levels. So about half of the time (53 %) in January, it is not possible to deduce if there are high clouds. In the summer it is only about 15 % of the time that observations of high clouds are hindered by lower clouds.
Column three (Ch = 0) show the frequency when NO high clouds are reported and this assessment has been possible to do, that is the lower clouds are not forming an overcast cloud‐deck.
Column four lists the frequencies of the TOTAL cloud cover being less than 3 octas, which is more or less clear skies (0,1 or 2 Octas).
Column five show the distance in hPa between the tropopause and the average height of the ISSL for Visby. The fifth column shows the distance between the tops of the ISSL layers to the lower layer of the cruising air traffic (FL330 corresponding to 260 hPa).
It is clear that cruising aircraft fly closest to the ISSLs between July and September.
Table 6 Top of ISSL, Distance to traffic for 2006‐2010
June July Aug Sep Oct June July Aug Sep Oct
298.89 248.33 290.00 271.11 278.00 2006 38.89 -
11.67 30.00 11.11 18.00 287.14 257.50 247.00 - 317.33 2007 27.14 -2.50 -
13.00 - 57.33 308.33 299.00 273.75 260.00 302.50 2008 48.33 39.00 13.75 0.00 42.50 310.59 282.22 257.50 280.77 320.00 2009 50.59 22.22 -2.50 20.77 60.00 293.33 250.71 271.25 281.76 293.85 2010 33.33 -9.29 11.25 21.76 33.85 298.89 248.33 290.00 271.11 278.00 2006 38.89 -
11.67 30.00 11.11 18.00
Looking in more detail on the distance between en‐route traffic and the ISSL‐top of layer (see Table 6), we find that for individual summer‐months the ISSL‐top sometimes reaches into the flight corridor. In fact, this happened three out of the five investigated July months. Two of five months in August and a close call for one month in September. (Note that September 2007 has been excluded because a lack of data). This suggest that the lower parts of the flight corridor is in touch with the ISSR approximately 50 % of the time in July and August. So a very rough estimate based on a ISSR frequency of ~25 % and a 50 % chance of the traffic being in the region yields that 12.5 % to the time in July and August the lower part of the en‐route traffic will generate persistent contrails. Here we also
assume that the Appleman criterion is fulfilled. During the rest of the months the ISSL above Visby are only affected by air traffic that arrive and depart in Visby.
The en route traffic area occurs around the tropopause, in wintertime mainly above and in the summertime more in the center.
4 Discussion
4.1 The use of interpolated low resolution data
This method generates uncertainties of the thicknesses of the layers when the sounding changes from above 100 % to below 100 % between data points. This may not be at the exact point where the interpolated point is set and therefore peaks can be missed if they are thinner than 10 hPa. For the plots that have been scanned through manually there is a trend that if a point is interpolated to another value, it is most common that the lower values are overestimated and the upper values are underestimated. This problem can be solved by using an interpolation with steps of 1 hPa at the cost of calculations that would take a longer time and a need of a computer with higher capacity. Although the benefit of this method is that it bins the data, which was helpful for the statistical
calculations. It also gave the same start heights for all the soundings and all the soundings got the same measure points.
In future studies the problem of underestimated and overestimated
measurements caused by the interpolation could be avoided by using high‐
resolution data.
4.2 Calibration Constants
Vömel et al. (2007) has shown that the accuracy of the Vaisala RS92 radiosonde is good when the Ccal and Crad calibration constants are applied in the tropics.
Miloshevich et al. (2009) has developed a new way of expressing Ccal and Crad with a higher accuracy than in this study. It has not been used in this report because it needs additional input data from ground measurements. Miloshevich et al. (2009) also includes a derivation of Crad where the angle of the sun can be set manually for each calculation due to season dependence and geographic placement. It could be recommended to use the Ccal and Crad derived by M.
Miloshevich et al. (2009) in further studies.
Persistent contrails can form in the tropopause, but values for the tropopuase was not included in this study because of the Vaisala RS92 radiosonde property that it underestimates the humidity values of a factor 2.
It could be recommended that time should be devoted to solve this issue. A possible solution is to use the correction developed by Miloshevich et al. (2009).
5 Conclusions
This study shows a method how to use low‐resolution data from the Vaisala RS92 radiosonde to estimate the frequency of ISSL. Results are presented for the period 2006 to 2010 for the four Swedish stations Luleå, Sundsvall, Visby and Kallax.
The frequency of ice‐supersaturated layers in the Swedish airspace is 44
%. Slight geographical variation is visible in the average values for the stations, for Luleå 48 %, Sundsvall 50 %, Landvetter 42 % and Visby 47
%.
Maximum and minimum individual frequencies are represented by 73 % for Visby in November 2007 and a minimum value for 11 % Luleå in July 2006.
The average height of the ISSL for all months and stations is 339 hPa.
The thicknesses of the ISSL layers have a seasonal variation, with an average value for the whole year of 42 hPa.
The average monthly maximum thickness is found for Luleå in February with a value of 60 hPa corresponding to 1200 meter and the average minimum thickness is found for Sundsvall in August of 30 hPa
corresponding to 600 meter.
Spichtinger et al. (2003) conclude that thicknesses in the interval 10 to 70 hPa are ISSR and thicknesses above 80 hPa are clouds. These conclusions are supported when analyzing the dataset in this study.
By looking at the case study for Visby for the days 24th and 25th of September 2010, one can see that the contrail cover that was created stayed for a number of days.
