The problem of interim RMR (Reken en Meetvoorschrift Railkverkeerslawaai) methods adaptation for local railway conditions is analyzed. Cargo train noise levels were experimentally investigated on Latvian railway. The comparable analysis of measured and numerically calculated, using RMR method, cargo train noise levels in eight octave bands was performed. Comparable analysis results had shown that experimentally measured noise spectrum retains modeled spectrums shape, yet, having statistically constant level difference in all octave bands. The idea of RMR methods adaptation altering only one correction coefficient was proposed. More measurements are to be done to validate proposed approach.