Sicard, M. [et al.]. Determination of the mixing layer height from regular lidar measurements in the Barcelona area. A: International Symposium on Remote Sensing. "Remote
sensing of clouds and the atmosphere VIII: SPIE International Symposium - Remote Sensing Europe: Barcelona, Spain: september 9-12, 2003: proceedings". Washington:
International Society for Photo-Optical Instrumentation Engineers (SPIE), 2003, p. 505-516. ISBN 8-8194-5118-8. DOI 10.1117/12.511481.
© 2004 Society of Photo Optical Instrumentation Engineers. One print or electronic copy may be made for personal use only. Systematic electronic or print reproduction and
distribution, duplication of any material in this paper for a fee or for commercial purposes, or modification of the content of the paper are prohibited.
http://dx.doi.org/10.1117/12.511481
Determination of the mixing layer height from regular lidar
measurements in the Barcelona Area
Michaël Sicard*a, Carlos Pérezb, Adolfo Comeróna, José María Baldasanob, Francesc Rocadenboscha
a
Universitat Politècnica de Catalunya, Dept. de TSC, Jordi Girona 1, 08034 Barcelona, Spain
b
Universitat Politècnica de Catalunya, LMA, Avda. Diagonal 647, planta 10, 08028 Barcelona,
Spain
ABSTRACT
Regular aerosol backscatter measurements using an elastic-backscatter lidar were performed between May 2000 and
November 2002 in Barcelona (Spain), in the frame of EARLINET (European Aerosol Research Lidar Network). The
mixing layer height, required to understand the chemical and physical processes taking place in the low troposphere, was
one of the major parameters to be retrieved. Three analytic definitions of the ML height have been tested using the
range squared-corrected lidar signal: (1) the minimum of its first derivative, (2) the minimum of its second derivative,
and (3) the minimum of the first derivative of its logarithm. The strong coastal and orographic influences and the
climatological settling of Barcelona determine the complexity of its atmospheric boundary layer dynamics and the high
heterogeneity of the lidar signals. Therefore, single lidar analyses do not allow an unambiguous determination of the
mixing layer height in many cases and complementary data are needed, such as synoptic maps, backtrajectories,
radiosoundings and solar irradiance profiles. The resulting mixing layer heights were compared to radiosoundings, and
the second method was found to give statistically the best results. This definition was used to process the whole dataset.
A number of 162 days and 660 profiles were examined. The mixing layer height was inferred in cases such as low
clouds, Saharan dust events and sea breeze and mountain induced recirculation. Variations between 300 and 1450 m
were observed over the three years.
Keywords : mixing layer height, lidar profiles, radiosoundings, complex orography
1. INTRODUCTION
The atmospheric boundary layer (ABL) can be defined as the part of the troposphere that is directly influenced by the
presence of the earth’s surface. The mixing height (MH) is the height of the mixing layer (ML) which is the layer
adjacent to the ground over which pollutants or any constituents emitted within this layer or entrained into it become
vertically dispersed by convection or mechanical turbulence within a time scale of about an hour1. The MH is a key
parameter to characterize the structure of the ML. Different ways exist to determine or estimate the MH.
Radiosoundings are the most common data source to retrieve the MH based on wind, temperature and humidity profiles.
However, each balloon is lost in the atmosphere and because of its high cost, measurements are only taken twice daily at
specified synoptic times (00 UTC, 12 UTC).
Active remote sensing systems such as lidars use aerosols as tracers of the ABL dynamics. The optical power measured
by a lidar device is proportional to the aerosol content of the atmosphere. For example, the greater aerosol and moisture
content in the ML than in the free troposphere (FT) causes more laser light scattering and the boundary between the two
layers can be easily detected2. Thus, lidars have been increasingly used to estimate the MH2-5. However, interpreting
data from lidars is often not straightforward, because the detected layers are not always the result of ongoing vertical
mixing, but may originate from advective transport or past accumulation processes6. While over land surfaces in highpressure regions the ML has a well-defined structure that evolves with the diurnal heating cycle, in coastal regions with
complex orography as Barcelona, mesoscale phenomena as sea-land breezes and mountain-induced winds modify the
ML flow generating circulations in conjunction with diurnal heating cycles. The more usual synoptic situations
affecting the Barcelona area are westerly and northwesterly flows, and typical summertime weak pressure gradient
conditions7-8. During these summertime episodes, layering and accumulation of pollutants occur all over the region9-10.
