Statistical Characterisation of Cosmic Ray Neutron Count Rate Fluctuations at a Fixed CRNS Station

Project Description: 

Cosmic Ray Neutron Sensing (CRNS) infers soil moisture content from the intensity of epithermal neutrons measured at the land surface. These neutrons originate from cascades initiated by primary cosmic ray particles entering the upper atmosphere, and their flux at ground level is modulated by the hydrogen content of the surrounding environment. Because the cosmic ray source itself is subject to temporal variability — driven by solar activity, geomagnetic conditions, and atmospheric pressure and humidity — the raw neutron count rate measured by a CRNS station reflects a superposition of the environmental hydrogen signal of interest and multiple external forcing factors. Disentangling these contributions requires a rigorous understanding of the statistical properties of the neutron count rate time series, including the nature of its random fluctuations and how these vary with measurement averaging interval. Radioactive decay and the stochastic arrival of cosmic ray secondaries are fundamentally random processes, and the count rates produced by nuclear detectors are expected to follow Poisson statistics under idealised conditions. In the Poisson framework, the variance in the number of detected events equals the mean count rate, and the relative uncertainty in a measurement decreases as the square root of the total counts accumulated. For CRNS applications, this has direct practical implications: longer counting intervals reduce statistical noise but sacrifice temporal resolution, while shorter intervals preserve the ability to track rapid changes in soil moisture at the cost of increased measurement uncertainty. Determining the optimal counting interval for a given site and application therefore requires empirical characterisation of the actual statistical behaviour of the signal, which may deviate from pure Poisson expectations due to detector noise, electronic dead time, atmospheric variability, and non-stationarity in the cosmic ray flux. Atmospheric correction procedures — primarily for air pressure, atmospheric water vapour, and incoming cosmic ray flux — are routinely applied to CRNS data to remove the dominant external signals before soil moisture retrieval. However, the residual statistical properties of the corrected signal, and whether they conform to the theoretical expectations for a Poisson process, have not been systematically examined for stations operating under southern African atmospheric and geomagnetic conditions. Such a characterisation is important both for establishing the achievable precision of soil moisture estimates and for identifying any systematic artefacts introduced by the correction procedures themselves. This project will analyse an extended time series of raw and corrected neutron count rate data from the group's fixed CRNS station, testing the conformity of count rate distributions to Poisson and Gaussian expectations across a range of averaging intervals. The student will apply standard atmospheric correction algorithms and evaluate the effect of each correction step on the residual statistical properties of the signal. Optimal averaging intervals will be determined based on quantitative signal-to-noise criteria, and the results will be translated into a practical data processing protocol for the group's CRNS network. The findings will directly improve the reliability of soil moisture products supplied to the group's hydrological modelling partners and contribute to best-practice guidance for CRNS operations in semi-arid southern African environments.(NWU)

Supervisor

North-West University (NWU)

Co-Supervisor