fasa – fire airborne spectral analysis of natural disasters...table i. technical data of the miror...

117

ANNALS OF GEOPHYSICS, VOL. 49, N. 1, February 2006

Key words remote sensing – wild fires – volcanoemissions – Fourier spectrometry – data-fusion

1. Introduction

Fires change ecosystems in all vegetationzones. The concern of the public, politicians andscientists about climate change and environmen-tal hazards has stressed the need for a better un-derstanding of the dynamics of biomass burningand volcanic activities at all scales from local toglobal, in order to estimate the role of fire as acontributor to that change. Recent publicationsindicate that biomass burning may be a majorsource of greenhouse gases (Page et al., 2002).

However, large uncertainties in the quantifica-tion of these emissions exist.

Airborne remote sensing is the predestinedmethod to investigate natural disasters. It per-mits riskless measurements and coverage oflarge and otherwise often difficult to access ar-eas. However, by now mainly space borne sen-sors that were not developed for HTE investiga-tions were used. They often suffer from detec-tor saturation and low spatial resolution. Usual-ly they permit only the detection of larger firesand provide only surface temperatures and thesize of the burnt areas; but no direct informationwith regard to the emitted gases.

The Italian Instituto Nazionale di Geofisica eVulcanologia (INGV) and the German Aero-space Center (DLR) cooperate in developingnew airborne instrumentation specifically forHTE analyses. This system is called FASA (FireAirborne Spectral Analyser). The objective is todetermine not only temperatures and areas; be-yond these parameters the quantification of thereleased gases and aerosols is to be achieved.

FASA – Fire Airborne Spectral Analysis of natural disasters

Erwin Lindermeir (1), Hermann Kick (1), Volker Tank (1), Peter Haschberger (1),Maria Fabrizia Buongiorno (2), Giuseppe Distefano (2), Stefania Amici (2), Dieter Oertel (3),

Friedrich Schrandt (3), Wolfgang Skrbek (3) and Holger Venus (3)(1) DLR – German Aerospace Center, Remote Sensing Technology Institute, Wessling, Germany

(2) Istituto Nazionale di Geofisica e Vulcanologia, Roma, Italy(3) DLR – German Aerospace Center, Institute of Space Sensor Technology

and Planetary Exploration, Berlin, Germany

AbstractAt present the authors are developing the system FASA, an airborne combination of a Fourier Transform Spec-trometer and an imaging system. The aim is to provide a system that is usable to investigate and monitor emis-sions from natural disasters such as wild fires and from volcanoes. Besides temperatures and (burned) areasFASA will also provide concentration profiles of the gaseous combustion products. These data are needed to im-prove the knowledge of the effects of such emissions on the global ecosystem. The paper presents a descriptionof the instrumentation, the data evaluation procedure and shows first results of retrieval calculations based onsimulated spectra.

Mailing address: Dr. Erwin Lindermeir, DLR – Ger-man Aerospace Center, Remote Sensing Technology Insti-tute, P.O. Box 1116 D-82230 Wessling, Germany; e-mail:[email protected]

118

Erwin Lindermeir et al.

The task requires the use of a sensor systemthat provides high spatial and high spectral res-olution simultaneously. The following sectionsfirst present the choices made for the instru-mentation and secondly the principles of the da-ta fusion algorithms as well as a first simulationresults.

2. Instrumentation

2.1. Concept

At first glance one might be tempted to con-sider the use of an imaging spectrometer, whichprovides spatial and spectral resolution simulta-neously. However, so as to retrieve concentra-tion profiles it is necessary to resolve the singleemission lines of the gases in the spectra. Theirline widths are in the order of 0.1 cm−1. In addi-tion, burning areas exhibit a quite high degree ofspatial variation which leads to the demand foran accordingly high spatial resolution (∼1 m) ofthe sensor. As the retrieval algorithms which areused to derive concentrations and temperaturesfrom the spectra need a certain SNR level toproduce high quality results, the use of an imag-ing spectrometer for this application will sufferfrom a distribution problem: the received energycan either be used to obtain low noise spectra athigh spectral resolution, or low noise images athigh spatial resolution.

Therefore it was decided to use two special-ized instruments for the two tasks: a three chan-nel imaging system and a Fourier transformspectrometer. The next two sections describethese instruments in more detail.

2.2. MIROR Fourier transform spectrometer

The airborne Michelson Interferometer withRotating Retro-reflector (MIROR) was origi-nally designed and built by DLR to investigateemissions of aircraft engines in flight (Hasch-berger and Lindermeir, 1996). It is thereforevery well suited to record spectra of hot com-bustion products. The spectra provided byMIROR have already been successfully used todetermine temperature and concentration pro-

files in the plumes of aircraft jet engines (Lin-dermeir et al., 2001).

