LIDAR Technique for Atmospheric Monitoring in The Pierre ...personalpages.to.infn.it/~sequeiro/pdf/pre-Tesis.pdf6 replacing mechanics and electronics parts in manner to do them operatives.

1

LIDAR Technique for Atmospheric Monitoring in The Pierre

Auger Observatory

Guillermo Manuel Sequeiros

Corso di Dottorato in Fisica – XVIII Ciclo

May , 2005

1. Introduction

For about more than 100 years the fascination of cosmic rays had excited the

physicists. Victor Hess in 1912 was one of the pioneers in observing highly energetic

particles from outer space. In his words: “a radiation of very high penetrating power

entering our atmosphere from above”[1].

The cosmic ray composition and its shape is well known up to energies of 1019

eV.

Nonetheless there are open question about this energy value, such as the right shape of

the spectrum, the existence of the GZK cutoff, their mass composition, the accelerate

sites and mechanisms able to accelerate particles to such high energies, the anisotropy or

isotropy their direction distribution.

With the Pierre Auger Observatory we want to study cosmic rays at the upper end of

the known energy spectrum, i.e. events with energies above 1018

eV, using a combination

of the two techniques used by the previous largest experiments: the Ground Array

technique and the Fluorescence technique.

Measurements of the cosmic-ray air-shower fluorescence at extreme energies requires

precise knowledge of atmospheric conditions. The absolute calibration of the cosmic-ray

energy requires a good estimate of the absorption coefficient [2].

In my PhD pre-thesis I introduce a brief description of The Pierre Auger Observatory,

then the atmospheric techniques used actually in this observatory, describing with details

the LIDAR technique, the hardware and software of the LIDAR systems that our group

have mounted , developed and put in operation the different sites. I described the LIDAR

data as well as the online monitor tools for the control of the data acquisition. Molecular

and aerosol contributions to the LIDAR equation are discussed and an algorithm for

cloud height detection is thoroughly described.

2

2. The Pierre Auger Observatory

The Auger Observatory consists of two parts: the first system is currently installed in

the southern hemisphere in Argentina and the second will be built in the next years in the

northern hemisphere, both sites covering a total surface of 6400 km2

will able to detect

6000 events/year above 1019

eV and 60 events/year above 1020

eV. Both sites are located

at around 35 – 40º latitude allowing altogether a full sky exposure.

The particles detectors are place on an area of about 3000 km2. The site was chosen

according to the necessity of a large almost flat area , 1400 m a.s.l. with sparse human

inhabitants, but with some infrastructure near the array, no significant sources of light

and cloud cover of less than 15 %. The system is described in the following.

2.1. Fluorescence Detector

These detector type used a detection technique which consist in detecting fluorescence

light emitted by de-excitation of nitrogen molecules and nitrogen ions excited from

charged particles of the showers.

The atmosphere on the array will be observed by four fluorescence detector telescopes

each of them covering a 180º x 30ºfield of view (see Figure 1). The duty cycle of these

detectors is ∼10 % considering that the system requires illuminated moon fraction below

50 % and clear sky.

Figure 1: cover zone of the Pierre Auger Observatory in Malargüe Mendoza, Argentina. The dots

represent the array of the 1600 water Cherenkov tanks, are signed with red dots the four fluorescence

telescope stations, Los Leones, Coihueco, Morados and Pampa Amarilla (Norte).

3

Each fluorescence detector is composed of six telescopes, each telescope being made

of segmented spherical mirrors with 3.4 m curvature radius, a corrector ring, a UV filter

and 440 photomultipliers (PMTs) camera placed in the focal plane. Each camera inside a

fluorescence detector is read out separately (see Figure 2).

Figure 2: Scheme of a fluorescence telescope system

The array of PMTs sees the trace of the Extensive Air Shower (EAS) as a light spot

crossing the atmosphere at the speed of light along a line. The pixels on the camera will

be hit sequentially by fluorescence photons coming from these light points.