From Table 6 it is shown that in 12 % of the time in July and August the lower part of the en‐route traffic will generate persistent contrails where assumed that the Appleman criteria is fulfilled. For the other months there is a possibility that persistent contrails can appear occasionally, but not as frequent as for July and August.
6 Acknowledgements
This work has partly been funded by the European Network for Environmentally Compatible Air Transport System (ECATS). ECATS is a European Network of Excellence for research into the impact of air transport on the environment. It is supported by the European Comission (DG Research/Aeronautics ANE‐CT‐2005‐
012284).
Tomas Mårtenson, thank you for excellent mentoring and support.
Hans Bergström and Cecilia Johansson, thank you for valuable discussions during the study.
Pontus Von Schoenberg at FOI for providing the basic structure of the Matlab code and lending of the python script.
Peter Löfwenberg at METOCC for providing the synoptic statistics of cloud cover over Visby.
Luftfartsverket in Malmö for providing the statistics of planes that fly through the airspace above Visby.
7 References
Appleman, H., 1953: The formation of exhaust condensation trails by jet aircraft.
Bull. Amer. Meteor. Soc., 34, pp. 14–20
Lee, D. S., Fahey, D. W., Forster, P. M., Newton, P. J., Wit, R. C.N., Lim, L. L, Owen, B., Sausen, R., Atmospheric Environment, Vol. 43, No. 22‐23. (July 2009), pp. 3520‐
3537. doi:10.1016/j.atmosenv.2009.04.024
Gierens, K., Schumann, U., Helten, M., Smit, H., and Marenco, A.: A distribution law for relative humidity in the upper troposphere and lower stratosphere derived from three years of MOZAIC measurements, Ann. Geophys., 17, pp. 1218‐1226, doi:10.1007/s00585‐999‐1218‐7, 1999.
Vömel, H., Selkirk, H., Miloshevich, L., Valverde‐Canossa, J., Valdés, J., Kyrö, E., Kivi, R., Stolz, W., Peng, G. and Diaz, A. Radiation dry bias of the Vaisala RS92 Humidity Sensor. pp. 953–963, June 2007.
Miloshevich, L. M., H. Vömel, D. N. Whiteman, and T. Leblanc (2009), Accuracy assessment and correction of Vaisala RS92 radiosonde water vapor
measurements, J. Geophys. Res., 114, D11305, doi:10.1029/2008JD011565.
Moss, S. J., 1999: The testing and verification of contrail forecasts using pilot observations from aircraft. Meteorol. Appl. Volume 6, Issue 3, pp. 193–200, September 1999
Murphy, D. M. and Koop, T. Review of the vapour pressures of ice and
supercooled water for atmospheric applications. Q. J. R Meteorol. Soc. (2005), 131, pp. 1539–1565
Dickson, N. C., Gierens, K. M., Rogers, H. L., and Jones, R. L.: Probabilistic description of ice‐supersaturated layers in low resolution profiles of relative
humidity, Atmos. Chem. Phys. Discuss., 10, pp. 2357‐2395, doi:10.5194/acpd‐10‐
2357‐2010, 2010.
Spichtinger, P., Gierens, K., Leiterer, U., and Dier, H.,. Ice supersaturation in the tropopause region over Lindeberg, Germany. Meteorologische Zeitschrift, Vol. 12, No.3, pp. 143–156, June 2003.
Schumann, U., J. Ström, F. Arnold, T. K. Berntsen, P. M. de F. Forster, J.‐F. Gayet, and D. Hauglustaine, 2001: Aviation impact on atmospheric composition and climate, in: European Research in the Stratosphere 1996‐2000, Chapter 7, EUR 19867, pp. 257‐307, European Commission, Brussels.
P. Spichtinger and K. Gierens. Ice‐supersaturated regions.
www.pa.op.dlr.de/issr/, March 2003.
Tidigare utgivna publikationer i serien ISSN 1650-6553
Nr 1 Geomorphological mapping and hazard assessment of alpine areas in Vorarlberg, Austria, Marcus Gustavsson
Nr 2 Verification of the Turbulence Index used at SMHI, Stefan Bergman
Nr 3 Forecasting the next day’s maximum and minimum temperature in Vancouver, Canada by using artificial neural network models, Magnus Nilsson
Nr 4 The tectonic history of the Skyttorp-Vattholma fault zone, south-central Sweden, Anna Victoria Engström
Nr 5 Investigation on Surface energy fluxes and their relationship to synoptic weather patterns on Storglaciären, northern Sweden, Yvonne Kramer
Nr 212 T-Box Genes and Arthropod Development . Mette Lundgren, February 2011
Nr 213 Avgränsning av tillrinningsområden till grundvattenmagasin – vilken information ger berggrundens överyta, Kajsa Bovin, Mars 2011
Nr 214 Simulation and Analysis of Wind Characteristics in the Region Hosting the Sailing Competitions of the 2012 Olympic Games, Magnus Baltscheffsky, Mars 2011 Nr 215 Cubanite and Associated Sulphide Minerals from some Central Swedish Sulphide Ores, Anders Hallberg, April 2011
Nr 216 A Bio‐geochemical Survey of ThreeWatercourses in Uppland, Erika Rönnbäck, April 2011
Nr 217 Mineralogy and Lithogeochemical Signature of a Stratigraphic Profile through the Gränsgruvan Zn-mineralization, Bergslagen, Sweden, Erik Björklund, April 2011 Nr 218 Belizean Barrier Reef showing signs of recovery twelve years after the 1998 disturbances, Joaquin David Magaña, April 2011
Nr 219 Turbulence Intensity in Complex Environments and its Influence on Small Wind Turbines, Nicole Carpman, May 2011