Water quality modelling using combination of support vector regression with sequential minimal optimization for akkopru stream in Van, Turkey

Aldemir A.

Fresenius Environmental Bulletin, vol.30, no.2, pp.1518-1526, 2021 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 30 Issue: 2
  • Publication Date: 2021
  • Journal Name: Fresenius Environmental Bulletin
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Aerospace Database, Aqualine, Aquatic Science & Fisheries Abstracts (ASFA), CAB Abstracts, Chemical Abstracts Core, Communication Abstracts, Environment Index, Geobase, Greenfile, Metadex, Pollution Abstracts, Veterinary Science Database, Civil Engineering Abstracts
  • Page Numbers: pp.1518-1526
  • Keywords: Water quality modelling, Support vector regression, Sequential minimal optimization, Akkopru stream, NEURAL-NETWORK, RIVER, ALGORITHM, RESERVOIR, REGION
  • Van Yüzüncü Yıl University Affiliated: Yes

Abstract

© by PSPWater sources pollution is a great environmental problem that negatively effects on life. The first step to prevent this environmental problem should be determined pollution by monitoring the water quality in sources. In this research, studies on the field and in the laboratory, water quality parameters of Akkoprti stream located in the province of Van, were analyzed. The in situ measurements and analyses were made on water samples taken from nine different sampling points on the Akkoprti stream monthly in 2018. Measurements and analysis of water quality parameters and pollution loads (ammonium, nitrite, nitrate, phosphate, chlorine, COD, DO, pH, conductivity vs.) were made and the model based on the four important variables (ammonium, nitrate, phosphate, COD) were simulated. Along the stream length, the mean absolute percentage errors were calculated to compare simulation data with the experimental data. According to the results, model predictions with experimental data has been observed that a significant degree of alignment. The important indicators of stream ammonium nitrogen, nitrate nitrogen, phosphate and chemical oxygen demand in the average absolute error values, for these variables, calculated. It is observed that the water flow regime of the stream is irregular and this is reflected in the water quality. The quality of stream water could be ranked as I and II classes and it has suitable quality concerning aquaculture and irrigation.