Testing a proposed transcontinental great circle alignment: Land-constrained Monte Carlo baselines, geographic explanations, and localized monument divergence at the Memphis necropolis

v7: Substantially revised for Journal of Computer Applications in Archaeology (JCAA) following desk rejection from AAS (Springer) on scope. Abstract and introduction reframed around land-constrained Monte Carlo methodology as primary contribution. New analyses: Mediterranean coastal benchmark (demonstrating bidirectional bias correction), jitter bandwidth sensitivity (σ = 1°, 2°, 5°), leave-Memphis-out analysis (D = 5.21), classification audit (87% strict / 99.3% lenient agreement across 150 sites), Monte Carlo diagnostic histograms confirming normality (Shapiro-Wilk p > 0.12 all databases). D explicitly defined as descriptive contrast statistic. Multiplicity table, pseudocode appendix, and MC diagnostics added as supplementary materials. Post-hoc mitigation strategy formalized in Methods. Auxiliary falsification sections compressed. Full citation audit: 33 references, 27 DOI-verified, 4 orphaned references removed. No statistical results changed from v6. Earlier versions: v6 reformatted for AAS. v5 submitted to JAS:Reports. v1–v4 tested the Alison great circle alignment with progressively refined baselines. All code and data: github.com/thegreatcircledata/great-circle-analysis