- Rating
- 3
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 3
- Soundness
- 2
Read review
Paper summary
This paper studies how to adapt univariate time-series foundation models to multi-station weather forecasting, where both spatial dependencies and foundation-model-specific forecasting errors should be considered. The authors propose WxFM-XL, which constructs frequency-specific cross-station error correlation graphs and adaptively fuses them with a spatial correlation graph to guide a lightweight graph residual adapter. Experiments on several multi-station weather datasets using Timer and Sundial demonstrate improvements over various forecasting and foundation model adaptation baselines.
Critical strengths
- Clear task-specific motivation. The paper identifies a meaningful limitation of applying univariate TSFMs directly to multi-station weather forecasting, where geographic relationships and station-specific prediction errors are not explicitly modeled. The observation that geographically distant stations can exhibit correlated forecasting errors provides an interesting motivation for going beyond conventional spatial proximity graphs.
- Reasonable and coherent adaptation framework. The integration of spatial correlations, frequency-specific error correlations, and residual correction is well aligned with the problem formulation. Freezing the pretrained foundation model while introducing a graph-based adapter is a practical strategy for exploiting existing temporal priors while learning task-specific dependencies.
- Competitive empirical performance. The paper evaluates the proposed approach on multiple regional and global weather datasets, with comparisons against traditional spatiotemporal forecasting models and foundation model adaptation approaches. The ablation results provide initial evidence that frequency decomposition and error correlation graphs contribute to the overall performance.
Critical weaknesses
- Incomplete comparison with recent multivariate time-series foundation models. The evaluation mainly considers Timer, Sundial, Moirai, Timer-XL, and AdaPTS, but omits recent multivariate-capable TSFMs such as Chronos-2. Since the paper explicitly targets the limitations of extending univariate foundation models to multivariate or multi-station forecasting, comparisons with stronger recent models that natively support multivariate dependencies are important for establishing the competitiveness and practical relevance of WxFM-XL. Without such comparisons, it is difficult to determine whether the proposed adaptation framework offers advantages over modern multivariate foundation models.
- Insufficient evidence for the unique value of foundation-model-specific error correlations. The paper shows that adding error correlation graphs improves performance, but does not sufficiently demonstrate that these graphs provide information beyond conventional cross-station dependencies. In particular, comparisons with graphs constructed from raw observations, foundation model predictions, or directly learned station relationships are missing. Consequently, it remains unclear whether the improvement is specifically attributable to modeling foundation-model-specific error priors or simply to introducing additional cross-station connectivity.
- The necessity of frequency-specific graphs and dynamic fusion is not fully validated. Removing frequency decomposition or error correlation graphs degrades performance, but these ablations change multiple aspects of the model simultaneously. Comparisons between a shared error correlation graph and separate frequency-specific graphs are needed to establish the necessity of frequency-dependent error structures. Furthermore, the proposed fusion primarily relies on learned frequency-specific coefficients rather than genuinely input-dependent dynamic relationships, making the claimed dynamic adaptation somewhat limited. Comparisons with simpler fixed fusion strategies and more expressive dynamic gating mechanisms would strengthen the justification.
- Limited analysis of the stability and transferability of error correlation priors. Error correlation graphs are estimated from historical training residuals, yet these relationships may vary across seasons, weather regimes, and geographic regions. The paper does not sufficiently examine whether the estimated graphs remain reliable under temporal distribution shifts or when adapting to new regions with limited labeled observations. This raises concerns about the robustness and generalizability of the proposed error-prior-based adaptation mechanism.
- The efficiency advantages of lightweight adaptation are not fully established. Although the foundation model is frozen, constructing cross-station error graphs and introducing graph-based residual adapters may incur additional computational and memory costs, particularly for datasets containing thousands of stations. A systematic comparison of training cost, inference latency, memory consumption, and parameter efficiency would be useful to substantiate the claimed practicality of the approach.
Questions for authors
See weakness
4 comments
Response to Reviewer Q2UQ (Part 2 / 2): W4–W5
In this second part, we address W4–W5 by evaluating robustness across seasons and weather regimes, adaptation under sparse regional station coverage, and computational efficiency.
W4: Robustness across seasons, weather regimes, and geographic regions.
We supplemented the evaluation using WxFM-XL (Timer) on the Global dataset (24h; MSE; seed 2021).
Across seasons. We compared the static training graph with seasonally re-estimated graphs, each constructed from all observed foundation-model forecast errors within a seasonal test block and used to evaluate that block. The August–December period covers summer/autumn/winter in the Northern Hemisphere and winter/spring/summer in the Southern Hemisphere.
Season-specific re-estimation changes overall MSE by only +0.52% for temperature and +0.02% for wind. The small wind improvements in two blocks show limited benefits from seasonal re-estimation, supporting the robustness of the static training graph under the observed seasonal changes.
| Season: Northern / Southern Hemisphere (months) | Temp: static | Temp: re-estimated | Wind: static | Wind: re-estimated |
|---|---|---|---|---|
| Summer / Winter (August) | 8.7724 | 8.8580 | 2.9255 | 2.9342 |
| Autumn / Spring (September–November) | 7.0521 | 7.0739 | 3.6324 | 3.6320 |
| Winter / Summer (December) | 8.4611 | 8.5241 | 3.9483 | 3.9471 |
Across weather regimes. We evaluated cold/normal/hot temperature and calm/normal/strong wind conditions, using thresholds determined from training data. Here, the re-estimated graph is periodically updated using recent foundation-model forecast errors whose corresponding observations are available before each prediction, whereas the static training graph remains unchanged.