*
msicard@tsc.upc.es; phone 34 934 017 758; fax: +34 934 017 200; http://www.tsc.upc.es
Remote Sensing of Clouds and the Atmosphere VIII, edited by Klaus P. Schäfer,
Adolfo Comerón, Michel R. Carleer, Richard H. Picard, Proc. of SPIE Vol. 5235
(SPIE, Bellingham, WA, 2004) · 0277-786X/04/$15 · doi: 10.1117/12.511481
505
In addition to these regional recirculation, synoptic scale meteorology induces frequent outbreaks of Saharan dust in
summer11.
In the frame of EARLINET12 (European Aerosol Research Lidar Network), 21 lidar stations performed regular lidar
measurements from May 2000 to November 2002 to provide a climatological database of the vertical and horizontal
distribution of aerosols over Europe. Among other subjects, the network studies the aerosol properties in the lower
troposphere and more particularly in the ABL at different time scales (diurnal and seasonal cycle). In this context, the
MH was one of the major parameters to be retrieved. Several retrieval methods have been tested in order to estimate the
MH from the whole climatological database.
2. THE BARCELONA AREA: OROGRAPHY AND METEOROLOGICAL SETTLING
Barcelona is located on the shores of the Mediterranean Sea, on the northeastern corner of the Iberian Peninsula (IP).
The meteorology and the origin of the air masses arriving at the IP are highly influenced by the Azores high pressure
system which is located over the Atlantic Ocean and that intensifies during the warm season inducing very weak
pressure gradient conditions all over the region. The major orographic features that influence the flows arriving at the
Barcelona air basin are the Pyrenees and the Ebro valley. The Pyrenees range from 2000 m to 3000 m, acting as a
natural barrier of the flows and producing important orographic forcings into the low troposphere. The Ebro valley has a
length of 350 km, channelling the flows of the Cantabric sea to the Mediterranean or vice versa. The nearby orography
of the region is dominated by four main features arranged parallel to the coastline: the coastal plain, the coastal mountain
range (500-700 m) the pre-coastal depression, and the pre-coastal mountain range (700-1700 m).
3. THE BARCELONA LIDAR SYSTEM AND EARLINET
A transportable, steereable lidar system allowing three-dimensional scans has been developed at the Universitat
Politècnica de Catalunya (UPC)13. A Raman channel is currently being added to the system14. The present system is
based on a Nd:YAG laser working at the 1064-nm fundamental wavelength and at the 532-nm second harmonic,
delivering pulses of equal energy (160-mJ) and 6-ns duration with a 20 Hz PRF. The photoreceiver is based on an
avalanche photodiode (APD) with a wide spectral response (its quantum efficiency is about the same at both
wavelengths). One changes the wavelength by placing manually the corresponding interference filter in front of the
APD. The emission and reception axes are different so that a blind zone is observed between 0 and 250 m. However the
overlap factor reaches 1 rapidly.
The Barcelona lidar station forms part of the EARLINET. Regular lidar measurements were undertaken from May 2000
to November 2002 on preselected dates regardless of the weather conditions to provide an unbiased climatological
database of the horizontal and vertical distribution of aerosols over Europe12. For this reason a common schedule of
three measurements per week was agreed. Measurements were performed on Monday at 1400 LST (local solar time) +/1 hour and at sunset –2/+3 hour and on Thursday at sunset –2/+3 hour. Furthermore, the network performed diurnal
cycle measurements under unperturbed weather conditions, ideally under high-pressure systems to allow simultaneous
observations at different stations and quantify the behaviour of aerosols at the regional scale.
All lidar measurements were made at the UPC campus, south west of Barcelona (41.39 N, 2.12 E, 115 m above sea
level) at the wavelength of 532 or 1064 nm. Sequences of 1-minunte duration (1200 shots) were recorded. The 30minute integrated profiles used in this analysis are the average of 30 consecutive 1-minute profiles.
