Data Processing and Mapping in Airborne Radiometric Surveys

SCHWARZ, G.F., KLINGELE, E.K., RYBACH, L., Institut für Geophysik ETHZ, CH-8093 Zürich. Switzerland

Abstract

Helicopter surveys permit rapid evaluation with complete areal coverage of the terrestrial gamma radiation from natural and artificial radioisotopes in the topmost layer of the ground. The gamma spectrometric measurements (256 channels) are performed by a 16.8 liter NaI detector.
        In order to identify possible radiation level changes around the five Swiss nuclear installations (four power plants and one research facility) the surrounding regions of each site are surveyed annually. In addition, regions with elevated natural radioactivity are mapped within the framework of the Swiss National Geophysical Survey. In case of accidents with radioactive material, or from debris of nuclear-powered satellites, the system would be used to locate the radioactive sources.

Zusammenfassung

Helikoptermessungen erlauben eine schnelle und flächendeckende Erfassung der terrestrischen Radioaktivität der natürlichen und künstlichen Radioisotopen in der obersten Erdschicht. Die gammaspektrometrischen Messungen (255 Kanäle) wurden mit einem NaI-Detektor mit 16.8 Liter Volumen durchgeführt.
        Um mögliche Veränderungen des Strahlungspegels in der Umgebung der fünf schweizerischen Kernanlagen (vier Kernkraftwerke, ein Forschungsinstitut) zu erfassen, wird jedes Jahr eine Befliegung durchgeführt. Zusätzlich wurden die Gebiete der Schweiz mit erhöhter natürllicher Radioaktivität im Rahmen der geophysikalischen Landesaufnahme vermessen. Das Messsystem ist auch für Notfalleinsätze zum Beispiel zur Lokalisation von Bruchstücken von radioaktiven Satelliten oder verlorenen radioaktiven Quellen vorgesehen.

Résumé

Les levés par hélicoptère permettent une évaluation rapide avec une couverture aérienne complète de radiation terrestre gamma à partir d’isotopes radio naturels et artificiels dans la couche la plus superficielle du sol. Les mesures spectrométrique gamma (256 canaux) sont réalisées par un détecteur NaI 16.8 litres.
        Afin d’identifier des changements possibles de niveau de radiation autour des cinq installations nucléaires de la Suisse, (quatre unités de production et une installation de recherche) les régions entourant chacun de ces sites sont surveillées annuellement. De plus, des régions ayant un taux de radioactivité naturelle élevé sont cartographiées dans le cadre de l’institut national suisse de géophysique. En cas d’accident avec un matériel radioactif ou avec des débris de satellites à énergie nucléaire, le système pourrait être utilisé pour localiser les sources de radioactivité.

Introduction

Regional and local gamma-ray radiation maps are of key importance for a variety of purposes: location and monitoring of contamination, basic data for radiation biology (variation of the natural exposure rate in the context of the effects of low doses), information relevant to prospecting for raw materials (for example potassium alteration) and geological mapping.
        Since areal radiometric surveys are very expensive and time consuming, they are advantageously carried out airborne. Airborne surveys permit rapid evaluation of radiation levels of large areas. In inaccessible regions, surveys with complete areal coverage are only possible from the air. Because of the larger ground clearance and the higher speed, the coverage per unit time of an airborne system is about 2500 times larger than of a comparable ground system. This is of key importance in radiological accidents. Although the costs for the measuring instruments and the flights are relatively high, the resulting cost per surveyed area is clearly lower than for a comparable terrestrial survey. Three projects actually make use of airborne radiometric measurements in Switzerland.
        The first project aims at a better evaluation of the natural radiation level in Switzerland. The Swiss Geophysical Commission (SGPK) decided to map selected regions with elevated natural radioactivity within the framework of the Swiss National Geophysical Survey (Schwarz et al., 1992). The main attention was given to the crystalline rocks of the Central Massifs of the Swiss Alps because of their relatively high natural radioactivity. The area covered by this survey is about 3000km².
        The second project financed by the Swiss Federal Nuclear Safety Inspectorate (HSK) concerns on artificial radioactivity. The surrounding regions (approx. 50km²) of the four nuclear power plants (Beznau, Gösgen, Leibstadt and Mühleberg) and the Swiss nuclear research facility (Paul Scherrer Institute) are surveyed annually since four years. The measurements aim to monitor the dose-rate distribution and to provide a documented reference base (Schwarz et al., 1989, 1990, 1991 and 1992). From 1994 on the surveys will be carried out biannually.
        In addition to the mentioned two projects the airborne measuring system will be mobilized in case of accidents with radioactive material or from debris of nuclear-powered satellites. The emergency measurements will be done by military helicopters under the control of the Swiss National Emergency Operation Center (NAZ).
        The methodology developments (Schwarz, 1991) as well as the measurements were carried out by the Institute of Geophysics of the Swiss Federal Institute of Technology (Zürich).