Operating a Michelson interferometer in aharsh environment such as on board an aircraftis a challenge for the optical set-up of the in-strument. The quality of the interferometer’sdetector signal strongly depends on the preciseposition of the optical components. Permissibletolerances are below the wavelengths of the in-cident radiation, i.e. less than a micrometer forinfrared applications. From the mechanicalpoint of view the necessary movement of one ofthe optical components further complicates theproblem. Consequently, Michelson interferom-eters are routinely used mainly under laborato-ry conditions.

With the MIROR design a very rugged opti-cal set-up was realized which additionally over-comes the stop-and-go movement of other solu-tions. The most advanced of the MIROR set-ups is depicted in fig. 1 from which the her-itage from the classical Michelson set-up is stillobvious. There are the beamsplitter, the twomirrors, which are labelled ‘mirror with hole’,and the detector with appropriate optics. In ad-dition to these classical components two mir-rors direct the two beams which the beamsplit-ter generates into a retro-reflector. This is anoptical element consisting of three plane mir-rors which are perpendicular to each other, i.e.they form the corner of a cube. The retro-reflec-tor property, which is exploited by the MIRORprinciple, is that beams are reflected in thesame direction in which they entered the reflec-tor. The reflector only introduces a shift to theside. This effect is independent of the angle ofincidence. The two beams from the beamsplit-ter are therefore reflected back to the mirrorswith the holes from where they are sent back tothe beamsplitter. Here they recombine and theinterference pattern finally reaches the detector.

During a measurement the reflector is spin-ning at constant angular velocity around the ax-is. As indicated in fig. 1 the reflector’s centre islocated at a distance L from the axis of rotation.As shown in (Haschberger, 1992) this causes acontinuous change of pathlengths for the twobeams and thus the interferogram is generated.

As mentioned above the ruggedness of theDLR MIROR spectrometer was already shown.

119


It provided high quality spectra not only underflight conditions but also during jet engine runsin test rigs on the ground. These latter testswere extreme, as the spectrometer was not onlysubject to vibrations but also to high levels ofacoustic noise. Figure 2 shows a photograph of

the optical set-up installed in a DLR researchaircraft. The technical data of MIROR are sum-marized in table I.

For the FASA experiment MIROR will beused in a NADIR configuration. Table II pro-vides the footprint diameter on the ground and

Fig. 2. Optical set-up of the MIROR spectrometer installed in one of DLR’s research aircrafts.

Fig. 1. Optical set-up of the MIROR Fourier transform spectrometer.

120


the smear for some flight altitudes. Typicallythese will be in the range 3000 to 4000 m.

2.3. ABAS Advanced Bird Airborne Simulator

ABAS consists of the MIR/TIR Hot SpotRecognition Sensor (HSRS) and VIS/NIRWide-Angle Airborne Camera (WAAC).

HSRS is a bi-spectral push-broom scannerwith spectral bands in the MWIR at 3.9 nm andTIR at 8.8 nm. The sensitive devices are 2MCT photodiode lines. The lines – with identi-cal layout in the MWIR and TIR – comprise 2 ××512 elements each in a staggered structure.

These arrays have to be cooled to 100 K in theMWIR and to 80 K in the TIR. The maximumachievable at 80 K TIR photodiode cut-off wave-length of about 10 nm and the atmospheric ozoneband at 9.6 nm make it necessary to use the 8.5-9.3 nm band for TIR channel of the HSRS in-stead of the commonly used 10.5-11.7 nm band.The cooling must be conducted in a satellite mis-sion by small Stirling cooling engines. HSRS

sensor head components of both spectral chan-nels are based on identical technologies to pro-vide good pixel co-alignment. Both spectralchannels have the same optical layout but withdifferent wavelength-adapted lens coatings.

The HSRS sensor data are read out continu-ously with a clock time which is exactly onehalf the pixel dwell time. This time controlled«double sampling» and the staggered line arraystructure allow a geometric resolution enhance-ment compared to a classic line array.

Radiometric fire measurements require alarge dynamic range, as outlined above. To fulfilthis requirement, a second scene sampling has tobe performed with reduced exposure time (with-in the same sampling clock time interval!) if hotareas are identified during the real-time process-ing of the data from the first scene sampling. Thissecond sampling is only performed if the real-time processing of the first sampling set indicatesthat detector elements are saturated or close tosaturation. These supplementary sampled pixelsare added to the normal thermal scene measure-ment data as the so called «Hot Area (HA)» datasub-set. All these sampled data are transmitted tothe HSRS data acquisition controller via a specialserial interface. In the result, the HSRS has anadaptive radiometric dynamic range which al-lows the recognition of High Temperature Events(HTE) such as wildfires, volcanic activities, orcoal seam fires without sensor saturation.

