Adedeji Itunu C, Ahmadisharaf Ebrahim, Sun Yanshuo
Department of Civil and Environmental Engineering, Resilient Infrastructure and Disaster Response Center, Florida A&M University-Florida State University College of Engineering, 2525 Pottsdamer St., Tallahassee, FL 32310, USA.
Department of Industrial and Manufacturing Engineering, Resilient Infrastructure and Disaster Response Center, Florida A&M University-Florida State University College of Engineering, 2525 Pottsdamer St., Tallahassee, FL 32310, USA.
J Contam Hydrol. 2022 Dec;251:104078. doi: 10.1016/j.jconhyd.2022.104078. Epub 2022 Sep 15.
Predicting in-stream water quality is necessary to support the decision-making process of protecting healthy waterbodies and restoring impaired ones. Data-driven modeling is an efficient technique that can be used to support such efforts. Our objective was to determine if in-stream concentrations of contaminants, nutrients-total phosphorus (TP) and total nitrogen (TN) -total suspended solids (TSS), dissolved oxygen (DO), and fecal coliform bacteria (FC) can be predicted satisfactorily using machine learning (ML) algorithms based on publicly available datasets. To achieve this objective, we evaluated four modeling scenarios, differing in terms of the required inputs (i.e., publicly available datasets (e.g., land-use/land cover)), antecedent conditions, and additional in-stream water quality observations (e.g., pH and turbidity). We implemented five ML algorithms-Support Vector Machines, Random Forest (RF), eXtreme Gradient Boost (XGB), ensemble RF-XGB, and Artificial Neural Network (ANN) -and demonstrated our modeling framework in an inland stream-Bullfrog Creek, located near Tampa, Florida. The results showed that, while including additional water quality drivers improved overall model performance for all target constituents, TP, TN, DO, and TSS could still be predicted satisfactorily using only publicly available datasets (Nash-Sutcliffe efficiency [NSE] > 0.75 and percent bias [PBIAS] < 10%), whereas FC could not (NSE < 0.49 and PBIAS >25%). Additionally, antecedent conditions slightly improved predictions and reduced the predictive uncertainty, particularly when paired with other water quality observations (6.9% increase in NSE for FC, and 2.7% for TP, TN, DO, and TSS). Also, comparable model performances of all water quality constituents in wet and dry seasons suggest minimal season-dependence of the predictions (<4% difference in NSE and < 10% difference in PBIAS). Our developed modeling framework is generic and can serve as a complementary tool for monitoring and predicting in-stream water quality constituents.
预测河流水质对于支持保护健康水体和恢复受损水体的决策过程至关重要。数据驱动建模是一种可用于支持此类工作的有效技术。我们的目标是确定基于公开可用数据集，使用机器学习（ML）算法能否令人满意地预测河流中污染物、营养物质（总磷（TP）和总氮（TN））、总悬浮固体（TSS）、溶解氧（DO）和粪大肠菌群（FC）的浓度。为实现这一目标，我们评估了四种建模方案，这些方案在所需输入（即公开可用数据集（如土地利用/土地覆盖））、前期条件和额外的河流水质观测值（如pH值和浊度）方面存在差异。我们实施了五种ML算法——支持向量机、随机森林（RF）、极端梯度提升（XGB）、集成RF-XGB和人工神经网络（ANN），并在佛罗里达州坦帕附近的一条内陆溪流——牛蛙溪展示了我们的建模框架。结果表明，虽然纳入额外的水质驱动因素可提高所有目标成分的整体模型性能，但仅使用公开可用数据集仍可令人满意地预测TP、TN、DO和TSS（纳什-萨特克利夫效率 [NSE] > 0.75且偏差百分比 [PBIAS] < 10%），而FC则不能（NSE < 0.49且PBIAS > 25%）。此外，前期条件略微改善了预测并降低了预测不确定性，特别是与其他水质观测值结合使用时（FC的NSE增加6.9%，TP、TN、DO和TSS增加2.7%）。而且，所有水质成分在湿季和干季的模型性能相当，表明预测对季节的依赖性最小（NSE差异<4%，PBIAS差异<10%）。我们开发的建模框架具有通用性，可作为监测和预测河流水质成分的补充工具。
