Lidar monitoring of aircraft emissions for environmental air quality


22 Φεβ 2014 (πριν από 3 χρόνια και 5 μήνες)

197 εμφανίσεις

Lidar monitoring of aircraft emissions for
environmental air quality

S Christie
, M
A Graham

and D Raper

School of Chemical Engineering and Analytical Science, University of
Manchester, Sackville Street, PO Box 88, Manchester. M60 1QD. UK.

Centre for Air Transport and the Environment, Manchester Metropolitan
University, Chester Street, Manchester. M1 5GD. UK.

Corresponding author:

. The Univ
ersity of Manchester RASCAL Lidar facility has been
deployed at Heathrow and Manchester Airports to collect backscatter data from
exhaust emissions in the wakes of several hundred flights. The Lidar uses a
tripled eye
safe Nd:YAG at 30 Hz and 33
mJ/pulse. Although Lidar
measures the backscatter from airborne aerosols in the exhaust plume, the
dispersion of pollutant gases can also be inferred since the composition of
exhaust gases is known to a first approximation.

The size range of aerosols
ed by aircraft engines is well matched to light scattering at 355 nm
wavelength, and L
idar is well suited to studying the dispersion of aircraft
emissions. Aerosols associated with the exhaust from engines and tyre smoke
generated on touchdown were observe
d up to
90 s after emission. Modern
engines are so efficient, however, that only minimal soot is emitted. The
backscatter from the plume behind an aircraft in take
off roll showed an
enhancement of typically only a few 10’s of percent relative to that from


air. Interestingly, the signal was more clearly visible at low ambient temperatures
and high relative humidity, implying a significant hygroscopic component to the
aerosol. Tyre smoke was easily visible, giving backscatter an order of magnitude
eater than that from the engines.

Introduction and context

Aviation has become an integral part of the social and global economy. Furthermore, the
expansion of the air transport industry is expected to continue. The growth in air traffic,
coupled with the

ever increasing urbanization of airports as our cities expand, has
resulted in increasing concerns regarding the air quality in the vicinity of airports.

The principal emissions from aircraft gas turbine engines are carbon dioxide (CO
and water vapour
0) from the combustion of kerosene. Secondary emissions from the
combustion process consist of nitric oxide, nitrogen dioxide (termed together as NO
sulphur oxides (SO
carbon monoxide (CO),
burnt hydrocarbons (UHC) and soot
particles. Other ter
tiary exhaust constituents together with a discussion of the complex
chemical and dynamic processes they undergo are given in reference [1]. With the
exception of SO
, which is largely dependent on fuel sulphur content, the relative
concentrations of these

secondary emissions are dependent upon the engine operating
mode and the combustor design. While taxiing at low engine settings, for example,
where efficiency is less than optimal, the combustor operates at a lower temperature,
leading to the concentratio
n of UHC being relatively high and the concentration of NO

relatively low. Furthermore, reductions in aircraft exhaust emissions over recent years
have largely arisen from improvements in combustor design, with better control of the
air/fuel ratios and te
mperatures of the gases at different points in the combustor [2].

Under set conditions, the
rate at which

aviation fuel is burned by a given aircraft
engine, as well as its emissions, are well characterized. These data are documented in
the ICAO engine e
xhaust emissions data bank [3]. Less well established, however, is
how the exhaust emissions mix and disperse. Of particular importance are ‘ground
based’ exhaust plumes as these may directly affect local air quality. To put this into
context for the later

discussion of the Lidar measurements, let us consider the example
of a large aircraft such as a 747
400er powered by four General Electric CF6 engines
about to take off. The maximum take
off weight is 405 t and, at the
100% power
setting used for
off, each engine will burn 2.4 kg s

of kerosene, corresponding to a
total production of approximately 29 kg s

, 13 kg s

O, and 0.24 kg s




is in the form of
limit for NO

40 μg m

of air [4]
, since high


levels have been associated with wheezing and
shortness of breath


the consequences for people with lung conditions like asthma
being especially severe. Accordingly, an exhaust emission of this magnitude that does
not rapidly
disperse can have a significant impact on local air quality.

