Available via license: CC BY-NC-ND 4.0
Content may be subject to copyright.
*Corresponding Author: etturgut@anadolu.edu.tr
Anadolu Üniversitesi Bilim ve Teknoloji Dergisi A- Uygulamalı Bilimler ve Mühendislik
Anadolu University Journal of Science and Technology A- Applied Sciences and Engineering
2016 - Volume: 17 Number: 4
Page: 830 - 844
DOI: 10.18038/aubtda.279852
Received: 13 July 2016 Revised: 31 October 2016 Accepted: 11 November 2016
AN ANALYSIS OF VERTICAL PROFILES OF WIND AND HUMIDITY BASED ON
LONG-TERM RADIOSONDE DATA IN TURKEY
Enis T. TURGUT 1, *, Öznur USANMAZ 2
1 Aircraft Airframe and Powerplant Department, Faculty of Aeronautics and Astronautics, Anadolu University,
26470 Eskişehir, Turkey
2 Department of Air Traffic, Faculty of Aeronautics and Astronautics, Anadolu University,
26470 Eskişehir, Turkey
ABSTRACT
The prediction of atmospheric variables is fundamentally important for flight, and efficiency with environmental impact
analyses of aircraft. However, certain variables are less predictable leading to inefficient utilization of limited resources. To
maximize the efficiency of aviation systems more accurate approaches are required to increase the predictability of these
variables. One of these variables, wind aloft is analyzed in this study, with altitude, season and location, based on data
obtained from eight radiosonde stations operating in Turkey. It is found that as altitude increases, wind direction
approximates to 270°. Wind speed appears to be quadratically (or cubic for higher accuracy) proportional to altitude and
maximum average wind speed is observed in March. In addition, relative humidity decreases linearly with an increase in
altitude at an average of 4% per kilometer.
Keywords: Wind speed, Wind direction, Humidity
1. INTRODUCTION
Understanding airborne ambient conditions is of paramount importance for aviation. Of these, wind
speed and direction are essential for the safe and efficient use of an air space, as well as being the most
unpredictable. Others, viscosity, temperature, pressure, density and humidity are important for aircraft
flight, engine performance, fuel burning and emissions. Flying at higher cruise altitudes as much as
possible, or making use of tailwinds (when available) and avoiding headwinds (except during takeoff
and landing phases) usually leads to fuel savings. Generally, proportional to fuel consumption,
emissions of aircraft engines contribute significantly to harmful pollutants in the environment.
Moreover, these impacts show a distinct feature that other anthropogenic emissions sources do not.
This being emissions at various levels of the atmosphere, mainly at troposphere and lower
stratosphere.
Vertical profiles of temperature, humidity and wind are investigated by radiosonde devices [1,2],
research aircraft [3,4], unmanned aerial vehicle (UAV) [5,6] generally within the atmospheric
boundary layer or by satellites. These observations are mostly used to understand chemistry of the
atmosphere and trends in climatic characteristics.
These data and their vertical profiles are also important for the aviation. For instance, wind speed and
direction strongly affects the fuel consumption and flight time of a large aircraft. According to a study,
the effect of wind (combined with temperature) on fuel consumption can be as high as +1.4% and
+2.3% for the westbound (east to west) and +1% for the southbound flights, for a B747-400 [7].
While, many studies have been dedicated to mapping a ground wind profile for wind energy purposes
[8–10], there have also been a number of studies focusing on wind effect estimations for various
Turgut and Usanmaz / Anadolu Univ. J. of Sci. and Technology – A – Appl. Sci. and Eng. 17 (5) - 2016
831
situations, such as the descent phase [11–13] or for flight range [14]. These studies help to minimize
uncertainties encountered particularly during trajectory prediction investigations.
Accurate wind estimations also improve the capabilities and the navigation and guidance performance
of a UAV, such as a better geolocation of a way point and a better estimation of crab-angle [15]. Wind
is referred to be an important factor that affects the UAV trajectory and thereby the performance of the
mission of the UAV [16].
This study has two objectives. The first relates to wind aloft. Wind is a key parameter for instrument
flight procedure designers. The International Civil Aviation Organization (ICAO) suggests using,
when available, statistical wind data for instrumental flight procedure designs. Therefore, for countries
which have their own long-term wind statistics, procedures can be based on such wind data (e.g., the
United Kingdom and France). For others, the ICAO recommends a wind model, w=2h + 47, which is
based on only altitude, where w is wind speed in knots and h denotes altitude in thousands of feet [17].
Nonetheless, since the ICAO wind model is constructed to cover a relatively large region, there may
be serious overestimations on wind speed for specific regions, due to concerns for flight safety. While
these overestimations lead to inefficient utilization of air space, they do not provide additional safety.