Data acquisition

The measuring system consists of a helicopterborne gamma ray spectrometer with data control and storage and flight positioning instrumentation. The system is to its main part self-developed. Figure 1 shows the block diagram of the complete system. It is based on a modern airborne spectrometer designed for uranium exploration.

Figure 1: Block diagram of measuring system

The spectrometer covers the gamma ray energy range of 40keV to 3000keV with 256channels. An additional channel is used for the registration of high energy cosmic radiation. The detector used for the survey consists of a package of four 4"x4"x16" prismatic, thallium-doped, sodium iodide crystals (total volume 16.8 liters). Each crystal is equipped with its own photomultiplier tube (PMT). The whole package is heat and shock isolated and includes the high voltage power supplies and controls for the PMT's.
        The main advance of the spectrometer is the automatic gain control. During operation the system monitors each of the crystals and a separate spectra is accumulated in memory for each crystal. Based on the ⁴⁰K peak the gain of each spectrum are aligned and afterwards stacked together.
        A PC-based data acquisition system equipped with rack keyboard and 9"-monitor synchronizes and controls the measurements. The spectrometer data are collected every second together with radar altitude, time, barometric pressure, outside air temperature and aircraft attitude. During operation the system is controlled by a remote control console. The measured data are stored on JEIDA memory cards with a storage capacity of 2 Mb. This new storage medium in credit card size works without moving mechanical parts and is therefore insensitive against vibrations, humidity and temperature changes.

Figure 2: Flight path in the area KKB/PSI of the 1991 survey derived from GPS-data.

Positioning is done with the satellite navigation system GPS with a precision of ±25m. The positions are displayed to the operator for navigation and stored together with the spectrum data. If no satellite signal is available a vertically mounted camera is used for flight path recovery. An example of the positioning can be seen in figure 2.
        At a base station barometric pressure, temperature and humidity are recorded simultaneously with the flights. The quality check takes place directly after landing on a personal computer.
        To have uniform coverage of the surveyed area the flights are carried out in a regular grid. Figure 2 shows the flight path in the area around the nuclear power plant Beznau (KKB) and the Paul Scherrer Institute (PSI). The spacing between flight lines is 250m, flight height is 90m above ground. Along the flight lines a measurement is taken every second, which corresponds to a distance of 25m between two measurements at a flight speed of 90km/h.

Processing of raw data

Since NaI-detectors have a relatively poor energy resolution the data processing is based on so called energy windows in which the registered counts are integrated. The energy windows are centered on spectral regions of special interest. The energy windows have to be sufficiently separated from each other to keep interactions small.

Figure 3: Spectrum of the g-radiation in the Magadino plain (TI) as measured in july 1991.

The region centered at the 1460keV peak of ⁴⁰K is used for the determination of the potassium content. Uranium is detected using the 1765keV line of the daughter product ²¹⁴Bi. For thorium the 2615keV line of ²⁰⁸Tl is used. The energy windows are located at higher gamma ray energies where the absorption by air is less important. The artificial isotopes ¹³⁷Cs and the ⁶⁰Co are determined using energy windows centered at 660keV and 1250keV. Figure 3 shows a spectrum measured in the Magadino plain (TI). The peak at 660keV which can be identified in the spectrum is mainly caused by release of the Chernobyl accident.
        Three additional windows cover larger parts of the spectrum. The total count window covers the complete spectrum and is representative for the total amount of gamma radiation. The low (below 1400keV) and the energy part of the spectrum are registered in separate windows as an indicator for artificial radiation (see below).
        The gamma radiation registered by the detector in a helicopter is composed of the contributions from soil, atmosphere, aircraft and cosmic radiation (see figure 4):

Figure 4: Situation encountered during airborne radiometric measurements

• Cosmic radiation: During the absorption process of the cosmic high energy particle radiation gamma radiation is produced, which is registered by the spectrometer. The intensity of the cosmic radiation depends mainly on the thickness of the atmosphere above the measurement location and therefore on air pressure and altitude above sea level.