Lidar is a remote sensing technique which has been used for many years to map the
distribution of atmospheric aerosols [5]. It is the optical equivalent of radar: a pulse of
light emitted from
the Lidar undergoes absorption and scattering; a small fraction of this
scattered light is reflected straight back; and this backscattered light is collected using a
telescope and focused onto a detector. The detector signal is digitized as a function of
ime, and so the range an
d distribution of the aerosol may mapped.
The enhanced
concentration of aerosol in the wake of aircraft is principally derived from engine
exhausts. Although Lidar strictly measures the backscattered light from airborne
aerosols in
the exhaust plume, the dispersion of pollutant gases can be inferred. The
dispersion of aircraft exhaust plumes at ground level i
s dependent upon many dynamic
factors. A goal for the present study has been to observe the dispersion and identify the
key par
ameters and limits. This information is being used to develop

models of exhaust

dispersion, so that air quality may be better

Within the aviation industry, Lidar has previously been used in many applications
including: the stu
dies of wing tip vortices [6], contrail formation and dispersion [7] and
particulate emissions and dispersion [8]. Lidar has also been used for the study of
exhaust dispersion from rocket engines [9]. Very little has been reported, however, on
the use of L
idar for determining exhaust plume characteristics for dispersion modelling of
air quality in the vicinity of airports [10]. In this study we have scrutinized the dispersion
of ground based aircraft emissions further, but for the sake of brevity we must co
our discussion here to just a single mode of operation: the ground run towards take
A detailed discussion of other modes including rotation, climb
out, final approach,
landing and taxiing will be published elsewhere in due course.


A backscatter Lidar has been deployed at Heathrow and Manchester airports to map the
dispersion of aerosols from aircraft exhausts in the wakes of several hundred flights. The
system is mounted on a commercial vehicle with onboard power generation and so

fully autonomous and mobile. The system was originally developed by the Central
Electricity Generating Board for studies of dispersion from power stations [11], but was
transferred to Manchester following the privatization of the electricity supply ind

The Lidar now uses a frequency
tripled eye
safe Nd:YAG laser at 355 nm. It has a
pulse duration of 8 ns, a repetition rate of 30 Hz and an output power of 33 mJ/pulse.
The laser beam is expanded to a diameter of approximately 30 x 37 mm (elliptical
) that
diverges at 1 mrad as it exits the optics. This output beam is finally targeted using a
plane mirror positioned by a stepper motor under computer control: this allows scanning
either in elevation (up to a maximum of 60
), or in azimuth. The pointing

precision for the
mirror is 0.2

in elevation and 0.1

in azimuth. Calibration of the a
zimuth is established
from the line of sight to a convenient fixed target, such as a distant tower or chimney.
The orientation of this line is then established from an
Ordnance Survey map and
verified with a sighting compass.

The light backscattered from the atmospheric aerosol is collected using a telescope
with a 10” parabolic mirror that is collinear with the output beam. The collected light is
split into two com
ponents, which each pass through a 20 nm interference filter before
being incident on a photomultiplier tube (PMT). The use of two detectors allows the
signal to be range gated into near field and far field components. This increases the
measurable dynamic

range and hence the useful operating range of the system. In
addition it also allows spurious features and noise present in one PMT signal to be
identified through comparison with the signal from the second PMT. The signal from the
PMTs is pre
amplified b
efore being presented to a 60 MHz digitizer with 10

resolution. A digitizer with 125 MHz sampling rate is also available to the system,
implying a 1.2 m range resolution on a single channel. A schematic diagram of the
optical configuration and data cap
ture system is shown in figure 1.

The process of firing the laser, scanning the input/output mirror, digitizing the signals
and capturing the data is fully automated under computer control when the system is
operational, while all parameters are rapid
ly reconfigurable from software menus when
the system is off


Schematic diagram of the optical and data capture systems

For aircraft exhaust measurements, a range resolution of 5 m was chos
en: this gave
adequate detection of the rather tenuous plumes. The maximum range (determined by
the geometric loss of signal with distance) was no more than 1200 m. The minimum
range was 200 m, determined by the overlap of the outgoing and return optics.


scan (or sweep) is assembled from a number of individual Lidar
shots. Typically the
spacing in the
elevation between
successive shots in a vertical scan
would be 0.5°. Given the pulse repetition rate of 30 Hz, and useful elevations up to

