WSP05303 - Laserfiche WebLink

Home Browse Search

2 o NATURAL INFLOW RANDOM VARIABLES CJl 00 The major simplification accomplished by this method is the use of CO only three random variables. These random variables represent monthly inflows in the spring runoff, summer and winter seasons, The spring runoff random variable, denoted by R, covers the months of April, May and Jtme. The summer random variable (S) covers July, August and September, The winter random variable (W) covers the 6 months from October through March, S is dependent upon R because the melting snowpack provides IIlUch of the summer inflows. However, summer rains provide considerable variation in S which is not dependent upon R. W is independent of R because the snowpack has melted before October. However, some correlation exists from R onto W due to long tenn weather cycles, Since the winter season is frequently in a frozen condition in the Rocky Motmtain region, the range of W inflows is quite narrow. In the White River data, the variation in W due to both R and the weather cycle correlation (which cannot easily be separated) is small, Year to year correlation in R exists, but it is only a small part of 1he total variation in R. This correlation is only significant when a regulation scheme attempts to maintain carryover storage which is large in comparison to annual inflows. This correlation does not affect more modest regulation schemes, In this analytic method this annual correla- tion is ignored (making R independent year to year). However, it should be considered (requiring a large safety stock of carryover storage) whenever the attempted regulation involves a high percentage of the annual inflows. Simulation methods using a time series of historic inflows include the correlation effect. However, it is only one short realization of this effect.