Abstract
Abstract
This study presents a rigorous least-squares adjustment and quality assessment of the Mzuzu University geodetic control network, comprising 36 stations, including 5 fixed and 31 unknown points. Using 630 simulated distance observations (σ = 0.02 m), the adjustment converged with an a-posteriori variance factor of 0.97, confirming appropriate stochastic specification. Mean horizontal and vertical precisions were 7.4 mm and 6.5 mm, respectively, while residual RMS was 0.0198 m. The global Chi-square test (p = 0.999) validated the stochastic model, and Baarda data snooping successfully identified injected outliers. Internal reliability analysis yielded a mean redundancy number of 0.61 and Minimal Detectable Bias of 0.067 m, indicating detection of errors exceeding 7 cm at 80% power. External reliability showed maximum coordinate shifts of ≤ 3.1 cm from undetected blunders. A mean PDOP of 2.08 confirmed strong network geometry. Precision mapping, correlation, and influence analyses indicated homogeneous uncertainty, weak parameter correlations (|ρ| < 0.3), and balanced observation leverage. Overall, the network demonstrates suitability for high-precision engineering applications, including deformation monitoring and infrastructure alignment. The integrated diagnostic framework provides a robust basis for geodetic network quality assurance, with field validation, free-network adjustment, and multi-epoch surveys recommended for operational verification.