30°, it
follows that the data to construct a Lidar scan are collected in approximately 2 s. In
certain circumstances this finite time can raise some practical problems in the
interpretation of the measurements. For example, an aircraft on final approach t
o land or
in the final stages of take
off will be travelling at 70 m s

and could thus move by 140 m
over the course of the scan. In this case the observed exhaust plume structures may
change across a scan and some care is required in the analysis. The di
spersion of the
aircraft exhaust plume may be reconstructed from

a time series of 2
sections. The time series is built up from a continuous repetition of the same scan
geometry while the plume advects through the Lidar beam. The system must be
however, between each of the scans and the mirror returned to its original position. As a
result the cycle time between scans is approximately 5 s.

The backscatter intensity and the attenuation coefficient in the Lidar equation
independent func
tions of

the size spectrum and refractive index of the aerosol target
[12]. Neither of these, however, is generally known.
The signal processing of our
measurements integrates a number of subsystems in an attempt to overcome this
problem as well as to corr
ect for the ambient background light level. Firstly, 30 samples
are recorded before the laser is triggered. These provide a level for the ambient
background, which may be subtracted from the return signal. The signal is then scaled
for the monitored energy

of each shot, multiplied by (range)

to allow for the geometric
spreading of the return signal, and a best
fit exponential decay estimated over the entire
shot to correct for scattering from ambient aerosol. Dividing the received signal by this

then gives us the signal from the aircraft emissions relative to the
backscatter from clear air. Typically, this would have a value of a few 10’s of percent for
engine emissions, though in the case of tyre smoke the backscatter might be many
times that fr
om the clear air. Programs have been developed over the years to derive
plume parameters (height, spread etc.) from these measurements [11].


The University of Manchester’s RASCAL Lidar facility was deployed at Heathrow airport
in May 2005, and
at Manchester airport in September 2005 and April 2006. Data that
sample the dispersion of aerosols from aircraft exhausts in the wakes of several hundred
flights have been collected. These data cover a variety of aircraft types and classes.
Most observati
ons were made with the beam swept in
the vertical and oriented at


to the runway. A limited number of observations with the
beam being swept in azimuth at a low angle of elevation were also made.

Data were acquired in early, inte
rmediate and advanced stages of the take
off ground
run; from airborne aircraft in both departure and arrival; and on touchdown and over the
landing ground run. There was substantial data coverage in conditions ranging from near
calm to 10



wind speed
; neutral to moderately unstable; relative humidity from 40

100%; and air temperatures of 4°

19° C. An onboard meteorological station with an
extendable mast mounted on the RASCAL vehicle logs wind speed, wind direction,
temperature, humidity, and shor
twave insolation at 10 s intervals. These data are used
as an aid in the interpretation of the aircraft exhaust dispersion data.

Results and discussion

The size range of aerosol emitted by aircraft engines is well matched to the scattering of
UV radiation

at 355 nm wavelength. The quantity of aerosol emitted, however, is fairly
small: aviation spirit has a sulphur content of typically only 0.1 % [1] and modern engines
emit minimal soot. The backscatter from the plume behind an aircraft in take
off roll

wed an enhancement of typically only 50% relative to that from clear air. A time
series of Lidar scans from an aircraft at the start of take
off roll and with a typical
emission profile is shown in figure 2. The aircraft is a Boeing 767

375er, with two win
mounted turbofan engines 20 m apart. The aircraft is initially
positioned on the runway
centre line, 550 m from the Lidar, being at rest awaiting clearance. As the engines power
up for take
off a region of enhanced backscattering is seen to develop in th
e images.
The aerosol in this young exhaust is relatively concentrated and induces a marked
excursion from ambient backscatter levels.
the aircraft will have move
d off
shortly after the first image, the Lidar scanning direction is maintained through
out the
image sequence. Hence this time series illustrates the dynamics of the exhaust plume
during take
off roll. However, the advection of the exhaust by the wind relative to the
sweeping plane of the Lidar must also be considered in the analysis.

ure 2. A time series of
Lidar scans from an aircraft at the start of roll out and
with a typical emissions profile. Time increases by 5 s between successive scans

Lidar beam intersects the runway
at an angle of 17
. Surface wind 2.3 m

at 045

Note that the j
et streams issuing from individual engines are not resolved. It is presumed
that a region of integrated flows quickly develops in the highly turbulent wake to merge
into a single downstream flow. An initial analysis of this dat
a suggests that the
downstream flow may best be modelled as a wall jet [13].