2.2. Surface Detector

The surface is made of water Cherenkov tanks, the final array will we conform of 1600

tanks separated 1.5 km distance one each other. These detectors observe the Cherenkov

light emitted by secondary particles (muons and electrons) of an EAS when they cross the

water in the tank and their velocity is above the Cherenkov threshold in the water. The

tanks have a sensitive water volume (1.2 m height, 3.4 m diameter) and contain 3 PMTs

which detect the Cherenkov light (see Figure 3 and 4). The duty cycle of these detectors

is 100 %.

4

Figure 3: photograph of a Cherenkov tank Figure 4: Schematic drawing of a tank

installed on the site.

3. Atmospheric Monitoring for the Auger Observatory

As we mention before the detection technique used by the Pierre Auger Observatory

required continuous monitoring of the light attenuation between the fluorescence source

and the detector. A short description of the atmospheric monitoring systems that have

been installed in the observatory is given below.

• Horizontal Attenuation Length Monitor: composed of three systems which

observed tree different light trajectory through the site recording the horizontal

attenuation lengths within or close to the wavelengths values to which the

fluorescence detectors are sensitive. The principal components are a light source

and a CCD camera. These measurements are made at 4 different wave length

throw filters (365 nm, 404 nm, 436 nm, 542 nm).

• Cloud Cameras: clouds can interfered seriously in the fluorescence detection, for

this is necessary to have a sky cloudiness control during the data taking. For this

objective are used infrared cameras installed one in each site of the array which

scan the sky during all the night including the field of view of the Fluorescence

Detector. The spectral range is from 7 – 14 µm and the field of view 45º x 36º.

The cameras give a full sky scan every 15 minutes.

• Weather Stations: There are 3 weather stations one at each fluorescence site and

one at the central site. They monitor daily temperature, wind speed and direction,

pressure, solar radiance and relative humidity.

• Central Laser Facility: The so-called Central Laser Facility (CLF) is based on a

remotely controlled steerable laser (main wavelength: 355 nm, 6 mJ typical beam,

7 ns with). It has been setup at the center of the array, and it produces a calibrated

test beam for the FD. The CLF was conceived to meet a number of atmospheric

monitoring and detector (FD and SD) performance needs, including: 1)

measurement of the aerosol vertical optical depth versus height in the center of

the array with different systematical uncertainties from the backscattered LIDAR;

5

2) measurement of the horizontal uniformity across the aperture of the array; 3)

monitor of the relative timing of the FD, the relative calibration (photons to ADC

values) of the FD; and the trigger efficiency of the FD; 4) monitor of the relative

timing between the FD and SD; 5) check the FD mirror alignment [3].

• LIDAR stations: this system uses an steerable UV laser beam to probe an

specific region of the atmosphere. The beam backscatters on haze and aerosols,

while the reflected light is collected with a mirror onto a photomultiplier read by a

computer (see Fig. 4 and 5). The pulse shape analysis gives the absorption-

coefficient map of the sky.

Campo di vista

dello specchio Fascio laser

Figure 4: Schematic of a LIDAR system

Figure 5: Typical LIDAR signal, and the

overlap produced between the laser beam and

the telescope mirror

4. LIDAR Telescope, Hardware and Software

4.1. The Telescope Components

We have already installed three LIDAR systems in the stations of Los Leones,

Coihueco and Morados(two of them fully operative). These systems consist basically

of recycled EAS-TOP experiment telescopes [5] doing appropriate modifications and

6

replacing mechanics and electronics parts in manner to do them operatives. A short

description of the components parts follows.

• Telescope Structure: it is an steerable alt – alt azimuthal mount that allows

movimentation on two coordinates: the zenith axis rotates the entire structure

within a range of 180º (see Figure 6). The second one “azimuth” permits to the

structure which support the mirrors and the laser 180º movements in azimuth (see

Figure 7). The combination of these two movements allows a complete coverage

of the sky. The movements in the zenith axis are transmitted trough a geared

wheel and a geared belt. The transmission movements in the azimuth axis are

made through a wheel and an iron cable. The telescope structure holds both

motors (and their covers,) both encoders, limit inductive sensors, extra-limit

mechanical switches and the inductive zero position indicators.