Therefore, investigations on relatively local wind variations with altitude are useful in order to use air
space more efficiently.
Knowing the relationship between wind and altitude contributes to a better determination of the
boundaries and protection areas around an aircraft’s flight path. In this respect, in a previous study by
the same authors, variation of maximum wind speed with altitude was investigated for Turkey [18].
Considering maximum wind speed values at each altitude and for each season, a linear regression
model for wind based on altitude was developed; this being an accurate alternative for other existing
wind model of the ICAO. Accordingly, up to 25,000 ft (inclusive) wind speed can be found by the
equation 2h+26, whereas above 25,000 ft, it can be found by the equation h+50.
In addition to wind speed, wind direction is also important. Unlike wind speed, changes in wind
direction might have greater effects on protection areas around a flight track. Although the prevailing
wind direction for a specific route is generally known, wind direction can be highly variable.
Therefore, in instrumental flight procedure designs, wind direction is assumed as omnidirectional for
safety concerns. For the same reason, the results of this study do not suggest any specific wind
direction.
A second objective is to determine the variation of humidity at different altitude conditions. Bearing in
mind that the reliability of certain humidity sensors that are used in radiosonde devices at higher
altitude levels can be questionable [3], radio sounding provides cost effective, long-term and stable
observations. Humidity not only plays an important role in climate, it may also affect engine power
output, and in turn NOx emissions, by absorbing combustion heat. Considering that the specific heat of
water vapor is greater than air, combustion with humid air tends to produce lower flame temperatures.
In addition, at higher equivalence ratios, water vapor tends to react with carbon to produce carbon
monoxide and hydrogen [19]. However, the effect may be different depending on the combustor type,
therefore, it might be difficult to ascertain the net effect of humidity. In leaner combustors, humidity
may have no effect on engine power output or NOx emissions, whereas for other types of combustors,
flame temperature may vary significantly or NOx emissions may reduce with an increase in humidity
for the same combustor inlet temperature. In modern aircraft engines, fuel control units alter fuel flow
to the combustion chamber in order to restore engine thrust when the humidity levels become
significant. Therefore, knowing the humidity variation with altitude contributes to aircraft engine
performance and NOx emission models for flight conditions. Humidity effect on NOx emissions is
discussed in Boeing Fuel Flow Method 2 [20], which is one of the well-established and extensively
used emission calculation methods [21]. In this method, a parameter called as humidity ratio is
Turgut and Usanmaz / Anadolu Univ. J. of Sci. and Technology – A – Appl. Sci. and Eng. 17 (5) - 2016
832
required to calibrate sea level NOx measurements for a given altitude level. This ratio can be
calculated by a number of parameters including ambient temperature, saturation vapor pressure and
relative humidity. Accordingly, if there is no humidity measurements for a given altitude level, the
method suggests to use standard atmosphere sea level value of 60%.
At the end of the study, correlations for wind speed and humidity are also developed.
2. MATERIALS AND METHODS
The data of this study was obtained from long-term radiosonde measurements. In radiosonde stations,
a radiosonde device, equipped with a transmitter, ascends with a weather balloon and simultaneously
measures and transmits parameters, such as temperature, relative humidity, wind and pressure, to a
ground station. Measured values cover a vertical distance of the troposphere up to the 10 hPa, however
the ascend frequency can be 2-4 per day [6]. Turkish radiosonde stations and their World
Meteorological Organization identifications are given in Figure 1.
Figure 1. Turkish radiosonde stations (circle markers indicate stations close to the sea)
As can be seen from Table 1, the meteorological data, obtained from the Turkish State Meteorological
Service, cover around 40 years of the measurements, with two exceptions of stations 17095 (5 years)
and 17351 (26 years). There are total of 10,927,597 rows of data with measurement times concentrated
at 0:00 h (51.17%) and 12:00 h (48.44%), while the remaining items of data were measured at various
times. The dataset is evenly distributed among months and days. The statistical calculation were
performed by a software package, IBM SPSS Statistics (version 22.0).