In general it was possible to observe the dispersion of exhaust emissions from take
off roll over time scales of order 1

s after their release. The maximum time represents
e point at which the backscattered light and therefore concentration of aerosol is
indistinguishable from ambient levels
It was also observed that the lifetime and
concentration of aerosol in the exhaust plume is sensitive to the air temperature and
ive humidity.

implies that much of the detectable aerosol is sulphate,
despite the low sulphur content of the fuel.

It should be noted, however, that chemi
and nitric acid have also been observed in aircraft engine emissions [1].

ly the exhaust plume for an aircraft in take
off roll may extend over some 100’s
of m horizontally and
from the lowest elevation sampled up to one 100 m vertically. Late
in the sequence shown in figure 2 there is the development of an off
ground peak in th
concentration of scatterers that is consistent with the action of buoyancy. Nevertheless,
buoyancy was not observed as a dominant effect in younger plumes although some
evidence of enhanced buoyant action was observed in light wind conditions.


Lidar has long been proven as a fast and efficient remote sensing technique. It has now
been shown that
Lidar is well suited to study the dispersion of aircraft emissions. Data
that sample the dispersion of aerosols from aircraft exhausts in the wakes of
hundred flights have been collected. Aerosols associated with exhausts from engines
were observed for up to
90 s after emission. Furthermore, t
he lifetime and concentration
aerosol in the exhaust plume
correlate with

the air temperatu
re and relative
An initial analysis of the dispersion data suggests that during rollout the
downstream flow may best be modelled as a wall jet.

his study has given significant insight into the initial plume characteristics from
aircraft exh
austs. The results will be used to increase the accuracy of current exhaust
dispersion models.


We would like to thank EPSRC and the Department for Transport for funding and supporting
these studies. We are also grateful to Manchester Airpo
rt plc, Heathrow Airport, BAA, and NATS
for their help and assistance. Special mention for Chris Paling, Nita Easy, Howard Coney, Kevin
Morris and
Tsvetina Evgenieva

whose contributions have been significant



IPCC ‘Aviation and the global enviro
nment’ Cambridge university press, ch 2, 1999.


‘The Jet Engine’ Rolls Royce, 2005.




SR Ahmad “Application of lidar to atmosphere pollution mapping: a review”, Proc. SPIE,
Vol 3127, pp1
13, (1997).


F Holzäpfel,

T Gerz, F Köpp, E Stumpf, M Harris, R Young, A Dolfi
for Circulation Evaluation of Aircraft W
ake Vortices Measured by Lidar’

J. Atmos & Ocean
Tech. 20, 8, pp1183
1195, (2003).


R Sussmann, K Gierens, ‘Differences in early contrail evolution of 2
engined versus 4
engined aircraft. Lidar measurements and numerical simulations’ J. Geophys. Res, 106, p
4911, (2001).


R DeCoursey, L Poole, C Hostetler, G Kent, G Hansen ‘Aircraft exhaust particle measure
ment with multiple ground
based lidar systems’ Proc SPIE. Vol 3065, pp9
20, (1997).


D Dao, A Dentamaro, ‘Lidar measurement of rocket exhaust plume d
ispersion and
in the stratosphere’ US Airforce research lab.
Rep. No: AFRL


R Wayson, G Fleming, B Kim, W Eberhard, W Brewer, Final report: the use of LIDAR to
characterize aircraft initial plume characteristics. US Dept Tr
ans, FAA.



M Bennett, S Sutton, D Gardiner ‘An analysis of Lidar measurements of buoyant plume rise
and dispersion at five power stations’ Atmos. Environ., 26A, pp3249
3263, (1992).


M Bennett ‘The use of a rapid
scanning backscatter LIDAR to validat
e dispersion models’
Proc SPIE, Vol 4397, pp431
439, (2001).


A Graham, D Raper, S Christie, M Bennett ‘Air quality at Heathrow airport: impact of
emissions from aircraft in ground run and flight, First Tech Report, DfT (2005).