Figure 6: LIDAR telescope structure Figure 7: LIDAR telescope structure

scheme movement in zenith axe scheme movement in azimuth axe

• Laser System: the laser is mounted on the telescope. The laser shots into the

atmosphere in a desired direction and the receiving system measure the back-

scattered light as a function of time. At the moment for economic reasons we have

installed two different laser systems in both LIDAR. Their principal

characteristics are: Los Leones laser system: manufacturer: Big Sky

Technologies, model: “Ultra”, type: Nd:YAG , main wavelength: 355 nm, pulse

energy (max): 6 - 7 mJ, pulse rep. freq.(max): 0 to 20 Hz (used at 10 Hz).

Coihueco Laser system: model: DC30-351, manufacturer: Photonics Industry,

laser type: Nd:YLF, main wave length: 351 nm, pulse Energy: 150 uJ, pulse with:

10-25 ns, repetition rate: 0 to 10 kHz (used at 333 Hz).

• Telescope Optic: each LIDAR detector hold up 3 parabolic mirrors (80 cm

diameter and 41 cm focal length) made of Pyrex glass (O2Si) supported in the

modified alt-azimuthal mounting. The pointing accuracy of these steer able

telescopes is 0.05º. The full field of view of these detectors is ~ 1.76º x 1.76º.

7

• Photomultiplier tubes: the signal is received for three Hammamatsu R7400

photomultipliers tubes each one with a module size 60.7 x 25 mm, including a

UVG filter.

• Telescope Cover: all LIDAR telescopes are protected from atmosphere agents

and solar light with a cover (see Figure 8). Both parts of the cover are opened and

closed at the beginning and at the end of the data acquisition through two robust

motors, steel cable and pulleys system.

Figure 8: photograph of the LIDAR

system with the cover closed

• The Acquisition System: The signal is digitized using a three-channel LICEL

transient recorder TR40-160 with 12 bit resolution at 40MHz sampling rate

with 16k trace length combined with a 250MHz single photon counting

system. Maximum detection distance of the hardware is thus, with this

sampling rate and trace length, set to 60 km. However, in real measurements,

atmospheric features up to 30 km only are observed. LICEL is operated using

a PC-Linux system through a National Instruments digital input-output card

(PCI-DIO-32HS) (see Figure 9). The data acquisition system is managed

using a ROOT graphic interface [6] .

8

Figure 9: The LICEL TR40-160 receives the trigger from

the laser and the signal from Hammamatsu R7400 phototube.

The Linux-PC controls the LICEL digitizer through

PCI-DIO-32HS Digital Input/Output card.

4.2. Telescope Control Movement

4.2.1. Hardware

• Control Panel and Electronic: the panel is divided in two parts (see Figure

10), the left one which contains the components powered at 220 V and 110 V

while the right one contains the low voltage components ( 24 V and 5 V).

Figure 10: Schematic of the Control Panel indicated the new components that have been included in

order to upgrade the telescopes

Maestro

modules

transformers

Motion

Coordinator

Module(MC204)

CAN 16 I/O

and CAN 8

Analog Inputs

9

Motion Coordinator Module (MC204): this Trio Technology module is

capable of driving a multi-axis machine and its auxiliary equipment, in our

telescope the module commands 2 axes. The controller was programmed

using the TRIO BASIC programming language. This can be used to build

"standalone programs" or commands can be sent from an external computer

[7]. This system allows adding digital I/O and additional equipment. MC204

contains the processor and logic power supply.

Each MC204 module contains two daughter boards with respective serial

ports used for the encoder communication.

The Motion Coordinator is connected to our computer via an RS-232-C serial

port that we used to both load programs and execute commands from remote.

