|
<br />points should yield the better result. Unless a
<br />multivariate optimization method can be resorted to
<br />minimize the expression as shown on the right-hand
<br />side of Eq, 61 or 62 in which the "objective"
<br />function is now a function not only of "I, but also of
<br />a, b. and c, to equalize the areas under the measured
<br />and synthetic hyetographs appears to be a better
<br />means and optimization criterion for best fitting the
<br />design hyetograph to the recorded hyetograph, The
<br />latter optimization techniques should be explored in
<br />future study.
<br />
<br />The data points used in the present analysis are
<br />only limited to those places in a recorded hyetograph
<br />where either time or intensity changes in value.
<br />Therefore, the limited number of data points may
<br />often result in big errors in the estimation of the a, b,
<br />and c values in the recorded hyetograph and hence
<br />the "I value.
<br />
<br />Major storms for six stations in ARS
<br />experimental agricultural watersheds (ARS Black
<br />Book Series, 1933-67) as well as those collected by
<br />six automatic recording rain gages in each of the two
<br />urban highway watersheds in the Salt Lake City area
<br />(Fletcher and Chen, 1975) are selected and analyzed.
<br />Actual hyetographs of intense bursts may have
<br />double peaks, sometimes even more (Le., multiple
<br />peaks), but they are all treated as a single peak and
<br />then applied with the equations, One of the most
<br />difficult problems encountered in the analysis is the
<br />selection of td for the "isolated" burst. Some storms
<br />are very distinctively isolated, but others are not,
<br />depending upon the intensity of the antecedent and
<br />the trailing parts. Especially some prolonged storms
<br />with the intensity less than 0.1 in./hr in the trailing
<br />part lasted for tens of hours, For maintaining
<br />consistence and the range of interest, storm bursts
<br />with the duration longer than 24 hours are excluded
<br />from this analysis.
<br />
<br />Major Storms for Selected Stations from
<br />ARS Experimental Agricultural Watersheds
<br />
<br />In the design of a drainage inlet, criticism may
<br />arise if all the significant factors which affect runoff
<br />from the urban highway watershed are not taken into
<br />account. One such factor is the infiltration capacity
<br />that varies due to different species of turf planted on
<br />the urban highway sideslopes.
<br />
<br />The urban highway watersheds may be classi.
<br />fied according to the following six major species of
<br />turf planted on the highway sideslopes: Bermuda
<br />grass (Cynodon dactylon), Crested wheat grass (Agro-
<br />pyron dactylon), Fescue grass (Festuca elation var
<br />arundimlcea), Kentucky blue grass (Poa protensis),
<br />Red top grass (Agrosris palustris; Agrostis alba) and
<br />Rye grass (Lolinum perenne; Lolinum multijlorum).
<br />
<br />Six grass zones were delineated by Fletcher I in
<br />accordance with the adaptability of a species of grass
<br />for highway sideslopes In an area under study (see
<br />Fig. 9). Note that this delineation is not unique
<br />because other species may also be planted on the
<br />same areas. Unspecified areas in Fig. 9 are either
<br />planted with other than the six species of grass or are
<br />not taken into consideration in the delineation for
<br />lack of available data.
<br />
<br />The U.S. Weather Bureau Technical Paper No.
<br />40 was consulted for the selection of most intensive
<br />stations. one from each grass zone, from the ARS
<br />experimental agricultural watersheds (ARS Black
<br />Book Series, 1933-67) at which official records of
<br />actual hyetographs have been kept, The six stations
<br />representing approximately the points of the most
<br />intensive rainfall for any duration and frequency in
<br />the respective six grass zones are marked in Fig. 9.
<br />Major storm bursts at the six stations were obtained
<br />from the ARS Black Book Series and then analyzed,
<br />and the storm parameters for each storm were
<br />calculated and tabulated in Table 12.
<br />
<br />Typical hyetographs for the six stations, one
<br />from each grass zone, with their best -fitted
<br />counterparts are shown in Figs. 10 through 15 for
<br />comparison.
<br />
<br />Analysis of the hyetograph equations, Eqs. 24
<br />through 34, has resulted in the derivation of the
<br />important dimensionless parameters, bltd, c, and "I,
<br />for the design storm formation. The values of these
<br />parameters were examined in an effort to establish
<br />the relationships among them for each station
<br />investigated. However, plots of these parameters on
<br />linear, semi-log, or log-log paper failed to reveal
<br />the existence of such relationships. For example, the
<br />computed "I value for each station varies almost
<br />independently of any of the other parameters. No
<br />attempt was made to calculate the statistical mean "I
<br />value of each station for those actual hyetographs
<br />analyzed. It would probably require more data points
<br />in the analysis in order for any statistical mean "I
<br />value to be meaningful.
<br />
<br />Major Storms for Two Salt Lake City
<br />Urban Highway Watersheds
<br />
<br />Major storms for two Salt Lake City urban
<br />highway watersheds (Fletcher and Chen, 1975) were
<br />analyzed in similar ways as for ARS watersheds and
<br />the values of the storm parameters calculated
<br />accordingly. There were six recording gages for each
<br />station and hyetographs measured at the six gages
<br />
<br />IPletcher, J. E., Storm design criteria for highway in-
<br />lets, Unpublished Interim Report for Federal Highway Ad-
<br />ministration, Utah Water Res. Lab., Utah State University,
<br />Logan, Utah, 1973.
<br />
<br />30
<br />
|