3:00 PM - 3:15 PM
[ACG42-06] Storm runoff at hillslopes and channels in mountain catchment
Keywords:mountain river, peak propagation, velocity, celerity, hillslope, channel
It is important to clarify the dynamics of water and sediment in mountainous watersheds to understand the dynamics of water and sediment continuously from land to coastal zones. In Japan, about 70% of the land is mountainous, thus, water and sediment from mountain watershed transported to downstream plains and used for human activities, and the remaining will be brought to the coastal zones. Therefore, the amount of water and sediment transported from mountain watersheds is one of the most important conditions that determine the dynamics of water and sediment not only on land but also in coastal zones. However, information is still limited in mountain watershed because water and sediment dynamics changes dramatically between normal and storm condition and measurements during heavy storm is very difficult. In this study, we introduce the recent studies on flood flow from hillslopes and channels in mountainous watersheds based on field observations and literature review.
In mountainous watersheds, where the watershed area is steep and relatively small, it has been thought that the rainfall hyetographs transformed to discharge hydrographs mainly in hillslopes because water movement in hillslope soil should have greater effect on the waveform transformation and/or peak delay than that in the open channel flow. Researches have been conducted in hillslopes and the processes that control flood flow during small to middle sized storms have been clarified. However, responses and processes that occur during heavy storms have not been well understood. In addition, because mountain rivers have rough bed morphology consisting of gravels of various sizes, water depth and velocity distribution changes greatly during storms. However, even if the discharge is known, prediction of the water depth and velocity is still difficult.
To understand flood flow from hillslopes, we installed 16 water level recorders within a 4.5 km2 watershed in the University Forest of The University of Tokyo in Minami-Izu, Shizuoka and measured water levels for every minute. Because there were linear relationships between the channel flowpath length and the peak lag time during heavy storms, the peak propagation speeds in the channels were estimated from the slopes of these linear relationships, and the peak lag times in the hillslope flowpath were estimated from the intercepts. During heavy storms, that occur once every year to several years, the peak lag times in the hillslopes were -1 to 2 minutes, and the peak propagation speeds in the hillslopes were as fast as a few m/s. While the peak propagation speeds in the channels were 2.9 m/s at maximum. These small contributions of lateral flows in hillslopes on the catchment peak lag times have been demonstrated by other experimental and theoretical studies.
For flood flow in channels, we found the drastic changes in channel resistance during heavy storms. In practice, Manning's roughness coefficient has been used for flood analysis even in mountain rivers. The channel resistance decreased significantly as the water level rises, from more than 0.5 to less than 0.1 in terms of Manning's roughness coefficient. We also found that the change in the channel resistance with water depth became smaller when water depth increased to certain depth as most of gravel submerge, but that the measured minimum values of the channel resistance were often larger than the range that has been used as the reference.
In mountain watershed, when hillslopes become sufficiently wet during heavy storms, peak propagation speed in hillslopes increases and rainwater is expected to discharge to channel simultaneously and quickly from all hillslopes. While in mountain rivers with rough bed, the flow speed does not become very fast due to the high resistance. Therefore, the delay between the peak of rainfall and the peak of runoff during heavy storms is small on lateral water movement in hillslopes and occurs mainly in the channel.
In mountainous watersheds, where the watershed area is steep and relatively small, it has been thought that the rainfall hyetographs transformed to discharge hydrographs mainly in hillslopes because water movement in hillslope soil should have greater effect on the waveform transformation and/or peak delay than that in the open channel flow. Researches have been conducted in hillslopes and the processes that control flood flow during small to middle sized storms have been clarified. However, responses and processes that occur during heavy storms have not been well understood. In addition, because mountain rivers have rough bed morphology consisting of gravels of various sizes, water depth and velocity distribution changes greatly during storms. However, even if the discharge is known, prediction of the water depth and velocity is still difficult.
To understand flood flow from hillslopes, we installed 16 water level recorders within a 4.5 km2 watershed in the University Forest of The University of Tokyo in Minami-Izu, Shizuoka and measured water levels for every minute. Because there were linear relationships between the channel flowpath length and the peak lag time during heavy storms, the peak propagation speeds in the channels were estimated from the slopes of these linear relationships, and the peak lag times in the hillslope flowpath were estimated from the intercepts. During heavy storms, that occur once every year to several years, the peak lag times in the hillslopes were -1 to 2 minutes, and the peak propagation speeds in the hillslopes were as fast as a few m/s. While the peak propagation speeds in the channels were 2.9 m/s at maximum. These small contributions of lateral flows in hillslopes on the catchment peak lag times have been demonstrated by other experimental and theoretical studies.
For flood flow in channels, we found the drastic changes in channel resistance during heavy storms. In practice, Manning's roughness coefficient has been used for flood analysis even in mountain rivers. The channel resistance decreased significantly as the water level rises, from more than 0.5 to less than 0.1 in terms of Manning's roughness coefficient. We also found that the change in the channel resistance with water depth became smaller when water depth increased to certain depth as most of gravel submerge, but that the measured minimum values of the channel resistance were often larger than the range that has been used as the reference.
In mountain watershed, when hillslopes become sufficiently wet during heavy storms, peak propagation speed in hillslopes increases and rainwater is expected to discharge to channel simultaneously and quickly from all hillslopes. While in mountain rivers with rough bed, the flow speed does not become very fast due to the high resistance. Therefore, the delay between the peak of rainfall and the peak of runoff during heavy storms is small on lateral water movement in hillslopes and occurs mainly in the channel.