Bridging the Gap in Tsunami Runup Observations under Rough Sea States: High-Frequency Automated Image Analysis during the 2025 Kamchatka Event

Abstract Following the 2025 Kamchatka Earthquake (Mw 8.8), tsunamis were observed along the Pacific coast of Japan, as well as in many other coastal regions. This study applies an automated image analysis method to video footage from the Hasaki Oceanographical Research Station to perform continuous high-frequency wave runup monitoring over a five-day daylight period spanning from the day preceding to three days following the tsunami onset. The analysis successfully extracted water-level fluctuations in the tsunami-period band (600–6000 s) from the runup time series. Error propagation analysis confirmed vertical uncertainties in the tsunami-period band within 0.05–0.1 m. Comparison with offshore wave gauge data also showed good agreement in phase and temporal variation, supporting the validity of the method. Results suggest that an approximate linear separation of the tsunami-band component and swash components was possible in the present case, since the time scale of the tsunami-band component was sufficiently longer than those of the short-period runup components. Importantly, a negative correlation between tsunami-band water-level fluctuations and swash heights was identified. This is consistent with the mechanism proposed in a recent study, where tsunami-induced water-level fluctuations alter the effective foreshore slope, thereby modulating swash dynamics. Correspondingly, 10-min time-averaged imagery systematically underestimated tsunami-band amplitudes, whereas our high-frequency approach enabled direct extraction of the tsunami-band component via band-pass filtering. These findings demonstrate that high-frequency, automated camera-based monitoring is an effective tool for bridging the observational gap in coastal tsunami dynamics.