climb rate – altitude correlation from analysis of igc files
TRANSCRIPT
Christof Maul, OSTIV Meteorology Panel Workshop, Pfaffstätten/A, 26.9.09 Climb Rate Altitude Correlation from Analysis of IGC Files
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
abun
danc
e, %
vertical speed, m/s
Climb Rate – Altitude Correlation from Analysis of IGC Files
Christof MaulTechnische Universität BraunschweigJohann Wolfgang Goethe-Universität Frankfurt (Akaflieg Frankfurt)
Christof Maul, OSTIV Meteorology Panel Workshop, Pfaffstätten/A, 26.9.09 Climb Rate Altitude Correlation from Analysis of IGC Files
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
abun
danc
e, %
vertical speed, m/s
Source data: Competition flights
Advantages: Open accessWell-defined pilot strategyWell-defined terrain and weather conditions
B1400235136607N01245193EA016980181000208173008B1400475136514N01246159EA016790178000207169010
Convert data string to 5-dimensional flight vector V(t, y, x, h, w=dh/dt)
14.0064, 51.6101, 12.7532, 1698, -2.6714.0131, 51.6086, 12.7693, 1679, -0.79
Christof Maul, OSTIV Meteorology Panel Workshop, Pfaffstätten/A, 26.9.09 Climb Rate Altitude Correlation from Analysis of IGC Files
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
abun
danc
e, %
vertical speed, m/s
Competition data pre-treatment
removal of pre-start and post-finish data
Christof Maul, OSTIV Meteorology Panel Workshop, Pfaffstätten/A, 26.9.09 Climb Rate Altitude Correlation from Analysis of IGC Files
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
abun
danc
e, %
vertical speed, m/s
Competition data pre-treatment
removal of pre-start and post-finish data
Christof Maul, OSTIV Meteorology Panel Workshop, Pfaffstätten/A, 26.9.09 Climb Rate Altitude Correlation from Analysis of IGC Files
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
abun
danc
e, %
vertical speed, m/s
Analyzed competitions (2007):
1) Lilienthal Glide, Lüsse, Germany, July 14 to 2797 competitorsJuly 16, 23, 26254 flights127,000 data points
2) European Gliding Championships, Issoudun, France, Aug 6 to 19 91 competitorsAugust 9, 12, 13, 16, 18386 flights177,000 data points
3) Junior World Gliding Championships, Rieti, Italy, July 28 to Aug 1153 competitorsJuly 29, 30, August 4, 6, 7221 flights95,000 data points
Christof Maul, OSTIV Meteorology Panel Workshop, Pfaffstätten/A, 26.9.09 Climb Rate Altitude Correlation from Analysis of IGC Files
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
abun
danc
e, %
vertical speed, m/s
Basic results – tracks – y(x):
well defined geographical area with homogenous samplingLilienthal Glide
Christof Maul, OSTIV Meteorology Panel Workshop, Pfaffstätten/A, 26.9.09 Climb Rate Altitude Correlation from Analysis of IGC Files
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
abun
danc
e, %
vertical speed, m/s
From variograms w(t) to vertical speed distributions f(w):
projection of variogram onto w axis
-6-5
-4-3
-2-1
01
23
45
60123456
ve
rtica
l spe
ed, m
/s
abundance, %23
07
Christof Maul, OSTIV Meteorology Panel Workshop, Pfaffstätten/A, 26.9.09 Climb Rate Altitude Correlation from Analysis of IGC Files
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
abun
danc
e, %
vertical speed, m/s
Climb and sink vertical speed distributions are general features
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
abun
danc
e, %
abun
danc
e, %
abun
danc
e, %
abun
danc
e, %
abun
danc
e, %
0908
1208
1308
1608
vertical speed, m/s
1808
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
vertical speed, m/s
abun
danc
e, %
abun
danc
e, %
abun
danc
e, %
abun
danc
e, %
abun
danc
e, % 2907
3007
0408
0608
0708
Rieti Issoudun
Christof Maul, OSTIV Meteorology Panel Workshop, Pfaffstätten/A, 26.9.09 Climb Rate Altitude Correlation from Analysis of IGC Files
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
abun
danc
e, %
vertical speed, m/s
From barograms h(t) to altitude distributions f(h):
projection of barogram onto h axis
050
010
0015
0020
00012345678
2307
abundance, %
altit
ude,
m
Christof Maul, OSTIV Meteorology Panel Workshop, Pfaffstätten/A, 26.9.09 Climb Rate Altitude Correlation from Analysis of IGC Files
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
abun
danc
e, %
vertical speed, m/s
Circling vs. straight climb
sink fs (h)w = dh/dt < 0
climb fc (h)w = dh/dt > 0
thermal/circling fcc(h)w = dh/dt > 0
(dx/dh)2 + (dy/dh)2
0
aligned lift fcs(h)w = dh/dt > 0
(dx/dh)2 + (dy/dh)2
0
total f(h)
Christof Maul, OSTIV Meteorology Panel Workshop, Pfaffstätten/A, 26.9.09 Climb Rate Altitude Correlation from Analysis of IGC Files
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
abun
danc
e, %
vertical speed, m/s
Altitude distributions of thermal fcc(h) () vs. aligned fcs(h) () lift. Significant variations, aligned lift accounts for as much as 50% of total lift!
