Enhancing machine learning-based seasonal precipitation forecasting using CMIP6 simulations
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
The limited availability of observational and reanalysis data presents a significant challenge in training machine learning (ML) models for seasonal climate forecasting. Here, we show that training ML-based seasonal forecasting models with a larger number of individual simulations from CMIP6 models enhances their generalization ability and improves precipitation forecasts over South America. Using TelNet, a sequence-to-sequence machine learning model, we assess the performance of models trained with different numbers of CMIP6 simulations compared to those trained with ERA5 reanalysis and the CMIP6 ensemble mean. The results reveal that models trained with only a few CMIP6 simulations perform worse than those trained with ERA5, primarily due to instability during ML model tuning and reduced generalization ability. However, as the number of CMIP6 models increases, performance improves and surpasses both ERA5- and ensemble-mean-based ML models. Reliability and sharpness diagrams analysis further demonstrate that ML models trained with more CMIP6 simulations yield more confident and calibrated forecasts. Moreover, CMIP6-based TelNet constantly outperformed state-of-the-art dynamical models across different initialization months and lead times. This study underscores the potential of leveraging large multi-model dynamical simulations for robust ML-based seasonal climate forecasting. • Data Limitation Challenge: ML models struggle due to limited observational and reanalysis climate data. • Benefit of Using Many Simulations: Training on many CMIP6 simulations improves ML accuracy, surpassing ERA5 or ensemble models. • Improved Reliability and Performance: More CMIP6 simulations yield confident, well-calibrated forecasts that beat leading dynamical models.
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 | 0.002 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.001 | 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) | 0.002 | 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