4. 2. 2. Software

The telescope control software was developed in TrioBASIC, a language similar of

BASIC with predefined functions for the control trough the “Motion Coordinator”.

TrioBASIC is a multitasking language and can run up to five programs

simultaneously .The code is recorded in the MC204 internal memory ( 122 kB flash

EPROM).

The different telescope functions and the associated programs are:

- initialization variables (STARTUP.BAS),

- open cover (OPEN_COVER.BAS),

- close cover (CLOSE_COVER.BAS),

- LIDAR initialization (INIT.BAS),

- manual control (PAD_SON.BAS),

- telescope movement (MOVETO.BAS),

- telescope park (PARK.BAS) ,

- telescope unblock (UNBLOCK.BAS).

Is available the telescope works local mode through a pad which can be connected

to the control panel.

In addition the telescope can work in remote manner, it means, can be driven

from the Central Building. The communication between the MC204 and the Linux

LIDAR computer on each site is established trough the serial port RS 232 that was

mention before. There is a program in C++ which permit to send instructions through

the Linux shell and in the same way permit receiving information from the MC204.

5. LIDAR Data, Data Control and Analysis

5.1. Data Structure

10

The data from Malargüe LIDAR stations is stored in root binary files [8] . Basically

these files contains objects inherited from ROOT TObject .

There are two main objects of concern, one is the object that belong to the class

TRunHeader which has the basic information about the LIDAR run: the acquisition

starting time, the total number of LIDAR events inside the file the number of operational

digitizer modules, the PMT voltages, the laser power, and some remarks or comments.

The other object is TLEvent, used as data container for the LIDAR signal provided by all

operating LICEL transient recorders. For each channel we recorded the number of single

photon counting and the flash ADC trace.

During the acquisition runs, the temporal sequence of LIDAR events is stored in the

file, following a TTree structure. In ROOT a TTree structure is made of a series of

TBranch (in his case assigned to TLEVENT objects) that can be read independently of

each other, as distinct “TTree entries”.

5.2. Strategies for scanning the sky

5.2.1. “AUTOSCAN” strategy

There are four different strategies for the data taking. The AUTOSCAN strategies are

running automatically during all the night covering a sky cone of approximately 50º

(outside of FD field of view), and doing both discrete and continuous scanning:

• Discrete sweeps strategy: in this case the shoot trajectory follow the

equation:

sec1sec ∆+= iθ

where 065.0sec =∆ and θ is the polar angle;

- for the zenith discrete sweep we use: ϕ = 0º, 180º ;

- for the azimuth discrete sweep we use: ϕ = +120º, -120º.

where ϕ is the azimuth angle centered.

• Zenith continuous sweep: the telescope shoots from 54º to -44º in zenith

with azimuth= 0 during the entire scan at 0.075 º/s velocity.

• Azimuth discrete sweep: in this strategy the telescope shoots from 43º to

-29º in azimuth with zenith= 0 during the entire scan at 0.058 º/s

5.2.2. Shoot the Shower (StS) strategy

Through this strategy will be possible to monitor the attenuation coefficients from the

shower path to the Fluorescence Detectors. The intention is that clouds and patches

would be accounted properly, and their effects on the FD reconstruction, corrected.

11

These strategy works only with an important event is seen by at least two of the

fluorescence detectors (see Figure 10), or when the event is seen by both the

fluorescence and the surface detector (“stereo event”). The sequence can be described as

follow:

- the FD generates a trigger (T3) of third level (T3 contains information about the

geometry of the shower, the SDP vector);

- LIDAR receives the T3 from FD;

- Run Control (client of the LIDAR server) calculates the shower track points vector

position;

- are setting new LIDAR steering parameters;

- LIDAR request to enter in the FD field of view (FOV), so the FD toggles Lid Veto ON;

- FD Data Acquisition stops, with the photomultipliers on;

- LIDAR init the Shoot the Shower scanning (see Figure 11);

- when the scanning is done (approximately takes 2 minutes), LIDVeto is put in OFF;

- after this AUTOSCAN strategy restart with a Zenith Discrete sweep;

Figure 10: An event seen by the FDEyeDisplay, this produce

a T3 which is send it to the LIDAR to start the StS strategy.