• Atmospheric radiation: The most important radiation source in the atmosphere is the isotope ²²²Rn of the inert gas radon. It originates from the decay of uranium. The concentration of radon in the air depends on meteorological conditions and is strongly variable (Kogan et al., 1969).

• Background radiation from aircraft and detector: In the aircraft as well as in the detector material itself small amounts of radioactive material are present. The helicopter background is assumed to be constant.

• Terrestrial radiation: Terrestrial radiation is mostly produced by the decay of the four natural radio isotopes ²³⁵U, ²³⁸U, ²³²Th and ⁴⁰K. Sometimes contributions of long living decay- and activation-products from nuclear weapon tests and/or nuclear facilities as ¹³⁷Cs and ⁶⁰Co add to the natural contributions.

The aim of airborne radiometric measurements is the determination of the radionuclide content of ground using the information of the direct terrestrial gamma radiation. All other contributions are perturbing and have to be removed.
        These contributions are determined by flights above large lakes, where the ground radiation is completely absorbed by water. Whereas the helicopter background and atmospheric radiation is assumed constant, the contribution of the cosmic radiation increases with increasing height. Therefore the calibration flights are carried out at different altitudes above sea level.
        The gamma ray spectrum received during field measurements is very complex. It is composed of the contributions of several radio isotopes. The spectrum is further complicated by scattering and absorption, which can occur in the ground, in the air and in the detector itself. To reduce the radiation intensities measured from the air to the ground activity a second set of corrections is needed. These corrections are called normalization corrections and are carried out applying the following steps:

• Spectral stripping: Gamma rays are scattered in the detector crystals, in the air and also in the soil. The scattering changes the energy of the radiation and therefore the original spectrum. The spectral stripping reduces the data to the primary (direct) radiation contribution by removing the effects of scattering and contributions from other isotopes.

• Altitude correction: The effects of a varying flight altitude are removed with the altitude correction. These variations represent the most significant perturbation to airborne radiometric measurements. The gamma ray intensity decreases rapidly with increasing ground clearance because of absorption of gamma rays in the air.

• Topographical correction: The effects of topographic relief can be neglected in flat or slightly hilly areas. In narrow valleys, which are abundant in the Swiss Alps, topography can produce a strong increase of the count rates by as much as 100%. These effects can be corrected adequately in most practical circumstances with the topographical correction (Schwarz et al. 1992).

• Conversion to general units: In a last step the detector-specific count rates are converted to the corresponding soil activity or radionuclide concentrations. For this conversion the detector sensitivity is needed.

Contrary to the background corrections, the normalization corrections need assumptions on the distribution of the radioactive source(s). The topographical and the altitude corrections as well as the conversion to general units can only be carried out when the activity distribution is known. Normally the natural radioisotopes are homogeneously distributed in the ground. Their activity distribution can therefore be considered as known, which is not the case regarding artificial isotopes. This is the reason why the normalization corrections with exception of the spectral stripping and the altitude correction generally apply only to the natural isotopes.
        The methods developed for processing and correction allow a routine processing of airborne radiometric data acquired even in areas with high topographic relief. The complete processing software including that for corrections and map outputs has been implemented on a transportable microcomputer. This enables the acquired raw data to be processed directly in the field, a facility of great importance in the event of a radiological accident. The programs contain a total of 32'000 lines of source code (Schwarz, 1991).

Calibration

The correction factors for the normalization corrections, the spectral stripping factors, the attenuation coefficient of air and the detector sensitivity are determined ideally on concrete calibration pads with a precisely known radioisotope content (Purvance and Novak, 1983). Unfortunately there are no calibration pads available in central Europe. The factors had to be therefore determined using radioactive point sources and to be mathematically corrected for the different source geometry (Grasty, 1975; Schwarz, 1991).

Figure 5: Experimentally determined detector sensitivities

The results of this calibration procedure where controlled by high precision in-situ gamma spectrometric measurements (Leupin, 1990 and Murith et al., 1990) together with gammaspectrometric measurements on rock samples (Kissling, 1976; Labhart and Rybach, 1971). The corrected count rates measured with the airborne system are plotted against the determinations of ground activity in figure 5. As can be seen from figure 5 the data show a reasonably good correlation between the airborne and the ground measurements.

Table 1: Summary of the detector sensitivity determinations (100m above ground)

The ratio countrate/groundactivity allows to determine the detector sensitivity for the specific flight altitude (100m in this case) for each energy window. As can be seen from table 1 the experimentally found values are in good agreement with the values derived from point sources.

