Boundary-consistent prospective appraisal of an unopened metro line: mobile-OD evidence from Shenzhen

Abstract

Prospective metro appraisal requires evidence that supports both comparisons between alternatives and the scope of carbon claims. Using 4.82 million mobile OD records representing 159.1 million provider-inferred trips, we assess unopened Shenzhen Metro Line 17 Phase I. Seven-mode associations with 800-m walking-network access at both trip endpoints are transferred under fixed observed demand to line addition alone and an optimistic permeability-screening scenario with reduced walking detours, not an implementable works package. Both-end access is associated with 9.21 percentage points more metro use but 4.77 points less car use. The screening scenario yields a larger annual partial operating-carbon reduction in 99.46% of tested parameter settings. Prediction improves in only one of two withheld-line tests. Comparing unexpanded-sample avoided-car emissions with full service-plan traction emissions diagnoses limited evidence coverage. It does not establish the project’s net carbon balance. The case shows that robust scenario ranking and credible claims about carbon-benefit magnitude and scope require different evidence. Fixed-demand comparisons should therefore be accompanied by tests of behavioral transferability and explicit documentation of population coverage and inventory scope.