Laserfiche WebLink
<br />i)" t") --"1 <br />...1...:-' <br /> <br />. <br /> <br />. <br /> <br />5. Storage requirements and Yield Failure Probabilities <br /> <br />In this section the potential yield failure probabilities and accumulated deficit statistics for <br />various target drafts from the river system are assessed. This is a hydrologic quantification of the <br />implications of the differences in local mean flow and variance, A target draft, D, is set at (a) the mean <br />annual 1929-1993 flow, (b) 90% of that flow, (c) 75% of that flow and (d) 50% of that flow, reflecting <br />different levels of river development for water supply. The accumulated deficit or storage, S, required to <br />meet this draft is computed using the sequent peak algorithm: <br />St= max(O,S,+ D - QJ <br />where Q t is tile flow in the river and S t is the accumulated deficit for year l. <br />The probability of not being able to meet the target draft in any year, I.e., P(D:>Q J, is estimated <br />by defining the indicator variable I I, as I ifD>Q t and 0 else, and then computing the moving average of! <br />t using lowess with a moving window of 20 years, <br />The results of these analyses with the data from the Archuleta gage are shown in figure 6. As the <br />level of development increases, the storage needed to meet the accumulated deficit increases dramatically <br />both in magnitude and the duration of time over which the accumulated deficit or drought is seen, relative <br />to the last 20 years. For the lower target yields (50 and 75% of mean annual flow), the accumulated deficit <br />during the last 20 years is the same as it is with the full record. The relative severity of the droughts <br />experienced in the 1950's and the earlier record exceeds that of anything experienced in the last twenty <br />years at the 75% of mean annual flow target draft. Interestingly, for the 50% draft the maximum <br />accumulated deficit is e~l'erienced in the last twenty years, This occurs because 1977 is the driest year in <br />the record, and it fails in this period. Thus, the 1974-93 period has a rather severe single year event but is <br />a wet period otherwise that is not representative of severe multi-year droughts. The annual probability of <br />draft failure is the lowest during the 1974-93 period for any level of draft. This is another way of <br />recognizing that this is a relatively wet period. Since a purpose of river system simulation is to assess <br />e~l'osure to a severe run of low flow events, there are good reasons in this data set to use the longer <br />record. <br /> <br />May 25, 1997 <br /> <br />7 <br />