FLOOD05499 - Laserfiche WebLink

Home Browse Search

f(t)<Ks and Lr(t)<~from both Eqs, 16 and 17, respectively, If the .p(z, t)- or O(z, t)-distributions computed are accurate, the f(t) values computed from Eqs, 10, 11, and 16, respectively, should not deviate very much from each other. However. given h(t), any error in the numerical computation of the 0 value would cause errors in tbe computation of Lr(t) by using Eq. 17, which in turn makes the f(t) calculation inaccurate by using Eq. 16, It will be shown later at some l~ngth in this report that oscillation in the computation of f(t) at the beginning of ponding is actually induced by this inaccuracy in the (J -distribution computation, and aggravated, especially, by the larger finite-difference space.step size, the higher rainfall intensity, the higher ponding Or(t) Water surface, z = h(t) Soil surface, z ~.: ~~..:;~}~~~~~~~~/: f(t) ~ r(t) L f Saturation front, z = -Lf(t) ,1 -----lJ---- f(t) ~ r(t) depth, and/or the smaller initial moisture content. Nevertheless, the computational oscillation damps out quite quickly with the advancing saturation front as it proceeds deeper in the soil column. After ponding, complications in the computa~ tion of the >/I.z, t)- or O(z, t)-distributions and hence the f(t) value have also been recognized by some investigators (Freeze, 1969; Smith and Woolhiser, 1971), Freeze imposed the upper boundary condition at the first unsaturated node while Smith and Woolhiser incorporated Eq, 15 into their implicit finite-difference scheme at the last saturated nodal point. To apply the upper boundary condition, either Eq. 3 before ponding or Eq, 4 after ponding, at the z v o .~~.'.;g~':~~~-: . . .. c\J Saturated zone _____~nn Unsaturat,ed zone c\1-d i st ribution Figure I. DefInition skeleh of rain infUtralion afler ponrling, 17