Table III shows the main technical data ofABAS and table IV lists the swath widths andpixel sizes on the ground for various flight alti-tudes.

Table I. Technical data of the MIROR Fouriertransform spectrometer.

Spectral range 600-3000 cm-1,(3.3-16.6 nm)

Spectral resolution 0.2 cm−1

Registration time 0.22 sper interferogramMeasurement rate 3 Hz

Field of View 20 mradNESR (single scan) 6.2×10−8 W/(cm2 sr cm−1)

Table II. Footprint diameter and smear for variousflight altitudes above ground.

Flight altitude MIROR footprint Footprint smearabove ground diameter for 75 m/s speed

(m) (m) over ground (%)

2000 40 27.53000 60 144000 80 7

Table III. ABAS technical data.

Spectral range VIS/NIR: 0.4-0.6 nmMIR: 3.4-4.1 nmTIR : 8.5-9.3 nm

Field of View VIS/NIR: 80°MIR/TIR: 19°

Detector elements VIS/NIR: 5184number (no. of pixels/line) MIR/TIR: 2×512

(staggered)

Instantaneous VIS/NIR: 0.27 mradField of View MIR/TIR: 0.66 mrad

121


(co-alignment). Furthermore, synchronisationof the recording of spectra and images in timeis required.

The latter task is solved by establishing acommon time base, i.e. the GPS time of ABAS.In addition, MIROR is provided with an electri-cal signal that permits counting of the currentlyrecorded image line of ABAS’VIS/NIR channel.This line counter is stored together with the im-ages and the interferograms. Thus the correct as-signment of the data from both systems is guar-anteed.

Co-alignment is established by also using theVIS/NIR channel: a small video camera equippedwith the same optical filter is fixed at the opticalblock of the spectrometer. Its line of sight is di-rected in parallel to the spectrometer’s line of

Table IV. ABAS-Swath width and pixel sizes for various flight altitudes above ground.

Flight altitude above ground (m) Swath width (m) Linear pixel size (m)VIS/NIR MIR/TIR VIS/NIR MIR/TIR

2000 ∼3300 ∼660 ∼0.7 ∼1.43000 ∼5000 ∼1000 ∼1.0 ∼2.04000 ∼6600 ∼1320 ∼1.4 ∼2.8

Fig. 3. Block diagram of the BIRD airborne simulator.

Fig. 4. ABAS sensor platform.

Figure 3 depicts a block diagram of theABAS components. In addition to the camerasthere are a GPS receiver and a gyro processingunit which aid in geo-referencing the recordedimages. ABAS is also connected to the air-craft’s ARINC bus which provides additional(house keeping) data from the aircraft such astemperature and pressure at flight altitude, trueair speed and the like. A photo of ABAS is pro-vided with fig. 4.

2.4. Co-alignment and synchronisation

As both systems are installed as mechani-cally independent instruments in the aircraft, itis necessary to align their measurement spots

3 4

122


sight. The camera’s FoV is about 10 times largerthan the spectrometer’s FoV. The area observedby the spectrometer in the video camera imagesis determined before flight in the laboratory.

During flight the video images are recordedsimultaneously with the interferograms and theABAS images. By applying well known tech-niques (e.g., correlation) the spectrometer FoVis located in the ABAS images.

3. Data fusion algorithm

Estimating High Temperature Event (HTE)emission characteristics based on NADIR FTIRmeasurements is a challenging problem.

First, the retrieval of temperature and con-centration profiles from spectroscopic measure-ments obtained by vertical sounding is wellknown to be an ill-posed problem, and ad-vanced numerical techniques have to be used tofind meaningful solutions. In addition, due tothe limited spatial resolution of Fourier trans-form spectrometers, HTE scenes observed bythe NADIR viewing FTS will be horizontallyinhomogeneous, i.e., in case of fire observa-tions the observed spectrum will be a superpo-sition of contributions of flaming, smouldering,smoky, and unperturbed areas.

For the solution of ill-posed problems thereare several algorithms available to obtain solu-tions. During the last years a method based onthe ‘Generalised singular value decomposition’was developed at DLR and will be applied forFASA. A detailed description of this method isavailable in Schimpf (1999).

The problem caused by scene inhomo-geneities can be solved by combining theMIROR spectra with the images provided bythe ABAS system. This ‘data-fusion’ will bebased on a sequential approach.

In a first step the data obtained by the imag-ing system is used to analyse and classify theFTIR footprint area. In this manner the imagesprovide essential information for the retrieval ofgas parameters of the HTE, i.e. surface compo-sition (distribution of flaming, smouldering,etc.) and temperature estimates. Based on thisknowledge an initial guess for the iterative non-linear least squares retrieval problem is ob-tained. This calculation finally yields the tem-perature and concentration profiles of the HTE.

