Laserfiche WebLink
Memorandum To: Ray Alvarado M~t~ch 10, 2003 <br />Page 5 <br />discussions with State personnel on daily modeling in the extremely water short Rio Gr de basin <br />have indicated daily baseflow calculations are required in order to achieve a reasonably alibrated <br />daily model. <br />When daily baseflows are calculated from daily terms, whether gaged or estimated, the luality of <br />input data and engineering estimates of water use efficiency and return flows become c lical. In a <br />water short system, poor input data can result in overestimating baseflows. This occurs Because <br />when the baseflow computation results in a negative value, StateMod sets the baseflow p zero since <br />the negative number has no physical meaning. For both monthly and daily models, setti negative <br />baseflows to zero enlarges the modeled water supply. With a daily model, the effect is a aggerated <br />because there are 30 or 31 "opportunities" each month to bias the baseflow estimate. In monthly <br />model, only one calculation is made, and it represents the net impact over the month of ~e various <br />baseflow components. This calculation is less likely to be negative than the calculation br any given <br />day. As recommended by the StateMod documentation, negative baseflows are an Indic ion of poor <br />data or engineering estimates and should be eliminated or avoided. However, in practic this might <br />be a time consuming calibration activity that is more difficult to achieve on a daily time step. <br />The daily pattem approach in the Yampa pilot study not only provided excellent results, it also <br />relieved the user of having to generate daily values for historical diversions, imports, re ervoir <br />contents, and gage flows. One caveat is that the hydrograph of the "pattern gage" shoul rsesemble <br />the gage being filled. For example, in the Yampa basin we observed that the mainstem age Yampa <br />River at Steamboat Springs was a very good pattern gage for all downstream baseflow pints, but <br />did not work as well for tributary baseflow sites. Ideally, the pattern gage should also b minimally <br />impacted by upstream operations. The extent to which potential pattem gages have thes <br />characteristics will influence overall results and should be considered. <br />4.2 Daily Demand Modeling <br />The Yampa daily model demonstrated that averaging monthly demand in combination ith <br />baseflows that vary daily can lead to underestimating diversions. Figure 3 illustrates the <br />phenomenon. The demand shown is average monthly demand, as in Figure 2, superimp sed on the <br />daily hydrograph for 1983. StateMod limits the diversion in each simulation step to the esser of <br />demand and supply. Therefore, in the first half of July, 1983, StateMod would limit div xsions to the <br />level of the average monthly demand, despite a plentiful water supply. In the latter part if the month <br />as the hydrograph drops, the modeled diversion would be limited by supply. Therefore, fbaseflow <br />is represented as varying daily, demand may need to be smoothed in some fashion for b st results. <br />StateMod has the ability to create a daily demand time series that passes through the av rage <br />monthly demand at mid-month, and preserves the monthly total demand. This option ( proach 4 <br />above) was used in the Yampa basin daily model. <br />Another option available in StateMod is to use the daily decrementing approach to esti ate <br />demands. This approach estimates daily demands by providing a monthly total that is d creased each <br />day in the month that a diversion occurs. For the Yampa, a system with a relatively ab iiant water <br />supply, this option resulted in demands being satisfied in the early part of a month, and o diversions <br />occurring later in the month. This representation was therefore unacceptable for the Y pa basin. <br />CS-B99-190-OOB /rASKMEM03PINAL.DOC I304+CE <br />