Abstract Insufficient low-frequency data can pose a problem when inverting fundamental-mode Rayleigh-wave phase velocities using two-channel (SASW) or multiple-channel (MASW) active-source methods. Depth of penetration and accuracy of the inverted models are particularly sensitive to the low frequencies. Ambiguities might be reduced with supplemental geologic or geophysical information about the near surface, such as passive-source seismic data and well information. Dispersion data from higher modes might complement the fundamental mode in resolving shear-wave velocities in some cases because higher modes are more sensitive to shear-wave velocities at greater depths. However, interpretation of higher-mode phase velocities poses a challenge because the higher modes are not always easy to separate from the fundamental mode and possibly from other scattered energy. The energy partition into fundamental and higher modes results in different frequency-dependent signal-to-noise ratios for the different modes. Furthermore, convergence of the solution’s data error to the hypothetical global minimum appears to be more complex when fundamental and higher modes are included in the data error to be minimized. Some of the difficulties encountered when inverting higher-mode surface waves are investigated for some simplified shallow (< 10-m) earth models of engineering significance: a sediment profile with a constant stiffness gradient, a low-velocity sediment overlying shallow bedrock, and a high-velocity layer overlying and underlying softer sediment layers.