Because storm runoff loading in a river is larger in comparison with loading in same period of dry weather day, the author proposes following regression models to estimate its loading. Plotting the relationships between sum of loading a unit area (ΣL
gross/A) and sum of flow a unit area (ΣQ
gross/A) in direct runoff period by storm events on logarithmic paper, the points are in a restricted range and are expressed as an exponential equation, ΣL
gross/A=a·(ΣQ
gross/A)
n, by regression analysis. Similarly, for net loading by direct runoff (ΣL
net) which is substracted sum of loading in same period at dry weather day (ΣL
base) from ΣL
gross, a regression equation is obtained with net discharge (ΣQ
net), ΣL
net/A=a·E (ΣQ
net/A)
n And also, net loading a unit area (ΣL
net/A) could be expressed as regression equations with ΣQ
net, observation period (T
dir) or rainfall duration period (T
rain) and area of basin (A) as following:ΣL
net/(A·T
dir) =a [ΣQ
net/(A·T
dir)]
n;ΣL
net/A=a·[ΣQ
net/(A·T
rain)]
n. Using observation data (28 cases) of influent rivers into the Lake Kasumigaura, regression equations of four types are obtained for almost all water quality indicators with high regression coefficients. Especially, these four regression models could be applied not only for total and particulate component loadings but also for dissolved component loading.
抄録全体を表示