Laserfiche WebLink
<br />n Un.J /.. 3 () <br />0. J..\) ~ <br /> <br />Director, Colorado Water Conservation Board <br />February 2p, 1981 <br />Page 2 <br /> <br />decade old. More recent data will be included in the 1981 Colorado Com- <br />prehensive Outdoor Recreation Plan which is soon to be released, but these <br />data will be less detailed with respect to recreational activity classes. <br />Nonetheless, this data set is the best by far of all of those employed in <br />the COWRISM program. <br /> <br />Most of the data files were derived from the 1978 recreation site survey <br />conducted by R. G. Walsh and his associates at Colorado State University. <br />This survey, due to budgetary limitations, was conducted in a single year <br />which was characterized by abnormal streamflow levels, over only a portion <br />of the recreational season, at only a few recreational sites, and with an <br />inordinately small sample size. Walsh and his associates did a good job <br />within these severe limitations, but the resulting data can only be used <br />by making major generalizations and extrapolations which no statistician <br />could defend. We believe that it would cost at least $200,000 to conduct <br />a statistically reliable survey and thus to improve upon Walsh's data. <br /> <br />Beyond the question of reliability, the Walsh data are incomplete in one <br />important respect. No consideration was given to the distribution of <br />recreational activity over time. The use of three recreational seasons <br />within the COWRISM model offsets a part of this difficulty, but only in <br />the formal sense, because it requires assumptions about seasonal shares <br />in place of empirical data. Just as important as interseasonal variation <br />is variation within seasons, and particularly within the week. It is <br />reasonable to suppose that activity levels are generally higher on week- <br />ends, and thus that the application of the crowding response function to <br />mean activity levels introduces another systematic bias. In this case, <br />the bias is upwards. COWRISM is bound to overestimate recreational ac- <br />tivity levels because it cannot allow for the effect of weekend crowding. <br /> <br />If the Walsh survey were to be repeated, it would be desirable and almost <br />costless to include questions on the interseasonal, intraseasonal, and <br />intrasemanal distribution of activity. In the absence of such a repeat <br />survey, there is little that can be done to rectify this discrepancy. <br />This discharges our responsibility under item l.b. of the contract. <br /> <br />We have conducted an evaluation of the conceptual model upon which COWRISM <br />is based. The model might be improved upon in two ways. First, it could <br />be marginally improved at little cost (perhaps $500 of programmer and <br />analyst time) to remove one current inadequacy. That inadequacy stems <br />from the fact that crowding is permitted in the model only up to the point <br />at which the benefits derived by an additional recreationist no longer <br />exceed the costs imposed upon previous users. While this is a socially <br />