Newly proposed missions to the Moon/Mars often require vehicles to operate at high speed for time-efficient surface explorations. Among wheel-terrain interaction models in the past being applicable for quasi-static locomotion, a granular Resistive Force Theory (RFT) is increasingly employed alternatively for its simplicity and applicability. The RFT has also been extended to the Dynamic RFT (DRFT), which considers the rate-dependent effects often seen in the high-speed locomotion. In this research, first, we validate the DRFT through single-wheel experiments. The results suggest that the original DRFT underestimates the influence of wheel rut and that the granular flow beneath the wheel may decrease the wheel traction force. We then propose an empirically modified DRFT (m-DRFT), which considers an appropriate evaluation of the wheel rut depth and a simplified model for the granular flow. The comparison with the experimental results confirms that the m-DRFT accurately predicts the wheel drawbar pull in relatively low slip conditions while the original DRFT overestimates it. Moreover, the m-DRFT reproduces the temporal oscillations of the drawbar pull generated by the wheel grousers. These results imply that the m-DRFT is applicable for transient locomotion state or even fast-rotating wheel where estimating the sequential changes of wheel forces becomes essential.