A multiseries stochastic model for synthetic monthly flows


Haktanir T., Kara M. B., AÇANAL HAKTANIR N.

HYDROLOGICAL SCIENCES JOURNAL-JOURNAL DES SCIENCES HYDROLOGIQUES, vol.67, pp.741-758, 2022 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 67
  • Publication Date: 2022
  • Doi Number: 10.1080/02626667.2022.2039662
  • Journal Name: HYDROLOGICAL SCIENCES JOURNAL-JOURNAL DES SCIENCES HYDROLOGIQUES
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Academic Search Premier, IBZ Online, PASCAL, Agricultural & Environmental Science Database, Aqualine, Aquatic Science & Fisheries Abstracts (ASFA), CAB Abstracts, Compendex, Geobase, INSPEC, Pollution Abstracts, Civil Engineering Abstracts
  • Page Numbers: pp.741-758
  • Keywords: stochastic models for monthly stream flows, Box-Jenkins models, stochastic hydrology
  • Erciyes University Affiliated: Yes

Abstract

A new multiseries stochastic model is developed that takes n-year-long monthly stream flows observed at as many as 20 stations in a group and computes N-year-long synthetic monthly flows for up to 10 000 years. Initially, any series that exhibit trends with a 99% confidence level are detrended. The monthly flows of each observed series are transformed to standard-normal variates, and a variate is related to those of the other series of the same month, those of the previous month of all series, the lag-one moving-average component, and the random component. 2 M parameters of each series are iteratively computed by the least squares method, such that the variance of the random component determined using the observed variates and those computed by the model equals the variance given by the model. The model was applied to five groups in Anatolia. The relevant peculiarities of the observed and synthetic flows were in harmony.