Rieti Issoudun
0 500 1000 1500 20000
2
4
6
8
0 500 1000 1500 20000
2
4
6
8
0 500 1000 1500 20000
2
4
6
8
0 500 1000 1500 20000
2
4
6
8
0 500 1000 1500 20000
2
4
6
8
circling climb straight climb
altitude, m
0908
1208
1308
1608
1808
abu
ndan
ce, %
abu
ndan
ce, %
abu
ndan
ce, %
abu
ndan
ce, %
abu
ndan
ce, %
0 500 1000 1500 2000 2500 3000 35000
1
2
3
4
5
6
0 500 1000 1500 2000 2500 3000 35000
1
2
3
4
5
6
0 500 1000 1500 2000 2500 3000 35000
1
2
3
4
5
6
0 500 1000 1500 2000 2500 3000 35000
1
2
3
4
5
6
0 500 1000 1500 2000 2500 3000 35000
1
2
3
4
5
6 circling climb straight climb
altitude, m
2907
3007
0408
0608
0708
abu
ndan
ce, %
abu
ndan
ce, %
abu
ndan
ce, %
abu
ndan
ce, %
abu
ndan
ce, %
Christof Maul, OSTIV Meteorology Panel Workshop, Pfaffstätten/A, 26.9.09 Climb Rate Altitude Correlation from Analysis of IGC Files
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
abun
danc
e, %
vertical speed, m/s
'dh)'h(f)'h(f)h(fh
0outin
cc
Description of circling climb altitude distributions fcc(h,w>0).
If w-h correlatins are absent, fcc(h,w>0) can be obtained by integrating over thermal entering and thermal exiting altitude distributions fin (h) and fout (h):
Consequently, fin (h) and fout (h) can be obtained by differentiating f(h):
)h(f)h(fdh
)h(dfoutin
cc
Alternatively, fin (h) and fout (h) can be directly obtained from identifying thermals from flight recorder data.
Both methods should give identical results!
(1)
(2)
Christof Maul, OSTIV Meteorology Panel Workshop, Pfaffstätten/A, 26.9.09 Climb Rate Altitude Correlation from Analysis of IGC Files
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
abun
danc
e, %
vertical speed, m/s
Excellent agreement between altitude distributions fcc(h) obtained from flight recorder data () and from integration of eq.1 ().