4. METHODS USED IN THE DETERMINATION OF THE MH
The optical power measured by a lidar is proportional to the signal backscattered by particles and molecules present in
the atmosphere. The lidar signal can be expressed as:
K
S(r ) = 2 ⋅ β a (r ) + β m (r ) ⋅ T (r ) 2 + S 0
(1)
r
where β a and β m are, respectively, particular and molecular backscatter coefficients, K is the system constant, T is the
atmospheric transmission, r is the range between the laser source and the target, and S0 is the background signal. The
range-squared-corrected signal (RSCS) is then defined as:
RSCS = (S − S0 )r 2
(2)
[
506
Proc. of SPIE Vol. 5235
]
The lidar signal shows a strong backscattering within the ML, which decreases through the entrainment zone and
becomes weak in the FT. These contrasts are the basis of the lidar estimations of the MH used in this paper. Since a
large climatological database of temporally integrated profiles needs to be analysed, the use of a method based on the
temporal variations of the profiles, as it is the case of the variance method15 is not possible. Only methods not requiring
temporal evolution of the ML are suitable for this kind of study. Three methods have been tested in order to estimate the
MH from the large database:
The first method is the gradient method (GM). It looks for the altitude, hGM, at which the minimum of the first derivative
of the RSCS, i.e. the minimum of the gradient profile, is reached:
∂RSCS
h GM = min
(3)
r
∂r
The second method is the inflection point method (IPM). It consists in finding the altitude, hIPM, at which the minimum
of the second derivative of the RSCS is reached. This minimum corresponds to the inflection point of the first
derivative. It is usually located below hGM:
∂ 2 RSCS
h IPM = min
(4)
2
r
∂r
The third method, so called logarithm gradient method (LGM), consists in finding the altitude, hLGM, at which the
minimum of the first derivative of the logarithm of the RSCS is reached:
∂ ln(RSCS)
h LGM = min
(5)
r
∂r
Figure 1 shows a typical averaged profile obtained on 16 October 2000 from 1202 to 1232 UTC. The MH retrieved by
the three methods are reported on the figure.
Fig. 1: 30-minute integrated RSCS profile on 16 October 2000 from 1202 to 1232 UTC.
However, things might be more complicated because for each method the minimum cannot be well defined or several
minima exist over an extended altitude range. Therefore, if possible one should correlate the altitude of the minima with
the temporal evolution of the aerosol layers. This is detailed in Section 6.1. Often the ML can be better identified if a
time series of profiles is available. Additional radio soundings or water vapour profiles can also help identifying the
Proc. of SPIE Vol. 5235
507
'real' ABL. In the morning, one often has to distinguish between the newly developing ML and the ABL from the day
before which is still present and in this case called residual layer (RL). This effect of two layers is described in Stull16.
If several layers exist that are clearly separated, only the lowest layer is labelled ML. In the morning, when both ML and
the RL on top of it may exist, these layers are typically well connected, but 2 local minima are observed.
The lidar dataset covers the years 2000, 2001 and 2002. The radiosoundings were launched at 0000 and 1200 UTC with
an incertitude known to be of –20 / +10 minutes. At 1200 UTC, the ABL can change quite rapidly and the 30-minute
integrated profiles eventually show large differences against 5-minute and 15-minute integrated profiles. Since the
incertitude on the hour of the radiosounding launches is not negligible, an analysis of the 5-minute integrated profiles
could lead to large errors in the comparison with the radiosoundings due to the eventual large time differences.
Therefore, 15-minute integrated lidar profiles with starting times between –20 / +10 minutes around 1200 UTC were
used for comparison with radiosoundings when possible.
5. COMPARISON OF THE MH FROM LIDAR MEASUREMENTS AND SOUNDINGS
Everyday the Meteorological Service of Catalonia performs 2 radiosoundings at 0000 and 1200 UTC. Measurements of
temperature, pressure, relative humidity, wind speed and direction are available. The potential temperature, θ, at the
altitude z is defined as:
P
θ(z) = T (z) s
P(z)
κ
(6)
where T is the temperature in Kelvin, Ps the standard surface pressure (Ps = 1.01325 bar), P the real pressure and κ = Rg /
Cp = 0.286 for a diatomic ideal gas.
5.1. The bulk Richardson number method
The bulk Richardson number method17 can be used both in convective conditions and in mechanical turbulence. The
Richardson number, Rib, is calculated as:
g (z − z 0 ) [θ(z) − θ(z 0 )]
R ib (z) =
(7)
θ(z) u (z) 2 + v(z) 2
where g is the acceleration due to gravity, z0 the height of the surface, and u and v the zonal and meridian wind
components, respectively. The ABL top is defined as the height where the Richardson number becomes equal or larger
than the so-called critical Richardson number, i.e. Rib > Ribc, where Ribc is the critical bulk Richardson number. A value
of 0.21 is taken for Ribc. Beyond this critical value of Rib the atmosphere can be considered fully decoupled from the
ABL.
