Validation and Performance Analysis of a Nonlinearpressure-Coupled Urban Flood Model Using a Cross-Domain Co2 Plume Operator

This study presents a validation and performance assessment of a nonlinear pressure-coupled urban flood model built on a cross-domain CO2 plume operator. The first part of this work set out the theoretical formulation and a risk-classification framework; here we test how well the model actually performs, looking at predictive accuracy, numerical stability and hydrodynamic realism. The test data are observed flood depths from Nairobi City County, taken at Kibera, Mathare, Mukuru, Eastleigh and the Central Business District, and we judge performance by the Mean Absolute Percentage Error, by spatial pattern, and by how the flood evolves in time. Observed and simulated depths come out close, with a Mean Absolute Percentage Error of 2.07%, and the pressure coupling tightens that agreement: it reproduces flood persistence, delayed drainage and subsurface redistribution. The DuFort-Frankel scheme converges, which confirms the solver is stable. In short, pressure-driven hydrodynamics borrowed from CO2 plume physics materially improves urban flood prediction in complex settings