Tracking the sources of dissolved organic matter under bio- and photo-transformation conditions using fluorescence spectrum-based machine learning techniques
Bibliographic record
Abstract
Dissolved organic matter (DOM) from various sources can lead to environmental issues such as eutrophication in agricultural watersheds. Effective source-tracking tools are needed to implement proper management practices. Fluorescence excitation–emission matrix (EEM) spectroscopy has been widely used to probe DOM composition. We explored optimal fluorescence EEM-based machine learning (ML) tools to quantify the proportions of different DOM sources in mixture samples under natural transformation conditions. Bulk DOM samples were prepared from soil and compost at various ratios and treated to simulate biogeochemical transformations. ML models based on all the EEM data outperformed those based on defined fluorescence indices. The trained support vector regression model (SVR) outperformed the conventional source tracking method of end-member mixing analysis (EMMA) with an R2 of 0.88 versus 0.83. Among the five suitable ML algorithms tested, SVR explained 90% and 85% of the variability in the proportions of soil and compost sources in the DOM mixture, with the mean squared errors of 0.004 and 0.007, respectively. The predicted capacity revealed a close relationship or causality between the specific mixing ratios of the bulk samples and the EEM spectra. The ML technique with EEM data was not constrained by the identification of all major sources, which is a required condition for the EMMA method. This study highlights the significant potential of EEM-based ML for tracing the source of DOM and establishes a basis for the future development of EEM data-driven models capable of tracking multiple DOM sources, even in the absence of all possible end-members.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".