Prediction of ambient PM2.5 chemical components in Southern California using machine learning
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Fine particulate matter (PM 2.5 , particulate matter with an aerodynamic diameter ≤2.5 μm) poses major public health and environmental risks, yet the toxicity of its chemical components remains poorly understood due to limited chemical speciation data. In this study we apply an extreme gradient boosting (XGBoost) machine learning framework to predict key PM 2.5 components including organic carbon, elemental carbon, nitrate, sulfate, ammonium, and metals, using readily available predictors: total PM 2.5 mass concentrations, meteorological variables, trace gas measurements, and indicators of exceptional events (e.g., wildfires, fireworks). Leveraging a decade of data from two monitoring sites in Southern California (Los Angeles and Rubidoux), the models achieved strong predictive performance, particularly for nitrate, ammonium, and elemental carbon. Among the most influential predictors across components were total PM 2.5 mass, relative humidity, and boundary layer height. This approach has promise for enhancing satellite remote sensing applications, improving chemical transport model inputs, and generating cost-effective estimates of PM 2.5 components during sampling gaps and in regions lacking frequent monitoring. Further research is needed to assess the generalizability of this framework across diverse geographic and climatic settings. • Machine learning models accurately predict daily PM 2.5 chemical components • Nitrate, ammonium, and organic carbon show the highest predictive performance • Relative humidity, PM 2.5 mass, and NO 2 are key predictors identified by SHAP • The framework addresses data gaps in chemical speciation monitoring networks • Results support satellite applications and cost-effective air quality assessment
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.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| 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 it