Modelling gamma spectrometry systems for use in beach monitoring near Sellafield

The response of large volume Airborne Gamma Spectrometry (AGS) detectors has been modelled using Monte Carlo methods as part of an assessment of ways to enhance monitoring of beaches and other coastal areas for radioactive particles. Airborne survey methods are capable of rapid surveys of large areas, and of covering diverse environments effectively, including areas where ground based work is difficult to implement safely. They have potential for increasing the effectiveness of ground based surveys, working in a complementary manner to identify areas where detailed work including recovery operations would be targeted. This study examined the theoretical sensitivities that might be obtained using airborne systems operated at 75 m (200 ft) ground clearance and a 70 knot (30 m s-1) survey speed, and also a “slow and low” design based on 15 m (50 ft) ground clearance and 15 knot (5 ms-1) forward velocity, or slower. Monte-Carlo methods were used to simulate the full spectral responses for a series of 137Cs and 241Am point sources placed in simulated natural background environments. The MCII Monte Carlo code, which has been extensively validated in the past and incorporates statistical estimation procedures to simulate spectra at airborne source-detector separations, was used. This code was set up for automatic sampling of the complex gamma ray spectra associated with the natural uranium and thorium decay series, with energy- intensity files collated from the JEFF 3.1.1 nuclear data for all lines from the natural decay series and 40K, without truncating low intensities. Thus a comprehensive description of the photon intensities of natural sources was achieved. The code was re-validated in this study using simulated and measured data for concrete calibration pads at SUERC and shown to be in excellent agreement with measured spectra in close coupled geometries. A brief assessment of the GEANT Monte Carlo package indicated that it was not ideally suited to the large source-detector geometries of airborne measurements. The geology and geochemistry of the Cumbrian coastline was reviewed, and data used to define mass absorption coefficients for substrates with a composition similar to St Bees Sandstone. For energies above 200keV, the mass absorption coefficients are insensitive to the range of major element compositions of different materials in the Sellafield area. These coefficients were incorporated in the Monte Carlo code, and used to simulate reference spectra for unit concentrations of 40K, Uranium and Thorium series sources at the two airborne survey ground clearances defined for the study. Uniform vertical and lateral activity distributions were used for these natural background reference spectra, representing a reasonable starting point for the work. These were combined and scaled to the natural radioelement concentrations determined previously from an area of beach south of St. Bees Head, and north of Sellafield, which had been surveyed in 2000 as part of a larger study commissioned by DEFRA. The agreement with the measured spectra from the earlier survey in the energy range associated with natural gamma rays was excellent, thus validating the simulation at airborne heights. Data from past surveys and other ground based measurements were also reviewed to assess the levels and variations of natural background sources in the coastal areas of West Cumbria. These were used to define three working levels (“Low”, “Medium” and “High”) of natural background concentrations, whose AGS contributions were simulated at the two airborne survey ground clearances, to form the natural background levels against which detection limits for point sources could be considered in the study. Additional simulations of the system responses to 137Cs sources in superficial locations, and at a burial depth of 10 cm, and of superficial 241Am sources were conducted for the two reference heights, and a series of source-detector offset distances. These were used to assess detection limits and signal significance levels for different survey scenarios, based on full energy peak count rates for the source spectra and the simulated backgrounds. The signal profiles corresponding to flight lines directly over and adjacent to the sources were also simulated for the first time. The results have shown that AGS surveys at 75m (200ft) ground clearance and 70 knots (30 m s-1), capable of surveying > 60km of shore in a 2h survey with multiple lines, should detect 137Cs sources of 5-10 MBq or greater in the presence of uniform natural radiation fields representing typical compositions for the area. More detailed surveys at 15m (50ft) ground clearance and 5 knots (2 m s-1) should detect 137Cs sources of 105 Bq at a depth of 10cm. Surveys at these reduced heights and speeds should also detect superficial 241Am sources of 106 Bq or greater. A high density survey, in which the source is registered in more than one measurement, would increase the significance of the summed signal by about 0.5σ. Reducing ground clearance would increase full energy peak count rates, by an order of magnitude at 5m compared to 15m. It would be expected that surveys at these ground clearances would significantly increase the significance of full-energy peak counts from sources, although the natural series spectra have not been simulated to quantify this. A data processing methodology that utilises the scattered part of the spectra (>80% of the total count rate from buried sources) might potentially further improve detection efficiency, and provide an estimate of source burial depth. For areas with enhanced background 137Cs activity, these detection limits are expected to be increased. So far the influence of 137Cs contamination in the environment has not been analysed to assess its impact on detection limits. A preliminary assessment confirms that this could have a significant impact on detection probabilities for 105 Bq 137Cs sources. Further work is thus required to review to complex distributions of background 137Cs in the coastal environments and their impacts limits of detection for radioactive particles. Additional work could also be contemplated to model surface roughness effects, which may also be significant, especially for low energy emitters such as 241Am, and to further explore the systematics of survey specification and detection efficiency for search and recovery operations, using the simulation methods developed here. This work has confirmed that airborne methods could in theory produced useful levels of sensitivity coupled with survey rates which could not be practically achieved at ground level. By demonstrating the absence of strong sources within large areas and identifying other locations where valuable ground-based resources should be focussed AGS has an important potential role. The use of HPGe detectors has not been considered, but the greater spectral resolution of these could improve detection efficiency, especially for 241Am. It is suggested that practical work be planned to verify the simulated detection limits, and to incorporate the method into programmes for assuring the environmental quality of coastal zones in the vicinity of the nuclear sites of the Irish Sea