Persistent Link:
http://hdl.handle.net/10150/300116
Title:
Conditional Streamflow Probabilities
Author:
Roefs, T. G.; Clainos, D. M.
Affiliation:
Hydrology & Water Resources, University of Arizona
Issue Date:
23-Apr-1971
Rights:
Copyright ©, where appropriate, is held by the author.
Collection Information:
This article is part of the Hydrology and Water Resources in Arizona and the Southwest collections. Digital access to this material is made possible by the Arizona-Nevada Academy of Science and the University of Arizona Libraries. For more information about items in this collection, contact anashydrology@gmail.com.
Publisher:
Arizona-Nevada Academy of Science
Journal:
Hydrology and Water Resources in Arizona and the Southwest
Abstract:
Streamflows of monthly or shorter time periods, are, in most parts of the world, conditionally dependent. In studies of planning, commitment and operation decisions concerning reservoirs, it is probably most computationally efficient to use simulation routines for decisions of low dimensions, as planning and commitment, and optimization routines for the highly dimensional operation rule decisions. This presents the major problem of combining the 2 routines, since streamflow dependencies in simulation routines are continuous while the direct stochastic optimization routines are discrete. A stochastic streamflow synthesis routine is described consisting of 2 parts: streamflow probability distribution and dependency analysis and a streamflow generation using the relationships developed. A discrete dependency matrix between streamflow amounts was then sought. Setting as the limits of interest the class 400-500 thousand acre ft in January and 500-600 thousand acre ft in February, and using the transforms specified, the appropriate normal deviates were determined. The next serious problem was calculating the conditional dependency based on the bivariate normal distribution. In order to calculate the joint probability exactly, double integrations would be required and these use too much computer time. For the problem addressed, therefore, the use of 1-dimensional conditional probabilities based on the flow interval midpoint is an adequate and effective procedure.
Keywords:
Water resources development -- Arizona.; Hydrology -- Arizona.; Hydrology -- Southwestern states.; Water resources development -- Southwestern states.; Probability; Stochastic processes; Streamflow; Decision making; Statistical models; Mathematical studies; Flow characteristics; Variability; Simulation analysis; Optimization; Reservoirs; Planning; Conditional probabilities; Flow intervals
ISSN:
0272-6106

Full metadata record

DC FieldValue Language
dc.language.isoen_USen_US
dc.titleConditional Streamflow Probabilitiesen_US
dc.contributor.authorRoefs, T. G.en_US
dc.contributor.authorClainos, D. M.en_US
dc.contributor.departmentHydrology & Water Resources, University of Arizonaen_US
dc.date.issued1971-04-23-
dc.rightsCopyright ©, where appropriate, is held by the author.en_US
dc.description.collectioninformationThis article is part of the Hydrology and Water Resources in Arizona and the Southwest collections. Digital access to this material is made possible by the Arizona-Nevada Academy of Science and the University of Arizona Libraries. For more information about items in this collection, contact anashydrology@gmail.com.en_US
dc.publisherArizona-Nevada Academy of Scienceen_US
dc.identifier.journalHydrology and Water Resources in Arizona and the Southwesten_US
dc.description.abstractStreamflows of monthly or shorter time periods, are, in most parts of the world, conditionally dependent. In studies of planning, commitment and operation decisions concerning reservoirs, it is probably most computationally efficient to use simulation routines for decisions of low dimensions, as planning and commitment, and optimization routines for the highly dimensional operation rule decisions. This presents the major problem of combining the 2 routines, since streamflow dependencies in simulation routines are continuous while the direct stochastic optimization routines are discrete. A stochastic streamflow synthesis routine is described consisting of 2 parts: streamflow probability distribution and dependency analysis and a streamflow generation using the relationships developed. A discrete dependency matrix between streamflow amounts was then sought. Setting as the limits of interest the class 400-500 thousand acre ft in January and 500-600 thousand acre ft in February, and using the transforms specified, the appropriate normal deviates were determined. The next serious problem was calculating the conditional dependency based on the bivariate normal distribution. In order to calculate the joint probability exactly, double integrations would be required and these use too much computer time. For the problem addressed, therefore, the use of 1-dimensional conditional probabilities based on the flow interval midpoint is an adequate and effective procedure.en_US
dc.subjectWater resources development -- Arizona.en_US
dc.subjectHydrology -- Arizona.en_US
dc.subjectHydrology -- Southwestern states.en_US
dc.subjectWater resources development -- Southwestern states.en_US
dc.subjectProbabilityen_US
dc.subjectStochastic processesen_US
dc.subjectStreamflowen_US
dc.subjectDecision makingen_US
dc.subjectStatistical modelsen_US
dc.subjectMathematical studiesen_US
dc.subjectFlow characteristicsen_US
dc.subjectVariabilityen_US
dc.subjectSimulation analysisen_US
dc.subjectOptimizationen_US
dc.subjectReservoirsen_US
dc.subjectPlanningen_US
dc.subjectConditional probabilitiesen_US
dc.subjectFlow intervalsen_US
dc.identifier.issn0272-6106-
dc.identifier.urihttp://hdl.handle.net/10150/300116-
dc.identifier.journalHydrology and Water Resources in Arizona and the Southwesten_US
dc.typetexten_US
dc.typeProceedingsen_US
All Items in UA Campus Repository are protected by copyright, with all rights reserved, unless otherwise indicated.