Estimating Truck Gate Demand Under Data Scarcity: A Provenance Audit and Profile-Based Disaggregation Framework for Shuwaikh Port, Kuwait

Abstract

Background: Authority-supplied daily port figures may be allocations of coarser records rather than measurements. Treating them as observed gate counts can produce spurious model accuracy. Methods: Using 365 daily container-discharge records for 2019 from the Kuwait Ports Authority (KPA), we apply four provenance diagnostics and a stochastic-count benchmark, formalize the allocation chain, distinguish discharge from pickup through a release-lag kernel, and explore assumed uncertainties by Monte Carlo simulation. Results: The daily series has a strong allocation signature (median within-month ratio coefficient of variation 2.1%; 141 of 308 operating days share repeated pairs). Under an assumed uniform pairing share, mean-preserving day-level noise, same-day pickup, and allocation of all modeled visits to six hourly slots from 11:00 to 17:00, peak-hour truck-visit medians are about 143, 185, and 259 for the three design levels. The maximum-day central 90% simulation interval reaches about 354 visits. Conclusions: The contribution is a provenance audit and conditional scenario framework for the import-discharge component. The intervals depend on uncalibrated assumptions and have no demonstrated coverage of observed gate traffic. Classified gate counts, container-to-truck linkage, and broader expert elicitation are required before operational use.