5.2 The simple parcel method
The simple parcel method18 has turned out to be most reliable in convective situations. The dry adiabate is followed
from the measured surface temperature (or an expected maximum temperature) to its intersection with the temperature
profile from the radiosounding. The MH is taken at the equilibrium level of an air parcel with the temperature. This
method was refined by several authors19-21 by adding an excess temperature (advanced parcel method). Since under
convective conditions a superadiabatic layer is usually found near the ground, the simple parcel method also implicitly
applies an excess temperature. In this section, and for the purpose of comparison with the bulk Richardson number
method, the simple parcel method is applied to the radiosoundings data.
5.3. Comparison of both methods
Figure 2 shows how both methods perform on the case of 16 October 2000. The 57-m difference is not significant on
this single case.
During the three years of measurements, the number of temporal coincidences between radiosoundings and lidar
measurements, as described in section 4, is 29. Figure 3 represents the MH retrieved by the Richardson method and the
simple parcel method for these 29 days. The agreement between both methods is good: the correlation coefficient is
0.997 and the standard deviation is 28 m.
Even though the comparisons were performed mainly at noon, the wide range of MH detected (150 – 1750 m) assesses
the reliability of such comparison. The two points around 150 m, even physically doubtful for a MH at 1200 UTC over
508
Proc. of SPIE Vol. 5235
Barcelona, were kept because both profiles of Rib and θ were similar in shape to the ones shown in Figure 2. The fact
that the simple parcel method gives slightly smaller heights probably comes from the missing excess temperature term in
cases dominated more by mechanical turbulences than by convective situations. Thus, the lidar profiles were then
compared to the MH retrieved by the Richardson number method.
Fig. 2: Richardson number and potential temperature profiles
from radiosounding measurements made on 16 October 2000
at 1200 UTC. The Richardson number method retrieves a
MH of 1084 m; the simple parcel method retrieves a MH of
1021 m.
Fig.3: Comparison of the MH retrieved from radiosoundings
by the Richardson number method and the simple parcel
method.
5.4. Comparison of the MH retrievals from radiosoundings and lidar measurements
From the 29 coincidences between lidar profiles and radiosounding launches, finally 20 lidar profiles were processed by
the three analysis methods. The reasons for discarding the other 9 cases are the following:
- Clouds occulting the ABL top,
- Aerosol layer on top of the ML coupled with the ML,
- The MH is underneath the useful lidar signal (before the overlap reaches 1),
- One of the methods exhibits too many negative peaks without crossing the y = 0 axis.
The results of the comparison between the lidar MH and the radiosounding MH are shown in Figure 4. For each of the
lidar analysis methods, good agreement with correlation coefficients of 0.963, 0.969 and 0.979, respectively, was found
for the GM, the IPM and the LGM. The smallest standard deviation is found for the IPM, with a value of 82 m, whereas
it is equal to 89 and 101 m, respectively, for the GM and the LGM. Both the GM and LGM seem to retrieve higher MH
than the IPM, which is normal since the inflexion point of the first derivative is usually found just before its minimum
(for the LGM, this is also true: only the distances change, not the relative position of the points between each other).
6. RESULTS AND DISCUSSION
6.1. Data process
To fully test the IPM and to identify its advantages and limitations, the analysis was extended to all the lidar data
available in Barcelona over the years 2000, 2001 and 2002. There are a total of 162 measurement days with 660 30minute integrated profiles. In many cases, and for the reasons cited in the introduction, a multi-layering is observed.
Therefore, an analysis based only on the lidar profiles is not sufficient to estimate the MH without ambiguity:
complementary data, such as solar irradiance profiles, backtrajectories, synoptic maps and radiosoundings, were used.
The diurnal cycle on 30 May 2002 (Figure 5) is a clear example of the difficulties that often appear when the previous
Proc. of SPIE Vol. 5235
509
Fig. 4: Comparison of hGM, hIPM and hLGM with the MH retrieved from radiosoundings by the Richardson number method. In the
middle of the figure is reported the correlation coefficient and the standard deviation between the measurements.
