Evaluation of continental-scale ensemble hydrological forecasts from Environment and Climate Change Canada: a comparison with forecasts from the Global Flood Awareness System (GloFAS)
Bibliographic record
Abstract
Environment and Climate Change Canada (ECCC) is in the process of deploying a continental-scale hydrological prediction system known as the National Surface and River Prediction System (NSRPS). Operating at a resolution of 30 arc seconds, NSRPS currently generates Ensemble Streamflow Predictions (ESPs) for over four million grid points. Issued once per day, the ensemble is composed of twenty members and provides 16-day forecasts. NSRPS differs from other hydrological forecasting systems in operational use in Canada by being continental in scope and by relying on an Earth System Modelling (ESM) approach for prediction. In order to assess the value of forecasts issued by NSRPS, a comparison is performed with a similar ESP system available over all of Canada: the Global Flood Awareness System (GloFAS) from the European Centre for Medium-Range Weather Forecasts (ECMWF). The evaluation focusses on the Great Lakes and St. Lawrence watershed as well as the Nelson and Churchill watersheds, each over one million km² in size. 393 streamflow stations are identified where NSRPS and GloFAS agree on the watershed delineation. The comparison is limited to the Spring, Summer and Fall of 2022 due to NSRPS forecast data availability. For most stations, NSRPS performs better than GloFAS in terms of Continuous Ranked Probability Score (CRPS), but the median of the potential CRPS across the 393 stations is very similar for days 3-16. Both systems suffer from a lack of spread, particularly for short lead times, but the problem is slightly more acute for GloFAS. Bayesian Model Averaging (BMA) is explored in order to obtain calibrated probabilistic forecasts that perform better than both NSRPS and GloFAS.
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 | 0.003 | 0.006 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| 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".