<br />SUITABILITY OF THE UPPER COLORADO RIVER BASIN FOR PRECIPITATION MANAGEMENT
<br />
<br />Hiroshi Nakamichi*and Hubert J. Morel-Seytoux**
<br />
<br />by
<br />
<br />INTRODUCTION
<br />
<br />1. Water needs of the basin. The Colorado River
<br />system is the largest in the United States that flows
<br />mainly through lands having a chronic water deficiency
<br />for cultivation of crops [1]. Since the 1940's, the
<br />basin's population has increased rapidly with an accom-
<br />panying growth in demand upon the region's water re-
<br />sources for irrigation, industrial, and domestic uses
<br />[2]. Over the decade from 1951 through 1960, the popu-
<br />lation of the five states comprising the Upper Colorado
<br />River Basin has increased by 40 percent, while over the
<br />same period the population of the nation as a whole has
<br />increased only by 20 percent [3].
<br />
<br />2. Precipitation management program. In an
<br />effort to reduce the severity of these demands, an
<br />atmospheric water resource project is currently pur-
<br />sued by the United States Department of the Interior,
<br />Bureau of Reclamation, Office of Atmospheric Water
<br />Resources. The goal of this project is to induce more
<br />precipitation from the atmosphere by winter cloud seed-
<br />ing operations over certain high altitude watersheds in
<br />the Upper Colorado River Basin. In the past, there
<br />was some controversy as to whether man could economi-
<br />cally increase precipitation in worthwhile amounts.
<br />There now exists evidence that this is possible at
<br />least in high mountain areas [4]. As of February 1969,
<br />plans of the Bureau of Reclamation called for a concen-
<br />trated experimental effort in two pilot areas of the
<br />Upper Colorado River Basin, to start in the fall of
<br />1969 [5]. This study was undertaken in connection with
<br />the Bureau's overall program in general and in connec-
<br />tion with this pilot program in particular.***
<br />
<br />3. Criteria of suitability. In the experimental
<br />or large-scale operational stage of the project, a
<br />site should be selected. At this point, one needs
<br />certain criteria in order to select suitable basins.
<br />These criteria should be considered both from a water
<br />resource and an evaluation standpoint [6]. The first
<br />standpoint requires a criterion of suitability for
<br />optimal water yield, and the second, a criterion of
<br />suitability for minimum time evaluation.
<br />
<br />Idea11y the criteria should be objective and
<br />simple. That is, they should be derived easily from
<br />available data rather than from theory. Though various
<br />aspects of research un cloud modification have been
<br />conducted sucessfully, it is still difficult to deter-
<br />mine its quantitative effect. Indeed, one of the
<br />
<br />Chapter I
<br />
<br />purposes of the pilot project is to deteTIlline the exact
<br />magnitude of the increase in precipitation on a large
<br />areal scale. Following this experiment, it may be
<br />possible to isolate the major factors that determine
<br />the magnitude of the increase in precipitation. Once
<br />precipi tation is induced, the increase in runoff, (llQ) ,
<br />caused by the increase of precipitation, (LIP), is esti-
<br />mated by a statistical relationship between precipita-
<br />tion and runoff, (Q = f(P)), often used when forecasting
<br />runoff:
<br />
<br />llQ = (Q+llQ) - Q = f(P+llP) - f(P)
<br />
<br />(1)
<br />
<br />Marginal criteria are defined in order to deteTIlline the
<br />relative suitability of many potential basins for mini-
<br />mum time evaluation, even if the type of statistical
<br />test and the design of the experiment are not known [6].
<br />One such criterion is derived from the "two-sample
<br />u-test. "
<br />
<br />The two-sample u-test is a test of the hypothesis
<br />that assumes that the mean of a statistical population
<br />(the values of annual runoff for a given basin over
<br />many years) has not changed significantly even though
<br />there were reasons to suspect it had. As the name
<br />implies, the application of the test requires the
<br />availability of two samples of data, one sample collec-
<br />ted prior to the suspected change and one collected
<br />afterward. If the suspected change is real but small,
<br />the records of many years may be necessary to determine
<br />its significance. If the change is large and the
<br />spread of the distribution is narrow, only a few years
<br />may be required.
<br />
<br />No statistical test is free of assumptions. The
<br />two-sample u-test assumes that only the mean of the
<br />population may have changed whereas the shape and the
<br />spread of the distribution have not. Assuming a normal
<br />distribution, the explicit expression [6] for the num-
<br />ber of years, N, necessary to guarantee the statistical
<br />significance of the observed or expected increase at
<br />the 95 percent confidence level is given by:
<br />
<br />N
<br />
<br />(1.96)2 x~ =
<br />(llQ) 2
<br />
<br />(2)
<br />
<br />3.84 0 2
<br />Q
<br />(llQ) 2
<br />
<br />* M.S. Graduate of Colorado State University, Civil Engineering Department, Fort Collins, Colorado, presently
<br />with Planning Division, Chugoku-Shikoku, Nosei kyota, 9-24 Tenjin-cho, Okayama-shi, Japan.
<br />Associate Professor, Civil Engineering Department, Colorado State University, Fort Collins, Colorado.
<br />***Since the initiation of this study the plans of the Bureau were modified. Currently (45) only one area
<br />is considered: the San Juan Mountains region.
<br />
|