TY - JOUR
T1 - Groundwater productivity potential mapping using frequency ratio and evidential belief function and artificial neural network models
T2 - Focus on topographic factors
AU - Kim, Jeong Cheol
AU - Jung, Hyung Sup
AU - Lee, Saro
N1 - Publisher Copyright:
© IWA Publishing 2018.
PY - 2018/11
Y1 - 2018/11
N2 - This study analysed groundwater productivity potential (GPP) using three different models in a geographic information system (GIS) for Okcheon city, Korea. Specifically, we have used variety topography factors in this study. The models were based on relationships between groundwater productivity (for specific capacity (SPC) and transmissivity (T)) and hydrogeological factors. Topography, geology, lineament, land-use and soil data were first collected, processed and entered into the spatial database. T and SPC data were collected from 86 well locations. The resulting GPP map has been validated in under the curve analysis area using well data not used for model training. The GPP maps using artificial neural network (ANN), frequency ratio (FR) and evidential belief function (EBF) models for T had accuracies of 82.19%, 81.15% and 80.40%, respectively. Similarly, the ANN, FR and EBF models for SPC had accuracies of 81.67%, 81.36% and 79.89%, respectively. The results illustrate that ANN models can be useful for the development of groundwater resources.
AB - This study analysed groundwater productivity potential (GPP) using three different models in a geographic information system (GIS) for Okcheon city, Korea. Specifically, we have used variety topography factors in this study. The models were based on relationships between groundwater productivity (for specific capacity (SPC) and transmissivity (T)) and hydrogeological factors. Topography, geology, lineament, land-use and soil data were first collected, processed and entered into the spatial database. T and SPC data were collected from 86 well locations. The resulting GPP map has been validated in under the curve analysis area using well data not used for model training. The GPP maps using artificial neural network (ANN), frequency ratio (FR) and evidential belief function (EBF) models for T had accuracies of 82.19%, 81.15% and 80.40%, respectively. Similarly, the ANN, FR and EBF models for SPC had accuracies of 81.67%, 81.36% and 79.89%, respectively. The results illustrate that ANN models can be useful for the development of groundwater resources.
KW - Artificial neural network
KW - Frequency ratio
KW - GIS
KW - Groundwater productivity potential
KW - South Korea
UR - http://www.scopus.com/inward/record.url?scp=85055508375&partnerID=8YFLogxK
U2 - 10.2166/hydro.2018.120
DO - 10.2166/hydro.2018.120
M3 - Article
AN - SCOPUS:85055508375
SN - 1464-7141
VL - 20
SP - 1436
EP - 1451
JO - Journal of Hydroinformatics
JF - Journal of Hydroinformatics
IS - 6
ER -