Turgut and Usanmaz / Anadolu Univ. J. of Sci. and Technology – A – Appl. Sci. and Eng. 17 (5) - 2016
833
Table 1. Yearly data count
Year
Samsun
17030
İstanbul
17062
Erzurum
17095
Ankara
17130
İzmir
17220
Isparta
17240
Diyarbakır
17281
Adana
17351
Row
total
1970
-
-
-
-
2,138
-
-
-
2,138
1971
2,104
4,308
-
3,712
2,181
1,684
1,874
-
15,863
1972
3,751
4,290
-
3,719
3,101
1,860
2,364
-
19,085
1973
3,810
4,236
-
3,566
3,933
1,998
2,132
-
19,675
1974
468
3,336
-
3,507
1,777
1,863
1,352
-
12,303
1975
3,372
1,855
-
3,743
3,257
2,250
2,154
-
16,631
1976
3,704
4,168
-
3,685
3,700
2,270
2,757
-
20,284
1977
3,874
3,896
-
3,585
3,443
2,248
2,045
-
19,091
1978
726
2,838
-
3,698
3,552
355
369
-
11,538
1979
1,978
2,282
-
515
2,613
2,234
996
-
10,618
1980
2,359
1,913
-
3,739
3,385
2,257
2,030
-
15,683
1981
4,129
1,905
-
3,775
3,645
3,292
3,176
-
19,922
1982
4,421
2,169
-
3,746
3,309
4,367
4,025
-
22,037
1983
4,483
1,946
-
3,738
4,188
4,315
3,843
-
22,513
1984
4,480
2,166
-
3,783
4,422
4,533
990
-
20,374
1985
4,503
2,170
-
3,744
4,429
4,534
554
255
20,189
1986
4,449
2,223
-
3,771
4,438
4,516
4,487
4,257
28,141
1987
4,294
2,220
-
3,778
4,436
4,486
4,517
4,469
28,200
1988
4,262
2,220
-
3,460
3,513
3,579
4,523
3,561
25,118
1989
4,488
2,190
-
3,616
450
4,332
4,495
4,413
23,984
1990
4,458
4,070
-
3,782
7
4,389
4,519
4,386
25,611
1991
4,498
4,472
-
3,846
4,215
4,021
4,391
4,388
29,831
1992
4,378
4,475
-
4,461
4,501
4,139
4,307
4,393
30,654
1993
4,482
4,448
-
4,431
4,130
4,162
4,456
4,452
30,561
1994
3,112
4,313
-
4,543
3,068
3,055
4,506
4,466
27,063
1995
77,303
81,406
-
96,798
86,134
75,759
80,774
97,781
595,955
1996
41,299
63,695
-
97,341
85,237
63,981
60,007
31,053
442,613
1997
54,959
68,228
-
81,363
84,387
45,830
42,416
52,185
429,368
1998
70,926
74,190
-
90,070
89,876
63,497
86,070
82,812
557,441
1999
79,590
78,851
-
88,636
102,538
78,207
73,592
87,724
589,138
2000
67,130
65,636
-
27,351
107,256
76,639
52,321
67,939
464,272
2001
112,236
106,778
-
98,144
114,821
92,551
105,945
109,087
739,562
2002
113,594
109,148
-
98,793
101,684
96,580
101,069
112,882
733,750
2003
113,997
121,179
-
106,883
114,002
103,114
108,398
118,823
786,396
2004
124,065
123,456
-
109,081
112,496
106,763
109,777
121,557
807,195
2005
122,805
122,766
-
107,224
122,331
105,364
110,056
120,322
810,868
2006
122,833
122,637
262
108,396
121,343
103,153
108,467
120,826
807,917
2007
116,694
121,191
81,041
107,861
119,975
99,479
104,416
118,891
869,548
2008
115,533
122,476
67,924
106,814
120,746
100,131
107,253
105,951
846,828
2009
118,445
119,908
88,903
103,974
115,708
101,671
105,508
116,427
870,544
2010
8,467
8,824
6,587
10,600
8,287
7,881
-
8,449
59,095
Total
1,546,459
1,584,478
244,717
1,527,272
1,688,652
1,397,339
1,426,931
1,511,749
10,927,597
Within the dataset, this study focuses on the variables of altitude, wind speed, wind direction, ambient
temperature, ambient pressure and relative humidity. Data was collected from radiosonde stations and,
during that time, there were eight active radiosonde stations distributed around Turkey. The analyses
show that there are extreme cases at both ends in temperature, humidity and wind direction variables,
which is believed to be due to technical logging faults on a minor scale. Therefore, a valid range for
these variables is assumed as follows: i) +60°C> temperature>-100°C; ii) 100%< relative
humidity<0%; and iii) 360°<wind direction. The removed cases are 551 for temperature, 483 for
humidity and 5546 for wind direction, while a number of cases are invalid in terms of more than one
property.
Wind speed and direction data is directly obtained from the dataset and investigated with altitude
which is treated as a categorical variable. While this is also true for relative humidity (RH), another
parameter, mixing ratio (MR), is also derived to account for ambient temperature and pressure on the
water vapor absorbing capacity of the ambient air. The MR, defined as the ratio of the mass of water
vapor mixed into the mass of dry air (g/kg), can be calculated as follows [22]:
Turgut and Usanmaz / Anadolu Univ. J. of Sci. and Technology – A – Appl. Sci. and Eng. 17 (5) - 2016
834
(1)
where the B value is 621.9907 g/kg for air, Pw is the water pressure (hPa), and Ptot denotes the total
ambient pressure (hPa);
(2)
where Pws denotes water vapor saturation pressure (hPa);
(3)
and where A, n and Tn are constants, T is ambient temperature (°C).