12

Figure 11: Trajectory that LIDAR described according

do to the shoot the shower strategy (as OLV display

can show).

5.3. Data Control

5.3.1. LIDAR Signal

A typical LIDAR signal (Figure 5) has a big peak, where the laser beam enter

completely in the mirror field of view, this happens at about 150 m. After the maximum

peak the signal decrease in an exponential way.

The overlap signal during the acquisition can suffer variations (even saturation) due to

the high voltage value set to the photomultipliers, the laser intensity, mirror

misalignment, etc. Due to this, we consider necessary to do a control of this data.

5.3.2. Functions implemented to OLV viewer

In in order to control the data quality, we implemented some functions to the existent

software viewer for shifters “OLV”.

“OLV” is a C++ software with a graphic interface QT. This program is launched at the

beginning of the data acquisition and is running during the entire data taking. “OLV”

permit to the shifters visualize the state of the run acquired (event by event), and is

possible select the visualization way: scale, PMT , bin averaging ,etc. (see Figure 12).

According to control the data we had implement to OLV the possibility of seeing the

evolution of the overlap function (peak signal) during all the data taking, the offset

variation and its σ2 (see Figure 13 and 14 ).

13

Figure 12: “OLV” graphic interface where we can visualize

the signal for each event in the “Event Display” window.

Figure 13 and 14: “OLV” graphic interface with the new add function which permit the visualization

of the overlap function (peak signal) during all the data taking, the offset variation and its σ2.

14

5.4. LIDAR Data Analysis

In an extensive air shower the cascade of particles along the shower axis is limited to a

narrow lateral distribution. In this way, fluorescence light is emitted with an intensity

proportional to the number of charged particles in the shower. The calorimetric measure

of the total electromagnetic (EM) component of the shower energy [6] is proportional to

the integral of EM particle density Nem along the shower direction x,

dxxNKE emem )(∫= (1)

with K ≅ 2.2 MeV cm2/g, where x is measured in units of longitudinal air density (g/cm).

E_ em is a lower bound for the energy of the primary cosmic ray. The lower portion of

shower development is usually obscured by the ground so that EM cascade reaching

below ground is included by fitting a functional form to the observed longitudinal profile

and integrating the function past surface depth. The number of photons Nph reaching

fluorescence detector (FD) is proportional to EM particle density Nem at the point of

production x ,so that in turn,

)(

)()(

2

xT

xrNxN

ph

em α (2)

with r(x) being distance between shower point and FD. Light originating within the

shower is certainly affected by the absorption and scattering on molecules and aerosols in

the atmosphere. The number of detected photons is thus reduced due to non-ideal

atmospheric transmission T(x) < 1, where:

)(

0

)(exp)( x

r

edrrrTτα −=

−= ∫ (3)

with α(r) volume extinction coefficient along the line-of-sight, and τ (x) the resulting

atmospheric optical depth (OD) to the shower point x.

In this sense, the atmosphere can be treated as an elementary-particle detector.

However, weather conditions change the atmospheric transmission properties

dramatically, resulting in a time-dependent detection efficiency. Therefore, an absolute

calibration system for fluorescence light absorption is an essential part of FD.

One of the most suitable calibration setups for FD is the backscattering LIDAR system,

where as we mention before, a short laser light pulse is transmitted from FD position in

the direction of interest. With a mirror and a photomultiplier tube, backscattered light is

collected and recorded as a function of time, i.e. as a function of backscatter distance.

