10:15 AM - 10:30 AM
[SSS09-06] On the relationship between the streamflow due to snowmelt flood and ambient seismic noise observed by Hi-net stations
Keywords:Snowmelt flood, Seismic ambient noise, Streamflow
Seki River basin is located in Niigata and Nagano prefecture of Japan. Snowfall is common over the basin during the winter season. Such accumulated snow starts melting from the beginning of the spring season each year that increase the rate of streamflow in the Seki River. The variation of streamflow is synchronized by the ambient noise level of nearby Hi-net seismograph stations. In this study, we surveyed the relationship between the ambient noise power of Hi-net stations located in the basin and discharge of the Seki River, in the period from March to May in 2012.
The N.MKOH and N.MKGH Hi-net stations are located in the middle basin (elevation: 240 m asl) and upstream (elevation: 1305 m asl) of the Seki River, respectively. Ambient noise of N.MKOH station shows a clear one-day periodicity associated with social activity, whereas N.MKGH station contains much less social noise as it is located deep in the mountains. The Futagojima streamflow observation point is located about 10 km downstream from N.MKOH station (elevation: 65 m). We computed power spectra of UD component of each Hi-net station every one minute, and calculated power time series for six octave-bandwidths from 0.5 to 32 Hz. To remove transient signals like earthquakes, we selected the minimum one-minute power among every continuous 10 minutes records as the representative value of the 10-minutes ambient noise power. Then we averaged the obtained representative values every one hour. The discharge record of the Futagojima station is also measured every one hour.
Since noise power of N.MKOH exhibits one-day periodic social noise, we used only the records observed from 8PM to 12PM, within which social noise is relatively calm, for the comparison to the discharge record. As a result, clear correlations with the discharge record are observed especially at 2 - 16 Hz. More precisely, the noise power does not depend on the discharge when the discharge is less than a threshold value. It increases proportional to about the first power of the discharge above the threshold value, and proportional to 1 to 2 power as the discharge increased further. Through the target period, the ratio of the noise power with respect to the discharge did not change drastically.
On the other hand, N.MKGH does not show clear correlations with the discharge in March, but the correlations become distinctive after April. The ratio of noise power with respect to the streamflow increases as time lapses after April. According to the nearby temperature and snow depth records of AMeDAS, snowmelt may not have proceeded until April. This indicates that streamflow at upstream is very small in March and it gradually increased from April to May. Consequently, noise at N.MKGH station is excited only after April and this may be why correlations with the dicharge of Futagojima station become distinctive only after April.
The N.MKOH and N.MKGH Hi-net stations are located in the middle basin (elevation: 240 m asl) and upstream (elevation: 1305 m asl) of the Seki River, respectively. Ambient noise of N.MKOH station shows a clear one-day periodicity associated with social activity, whereas N.MKGH station contains much less social noise as it is located deep in the mountains. The Futagojima streamflow observation point is located about 10 km downstream from N.MKOH station (elevation: 65 m). We computed power spectra of UD component of each Hi-net station every one minute, and calculated power time series for six octave-bandwidths from 0.5 to 32 Hz. To remove transient signals like earthquakes, we selected the minimum one-minute power among every continuous 10 minutes records as the representative value of the 10-minutes ambient noise power. Then we averaged the obtained representative values every one hour. The discharge record of the Futagojima station is also measured every one hour.
Since noise power of N.MKOH exhibits one-day periodic social noise, we used only the records observed from 8PM to 12PM, within which social noise is relatively calm, for the comparison to the discharge record. As a result, clear correlations with the discharge record are observed especially at 2 - 16 Hz. More precisely, the noise power does not depend on the discharge when the discharge is less than a threshold value. It increases proportional to about the first power of the discharge above the threshold value, and proportional to 1 to 2 power as the discharge increased further. Through the target period, the ratio of the noise power with respect to the discharge did not change drastically.
On the other hand, N.MKGH does not show clear correlations with the discharge in March, but the correlations become distinctive after April. The ratio of noise power with respect to the streamflow increases as time lapses after April. According to the nearby temperature and snow depth records of AMeDAS, snowmelt may not have proceeded until April. This indicates that streamflow at upstream is very small in March and it gradually increased from April to May. Consequently, noise at N.MKGH station is excited only after April and this may be why correlations with the dicharge of Futagojima station become distinctive only after April.