3. RESULTS AND DISCUSSION
3.1. Wind Direction
The variations of wind direction and speed with months and altitudes are illustrated in Figure 2. One
general observation from Figure 2 is that wind direction steadily approximates towards 270° with an
increase in altitude. The wind direction, at the lowest and the highest altitude levels for overall the
average of all of the stations, is found to be 169° (±111) and 260° (±39), respectively. The higher
standard deviation is due to relatively variable low altitude wind directions and tends to decrease with
an increase in altitude. In addition, as altitude level increases, the change in wind direction between
neighboring altitude levels decreases.
High altitude level wind directions are always closer to 270° (from west) in winter months compared
to summer months. For instance, in 17095, the winter average wind direction is closer to 270° than for
summer at an average of 25° throughout altitude levels. In certain stations (e.g., 17030 and 17062) and
at about 5-10 km of altitude level, summer average wind direction can be observed to be slightly
closer to 270° than for those in winter. In 17220, 17281 and 17351, on the other hand, the summer
average wind direction at low altitudes is observed to be closer to 270°, with an average of 28°,
compared to winter.
For a given altitude, wind direction does not change significantly between winter and summer months,
except for the first altitude categories in certain stations. However, average wind direction for spring
and summer months (04-09) appears to be statistically different than those for winter months.
For stations relatively far from the sea (see Figure 2e to h), wind directions for the first altitude
category show considerable variation compared to neighboring altitude categories or between months.
For the rest of the stations, which are close to the sea, the variability in wind direction appears to be
smooth.
Turgut and Usanmaz / Anadolu Univ. J. of Sci. and Technology – A – Appl. Sci. and Eng. 17 (5) - 2016
835
Figure 2. Variation of vector field of wind with month and altitude of the eight radiosonde stations
Referring to Figure 2c, an anomaly should be noted. For this station, the wind vectors patterns are
significantly different than those for the other stations and do not follow the expected tendency at
particular altitudes. The most abrupt change in wind direction is observed at around 9000 m of
altitude. To address this anomaly, the data for this particular station was investigated in detail. Results
indicate that: i) the wind direction data suddenly and substantially changes from 1994 to 1995 and the
average of the wind direction (throughout the twelve months and for all altitude categories) before
1994 is about 23°, whereas it is 232° after 1995; and ii) for altitudes higher than 10,000 m, there is no
wind direction data available before 1994. While the reason remains unknown, such a sharp increase
in wind direction from 1994 to 1995 suggests a technical error. In addition, it should be noted that no
such cases were detected for wind speed.
The first finding (i) explains why the wind direction vectors have lower degrees and show different
patterns than expected. This is because when the average of wind direction throughout the years (1970
to 2010) are considered, the overall average is found to be relatively lower compared to other stations.
The second finding (ii), on the other hand, can be used to justify why wind direction vectors above
10,000 m are restored, and how they are now similar to those for the other stations. This is because
there is no wind direction data available before 1994 (that is thought to be unreliable) for above 10,000
m. This discussion concludes that the change in wind direction with altitudes appears to be noticeably
similar for all of the radiosonde stations.
3.2 Wind Speed
Variation of the wind speed profile with altitude is illustrated in Figure 3. The first impression made is
that wind speed varies with altitude in a monotonic manner and that the effect of altitude may be
different for certain months. Maximum monthly average wind speeds are generally observed in
March, up to an altitude of 10 km, followed by December, January and February, respectively. Above
10 km altitude, the highest monthly average of wind speed is observed in July and August, except for
Turgut and Usanmaz / Anadolu Univ. J. of Sci. and Technology – A – Appl. Sci. and Eng. 17 (5) - 2016
836
station 17351. Consideration of monthly variation of wind speed in a vertical axis reveals different
patterns at different stations. In the following sections, wind speed is discussed for each station.
Station 17030 Samsun
Average wind speed changes between 8.5 m/s at ground level and 37.2 m/s at 13 km of altitude.
Maximum wind speeds are observed in March (up to 10 km of altitudes) and June (above 11 km). The
variation of average wind speed with per kilometer of altitude is higher at lower altitudes. The increase
in wind speed with subsequent kilometers of altitudes appears to be higher than 15% up to 5 km of
altitude, whereas it is recorded to be less than 5% above 11 km of altitude.
Station 17062 İstanbul
Average wind speed changes between 10.1 m/s at ground level and 35.1 m/s at 13 km of altitude.