In order to extract τ (x) from a backscattered elastic LIDAR signal we have to invert

the LIDAR equation (4):

15

)(2

2

0

0 )(2

)( rer

Ar

tcPrP τβ −= (4)

P(r) is the instantaneous received laser power (backscatter signal) at time t from a

distance r;

P0 is the instantaneous transmitted laser power at time t0;

t0 is the laser pulse width;

β (r) is the backscatter coefficient;

A is the effective receiving area of the detector , proportional to the area of the mirror and

proportional to an overlap between its field of view with the laser beam.

As we can see from Equation (2), the precision of measurement of α and the

corresponding integral τ directly influences the precision of primary particle energy

estimation.

Even if the LIDAR equation (4) can look simple, is difficult to solve it for the two

unknown variables α (r) and β (r), this leads to the ambiguity in their determination: the

equation can not be resolved without additional assumptions about atmospheric

properties.

Correcting for r2 and normalizing to r = r0, we can introduce an auxiliary S function:

[ ] );(2/)(ln)(

)(ln)( 002

00

2

rrrrrP

rrPrS τββ −== (5)

τ (r;r0) corresponds to the atmospheric optical depth between r0 and r, that we want to

estimate;

We consider separately the molecular and aerosol influence, and the coefficients α and

β can be written as the sum of the contribution of two independent components ( the

molecular and the aerosol), in this way:

)()()( hhh am ααα += and )()()( hhh am βββ += (6) and (7)

where αm ,αa, βm, βa correspond to molecular and aerosol attenuation and backscattering

coefficient respectively.

Both quantities are proportional to the local density and from:

)()(180

hd

dh ρ

σβ

Ω= (8)

)()( hh TOT ρσα = (9)

we have:

16

)()( 180 hd

d

hTOT

ασ

σ

βΩ

= (10)

where: º180Ωd

dσ = )º180(P (11)

Then, we can write )(hβ as:

)()º180()()º180()( hPhPh aamm ααβ += (12)

The aerosol phase function Pa(180º) for backscattering has, apart from the wavelength,

also a strong dependence on the optical and geometrical properties of the aerosol

particles. Nevertheless, at wavelength of 355 nm, values in the range 0.025 to 0.05 sr-1

can be assumed for aerosol phase function.

The angular dependence of molecular phase function is defined by the Rayleigh

scattering theory, where Pm(180º) = 3/8π sr-1

.

The molecular density variation with height is well known and have very little seasonal

fluctuations (see Figure 15).

Figure 15: Pure molecular optical depth variation in the different

season.

17

The proportion of clear skies is definitely lower than the percent of nights with aerosol

concentrated in clouds or aerosol layers, this is because the importance of the cloud

detection.

For clear sky the S function is similar to a line in which the pendent decrease lightly

with height (Figure 16), in fact assuming constant the attenuation and backscatter

coefficient we can write:

[ ] )()(2)]/()([ln);(2/)(ln)(

)(ln)( 01

0

00002

00

2

rraadrrrrrrrrP

rrPrS

r

r

−+=−=−== ∫αββτββ

The non linearity of the S function evidence the presence of aerosol or at higher levels

clouds. In this way we can delimit a zone of linearity defining the height in which start a

cloudy zone. In this terms according as we develop in the equations before, knowing or

estimating a αaerosol at this limit height, we can calculate the values of αa at any height

along the clear zone.

Figure 16: S function vs. height, in this graphic we can

observe two zones well define, the portion of clear sky and as

a big reflex ion the zone with cloud presence (RUN 1344 Los

Leones station).

For this propose we consider a method to calculate minimal height of a cloud layer.

The method consists in 1) calculate the derivative to the S function (see Figure 17)

applying a bin averaging chosen (we took a binavg. = 4); 2) consider the signal range in

which we analyze the presence of clouds, this is the range in which the S function has

acceptable error values, 3) put a threshold in the value of the maximum derivative

(according with different tests we chose dS = 0.5 ); 4) observing where the dS value is

evidence of

clouds at ≅

5Km height

linear function in

clear sky portion

18

superior to the threshold value and if this occur for some successive bins; 5) form a

vector with all the cases in which occur this, 6) do a comparison with the other

photomultipliers and take the coincidences, 7) take as cloud minimum height the

minimum value obtain in the coincidences.

