986 indicate that the corresponding parameter exerts a statistically significant influence on the response at the 95 percent confidence level. The output will be a ranked list of parameters identifying the dominant controls on predicted pressure evolution, which will both inform priorities for further site characterisation and identify which assumptions in the H1 model are most consequential for the planned falsification test. 3.5 Planned falsification test Once the sensitivity analysis is complete, the simulated pressure ensemble will be compared against observed piezometric data from the Mingomba site, following the falsification logic of the Popper-Bayes protocol introduced in Section 3.1. The reasoning is straightforward: if the observed data could plausibly have been produced by the model under the prior assumptions, the prior is consistent with reality and Bayesian inversion can proceed; if the observed data lies outside the range of responses the model can generate, the prior has been proven incorrect, and the modelling assumptions must be revised before any further inference is meaningful. Operationally, the test will compare the position of the observed data relative to the cloud of simulated responses. Because the simulated ensemble is high-dimensional, dimensionality reduction is required. Principal component analysis will be applied jointly to the simulated and observed data, projecting both onto the directions of greatest variability in the simulated ensemble. In this reduced space, the observed data is represented as a single point and the simulated realisations as a cloud of points; their relative position is then measurable. A statistical outlier test will be applied to determine whether the observed point lies within or outside the simulated cloud, using a distance metric (the robust Mahalanobis distance) that accounts for the shape and spread of the cloud rather than treating distance as a simple Euclidean measure. A distance value above the 95 percent confidence threshold will indicate that the observed data is statistically inconsistent with the simulated ensemble and that the H1 prior is therefore falsified. The outcome of the falsification test will determine the next step in the workflow. If H1 is not falsified, Approximate Bayesian Computation will be applied to generate posterior realisations conditioned on the observations, providing probabilistic forecasts of pressure evolution under the planned dewatering scenario with quantified uncertainty bounds. If H1 is falsified, the diagnostic information from the sensitivity analysis will be used to identify whether the prior parameter ranges are too narrow (in which case the prior is revised, the Monte Carlo ensemble regenerated, and the falsification test repeated) or whether the model is missing a physical mechanism, in particular delayed drainage from the rock matrix into the fracture system, requiring progression to hypothesis H2 with explicit dual porosity. 4. RESULTS The work reported here represents stages one through three of the Popper-Bayes protocol described in Section 3.1. The decision problem has been formulated, prior probability distributions have been specified for matrix permeability across twelve stratigraphic units, vertical anisotropy in the Bancroft Dolostone and Nchanga Member confining units, a fracture permeability multiplier, and three regional aquifer recharge boundaries (Tables 1 and 2). One hundred Monte Carlo realisations have been generated and forwardsimulated under the five-year dewatering scenario, producing a high-dimensional ensemble of predicted pore pressure responses across multiple monitoring locations and time steps. Visual inspection of the simulated ensemble confirms that predicted pressure responses span a wide range across the realisations, indicating that the prior uncertainty distributions encompass physically plausible outcomes for the Mingomba setting. This wide ensemble spread is itself a meaningful preliminary observation: it confirms that decisions about dewatering design at this site cannot be reliably made based on a single deterministic forecast, since the predicted response varies substantially with parameter combinations that are all consistent with current knowledge of the site. The next analytical stage is the application of DGSA over one thousand simulations to identify the dominant controlling parameters in this ensemble. The output of that stage will be a ranked list of parameters
RkJQdWJsaXNoZXIy MTM0Mzk2