Prediction of Photochemical Properties of Dissolved Organic Matter Using Machine Learning

Abstract

Apparent quantum yields (Φ) of photochemically produced reactive intermediates (PPRIs) formed by dissolved organic matter (DOM) are vital to element cycles and contaminant fates in surface water. Simultaneous determination of Φ_PPRI values from numerous water samples through existing experimental methods is time consuming and ineffective. Herein, machine learning models were developed with a systematic data set including 1329 data points to predict the values of three Φ_PPRIs (Φ_3DOM*, Φ_1O2, and Φ_·OH) based on DOM spectral parameters, experimental conditions, and calculation parameters. The best predictive performances for Φ_3DOM*, Φ_1O2, and Φ_·OH were achieved using the CatBoost model, which outperformed the traditional linear regression models. The significances of the wavelength range and spectral parameters on the three Φ_PPRI predictions were revealed, suggesting that DOM with lower molecular weight, lower aromatic content, and a more autochthonous portion possessed higher Φ_PPRIs. Chain models were constructed by adding the predicted Φ_3DOM* as a new feature into the Φ_1O2 and Φ_·OH models, which consequently improved the predictive performance of Φ_1O2 but worsened the Φ_·OH prediction likely due to the complex formation pathways of ·OH. Overall, this study offered robust Φ_PPRI prediction across interlaboratory differences and provided new insights into the relationship between PPRIs formation and DOM properties.