Predicting solar photovoltaic generation impacted by severe wildfire smoke
Bibliographic record
Abstract
Abstract The negative impact of wildfire smoke on solar photovoltaic (PV) generation by reducing the amount of solar irradiance reaching the modules has been observed worldwide. However, the predictive capability to capture the impact on solar electricity production still needs to be improved. For example, in the summer of 2023, smoke from Canadian wildfires spread to the northeastern U.S., impacting solar PV output in the region. The New York Independent System Operator (NYISO) day-ahead forecasts for this period significantly overpredicted PV output. This paper presents novel machine learning-based models for predicting the hourly solar capacity factor, focusing on improving predictive performance during periods of severe wildfire smoke. The results demonstrate a R 2 value of up to 0.85 for the severe wildfire periods (aerosol optical depth (AOD) above the 99.99th percentile) from our models, significantly outperforming NYISO’s R 2 value of 0.50 across six load zones included in the analysis. The greatly enhanced performance arises from two innovations. First, we adopted a series of data products, newly available in the public domain, from the high-resolution rapid refresh smoke (HRRR-Smoke) weather forecasting system. These include predictions of the AOD and the downward shortwave radiation flux incorporating aerosol impacts. Our study marks the first time the HRRR-Smoke wildfire AOD product has been used in solar electricity forecasts. Second, we employed upsampling strategies to address the data imbalance issues due to the inherently infrequent nature of wildfire events. As the data products are publicly available, our methodology can be readily adopted by power system operators to enhance predictions of solar electricity production during periods of wildfire smoke, ensuring the reliability of power grids with high penetration of solar energy.
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 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 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".