Re-estimation does not improve forecasting accuracy in any evaluated weather condition. These results support the robustness of the static training graph and suggest that frequent re-estimation is unnecessary under the observed weather changes.
| Variable / condition | Windows | Static training graph | Re-estimated graph |
|---|---|---|---|
| Temperature / cold | 18 | 10.7720 | 11.0133 |
| Temperature / normal | 3,017 | 7.5107 | 7.6319 |
| Temperature / hot | 283 | 8.1057 | 8.1897 |
| Wind / calm | 491 | 2.8567 | 2.8852 |
| Wind / normal | 2,670 | 3.6292 | 3.6771 |
| Wind / strong | 157 | 5.2358 | 5.3428 |
Across geographic regions. We further evaluated regional adaptation on an Australia-region subset of the Global dataset, containing 90 geographically dispersed stations with a median nearest-neighbor distance of 108.62 km. We kept the pretrained Timer frozen, constructed regional spatial and error correlation prior graphs using training-period data, and trained a newly initialized lightweight adapter.
WxFM-XL reduces temperature and wind MSE by 23.08% and 8.24%, respectively, compared with Timer alone. These results demonstrate effective adaptation under sparse station coverage by reusing pretrained temporal knowledge and learning region-specific residual corrections, without retraining the foundation model.
| Method | Stations | Temp-24 | Wind-24 |
|---|---|---|---|
| Timer | 90 | 10.557 | 5.256 |
| WxFM-XL (Timer) | 90 | 8.121 | 4.822 |
W5: Computational efficiency and practical feasibility.
Parameter efficiency. We added empirical comparisons of trainable parameters. WxFM-XL keeps Timer frozen and trains only 0.493M parameters, reducing trainable parameters by 99.4% relative to full fine-tuning.
| Adaptation method | Trainable parameters |
|---|---|
| Timer full-FT | 84.142M |
| Timer LoRA | 0.524M |
| WxFM-XL (Timer) | 0.493M |
Computational cost. We measured graph-construction time, training time, inference latency and peak memory across increasing station counts. Profiling on the Global dataset uses a 96h input and forecast horizon with a fixed top-k neighborhood size of 36.
Error-graph construction is performed offline and includes residual preparation, correlation computation and top-k selection. Training cost is measured per batch, and inference latency includes both the foundation model and adapter. These measurements demonstrate practical operation at thousands of stations.
| Dataset | Stations | Error graph (s) | Training (ms/batch) | FM+adapter inference (ms/window) | Training GPU peak (GiB) | Graph CPU peak (GiB) |
|---|---|---|---|---|---|---|
| Global | 250 | 0.640 | 46.1 | 15.4 | 0.63 | 1.38 |
| Global | 500 | 1.200 | 50.4 | 20.9 | 0.87 | 1.57 |
| Global | 1,000 | 2.625 | 45.7 | 26.9 | 1.36 | 1.94 |
| Global | 2,000 | 6.082 | 47.2 | 39.5 | 2.44 | 2.67 |
| Global | 3,850 | 19.017 | 93.5 | 77.2 | 4.82 | 4.11 |
The authors’ response strengthens the evaluation. The shared-graph comparison and the runtime measurements at 3,850 stations address my concerns about the graph design and practical cost. The new comparisons with Chronos-2 and alternative graphs are useful, though they cover one dataset, one forecast horizon, and one random seed. My overall assessment remains positive, and I retain my rating of 3 (Weak acceptance).
We are deeply thankful for the reviewer's timely acknowledgement and feedback. We are glad that our responses have addressed your concerns. Thank you again for your appreciation of our work and insightful comments.
Response to Reviewer Q2UQ (Part 1 / 2): W1–W3
Thank you for your constructive feedback. In this first part, we address W1–W3 with Chronos-2 comparisons, alternative station-relation graphs, and ablations of frequency-specific graphs and fusion strategies.
W1: Comparison with Chronos-2.
We added zero-shot and full fine-tuning comparisons with Chronos-2 on the Global dataset (24h; MSE; seed 2021). Both WxFM-XL variants achieve lower MSE than zero-shot and fine-tuned Chronos-2 on both variables, supporting their competitiveness against a recent multivariate foundation model.
W2: Comparison with alternative station-relation graphs.
We added comparisons with graphs constructed from (1) raw observations, (2) foundation-model predictions, and (3) directly learned station relationships, together with a geographic-proximity baseline. Results use WxFM-XL (Timer) on the Global dataset (24h; MSE; seed 2021).
The foundation-model-specific error correlation prior graphs achieve the lowest MSE on both variables. These comparisons support attributing the improvements to modeling foundation-model-specific error priors, beyond simply introducing additional cross-station connectivity.
W3: Frequency-specific error graphs and fusion strategies.
We compare frequency-specific error correlation prior graphs with a shared full-frequency error graph, retaining frequency decomposition to isolate the effect of graph construction. We additionally compare our learnable fusion with fixed coefficients and prediction-conditioned dynamic gating. Results use WxFM-XL (Timer) on the Global dataset (24h; MSE; seed 2021).
Prediction-conditioned dynamic gating generates fusion weights from the current foundation-model forecasts, allowing them to vary across samples and providing a more expressive fusion mechanism.
Frequency-specific graphs outperform the shared graph, supporting the value of modeling frequency-dependent error structures. Our learnable fusion also achieves lower MSE than fixed coefficients and prediction-conditioned dynamic gating on both variables. These results support learning a separate fusion weight for each frequency band to effectively integrate spatial relations and error correlation priors.