Abstract
In most watersheds, the concentration of dissolved solids in stream water decreases at high flows and increases at low flows. This phenomenon was first noticed by Durum (1953) and has been studied by worldwide simultaneous measurement of stream flow. As a result, various experimental and theoretical equations have been proposed relating concentration and stream flow (Hall, 1970). Although the concentration and stream flow equations stated above include two or four parameters, they can not fully represent the influences of dynamic hydrologic effects because they usually have only one independent variable, stream flow. At first, in this paper, hydrologic effects (dilution effect by precipitation, concentration effect by evapotranspiration, and mixing effect by the movement of water within a watershed) are examined using a watershed model. Secondly, a chemistry model for stream water is made incorporating mixing and adsorptionrelease relations within the watershed model. Finally, C1- simulation is made using this chemistry model. W-6, Hubbard Brook Experimental Forest of U. S. Forest Service, New Hampshire U.S.A. (Fig. 1) has been selected for the study, because many hydrologic, geologic, and chemical data have been accumulated so far. The data from May through November, or the non-snow covered season, of 1967, 1968, and 1974 are studied.
Federer (1972) presented a watershed model (Fig. 2) for Hubbard Brook, simplifying Stanford Watershed Model (Linsley and Crawford, 1960). In the present study, a base flow component is added to the watershed model, and optimizations of parameter values are carried out independently by a trial and error method.
The hydrologic effects are examined incorporating the calculations of amount and concentration of a material concerned into the watershed model for two cases; one represents when water is completely mixed in the watershed and the other when water is partially mixed. For both cases, it is assumed that the precipitation always has 1.0 unit concentration, that there are neither “sinks” nor “sources” within the watershed, and that there is no chemical reactions.
Variations of Cl- concentrations of stream water are smaller than those of precipitation and those expected by hydrologic effects in Hubbard Brook. Therefore, some buffering mechanism must be at work. It is estimated that the buffering mechanism is mainly due to adsorption-release mechanism by podzol. Among several adsorption equations, Langmuir adsorption relation is employed to represent the buffering mechanism. A computer subroutine is made which uses regula falsi method to numerically solve the amount of adsorption or that of release so that the Langmuir equation is always met. A chemistry model is made adding the subroutine into the watershed model. In other words, in the chemistry model, the stream water chemistry is defined by three factors: input data of chemistry of precipitation, the hydrologic effects, and the buffering mechanism by adsorption and release.
The stream flows are well simulated on a daily basis by the watershed model (Fig. 3). According to the simulation, the mean, the maximum, and the minimum values of water storage in W-6 are 164, 199, and 81mm, respectively. Mean residence time of stream water is calculated to be about 83 days.
When the hydrologic effects are expressed by a ratio of concentration of stream water to that of precipitation, the mean, the maximum, and the minimum concentrations of stream water are 1.91, 3.87, and 1.23 units, respectively for the data of 1967 in case of the complete mixing. When the partial mixing is assumed, larger hydrologic effects are estimated. In addition, it is confirmed that the variations of concentrations of stream water due to hydrologic effects are minimal when water is completely mixed in a watershed.
