An official website of the United States government
ABSTRACT Biochar is a promising soil amendment for enhancing soil organic carbon (SOC), but accurately predicting its effect under diverse environmental conditions remains challenging due to complex, nonlinear interactions among biochar properties, soil characteristics, climate, and management practices. To address this research gap, we developed an ensemble machine learning (ML) framework, combining Extremely Randomized Trees (ExtraTrees), Light Gradient Boosting Machine (LightGBM), and Categorical Boosting (CatBoost) regressors, to model SOC responses to biochar application using a globally curated dataset of 800 field observations. The ensemble model showed strong predictive performance (R2 = 0.86, RMSE = 0.11) and generalized well across a wide range of conditions. Shapley Additive exPlanations (SHAP) analysis identified biochar addition rates, crop types, soil type, and soil pH were the most influential predictors of SOC changes. The most effective biochar application rate was about 40 t/ha, and the saturation point was 121.7 t/ha. Partial dependence plots revealed nonlinear and threshold effects of pyrolysis temperature, initial SOC levels, and nitrogen content. To facilitate practical application, we also developed a user‐friendly graphical interface for SOC prediction under various biochar‐soil‐climate scenarios. This work highlights the predictive power and interpretability of ML tools in digital soil carbon modeling and supports data‐driven strategies for optimizing biochar use in climate‐smart agriculture.
more »
« less
Quantifying Multidimensional Effects of Physicochemical Parameters on PFAS Adsorption Using a Hybrid Response Surface Methodology-Machine Learning Approach
Per- and polyfluoroalkyl substances (PFAS) contamination has posed a significant environmental and public health challenge due to their ubiquitous nature. Adsorption has emerged as a promising remediation technique, yet optimizing adsorption efficiency remains complex due to the diverse physicochemical properties of PFAS and the wide range of adsorbent materials. Traditional modeling approaches, such as response surface methodology (RSM), struggled to capture nonlinear interactions, while standalone machine learning (ML) models required extensive datasets. This study addressed these limitations by developing hybrid RSM-ML models to improve the prediction and optimization of PFAS adsorption. A comprehensive dataset was constructed using experimental adsorption data, integrating key parameters such as pH, pHpzc, surface area, temperature, and PFAS molecular properties. RSM was employed to model adsorption behavior, while gradient boosting (GB), random forest (RF), and extreme gradient boosting (XGB) were used to enhance predictive performance. Hybrid models—linear, RMSE-based, multiplicative, and meta-learning—were developed and evaluated. The meta-learning HOP-RSM-GB model achieved near-perfect accuracy (R² = 1.00, RMSE = 10.59), outperforming all other models. Surface plots revealed that low pH and high pHpzc maximized the adsorption while increasing log Kow consistently enhanced PFAS adsorption. These findings establish hybrid RSM-ML modeling as a powerful framework for optimizing PFAS remediation strategies. The integration of statistical and machine learning approaches significantly improves predictive accuracy, reduces experimental costs, and provides deeper insights into adsorption mechanisms. This study underscores the importance of data-driven approaches in environmental engineering and highlights future opportunities for integrating ML-driven modeling with experimental adsorption research.
more »
« less
- Award ID(s):
- 2216148
- PAR ID:
- 10632024
- Publisher / Repository:
- American Chemical Society
- Date Published:
- Format(s):
- Medium: X
- Institution:
- North Carolina A and T State University
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Per- and polyfluoroalkyl substances (PFAS) are a type of emerging contaminant associated with significant health risks, such as carcinogenicity, endocrine disruption, and immunotoxicity. The persistent nature and widespread occurrence of PFAS, particularly perfluoroalkyl carboxylic acids such as perfluorooctanoic acid (PFOA) and perfluorononanoic acid (PFNA), necessitate the development of efficient remediation strategies. In this study, granular activated carbon (GAC) was modified with cetyltrimethylammonium chloride (CTAC) to enhance the adsorption capacity of PFAS. The surface morphology and physicochemical properties of unmodified and modified GAC (MGAC) were characterized using scanning electron microscopy and Brunauer–Emmett–Teller analysis, revealing significant surface and pore structure alterations following CTAC modification. Batch adsorption experiments demonstrated that MGAC exhibited superior adsorption capacities for PFOA (93.18 mg/g) and PFNA (130.94 mg/g) compared with unmodified GAC (90.53 mg/g and 103.71 mg/g, respectively). The observed increase in adsorption capacity is attributable to enhanced electrostatic and hydrophobic interactions facilitated by the quaternary ammonium groups and hydrophobic alkyl chains of CTAC. However, competitive adsorption tests indicated a reduction in PFNA adsorption efficiency, suggesting that PFAS cooccurrence affects adsorption dynamics. Adsorption isotherms were best described by the Langmuir model, indicating monolayer adsorption on homogeneous surfaces. The findings underscore the efficacy of surfactant-modified GAC in enhancing PFAS removal from aqueous environments and provide critical insights into the mechanisms governing PFAS adsorption under single and mixed-species scenarios. This study advances the development of adsorbents for PFAS remediation, addressing the challenges of complex environmental matrices and cooccurring PFAS species.more » « less
