Data from the passive teleseismic experiment TOR across the Trans-European Suture Zone (TESZ), which took place in 1996–1997, are used to study the random structure of the lithosphere NE of the TESZ (the Baltic Shield beneath southern Sweden) and beneath the north German/Danish Basin. P coda waves from deep teleseismic events give information on the small-scale heterogeneity of the near-receiver lithosphere on scales of one wavelength. Using an energy flux model, the time and frequency behaviour of the P coda has been interpreted in terms of scattering (Qs) and anelastic attenuation (Qi). Estimates of Qs were obtained for the frequency range 0.5–Hz for the north German/Danish Basin and 0.5–7 Hz for the Baltic Shield. The two areas show a different scattering behaviour. In Denmark/northern Germany the scattering (Qs ≈ 125ߝ200) is stronger than in southern Sweden (Qs ≈ 275–500) and the peak of Qs−1 occurs at higher frequencies NE of the TESZ. Anelastic attenuation is negligible for the Baltic Shield whereas for the north German/Danish Basin anelastic attenuation is present but still much weaker than scattering attenuation.

With a modified energy flux model for depth-dependent scattering, and assuming an exponential autocorrelation function, the depth variation of correlation length a and rms velocity fluctuation ε can in principle be resolved. For the Baltic Shield the scattering is mostly confined to the crust with a ≈ 1 km and ε 4 per cent. The subcrustal lithosphere only shows weak fluctuations. The data from the north German/Danish Basin cannot be explained by strong scattering within the crust only. Crustal parameters are a = 5–10 km and ε ≈ 8 per cent. For the subcrustal lithosphere, correlation lengths of 10–20 km and rms velocity fluctuations of 6–8 per cent are found. Additional tests confirmed that the thick sedimentary cover in this area has no significant effect on these results for the deeper structure. Correlation lengths for the sediments are smaller than 5 km and rms velocity fluctuations ε are 7–8 per cent.