Spatial bayesian methods of flow forecasting in the Black Volta river

Iddrisu Wahab Abdul, Kaku Sagary Nokoe, Frank Badu Osei, Eric Ofosu Antwi

Research output: Contribution to journalArticleAcademicpeer-review

10 Downloads (Pure)

Abstract

The use of Spatial Bayesian Vector Autoregressive (SBVAR) models for river flow forecasting is studied in this paper. SBVAR models based on both the First Order Spatial Contiguity (FOSC) and the Random-Walk Averaging (RWA) priors were estimated and compared in terms of forecast performance. Monthly data on river flows from January 2000 to December 2009 for the four gauge stations along the Black Volta River namely, Lawra, Chache, Bui and Bamboi was obtained from the hydrological services department of Ghana and used for model fitting. The estimation and forecasting procedure was conducted using the Econometrics Toolbox in MATLAB. Mean Absolute Percentage Errors (MAPEs) were calculated for all models considered. The results indicated very good forecasts for all the models considered. However, a comparison among them clearly indicated a much better performance by the SBVAR model based on the RWA prior which considered flows from only the immediate upstream gauge station as important while flows from all other gauge stations were considered unimportant.
Original languageEnglish
Pages (from-to)89-105
JournalEuropean journal of scientific research
Volume137
Issue number1
Publication statusPublished - 2016

Fingerprint

Dive into the research topics of 'Spatial bayesian methods of flow forecasting in the Black Volta river'. Together they form a unique fingerprint.

Cite this