Laserfiche WebLink
<br />OOlG43 <br /> <br />Director, Colorado Water Conservation Board <br />February 26, 1981 <br />Page 3 <br /> <br />desirable outcome, it is not likely to be a good approximation of what <br />will occur in reality. Significantly greater crowding could be expected <br />(this matter is discussed in greater detail in the attached paper de- <br />scribing the COWRISM model). If the program were to be used in a deci- <br />sion-making context, it would be advisable to first modify it to correct <br />this inadequacy by permitted crowding to increase to the point where the <br />net benefits derived by the marginal recreationist fall to zero. <br /> <br />Even with this change, however, the model would not be free of bias due <br />to the limited nature of the optimization assumptions upon which it is <br />now based. Essentially, COWRISM now assumes that recreationists will <br />decide where to carry out water-based recreational activities in such <br />a way as to maximize the net utility they receive from such activities, <br />and giving due consideration to the effect of streamflow level on the <br />utility produced by the recreational experience, but without simulta- <br />neously considering the effect of crowding on such utility. Thus, the <br />second way in which the COWRISM model could be improved would be to <br />completely reformulate it in order to permit simultaneous consideration <br />of location, streamflow level, and crowding. However, this would re- <br />quire complicated and expensive modeling due to the nature of the crowd- <br />ing problem. Distance and streamflow are externally determined vari- <br />ables which influence recreationists' behavior, but are not influenced <br />by it. Such one-way relationships are relatively easy to model. Crowd- <br />ing, however, is an endogenous variable which both influences and is in- <br />fluenced by recreationists' behavior. Either a mathematically determi- <br />nate simultaneous equations model or a complex and computationally ex- <br />pensive numerical approximation model (presently, COWRISM is a numerical <br />approximation model, but a simple one compared to what is now under dis- <br />cussion) would be required in order to deal adequately with this problem. <br />Either would be costly to devise and probably costly to use, as well. <br />There seems to be little reason to proceed along these lines unless the <br />model is to become a key decision-making device. Although inexpensive <br />by comparison with the empirical data collection needs described under <br />item l.b. above, it would still cost perhaps $20,000 to reformulate the <br />model. This discharges our responsibility under item l.c. of the <br />contract. <br /> <br />Late in 1980, Policy Sciences Associates' principal investigator held <br />discussions with staff members of the Colorado Water Conservation Board. <br />During those discussions, it was suggested that the hydrologic data base <br />available for use in the COWRISM program was deficient in two respects. <br />First, the inventory of Colorado stream mileage by region and stream <br />class was highly approximate, particularly with respect to the smaller <br />