Results

Short term and long-term streamflow observations

Discharge time series for the Gōno River observed at the studied stations are shown in Fig. 3. The Gōno River has a flashy regime, influenced considerably by the intensity of precipitation. Apparently, it can be realized that the discharge records reveal some years with low amount of flow rates. For example, 2002 to 2003, 2007 to 2009, and from 2015 to 2017. On the other hand, it can be distinguished that there is an upward trend in the maximum streamflow records induced by the large amount of precipitated rains. In other words, it is obvious that the maximum streamflow in 2006 peaked at 3300 m3/s approximately, and the maximum peaks observed at 2010 and 2014 were roughly 3500 and 3600 m3/s, respectively, Though it was not reported by the MLIT yet, in 2018 heavy rains documented the greatest flood that occurred in Hiroshima prefecture.
Figure 3 is here.
Fig. 3 Discharge time series during the study period (2002-2017) observed at Awaya (red), Minamihatachiki (green), Miyoshi (blue), and Ozekiyama (Black) stations.
Prior to information and complexity results, it is essential to briefly comment on the discharge measurements by FAT (QFAT ) compared to the RC (QRC ) estimates at Ozekiyama observation site. Demonstrated in Fig. 4, a comparison between QFATand QRC , it can be seen that both discharge methods show very good agreement. Furthermore, it can be noticed that FAT can capture the fluctuations that take place during very short times scales, this feature in particular, inspired us to profoundly explore this difference. Kawanisi et al., (2018, 2016) investigated the accuracy of river discharge of QFAT compared to other discharge computation methods (e.g. RC and ADCP) measured at Ozekiyama station and examined the error structure that impair the performance of FAT estimates, it was demonstrated that in the case of streamflow measurements by means of FAT at the Gōno River, the maximum potential error within low-flow circumstances is estimated as 15%.