methods are applied to the lidar profiles. Figure 5 shows the RSCS profiles, together with the profiles from the 3
methods at 0827, 1139, 1222, 1426, and 1543 UTC. Each profile shows a multi-layer structure. Figure 5a features the
development of the ML at 0827 UTC. An elevated layer is present at about 1000 m, which is fully decoupled from the
ML. At this time, if one first looks at the first derivative, there is no ambiguity in the retrieval of the MH: the highest
negative peak of the second derivative just below hGM determines the MH. Note that in this case, hIPM is not the absolute
negative peak of the second derivative (around 540 m) but the highest negative peak just below hGM (at 651 m). Figure
5b features the RSCS profile and the results from the 3 methods at 1139 UTC. The inflexion point corresponding to the
highest negative peak of the first derivative gives the MH, which is located at 630 m. The top of the first elevated layer
is located at 890 m. At 1222 UTC (Figure 5c), the IPM only would lead to a MH of 580 m. However, the GM indicates
without any ambiguity that hGM is 740 m, hence the highest negative peak from the IPM is just below at 704 m. Note
that here the MH is given not by the first highest negative peak from the second derivative but by the second one. At
1426 UTC (Figure 5d), the upper layer settles by subsidence and progressively gets connected to the ML. Like in the
0827 UTC case, the highest negative peak of the second derivative just below hGM gives a MH of 566 m. At 1543 UTC
(Figure 5e), the upper layer is almost completely coupled to the ML and even showing higher RSCS values, but a first
small decrease in the RSCS still allows the detection of the MH (524 m). This situation points out the importance of
following the temporal evolution of the ML along the day. However, in the climatological context, if we had considered
a unique 30-minute measurement as in Figure 5e, one could state that the MH is located at 832 m arguing that the lowest
minimum could be due to inhomogeneities within the ML. Thus, the diurnal cycle (Figure 6) allows following the
connection between the ML and the upper layer, and locating the proper minimum.
In all these examples, to be sure to present physically meaningful results, a cross-comparison of the temporal evolution
of the MH was made. This means that after the first evaluation of the temporal variations of the MH, the plot of Figure 6
is made to check the consistency between successive heights. Figure 6 shows the temporal variation of the lidar signal
on 30 May 2002. For example, the radiosounding indicates a MH of 725 m at 1200 UTC. The MH at 1222 UTC cannot
be the height of the first minimum of the second derivative at about 600 m, nor the second layer at about 1000 m. In that
case, the radiosounding considerably helps, but a cross-comparison of the temporal evolution of the MH (Figure 6)
would also indicate that these 2 alternatives (600 and 1000 m) are not possible regarding the previous MH.
In the particular case of 30 May 2002, the sequences from 1222 to 1252 UTC in Figure 6 seem to be split in 2: a first
510
Proc. of SPIE Vol. 5235
group of profiles with a clear MH at around 725 m, and a second group with a ABL height around 600 m. In that case,
the IPM is the only method able to detect both peaks (at 600 and at 725 m). The 2 gradient methods detect both peaks
but in between do not cross the x = 0 axis; therefore only one peak is the solution (the strongest one around 725 m).
Fig. 5: 30-minute integrated RSCS profiles on 30 May 2002 at (a) 0827, (b) 1139, (c) 1222, (d) 1426, and (e) 1543 UTC.
Looking at the solar irradiance profile (Figure 7a), a small decrease in the global solar irradiance is observed from 1200
to 1300 UTC. This decrease might explain the sudden fall of the MH around 1237 UTC (at half time of the
measurements).
Proc. of SPIE Vol. 5235
511
Fig. 6: 1-minute resolution diurnal cycle of the RSCS on 30 May 2002. The black dots indicate the MH retrieved by the IPM. The
white diamond indicates the MH measured by radiosounding at 1200 UTC.
(a)
(b)
(c)
Fig. 7: (a) Solar radiation over Barcelona versus time; (b) MSL pressure and 500 hPa geopotential height at 1800 UTC; (c)
Backtrajectories at 1300 UTC.
512
Proc. of SPIE Vol. 5235
This case highlights the drawback of the integration time on the MH retrieval accuracy. Within 30 minutes, the MH can
change significantly. However, the IPM will detect in the 30-minute integrated profile only one “averaged” height.
When applying the methods to single 30-minute integrated profiles without the diurnal evolution information, the
knowledge of the synoptic situation is essential to better estimate the MH. In Barcelona, the absence of large scale
forcing due to very weak pressure gradients and the development of mesoscale flows related to the daily heating and
cooling cycle (sea breezes, mountain-induced winds, valley winds and the Iberian thermal low) are summertime
recurrent weather patterns. During these episodes layering and accumulation of pollutants occur all over the region.
