Record Display for the EPA National Library Catalog

RECORD NUMBER: 2359 OF 2436

OLS Field Name OLS Field Data
Main Title Use of Stochastic Hydrology to Determine Storage Requirements for Reservoirs-a Critical Analysis.
Author Burge, Stephen J. ;
CORP Author Stanford Univ., Calif. Program in Engineering-Economic Planning.
Year Published 1970
Report Number EEP-34; DI-14-31-0001-3150; OWRR-C-1635; 01188,; C-1635(1)
Stock Number PB-195 691
Additional Subjects ( Reservoirs ; Water storage) ; ( Water storage ; Decision making) ; Stochastic processes ; Monte Carlo method ; Surface water runoff ; Requirements ; Water supply ; Correlation techniques ; Regression analysis ; Markon processes ; Theses ; Stochastic hydrology
Holdings
Library Call Number Additional Info Location Last
Modified
Checkout
Status
NTIS  PB-195 691 Most EPA libraries have a fiche copy filed under the call number shown. Check with individual libraries about paper copy. NTIS 06/23/1988
Collation 226p
Abstract
A study was conducted to locate areas of uncertainty when Markov runoff generation models are used in storage reservoir studies. A thorough examination of the basic method was made. The length of available record is critical to model parameter determination. Examination of the Monte Carlo generation technique showed that, for annual models and economic lives in the range 20 to 100 years, at least 1000 generated inflow traces are required to accurately define the distribution of storage to meet a specified demand. Storage is described by the extreme value type 1 probability distribution. Factors critical to storage requirements are the coefficient of variation and correlation of the inflow sequence, demand pattern and economic life. Except in the special case, where runoff is highly seasonal and the bulk of the demand occurs after the main runoff period, monthly generation models were found to be necessary. When a log-normal model is used, a method that preserves the properties of the observed sequence is necessary. Multiple-lag generation models were shown to be unjustified. Correlations in observed data at increasing lag result from short records. (Author)