-
Abstract Harmful phytoplankton blooms driven by climate warming and nutrient pollution are a growing threat to freshwater ecosystems worldwide. Predicting these blooms is critical for managing water resources. However, process‐based models often struggle to capture the complex nonlinear dynamics of phytoplankton. Although machine learning (ML) offers powerful predictive capabilities, its “black‐box” nature has limited its adoption for management. In this study, we applied four ensemble ML algorithms, Extreme Gradient Boosting (XGBoost), Random Forest (RF), Gradient Boosting Machine (GBM), and Categorical Boosting (CatBoost), to model phytoplankton dynamics in two adjacent drinking‐water reservoirs. All models performed similarly; however, the XGBoost algorithm achieved the best performance in both reservoirs, with a root mean square error (RMSE) of 2.4–6.6 μg/L chlorophyll a and a Pearson correlation coefficient (r) of 0.83–0.86. To enhance the interpretability of these ML models, we applied explainable artificial intelligence (XAI) techniques, including Shapley Additive Explanations (SHAP) and partial dependence analyses (PDPs). Our XAI analysis revealed reservoir‐specific dynamics, with deep dissolved oxygen and water temperature more influential in one reservoir, whereas seasonality and water column stability were more critical in the other. We also observed opposing effects of thermal stratification, with high water temperatures and strong stratification stimulating phytoplankton in one reservoir and suppressing it in the other. Stability analyses confirmed that the SHAP explanations were robust to perturbations, boosting confidence in their interpretability. This study demonstrates the application of a stability‐constrained XAI framework to provide transparent and trustworthy ecological insights, offering a robust approach for data‐driven water quality management and forecasting.more » « less
-
Abstract The application of machine learning (ML) techniques in materials science has revolutionized the pace and scope of materials research and design. In the case of metal–organic frameworks (MOFs), a promising class of materials due to their tunable properties and versatile applications in gas adsorption and separation, ML has helped survey the vast material space. This study explores the integration of reinforcement learning (RL), specifically Q‐learning, within an active learning (AL) context, combined with Gaussian processes (GPs) for predictive modeling of adsorption in MOFs. We demonstrate the effectiveness of the RL‐driven framework in guiding the selection of training data points and optimizing predictive model performance for methane and carbon dioxide adsorption, using two different reward metrics. Our results highlight the integration of RL as an AL method for adsorption predictions in MFs, and how it compares to a previously implemented AL scheme.more » « less
-
Land surface temperature (LST) is crucial for understanding earth system processes. We expanded the Advanced Baseline Imager Live Imaging of Vegetated Ecosystems (ALIVE) framework to estimate LST in near‐real‐time for both cloudy and clear sky conditions at a five‐minute resolution. We compared two machine learning (ML) models, Long Short‐Term Memory (LSTM) networks and Gradient Boosting Regressor (GBR), using top‐of‐atmosphere observations from the Advanced Baseline Imager (ABI) on the GOES‐16 satellite against observations from hundreds of observation sites for a five‐year period. Long Short‐Term Memory outperformed GBR, especially at coarser resolutions and under challenging conditions, with a clear sky R2of 0.96 (RMSE 2.31K) and a cloudy sky R2of 0.83 (RMSE 4.10K) across CONUS, based on 10‐repeat Leave‐One‐Out Cross‐Validation (LOOCV). GBR maintained high accuracy and ran 5.3 times faster, with only a 0.01–0.02 R2drop. Feature importance revealed infrared bands were key in both models, with LSTM adapting dynamically to atmospheric changes, while GBR utilized more time information in cloudy conditions. A comparative analysis against the physically based ABILSTproduct showed strong agreement in winter, particularly under clear sky conditions, while also highlighting the challenges of summer LST estimation due to increased thermal variability. This study underscores the strengths and limitations of data‐driven models for LST estimation and suggests potential pathways for integrating ML models to enhance the accuracy and coverage of LST products.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript
- Posted Content:
- https://doi.org/10.26434/chemrxiv-2025-3wl26