The temporal variation of the solar irradiance during that day (Figure 7a) allows discarding the presence of clouds on top
of the ML. Clouds can mask the top of the ML or can bias the measured height. The surface synoptic chart for the 30
May 2002 at 18 UTC features a very weak pressure gradient over the Iberian Peninsula and the Western Mediterranean
Basin, and the development of the Iberian thermal low (Figure 7b). The stagnant conditions appear clearly depicted in
Figure 7c where 4-day backtrajectories arriving at 700, 850 and 975 hPa show very short paths. These conditions
indicate the possible presence of upper aerosol layers coupled to the ML.
6.2. Mixing layer height in Barcelona between 2000 and 2002
Among the 660 measurements over 162 days, a comparison of the MH could be made only for a certain portion of the
diurnal cycle. We have chosen the period of maximum insolation running from 1000 to 1500 UTC, which corresponds
to the unstable thermal stratification. Figure 8 shows the MH in Barcelona retrieved by the IPM over the period 20002002. We have distinguished 2 seasons: summer (from April to September) and winter (from October to March). The
MH oscillates between 300 and 1450 m in summer and between 390 and 1420 m in winter. The standard deviation for
this portion of the day is 180 and 256 m, respectively in summer and winter. One could expect higher values in summer
due to enhanced convection but it is not the case. In Figure 8, no significant differences are observed between summer
and winter. Furthermore, the average MH in summer is even smaller than the average MH in winter. In coastal regions
in summer, a thermal internal boundary layer (TIBL) forms when cool marine air is advected onshore, and the lowest
portion becomes heated from below by the warm land surface. The TIBL generally does not extend all the way to the top
of the marine air associated with the intruding air mass, so the remainder of the cool air mass above the TIBL and below
the air flowing from the continent back to the sea, acts as a barrier for the TIBL vertical development.
1600
1400
S00
W00 S01
W01 S02 W02
Altitude (m)
1200
1000
800
600
400
200
0
15-3-00
1-10-00
19-4-01
5-11-01
24-5-02
10-12-02
Dates
Fig. 8: Mixing layer height as a function of day of the year between 2000 and 2002. Diamonds indicate summer measurements (S)
and crosses indicate winter measurements (W).
Proc. of SPIE Vol. 5235
513
6.3. Limitations of the method
We now turn to individual lidar profiles to further illustrate why and when the methods fail.
Figure 9a shows the 30-minute integrated profiles on 2 April 2001 at 1133 UTC. At 1200 UTC, the radiosounding
measurement indicates a MH of 589 m. It is obvious from Figure 9a that it is useless to look for the sharp decrease in
the backscatter signal that usually characterizes the MH4. The GM and the LGM are unable to determine the MH
because the ML is fully connected to the upper layers: the aerosol gradient between the ML and the upper layer is
inexistent. The IPM detects a peak at 674 m for an inflexion point of the first derivative that is positive, i.e. the signal
keeps increasing. In this particular case, the gradient of the increase at that altitude is much smaller, which could
determine the transition between both layers. But this is only an assumption and therefore, even the IPM is not suitable
for detecting the MH.
Another counter example for the proposed method is the case of the 14 May 2001 (Figure 9b and 9c). Figure 9b and 9c
show the 30-minute integrated profiles at 0856 and 0933 UTC on 14 May 2001, respectively. That day, the 30-minute
integrated results were biased by the return from some “cloudy” profiles. Figure 9b shows two shallow bumps around
850 and 1000 m, caused by the integration of a few “cloudy” profiles. The 3 methods work well and detect the MH at
about 600 m, which is consistent looking at the previous retrieved heights. However, a cloud considerably biased the
next half hour of measurement and the three methods fail: the MH is identified as the cloud height.
Fig. 9: Average RSCS profile over 30 minutes recorded (a) on 2 April 2001 at 1133 UTC, (b) on 14 May 2001 at 0856 UTC, and (c)
on 14 May 2001 at 0933 UTC.
As a summary, limitations of the IPM are found in the presence of elevated humid aerosol-laden layers whenever the
inversion capping of the mixed layer is weak. In this case, small aerosol gradients between the ML and the FT are much
harder to detect than those from the elevated layers, which exhibit large aerosol and humidity gradients with respect to
514
Proc. of SPIE Vol. 5235
their surrounding. This phenomenon is illustrated in Fig. 9 for the late morning where we see that the MH is less
marked than the top of the elevated moist layer. The IPM algorithm is designed to detect the second derivative peak,
which has the greatest magnitude. In this particular case, the 2 gradient methods also fail.