Maximum wind speeds are observed in March (up to 10 km of altitudes) and July (above 11 km),
albeit they are close to March wind speeds. The variation of average wind speed with per kilometer of
altitude is higher at lower altitudes. The increase in wind speed with subsequent kilometers of altitudes
appears to be higher than 13% up to 5 km of altitude, whereas it is recorded to be less than 5% above
10 km of altitude.
Figure 3. Effect of season on average wind speed – altitude relationship with stations (a-h)
Station 17220 İzmir
Average wind speed changes between 9.4 m/s at ground level and 39.0 m/s at 13 km of altitude. Maximum
wind speeds are observed in March (up to 10 km of altitudes) and July (above 11 km), albeit they are close
to March wind speeds. The variation of average wind speed with per kilometer of altitude is higher at lower
altitudes. The increase in wind speed with subsequent kilometers of altitudes appears to be higher than 14%
up to 5 km of altitude, whereas it is recorded to be less than 6% above 10 km of altitude.
Turgut and Usanmaz / Anadolu Univ. J. of Sci. and Technology – A – Appl. Sci. and Eng. 17 (5) - 2016
837
Station 17351 Adana
Average wind speed changes between 6.7 m/s at ground level and 39.8 m/s at 13 km of altitude.
Maximum wind speeds are observed in March (up to 6 km of altitudes) and February (above 7 km),
albeit they are close to March wind speeds. The variation of average wind speed with per kilometer of
altitude is higher at lower altitudes. The increase in wind speed with subsequent kilometers of altitudes
appears to be higher than 15% up to 6 km of altitude, whereas it is recorded to be less than 6% above
10 km of altitude.
Station 17095 Erzurum
Average wind speed changes between 9.7 m/s at 2 km (no data available for ground level and 1 km of
altitude) and 40.1 m/s at 13 km of altitude. Maximum wind speeds are observed in March (up to 10
km of altitudes) and July (above 11 km), albeit they are close to March wind speeds. The variation of
average wind speed with per kilometer of altitude is higher at lower altitudes. The increase in wind
speed with subsequent kilometers of altitudes appears to be higher than 13% up to 6 km of altitude,
whereas it is recorded to be less than 7% above 10 km of altitude.
Station 17130 Ankara
Average wind speed changes between 8.4 m/s at 1 km (highly variable data at ground) and 37.7 m/s at
13 km of altitude. The maximum wind speeds are observed in March (up to 5 km of altitudes), in
January (between 6 and 9 km of altitudes) and July/August (above 10 km), albeit they are close to
wind speeds in March. The variation of average wind speed with per kilometer of altitude is higher at
lower altitude. The increase in wind speed with subsequent kilometers of altitudes appears to be higher
than 13% up to 5 km of altitude, whereas it is recorded to be less than 6% above 10 km of altitude.
Station 17240 Isparta
Average wind speed changes between 7.8 m/s at 1 km (highly variable data for ground level) and 40.3
m/s at 13 km of altitude. Maximum wind speeds are observed in March (up to 9 km of altitudes) and
July/August (above 10 km), albeit they are close to March wind speeds. The variation of average wind
speed with per kilometer of altitude is higher at lower altitudes. The increase in wind speed with
subsequent kilometers of altitudes appears to be higher than 13% up to 5 km of altitude, whereas it is
recorded to be less than 6% above 10 km of altitude.
Station 17281 Diyarbakır
Average wind speed changes between 8.5 m/s at 1 km (highly variable data for ground level) and 42.2
m/s at 13 km of altitude. Maximum wind speeds are observed in June (at 1 km of altitude), in March
(between 2 and 8 km of altitudes) and February/August (above 9 km), albeit they are close to March
wind speeds. The variation of average wind speed with per kilometer of altitude is higher at lower
altitudes. The increase in wind speed with subsequent kilometers of altitudes appears to be higher than
15% up to 6 km of altitude, whereas it is recorded to be less than 7% above 10 km of altitude.
From the monthly wind speed analyses for individual stations, illustrated in Figure 4, it can be noted
that the wind speed variation with altitude exhibits two profiles. While both profiles indicate a strong
effect of altitude, for the summer months, (i.e., months 06-09), the wind speed increase with altitude
remains at a moderate level for several of the first low altitude categories compared to other months,
where the wind speed increase with altitude occurs at a rapid pace.
Turgut and Usanmaz / Anadolu Univ. J. of Sci. and Technology – A – Appl. Sci. and Eng. 17 (5) - 2016
838
Figure 4. Effect of station on wind speed – altitude relationship based on monthly measurements (a-l).
