TY - JOUR
T1 - Developing a data-driven modeling framework for simulating a chemical accident in freshwater
AU - Kim, Soobin
AU - Abbas, Ather
AU - Pyo, Jong Choel
AU - Kim, Hyein
AU - Hong, Seok Min
AU - Baek, Sang Soo
AU - Cho, Kyung Hwa
N1 - KAUST Repository Item: Exported on 2023-10-04
Acknowledgements: This work was supported by the Korea Environment Industry & Technology Institute (KEITI) through the Advanced Technology Development Project for Predicting and Preventing Chemical Accidents Project, funded by the Korea Ministry of Environment (MOE) (2022003620001). This work was supported by the Korea Environment Industry & Technology Institute (KEITI) through the Aquatic Ecosystem Conservation Research Program, funded by the Korea Ministry of Environment (MOE) (2020003030003).
PY - 2023/9/29
Y1 - 2023/9/29
N2 - Chemical accidents in freshwater pose threats to public health and aquatic ecosystems. Process-based models (PBMs) have been used to identify spatiotemporal chemical distributions in natural water. However, their computationally expensive simulations can hinder timely incident responses, which are crucial for minimizing negative impacts. Therefore, this study proposes a site-specific data-driven model (DDM) to supplement PBM-based chemical accident simulations. A convolutional neural network (CNN) was employed as the DDM because of its outstanding performance in capturing spatial patterns. Our model was developed to facilitate chemical accident simulations in the Namhan River, South Korea. The model datasets were generated using the PBM simulation outputs from toluene accident scenarios. Our DDM showed a Nash-Sutcliffe-efficiency of 0.94 and a root-mean-square-error of 0.023 μg/L for the validation set. Its computational time was approximately 64 times faster than that of PBMs. In addition, this study interpreted the DDM results using SHapley Additive exPlanations (SHAP). The SHAP findings highlighted the influential role of distance from the accident site in this study. Overall, this study demonstrated the applicability of our modeling approach in freshwater chemical accidents by providing rapid spatial distribution results complementing PBM simulations.
AB - Chemical accidents in freshwater pose threats to public health and aquatic ecosystems. Process-based models (PBMs) have been used to identify spatiotemporal chemical distributions in natural water. However, their computationally expensive simulations can hinder timely incident responses, which are crucial for minimizing negative impacts. Therefore, this study proposes a site-specific data-driven model (DDM) to supplement PBM-based chemical accident simulations. A convolutional neural network (CNN) was employed as the DDM because of its outstanding performance in capturing spatial patterns. Our model was developed to facilitate chemical accident simulations in the Namhan River, South Korea. The model datasets were generated using the PBM simulation outputs from toluene accident scenarios. Our DDM showed a Nash-Sutcliffe-efficiency of 0.94 and a root-mean-square-error of 0.023 μg/L for the validation set. Its computational time was approximately 64 times faster than that of PBMs. In addition, this study interpreted the DDM results using SHapley Additive exPlanations (SHAP). The SHAP findings highlighted the influential role of distance from the accident site in this study. Overall, this study demonstrated the applicability of our modeling approach in freshwater chemical accidents by providing rapid spatial distribution results complementing PBM simulations.
UR - http://hdl.handle.net/10754/694845
UR - https://linkinghub.elsevier.com/retrieve/pii/S0959652623030007
UR - http://www.scopus.com/inward/record.url?scp=85172261610&partnerID=8YFLogxK
U2 - 10.1016/j.jclepro.2023.138842
DO - 10.1016/j.jclepro.2023.138842
M3 - Article
SN - 0959-6526
VL - 425
SP - 138842
JO - Journal of Cleaner Production
JF - Journal of Cleaner Production
ER -