Another point is that the retrieval method of the MH cannot be automated: the cross-comparison of the temporal
evolution of the MH is a key aspect of the MH retrieval. In the global analysis described in Section 6.2., some cases
were discarded because the MH retrieval at a specific time did not match with the MH retrieved from the previous and
the next measurements.
7. CONCLUSION
The large orographic and coastal influences in Barcelona do not allow the retrieval of the MH by a classical 1-year
harmonic model (in sinus or similar) and well-established MH retrieval methodologies. Three methods applicable to
single lidar profiles were used to determine the ABL height. The methods are based on the determination of the
strongest decrease of the backscattered lidar signal. The inflection point method (the minimum of the second derivative)
gave the best results when comparing with radiosounding measurements. The two other methods give MHs a little
higher, corresponding to heights in the entrainment zone, i.e. between the ML top and the beginning of the FT.
Among the 660 measurements made over 162 days, the MH during the period of maximum insolation running from
1000 to 1500 UTC oscillates between 300 and 1450 m in summer and between 390 and 1420 m in winter.
Complementary data were used in almost all the cases: synoptic maps, backtrajectories, radiosoundings and solar
irradiance profiles. At this period of the day, the ABL changes quite rapidly and the analysis of the 30-minute integrated
profiles eventually show large differences against 5-minute and 15-minute integrated profiles. This highlights the
drawback of the integration time on the MH retrieval accuracy: the IPM will only detect an “averaged” height.
Limitations of the IPM are found in the presence of elevated humid aerosol-laden layers whenever the inversion capping
of the mixed layer is weak: small aerosol gradients between the ML and the FT are much harder to detect than those
from the elevated layers, which exhibit large aerosol and humidity gradients with respect to their surrounding. This is a
significant differential trend as compared to central Europe cities such as Paris or Hamburg.
The correlation coefficient of the MH cross-comparison between lidar and radiosounding measurements among all the
inter-compared methods may well be influenced by atmospheric decorrelating effects taking place over the 30-minute
lidar integration time, particularly between 1000 and 1500 UTC, when the boundary layer has a stronger evolution.
Thus, it is necessary that the methodological analysis includes a “discard” procedure to cope with time-punctual specific
uncorrelated scenes, such as those arising in cases where some clouds are occulting the ABL top, or an aerosol layer is
coupled to the ML on top of it. These situations are typical of the absence of large scale forcing due to very weak
pressure gradients in Barcelona. A methodological hint arising from Section 6.1. is to increase the correlation
coefficient by the temporal variance computed over the 30-minute time averaged profiles, since temporal decorrelation
of the MH parameter is a major error source in some of the analyzed profiles. In summary, the analysis of complex
patterns such as those occurring in Barcelona requires a progressive incorporation of new methodologies to the already
existing ones along the points suggested in the examples presented in this work.
ACKNOWLEDGEMENTS
The authors wish to acknowledge the following entities for partially supporting the research work and lidar systems
developed at UPC: European Union under the EARLINET contract UE EVR1-CT-1999-40003, CICYT (Spanish
Interministry Commission of Science and Technology) under the grants TIC 431/93, AMB96-1144-C02-C01, TIC991050-C03-01, REN2000-1907-CE and REN2000-1754-C02-02/CLI, Spanish Ministry for Education and Culture under
the Spanish-French Integrated Action HF1997-0212, and CIRIT (Interdepartmental Commission for Research and
Technological Innovation, Generalitat de Catalunya) under the contract IMMPACTE. ESA is also thanked for the
external postdoctoral fellowship allocated to M. Sicard.
REFERENCES
1.
P. Seibert, F. Beyrich, S.E. Gryning, S. Joffre, A. Rasmussen, and P. Tercier, Mixing layer depth determination for
dispersion modelling. In COST Action 710 - Final Report. Harmonisation of the pre-processing of meteorological
Proc. of SPIE Vol. 5235
515
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
516
data for atmospheric dispersion models, Luxembourg: Office for Official Publications of the European
Communities, 1992.
L. Menut, C. Flamant, J. Pelon, and P. H. Flamant, Urban boundary-layer height determination from lidar
measurements over the Paris area, Appl. Opt. 38, 945-954, 1999.
R. Boers, E. W: Eloranta, and R. L. Coulter, Lidar observations of mixed layer dynamics: tests of parametrized
entrainment models of mixed layer growth rate, J. Climate Appl. Meteorol. 23, 247-266, 1984.