The symbols represent data anomaly due to low data for the corresponding station
3.3 Relative Humidity and Mixing Ratio
In Figure 5, the variation of relative humidity and mixing ratio are illustrated for a randomly selected
year (2005) and month (May) for station 17130 at the 12:00h measurement period. Although, there is a
strong tendency toward lower RH at higher altitudes, the effect of altitude on daily RH can be
significantly variable. While on certain days a monotonic decrease in RH with an increase in altitude
can be seen, significantly different variations in RH can be also observed. It should be noted that
average standard deviations are around 5%, and are slightly higher for higher altitudes. The mixing
ratio in Figure 5 is calculated using Eqs.(1-3). Depending on ambient temperatures, two sets of
constants are used as the humidity, being assumed as water (between -19°C and +50°C) and ice
(between -70°C and -20°), with the maximum errors of these constants given as 0.083% and 0.052%,
respectively [22]. Unlike the RH, the daily average mixing ratio shows better correlations with
altitude. While the correlation indicates an almost linear variation at certain altitude ranges, from an
altitude around 5000-7000 m, the mixing ratios generally decrease to below 0.5 g/kg.
Turgut and Usanmaz / Anadolu Univ. J. of Sci. and Technology – A – Appl. Sci. and Eng. 17 (5) - 2016
839
Figure 5. Variation of mean relative humidity and mixing ratio with altitude. Error bars represent 1
The daytime and night-time RH observations are also compared. Usually, night-time values are found
to be higher than those for daytime. Independent t tests between these two periods suggest significant
differences in the average of RH at different altitudes and months, where the night-time values are
found to be an average of 3-4% higher compared to daytime values. Furthermore, the differences tend
to be higher at ground level.
The effects of altitude and month on RH are illustrated in Figure 6. The first observation made from
Figure 6 is that the humidity usually decreases with an increase in altitude. At certain stations, due to
the fact that ground level humidity is relatively lower, the change in humidity may not be monotonic.
Furthermore, the ground level humidity at stations located close to the sea are higher (ranging from
56% to 68%) than those for the other stations (ranging from 41% to 47%). Although it is far from sea,
in one exceptional station, 17281, relatively higher humidity (52%) is observed, which can be
attributed to the presence of local natural lakes and dam reservoirs around the station.
Turgut and Usanmaz / Anadolu Univ. J. of Sci. and Technology – A – Appl. Sci. and Eng. 17 (5) - 2016
840
Figure 6. Variation of relative humidity with altitude and month for different radiosonde stations (a-h)
As shown in Figure 6 e-h, the lowest RH levels are always observed in July and August for altitudes
equal to and greater than 7000 m. These results are also checked using the independent t test where the
test results do not express significant differences in humidity between these months. For stations close
to the sea, this result can even be lowered for altitudes of 1000-3000 m, depending on the station.
Nonetheless, it should be noted that generally, airborne RH is observed to be lowest for the hottest
summer months. This result can be extended as to the second and third lowest RH group for the
months of June and September, and May and October.
The variation of MR with altitude and month for each station is illustrated in Figure 7 e-h. It should be
noted that, the MR values steadily decrease with altitude up to 8000-9000 m, where it becomes
negligible above these altitude levels. Furthermore, compared to the RH curves in Figure 6, MR
curves exhibit a better and more predictable relationship with altitude and months. Unlike RH, the
summer months (06-09) show higher MR. Independent t test results reveal that, up to 1000 m of
altitude, the MR values at stations close to sea are higher than for other stations, whereas at higher
altitudes of up to 5000 m there is an opposite trend (p<0.001). For the remaining high altitude
categories, absolute MR values are so low that it is not possible to draw a statistically significant
conclusion.
Turgut and Usanmaz / Anadolu Univ. J. of Sci. and Technology – A – Appl. Sci. and Eng. 17 (5) - 2016
841
Figure 7. Variation of the mixing ratio with altitude and month for different radiosonde stations (a-h)
A data anomaly is detected for station 17062 and December. Referring to Figure 7b, it is observed
only at the altitude level categories below 2 km (±500 m). An investigation into this anomaly reveals
that it is caused by the data for 1997 only. When the 1997 data is excluded from the dataset, the
average mixing ratio is calculated as 3.75 g/kg and 2.88 g/kg for 1 km and 2 km of altitude categories.
However, for the dataset, including 1997 data, the same values increase to 3.97 g/kg and 4.17 g/kg,
respectively. To address this anomaly, the RH, temperature and pressure values are also checked.
While considerably lower values are observed for RH in 1997, the temperature and the pressure values
are found to be significantly different compared to data for the other years. Numerically, the averages
of temperature and pressure for all the years, except 1997, are calculated to be 1.4°C and 849 hPa, and
-0.5°C and 830 hPa for 1 and 2 km of altitude, respectively. The averages of temperature and pressure
only for 1997 are found to be 17.3°C and 459 hPa, and 9.2°C and 457 hPa for 1 and 2 km of altitude.