Map generation

Radiometric measurements always show a large statistical scattering. Additionally the corrections applied to a measurement contribute essentially to the total error. The total error of a corrected mesurement can be approximated by:

Where DI_corr is the error of the corrected measurement, I_corr is the corrected count rate and S½I_CT½ is the sum of the corrections applied. The error of single measurements typically ranges between 10% for the total-count-window and up to 50% in windows with low count rates like the cesium window.

Figure 6: Grid and flight path

Because of the large error of airborne radiometric measurements the classical data representation using isolines is not very well suited in this case. We decided to use the pixel representation instead. To reduce the error only an average of several single measurements is displayed on the radiometric maps. Usually a pixel size of 125x125m is used for the representation. Since the distance between two measurements is 25m a pixel represents the mean of five measurements. The pixel size of 125x125m corresponds also roughly to the field of view of the detector at a flight altitude of 90m (Pitkin and Duval, 1980).
        Since the line spacing is 250m, only half of the pixels will contain data without interpolation (see figure 6). Since the gamma ray field contains abrupt steps and the statistical error of airborne radiometric data is very large, the interpolation cannot be performed with standard methods. A simple method proposed by Green (1987) was therefore chosen for the interpolation.
        The remaining pixels are filled with the average of the surrounding neighbors if they have at least three already assigned neighbors. This procedure is repeated until the grid is completely filled. It has to be mentioned that the original values are not changed by this interpolation procedure.

Figure 7: Total count map of the KKB/PSI survey

Figure 7 shows the total activity map of the surroundings of the nuclear power plant Beznau (KKB) and of the Paul Scherrer Institut (PSI). The total intensity of g-radiation as measured from the helicopter is plotted in detector specific counts per second. The highest intensities (the two black spots) are measured over the two sites of the PSI. They are caused by the direct radiation of the storage areas for radioactive components (PSI-West) respectively for radioactive waste (PSI-East). The water of the river Aare strongly absorbs the ground radiation, which is why the course of the river stands out in white. Even differences in vegetation cover are visible; meadows (light grey) show generally higher values than forests (dark grey).
        Only the error of the count rate has been discussed so far. When converting the airborne count rates to ground activities additional perturbing effects have to be considered. For a constant ground activity a lower count rate is measured over forests than over meadows due to the additional absorption by trees. This attenuation varies between 5% and 25% depending on the biomass of the forest. Soil humidity and rain also influence the measurements up to 10%. This second effect is especially troublesome because the soil humidity varies with time.
        Both effects can distinctly reduced with the aid of ratios. The ratio of two windows is less affected by an additional absorber than the single window values, since the ratio is only affected by the difference of the attenuation coefficients of the two windows. Especially sensitive to artificial radiation is the so called Man Made Gross Count (V_MMGC) ratio (Hoover, 1988).

It uses the fact that the common artificial radio nuclides all radiate at gamma energies below 1400keV, the natural isotopes on the other hand emit gamma quanta of higher energy too. The ratio of the low energy part of the spectrum (MMGC1) to its high energy part (MMGC2) corresponds therefore to the ratio of artificial to natural radioactivity. An example of MMGC ratio map is shown on figure 8.
            Because of the strong correlation between the radioelements due to geochemical reasons, the maps of the natural radioisotopes are all very similar. Working with ratios of natural radioisotopes will therefore not only reduce disturbing effects, but also reduce the contrast of the image. A very good method to increase the contrast and with it the interpretability of an image is called histogram equalization (Haberäcker, 1985). After histogram equalization every color covers the same area on the map. The visual impression of such a representation is optimal. As a consequence the color scale becomes nonlinear.
            The histogram equalization is particularly useful producing ternary maps. In this representation the maps of potassium, uranium and thorium are merged into a single map. The color red is assigned to the potassium window. Uranium and thorium are colored green and blue respectively. Regions with a high relative potassium content will appear in red color shades on the map. Correspondingly areas with a high relative uranium or thorium content will appear in green or blue shades. Regions with equally balanced radioisotopes will be plotted as gray or white shades depending on the total activity. Ternary maps can almost be read like a geological maps, especially in regions with crystalline rocks (Schwarz et al., 1992).

Concluding remarks

The methods developed for processing and correction allow a routine treatment of airborne radiometric data acquired even in areas with high topographic relief. The results for the areas processed so far show a very good fit with geology and can be used for geological overview mapping.