The proceeding described above is visual-ized in fig. 5 and fig. 6.

3.1. Test calculation

Retrieval tests conducted at DLR havedemonstrated the feasibility of the proceedingoutlined above. Here the result of one such cal-culation is briefly described.

It was assumed that the spectrometer’s FoVis filled by a flaming, a smouldering, and anundisturbed zone plus smoke. The exact frac-tions of two cases are shown in fig. 7. The pa-

Fig. 5. Footprints of MIROR and ABAS. ABAS de-termines the distribution of the surface radiancewithin the MIROR footprint.

Fig. 7. Radiance spectra calculated for an undisturbed atmosphere and two different compositions of the spec-trometer footprint as indicated.

123


rameters describing the different zones are tak-en from various papers (Pereira and Setzer,1996; Stocks et al., 1996; Worden et al., 1997).With regard to gases this investigation consid-ered only H2O, CO2, and CO. The spectra in fig.

7 show high levels of contrast in the spectral re-gions in which these gases have emission linesas well as a generally increased signal level inthe window regions which is caused by radi-ance from the hot surface.

Fig. 6. Schematic visualizing the FASA data fusion algorithm.

124


Figure 8 shows retrieved CO2 profiles forthe «5% flaming zone» case. Obviously in thiscase the algorithm is capable of determining theconcentration profiles with high accuracy.

4. Conclusions

In order to improve the knowledge of the ef-fect of emissions by wild fires and volcanic ac-tivities on the ecosystem it is necessary to quan-tify the amount of emitted gases; especiallygreenhouse gases. The FASA system presentedin this paper aims at contributing to this task. Itconsists of the Fourier Transform spectrometerMIROR and the imaging system ABAS. Bothare operated in parallel on board a research air-craft. ABAS provides images of high spatialresolution. These are used to provide additionalinformation which is necessary to retrieve tem-perature and gas concentration profiles from theMIROR spectra. Results of a first calculation todemonstrate the feasibility of the retrieval algo-rithm were provided and show encouraging re-sults. Of course, further investigations are nec-

essary to describe the complete data evaluationchain. The aim of the FASA development is toprovide parameters such as gas column or pro-file information of e.g., CO2, H2O, CO, andSO2, temperatures, and also combustion effi-ciency, and aerosol parameters. As a first appli-cation of FASA a measurement campaign overMt. Etna, Italy is planned in 2003. It can be ex-pected that in the future the FASA capabilitieswill also permit improved evaluation of satellitedata.

REFERENCES

HASCHBERGER, P. (1992): Entwicklung eines Fourierspek-trometers mit rotierendem Retroreflektor, Ph.D. Thesis(Technical University Munich).

HASCHBERGER, P. and E. LINDERMEIR (1996): Spectrometricinflight measurement of aircraft exhaust emissions, J.Geophys. Res., 101, 25995-26006

LINDERMEIR, E., S. CLAUSEN, R.M. GEATCHES and E. VIVO

(2001): Profiling Spectrometery to simultaneously in-vestigate the spatial distribution of temperature andchemical species in aircraft exhausts (AEROPROFILE),Synthesis Rep. (European Commission, Project No.BE97-4467, Contract No. BRPR-CT-97-0509).

PAGE, S.E., F. SIEGERT, J.O. RIELEY, H-D.V. BOEHM, A. JAYA

Fig. 8. Retrieval results for the «5% flaming zone» case: CO2 profiles for the flaming and the smoulderingzones.

125


and S. LIMIN (2002): The amount of carbon releasedfrom peat and forest fires in Indonesia during 1997,Nature, 420, 61-65.

PEREIRA, A.C. and A.W. SETZER JR. (1996): Comparison offire detection in savannas using AVHRR’s channel 3 andTM images, Int. J. Remote Sensing, 17 (10), 1927-1937.

SCHIMPF, B.A. (1999): Analysis of far infrared spectra forthe determination of the middle atmospheric OH con-centration, Ph.D. Thesis (University of Bremen).

STOCKS, B.J., D.R. CAHOON, J.S. LEVINE, W.R. COFER andT.J. LYNHAM (1996): Major 1992 fires in Central andEastern Siberia: satellite and fire danger measure-ments, in Fire in Ecosystems of Boreal Eurasia, editedby J.G. GOLDAMMER and V.V. FURYAEV (Kluwer Acad-emic Publ., Dordrecht, The Netherlands), 139-150.

WORDEN, H., R. BEER and C.P. RINSLAND (1997): Airborneinfrared spectroscopy of 1994 western wildfires, J.Geophys. Res., 102, 1287-1300.

fasa – fire airborne spectral analysis of natural disasters...table i. technical data of the miror...

Documents