Open AccessCONTRIBUTION OF SEASONAL STOCHASTIC MODELS SARIMA TO THE RATIONAL WATER RESOURCES MANAGEMENT. THE CASE OF THE KRANIA ELASSONA KARST SYSTEM, THESSALY, GREECE
Author(s) -
Antonios Manakos,
George P. Demopoulos
Publication year - 2004
Publication title -
deltio tīs ellīnikīs geōlogikīs etaireias/deltio tīs ellīnikīs geōlogikīs etaireias
Language(s) - English
Resource type - Journals
eISSN - 2529-1718
pISSN - 0438-9557
DOI - 10.12681/bgsg.16700
Subject(s) - karst , autoregressive integrated moving average , hydrology (agriculture) , spring (device) , surface runoff , environmental science , stochastic modelling , groundwater , series (stratigraphy) , seasonality , groundwater flow , autoregressive–moving average model , moving average , flow (mathematics) , time series , autoregressive model , geology , aquifer , mathematics , statistics , engineering , geotechnical engineering , ecology , mechanical engineering , paleontology , geometry , biology
Several stochastic models, known as Box and Jenkins or SARIMA (Seasonal Autoregressive Integrated Moving Average) have been used in the past for forecasting hydrological time series in general and stream flow or spring discharge time series in particular. SARIMA models became very popular because of their simple mathematical structure, convenient representation of data in terms of a relatively small number of parameters and their applicability to stationary as well as nonstationary process.Application of the seasonal stochastic model SARIMA to the spring's monthly discharge time series for the period 1974-1993 in Krania Elassona karst system yielded the following results. Logarithms of the monthly spring discharge time series can be simulated on a SARIMA (4,1,1)(1,1,1)12 type model. This type of model is suitable for the Krania Elassona karst system simulation and can be utilised as a tool to predict monthly discharge values at Kafalovriso spring for at least a 2 year period. Seasonal stochastic models SARIMA seem to be capable of simulating both runoff and groundwater flow conditions on a karst system and also easily adapt to their natural conditions.Adapting the proper stochastic model to the karst groundwater flow conditions offers the possibility to obtain accurate short term predictions, thus contributing to rational groundwater resources exploitation and management planning