Rieti Issoudun
0 500 1000 1500 2000 2500 3000 35000
1
2
3
4
5
6
0 500 1000 1500 2000 2500 3000 35000
1
2
3
4
5
6
0 500 1000 1500 2000 2500 3000 35000
1
2
3
4
5
6
0 500 1000 1500 2000 2500 3000 35000
1
2
3
4
5
6
0 500 1000 1500 2000 2500 3000 35000
1
2
3
4
5
6 circling climb straight climb calculated circling climb
altitude, m
2907
3007
0408
0608
0708
abu
ndan
ce, %
abu
ndan
ce, %
abu
ndan
ce, %
abu
ndan
ce, %
abu
ndan
ce, %
0 500 1000 1500 20000
2
4
6
8
0 500 1000 1500 20000
2
4
6
8
0 500 1000 1500 20000
2
4
6
8
0 500 1000 1500 20000
2
4
6
8
0 500 1000 1500 20000
2
4
6
8
circling climb straight climb calculated circling climb
altitude, m
0908
1208
1308
1608
1808
abu
ndan
ce, %
abu
ndan
ce, %
abu
ndan
ce, %
abu
ndan
ce, %
abu
ndan
ce, %
Christof Maul, OSTIV Meteorology Panel Workshop, Pfaffstätten/A, 26.9.09 Climb Rate Altitude Correlation from Analysis of IGC Files
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
abun
danc
e, %
vertical speed, m/s
Result corroborated by absence of w-h correlation.
Rieti Issoudun
Mean vertical speeds (black) are essentially independent of altitude. In contrast, maximum and minimum vertical speed values (blue) occur at medium altitudes.
Christof Maul, OSTIV Meteorology Panel Workshop, Pfaffstätten/A, 26.9.09 Climb Rate Altitude Correlation from Analysis of IGC Files
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
abun
danc
e, %
vertical speed, m/s
Result is in contradiction to accepted theory, see e.g. "Meteorology and Flight"by T. Bradbury:
"Average rates of climb may be onlyhalf these values."
Christof Maul, OSTIV Meteorology Panel Workshop, Pfaffstätten/A, 26.9.09 Climb Rate Altitude Correlation from Analysis of IGC Files
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
abun
danc
e, %
vertical speed, m/s
Excellent agreement between altitude distributions fcc(h) obtained from flight recorder data () and from integration of eq.1 ().
Rieti Issoudun
0 500 1000 1500 2000 2500 3000 35000
1
2
3
4
5
6
0 500 1000 1500 2000 2500 3000 35000
1
2
3
4
5
6
0 500 1000 1500 2000 2500 3000 35000
1
2
3
4
5
6
0 500 1000 1500 2000 2500 3000 35000
1
2
3
4
5
6
0 500 1000 1500 2000 2500 3000 35000
1
2
3
4
5
6 circling climb straight climb calculated circling climb
altitude, m
2907
3007
0408
0608
0708
abu
ndan
ce, %
abu
ndan
ce, %
abu
ndan
ce, %
abu
ndan
ce, %
abu
ndan
ce, %
0 500 1000 1500 20000
2
4
6
8
0 500 1000 1500 20000
2
4
6
8
0 500 1000 1500 20000
2
4
6
8
0 500 1000 1500 20000
2
4
6
8
0 500 1000 1500 20000
2
4
6
8
circling climb straight climb calculated circling climb
altitude, m
0908
1208
1308
1608
1808
abu
ndan
ce, %
abu
ndan
ce, %
abu
ndan
ce, %
abu
ndan
ce, %
abu
ndan
ce, %
systematicallytoo small
systematicallytoo large
A second look to altitude distributions
Christof Maul, OSTIV Meteorology Panel Workshop, Pfaffstätten/A, 26.9.09 Climb Rate Altitude Correlation from Analysis of IGC Files
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
abun
danc
e, %
vertical speed, m/s
Excellent agreement between altitude distributions fcc(h) obtained from flight recorder data () and from integration of eq.1 ().
