Abstract
Effective groundwater planning and management should be based on the prediction of available water volume. The complex nature of groundwater systems makes this complicated and requires the use of complex methods. Data-driven models using computational intelligence are becoming increasingly popular in that field. The key issue in predictive modelling is the selection of input variables. Wrocław-Osobowice irrigation fields were a wastewater treatment plant until 2013. The monitoring of groundwater levels is being continued to assess the water relations in that area after the end of their exploitation. The aim of the study was to assess the Hellwig method for predictors’ selection in groundwater level forecasting with support vector regression models. Data covered the daily time series of groundwater level in the period 2015–2019. Obtained models with a root mean squared error (RMSE) of 0.024–0.292 m and r2 of 0.7–0.9 were considered as high quality. Moreover, they showed good prediction ability for high as well as low groundwater values. Additionally, the proposed method is simple, and its implementation only requires access to groundwater level measurement data. It may be useful in groundwater management and planning in terms of actual climate change and threat of water deficits.
Highlights
Precise groundwater level (GWL) forecasting is crucial for efficient groundwater planning and management
The following lines present the values of squared correlation (r2 ) between the observed and predicted time series and the size of models’ errors (RMSE, mean absolute error (MAE))
computational intelligence (CI) groundwater level forecasting constitutes a modern approach to supporting groundwater management and planning
Summary
Precise groundwater level (GWL) forecasting is crucial for efficient groundwater planning and management. Data-driven models using computational intelligence (CI) methods are becoming increasingly popular and have significant potential [1,2,3,4]. This results from the fact that, in contrast to creating physics-based models which are time-consuming and labor-intensive and require taking into account a large amount of data that describe the modelled phenomenon with the use of complex algorithms, they are much easier to implement. The use of artificial intelligence in creating data-driven models consists of predicting the size of the phenomenon by searching for links in sets of historical data and recreating the most frequently occurring patterns [5]. Literature indicates that the past time series of GWL bring the most important information to GWL time series modelling [9]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