These results indicate considerable high temperature and low pressure for corresponding altitude
levels, which may be due to technical errors or being an extreme case.
Analyses also show that the wind speed and RH can be predicted, though within a large tolerance,
depending on altitude, where average wind speed increases quadratically, while average RH decreases
linearly with an increase in altitude. Bearing in mind that this study was not focuses on the modelling
of the mentioned atmospheric parameters and comprehensive regression analyses were not performed,
though it would be useful to develop general equations for these variables.
In this respect, results show that the relationship between wind speed and altitude can be quadratic
(R2=0.996) or a cubic function (R2=0.999). The relationship may be even linear, if accuracy is not a
major concern. The regression equations for the average wind speed and RH (R2=0.981) is given as
follows:
(4)
Turgut and Usanmaz / Anadolu Univ. J. of Sci. and Technology – A – Appl. Sci. and Eng. 17 (5) - 2016
842
(5)
where the altitude variable is in meters. The first equation suggests that wind speed increases at an
average of 2.3 m/s for each kilometer of altitude, and this increment decreases with altitude. Regarding
the second regression equation, the ground level average RH can be noted at about 60%, as the
International Standard Atmosphere (ISA) assumptions at ground level, and decreasing by an average
of 4% with each kilometer increase of altitude. It should also be noted that the variation of RH with
altitude at each individual station reveals a 3.6-5.2% decrease for each 1 km increase in altitude. For a
typical cruise altitude of 11 km, it is found to be 16%. Similarly, a general regression equation for
average MR may be given as follows (R2=0.999):
(6)
where the altitude variable is in meters. However, it should be noted again that, despite highly
significant R2 values, these regression equations involve large tolerances, due to location, date and
time, and care should be taken when drawing conclusions. Comprehensive regression analyses will be
conducted in future studies.
4. CONCLUSION
Being two of the less predictable atmospheric parameters, wind speed and direction are of paramount
importance for instrument flight procedure designers, airlines, pilots and air traffic controllers.
Accurate wind estimations are also essential to improve the navigation and guidance performance of
unmanned aerial vehicles. They may affect flight fuel consumption, flight time, cruise altitude and
descent trajectory. In addition, to set protection areas around the flight trajectory of an aircraft, the
wind speed and direction at corresponding altitude levels must be known. There are certain wind speed
models providing wind speed prediction with significant overestimations to some extent. Nonetheless,
regional wind speed models are thought to reduce these overestimations and to provide better
utilization of air space. Humidity is also important, yet difficult to ascertain, particularly from an
engine performance and NOx emissions points of view.
This study describes variations of wind speed, wind direction and humidity, in Turkey, depending on
altitude, location and season. A dataset, including almost forty years of statistical data, has been
obtained from eight radiosonde stations operating in different regions of Turkey. It is found that
average wind direction changes from 169° at the lowest altitude to 260° at the highest altitude. Unlike
for the higher altitudes, at certain stations which are relatively far from the sea, wind direction at
ground level is observed to be largely unpredictable. However, predictability substantially increases
for these stations relatively close to the sea.
Maximum monthly average wind speeds are generally observed in March, up to an altitude of 10 km.
Average wind speed at 13 km of altitude can be between 35 m/s and 42 m/s, depending on location.
While daily relative humidity may be highly variable, there is a strong tendency toward lower relative
humidities at higher altitudes. The lowest relative humidities are observed in July and August for
altitudes equal to and greater than 7 km. Moreover, night-time relative humidity measurements are
found to be on average 3-4% higher than those for daytime measurements.
Since ambient pressure and temperature are also taken into consideration, a mixing ratio reveals a
better relationship with altitude. Unlike relative humidity, the summer months reveals higher mixing
ratios.
Turgut and Usanmaz / Anadolu Univ. J. of Sci. and Technology – A – Appl. Sci. and Eng. 17 (5) - 2016
843
Lastly, general regression equations for wind speed, relative humidity and mixing ratio are developed
with altitude being an independent variable. As a result, it should be noted that average wind speed
increases by 2.3 m/s, while average relative humidity decreases by 4%, for an increase of one
kilometer in altitude.
ACKNOWLEDGEMENT
The authors would like to thank the Turkish State Meteorological Service for providing the radiosonde
data for analyses.
REFERENCES
[1] Gaffen DJ, Elliott WP, Robock A. Relationships between tropospheric water vapor and surface
temperature as observed by radiosondes. Geophysical Research Letters 1992; 19 (18): 1839–1842.