Figure 17: derivative of S function (dS). The point line in blue

indicate the beginning of the cloud layer detected by the program, the

point line in red indicate the threshold use in the detection of the

cloud layer.

In order to repeat the calculation sequence for all the runs during a night, month, etc..,

we write a program in C++ named LDA2.cc, the one create an output file with diverse

information ( group of events, shoot angles, signal peak, minimum cloud height, etc…) .

Another program in C++ named histocloud2 make the different histograms and graphics

that represent the results of the night or month in consideration.

We took for our first test a cloudy night : the October 14, 2004 night (Los Leones

LIDAR station). Running LDA2 for the night (runs R2629 to R2649) the results (for all

the groups of events) are shown in Figure 18, the blue bars belong to the LIDAR range in

which is calculated the presence of clouds. We can obtain even a graphic of the cloud

height medium value for each run, during the same night (Figure 19).

As a clear night we analyze the October 10, 2004 night (Los Leones station). The

results are shown in Figure 20 . In Figure 21 a summary for all the data entries (October,

2004) and the respective cloud minimum height.

cloud layer

found at 5.4 Km

19

Figure 18: October 14 night, the blue bars represent the range in

which is calculated the presence of clouds, while the white points

represent the minimum height of the clouds, this is considering all

groups of events in the runs.

Figure 19: October 14, 2004 night, each point represent for each

run the cloud mean height, with the respective error bars.

20

Figure 20: October 10, 2004 night, showing just few points reveling

the cloud presence.

Figure 21: this histogram represent the proportion of cloud at

different altitudes height, taking as data all the events in October

2004 (Los Leones LIDAR station).

We represent even the LIDAR run data taking during all October month (see Figure

22), and the percent of cloud coverage sky according with the data obtain from our

software (see Figure 23)

In my thesis I will discuss the determination of cloud parameters during all the data

taking, the evaluation of the systematic error and its further development.

21

Figure 22: Run data taking during all October month, the

dark band represent the limit of the data acquisition during

twiglight time.

Figure 23: Percent of cloud coverage sky during October 2004(Los Leones station)

according with the data obtains with our method of the height layer cloud detection.

LIDAR run October 2004

% Cloud Coverage October 2004

0102030405060708090

100

6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

bad weather

22

References

[1] V. Hess, Phys. Zs, 13, 1912, 1084.

[2] A. Filipcic , M. Horvat, D. Veberic , D. Zavrtanik , and M. Zavrtanik, Analysis of

Lidar Measurements at the Pierre Auger Observatory, (2002).

[3] Brian Fick , James Matthews, John Matthews, Rishi Meyhandan, Megan McEwen,

Miguel Mostafa, Michael Roberts, Paul Sommers, Lawrence Wiencke, The First Central

Laser Facility, Auger GAP 2004-003 (2003).

[4] Iztok Arcon, Andrej Filipcic, Marko Zavrtanik, Jozef Stefan, UV LIDAR System for

Atmospheric Monitoring and Cloud Detection, Auger GAP 99-028 (999).

[5] R. Cester, M. Mostafa, R. Musa, LIDAR Telescopes for Atmospheric Monitoring,

Auger GAP 2001-052 (2001). [6] A. Filipcic, M. Horvat, D. Veberic, D. Zavrtanik, M. Zavrtanik, Scanning Lidar

Based Atmospheric Monitoring for Fluorescence Detectors of Cosmic Shower, Auger

GAP 2003-001 (2003).

[7] Trio Motion Technology Ltd., Motion Coordinator Technical Reference Manual,

(1998).

[8] Antonio R. Biral, Analysis of the 2002 Malargüe LIDAR data through Fernald’s

Method, Auger GAP 2003-008 (2003).

[9] F. G. Fernald, Analysis of Atmospheric LIDAR Observations, Optical Society of

America, 1984.