E. Dupont, J. Pelon, and C. Flamant, Study of the moist convective boundary layer structure by backscatter lidar,
Boundary-layer Meteorol. 69, 1-25, 1994.
V. Matthias, and J. Bösenberg, Aerosol climatology for the planetar y boundary layer derived from regular lidar
measurements, Atmospheric Research 63, 221-245, 2002.
P. Seibert, F. Beyrich, S.E. Gryning, S. Joffre, A. Ramussen, and P. Tercier, Review and intercomparison of
operational methods for the determination of hte mixing height, Atmospheric Environment 34, 1001-1027, 2000.
J. Martín-Vide, Característiques climatològiques de la precipitació en la franja costera mediterrània de la
Península Ibèrica, PhD. Thesis, Barcelona, edited by Institut Cartogràfic de Catalunya, 111-129, 1987.
O. Jorba, C. Pérez, J. M. Baldasano, F. Rocadenbosch, and M. A. López, Clúster análisis of backtrajectories
arriving at Barcelona air basin, in 1st EARLINET Symposium on the structure and use of the database derived
from systematic lidar observations, Hamburg (Germany), 11-12 February 2003.
M. Millán, R. Salvador, and E. Mantilla, Photooxidant dynamics in the Mediterranean basin in summer: results
from European research projects, J. of Geophys. Research 102, 8811-8823, 1997.
C. Soriano, J.M. Baldasano, W.T. Buttler, and K. Moore, Circulatory patterns of air pollutants within the
Barcelona air basin in a summertime situation: lidar and numerical approaches, Boundary-Layer Meteorol. 98,
33-55, 2001.
S. Rodríguez, X. Querol, A. Alastuey, G. Kallos, and O. Kakaliagou, Saharan dust contribution to PM10 and TSP
levels in Southern and Eastern Spain, Atmospheric Environment 35, 2433-2447, 2001.
J. Bösenberg, A. Ansmann, J.M. Baldasano, D. Balis, C. Böckmann, B. Calpini, A. Chaikovsky, P. Flamant, A.
Hagard, V. Mitev, A. Papayannis, J. Pelon, D. Resendes, J. Schneider, N. Spinelli, T. Trickl, G. Vaughan, G.
Visconti, and M. Wiegner, EARLINET: a European aerosol research lidar network. In Advances in Laser Remote
Sensing, Ecole Polytechnique, Palaiseau, 155-158, 2001.
F. Rocadenbosch, C. Soriano, A. Comerón, J. M. Baldasano, A. Rodríguez, C. Muñoz, and D. García-Vizcaíno, 3D
scanning portable backscatter lidar platform for atmospheric remote sensing: performance and architecture
overview. In Remote Sensing of Clouds and the Atmosphere V, Proc. SPIE 4168, 158-169, 2000.
F. Rocadenbosch, M. Sicard, A. Comerón, J.M. Baldasano, A. Rodríguez, R. Agishev, C Muñoz, M.A. López, and
D. García-Vizcaino, The UPC scanning Raman lidar: an engineering overview. In Proceedings of 21st
International Laser Radar Conference, Defence R& D – Valcartier, Val-Bélair, 69-70, 2002.
W. P. Hooper, and E. W. Eloranta, Lidar measurements of wind in the planetary boundary layer: the method,
accuracy and results from joint measurements with radiosonde and Kytoon, J. Climate Appl. Meteorol. 25, 9901001, 1986.
R. B: Stull, An introduction to boundary layer meteorology, Kluwer Academic Publishers, Dordrecht, 1988.
D.H.P. Vogelezang, and A.A.M. Holtslag, Evaluation and model impacts of alternative boundary-layer height
formulations, Boundary-Layer Meteorol. 81, 245-269, 1996.
C. G. Holzworth, Mixing depths, wind speeds and air pollution potential for selected locations in the United States,
J. Appl. Meteorol. 6, 1039-1044, 1967.
A. J. Garret, Comparison of observed mixed layer depth to model estimates using observed temperature and winds,
and MOS forecasts, J. Appl. Meteorol. 20, 1277-1283, 1981.
R. B: Stull, Static stability – an update, Bull. Amer. Meteorol. Soc. 72, 1521-1529, 1991.
G. Wotawa, A. Stohl, H. Kromp-Kolb, Parametrization of the planetary boundary layer over Europe: a data
comparison between the observation-based OML pre-processor and ECMWF model data, Contrib Atmos. Phys.
69, 273-284, 1996.
Proc. of SPIE Vol. 5235