[2] Christy JR, Norris WB, Spencer RW, Hnilo JJ. Tropospheric temperature change since 1979 from
tropical radiosonde and satellite measurements. Journal of Geophysical Research: Atmospheres 2007;
112(D6): D06102.
[3] Helten M, Smit HGJ, Sträter W, Kley D, Nedelec P, Zöger M, Busen R. Calibration and
performance of automatic compact instrumentation for the measurement of relative humidity from
passenger aircraft. Journal of Geophysical Research: Atmospheres 1998; 103(D19): 25643–25652.
[4] Khelif D, Burns SP, Friehe CA. Improved wind measurements on research aircraft. Journal of
Atmospheric and Oceanic Technology 1999; 16(7): 860–875.
[5] Reuder J, Brisset P, Jonassen M, Müller M, Mayer S. The Small Unmanned Meteorological
Observer SUMO: A new tool for atmospheric boundary layer research. Meteorologische Zeitschrift
2009; 18(2): 141–147.
[6] Martin S, Bange J, Beyrich F. Meteorological profiling of the lower troposphere using the research
UAV "M2AV Carolo". Atmospheric Measurement Techniques 2011; 4(4): 705–716.
[7] Baughcum Tritz G, Henderson C, Pickett C, Scheduled civil aircraft emission inventories and
analysis for 1992: Database, NASA, 1996.
[8] Solyali D, Altunç M, Tolun S, Aslan Z. Wind resource assessment of Northern Cyprus. Renew.
Sustain. Energy Rev 2016; 55: 180–187.
[9] Fant C, Gunturu B, Schlosser A. Characterizing wind power resource reliability in southern Africa.
Appl. Energy 2016: 161; 565–573.
[10] De Meij A, Vinuesa JF, Maupas V, Waddle J, Price I, Yaseen B, Ismail A. Wind energy resource
mapping of Palestine. Renew. Sustain. Energy Rev 2016: 56; 551–562.
[11] Stell L, Bronsvoort J, McDonald G. Regression Analysis of Top of Descent Location for Idle-
thrust Descents. In: Tenth USA/Europe Air Traffic Management Research and Development Seminar
(ATM2013); 2013; Chicago, Illinois USA.
[12] Franco A, Rivas D, Valenzuela A. Optimization of Unpowered Descents of Commercial Aircraft
in Altitude-Dependent Winds. Journal of Aircraft 2012: 49; 1460–1470.
Turgut and Usanmaz / Anadolu Univ. J. of Sci. and Technology – A – Appl. Sci. and Eng. 17 (5) - 2016
844
[13] De Jong PMA, van der Laan JJ, in’t Veld AC, van Paassen MM, Mulder M. Wind-Profile
Estimation Using Airborne Sensors. Journal of Aircraft 2014: 51; 1852–1863.
[14] Rivas D, Lopez-Garcia O, Esteban S, Gallo E. An analysis of maximum range cruise including
wind effects, Aerosp. Sci. Technol. 14 (2010) 38–48.
[15] Cho A, Kim J, Lee S, and Kee C, “Wind Estimation and Airspeed Calibration using a UAV with
a Single-Antenna GPS Receiver and Pitot Tube,” IEEE Transactions on Aerospace and Electronic
Systems, vol. 47, 2011, pp. 109–117.
[16] Petrich J and Subbarao K, “On-Board Wind Speed Estimation for UAVs,” AIAA Guidance,
Navigation and Control Conference, Portland, Oregon, 2011.
[17] ICAO. Procedures for Air Navigation Services–Aircraft Operations (PANS OPS). Construction of
Visual and Instrument Flight Procedures, Vol. 2, 6th ed., ICAO, Montreal, 2014
[18] Usanmaz O, Turgut ET. Wind Effect Analysis on Instrument Flight Procedures Using a Turkish
Wind Model 2012: 49; 2023–2032.
[19] Lewis GD. Prediction of NOx Emissions. In: ASME 1981 Int. Gas Turbine Conference and
Products Show; March 9-12, 1981; Houston, Texas, USA: American Society of Mechanical
Engineers. pp. V002T06A023–V002T06A023.
[20] DuBois D, and Paynter GC, Fuel flow method 2 for estimating aircraft emissions, SAE
International, 2006.
[21] Wasiuk DK, Lowenberg MH, and Shallcross DE, “An aircraft performance model
implementation for the estimation of global and regional commercial aviation fuel burn and
emissions,” Transportation Research Part D: Transport and Environment, vol. 35, 2015, pp. 142–159.
[22] Vaisala. Humidity Conversion Formulas. Helsinki, 2013. http://www.vaisala.com/Vaisala
Documents/Application notes/Humidity_Conversion_Formulas_B210973EN-F.pdf (accessed: May
10, 2016)
