Data assimilation for urban noise mapping with a meta-model

Accurately predicting dynamic noise levels in urban environments is non-trivial. This study aims to optimally combine both simulated and empirical data. Acoustic data from microphone arrays, traffic weather was merged with a map, created statistical emulator tool (meta-model). Each hour, map generated by the meta-model measured algorithmically readings form new composite map. The resulting analyzed best linear unbiased estimator under certain assumptions. performance evaluated leave-one-out cross-validation. of method depends on accuracy meta-model, input parameters structure error covariances between level errors. With 16 microphones over an area 3 km2, this achieves reduction 30% root-mean-square when compared only.