Rieti Issoudun
0 500 1000 1500 2000 2500 3000 35000
1
2
3
4
5
6
0 500 1000 1500 2000 2500 3000 35000
1
2
3
4
5
6
0 500 1000 1500 2000 2500 3000 35000
1
2
3
4
5
6
0 500 1000 1500 2000 2500 3000 35000
1
2
3
4
5
6
0 500 1000 1500 2000 2500 3000 35000
1
2
3
4
5
6 circling climb straight climb calculated circling climb
altitude, m
2907
3007
0408
0608
0708
abu
ndan
ce, %
abu
ndan
ce, %
abu
ndan
ce, %
abu
ndan
ce, %
abu
ndan
ce, %
0 500 1000 1500 20000
2
4
6
8
0 500 1000 1500 20000
2
4
6
8
0 500 1000 1500 20000
2
4
6
8
0 500 1000 1500 20000
2
4
6
8
0 500 1000 1500 20000
2
4
6
8
circling climb straight climb calculated circling climb
altitude, m
0908
1208
1308
1608
1808
abu
ndan
ce, %
abu
ndan
ce, %
abu
ndan
ce, %
abu
ndan
ce, %
abu
ndan
ce, %
GoodA second look to altitude distributions
Christof Maul, OSTIV Meteorology Panel Workshop, Pfaffstätten/A, 26.9.09 Climb Rate Altitude Correlation from Analysis of IGC Files
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
abun
danc
e, %
vertical speed, m/s
Strategy
fit 3rd order polynomial tomaximum climb rate data (black)
fix intercept at -0.5 m/s and compare to Bradbury
divide result by 2 (averagerates of climb may be only half these values) and compareto mean climb rate data (red)
A second look to w-h correlations
Christof Maul, OSTIV Meteorology Panel Workshop, Pfaffstätten/A, 26.9.09 Climb Rate Altitude Correlation from Analysis of IGC Files
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
abun
danc
e, %
vertical speed, m/s
3rd order polynomial fir tomaximum climb rate dataintercept -0.5 m/s
No fit! Result from abovedivided by 2, compared tomean climb rate data
Issoudun 20070809
Issoudun 20070809
Christof Maul, OSTIV Meteorology Panel Workshop, Pfaffstätten/A, 26.9.09 Climb Rate Altitude Correlation from Analysis of IGC Files
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
abun
danc
e, %
vertical speed, m/s
"Bradbury correlation" applied to mean thermaling climb ratesof Luesse and Issoudun competitions
Christof Maul, OSTIV Meteorology Panel Workshop, Pfaffstätten/A, 26.9.09 Climb Rate Altitude Correlation from Analysis of IGC Files
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
abun
danc
e, %
vertical speed, m/s
0
500
1000
1500
2000
2500
3000
3500
0 1 2 3 4 5
Iss1608 Iss1308 Iss1208 Iss0908 Iss1808 Lue1607 Lue2307 Lue2607
climb rate / ms-1
altit
ude
(m)
final comparison of statistical analysis to Bradbury's model
another excellent agreement!
Christof Maul, OSTIV Meteorology Panel Workshop, Pfaffstätten/A, 26.9.09 Climb Rate Altitude Correlation from Analysis of IGC Files
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
abun
danc
e, %
vertical speed, m/s
Another add-on to previous workconcerning vertical speed distributions
Gaussian or non-Gaussian?
-5 -4 -3 -2 -1 0 1 2 3 4 50
500
1000
1500
2000
sum
climb/sink rate / ms-1
0,0 0,5 1,0 1,5 2,0 2,5 3,0 3,5 4,0 4,5 5,0 5,5 6,00
200
400
600
800
1000
1200
Cou
nt
Bin Center (m/s)
f(w) straight f(w) circle
straight and circlingGaussian
straight flight onlyGaussian
same databut look non-Gaussian
Issoudun, 9.8.07
-5 -4 -3 -2 -1 0 1 2 3 4 50
500
1000
1500
2000
2500C
ount
climb/sink rate / ms-1
straight circle
Christof Maul, OSTIV Meteorology Panel Workshop, Pfaffstätten/A, 26.9.09 Climb Rate Altitude Correlation from Analysis of IGC Files
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
abun
danc
e, %
vertical speed, m/s
0 500 1000 1500 20000
2
4
6
8
altitude distributionstraight and circling
vertical speed distributionstraight and circling
more than 50% straight climbcircling climb rate distribution is Gaussian
Issoudun, 16.8.07
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
200400600800
10001200140016001800200022002400260028003000
Cou
nt
Bin Center
straight flight circling
Christof Maul, OSTIV Meteorology Panel Workshop, Pfaffstätten/A, 26.9.09 Climb Rate Altitude Correlation from Analysis of IGC Files
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 60
2
4
6
8
10
abun
danc
e, %
vertical speed, m/s
Summary of new results
1) vertical speed – altitude correlation
A close look reveals very good agreement of evaluated datawith correlation reported by Bradbury.
2) vertical speed distribution
No deviation from Gaussian normal distributions is observed.Holds also for circling only climb data.