Figure 8: 3D-representation of the MMGC-ratio of the region of Würenlingen/AG (looking from north west). The highest peak corresponds to a ground dose rate of 2mSv/a.

The airborne radiometric survey conducted so far covers about ten percent of the area of Switzerland. The results of the survey allow to describe the mean radioactivity level in Switzerland and its variation in broad outline (Schwarz et al., 1992). Together with in-situ ground measurements and data from rock samples the airborne survey can serve as a good base for the compilation of a radiometric (dose rate) map of whole Switzerland including the contributions of artificial isotopes and cosmic radiation.
        The measurements in the environs of the Swiss nuclear installations showed that all sites (with the exception of the Gösgen power plant) can be identified clearly on the MMGC-ratio maps. At sites of operating boiling water reactors the high energy radiation of the activation product ¹⁶N is clearly visible in the data.
        Figure 8 shows a three dimensional representation of the MMGC-ratio of the surroundings of Würenlingen/AG (looking from north east). The two locations of the Paul Scherrer Institute PSI-Ost and PSI-West show up particularly well. The signal is caused by the direct radiation of the storage areas for radioactive components (PSI-West) and radioactive waste (PSI-Ost). The smaller peak is caused by the Beznau nuclear power plant. The highest peak corresponds to a ground dose rate of 2mSv/a.
        No artificial radioactivity, that could not be explained by Chernobyl or nuclear weapon test, was detected outside of the fenced sites of the nuclear installations. The repeated measurements in the last four years (Schwarz et al., 1989, 1990, 1991, 1992) showed that the radioactivity level in the environs remained more or less constant within the measurement errors.

References

GRASTY, R.L., 1975: Uranium measurements by airborne gamma-ray spectrometry. Geophysics 40, 503-519.

GREEN, A.A., 1987: Leveling airborne gamma-radiation data using between-channel correlation information. Geophysics, 52, 1557-1562.

HABERAECKER, P., 1985: Digitale Bildverarbeitung, Grundlagen und Anwendungen. Hanser Studienbücher, München/Wien.

KISSLING, E., 1976: Radiometrische Untersuchungen im Rotondogranit, Diplomarbeit, Institut für Geophysik ETHZ, Zürich.

LABHART, T.P., RYBACH, L., 1971: Abundance and distribution of uranium and thorium in the syenite of Piz Giuv (Aar-Massif, Switzerland). Chemical Geology 7, 237-251.

LEUPIN, A., 1990: Vergleich aeroradiometrischer Messungen in der Umgebung von KKB/PSI mit In-Situ-Gamma-Spektrometrie-Messungen am Boden. Interner HSK-Bericht. Bundesamt für Energiewirtschaft.

MURITH, C., VOELKLE, H.R., SURBECK, H., PILLER, G., BAERISWIL, L., BEURET, P., FERRERI, G., GOBET, M., GURTNER, A., RIBODRY, L., 1990: Mesures in situ - mesures a‚roradiom‚trie dans le voisinage du PSI. Interner SUeR-Bericht. Bundesamt für Gesundheitswesen; Abteilung Strahlenschutz.

PITKIN, J.A., DUVAL, J.S., 1980: Design parameters for aerial gamma-ray surveys. Geophysics 45, 1427-1439.

PURVANCE, D., NOVAK, E., 1983: General procedure for calibration and reduction of aerial gamma-ray measurements: Specification BFEC 1250-B. U.S.Department of Energy, Rep. GJBX-16(83) DE84004974.

SCHWARZ, G.F., 1991: Methodische Entwicklungen zur Aerogammaspektrometrie. Beiträge zur Geologie der Schweiz, Geophysik Nr.23, Schweizerische Geophysikalische Kommission.

SCHWARZ, G.F., KLINGELE, E.K., RYBACH, L., 1989, 1990, 1991 and 1992: Aeroradiometrische Messungen in der Umgebung der schweizerischen Kernanlagen; Bericht zuhanden der Hauptabteilung für die Sicherheit der Kernanlagen. Interner Bericht, Institut für Geophysik ETHZ, Zürich.

SCHWARZ, G.F., KLINGELE, E.K., RYBACH, L., 1992: How to handle rugged topograpohy in airborne gamma-ray spectrometry surveys, First break 10, 11-17.

SCHWARZ, G.F., KLINGELE, E.K., RYBACH, L., 1992: Airborne radiometric mapping in Switzerland. Beiträge zur Geologie der Schweiz, Geophysik Nr.25, Schweizerische Geophysikalische Kommission.