Modeling High-Resolution Spatio-Temporal Wind with Deep Echo State Networks and Stochastic Partial Differential Equations

Kesen Wang; Marc G. Genton; Minwoo Kim; Stefano Castruccio

REVIEW 3 major objections 6 minor 58 references

Modeling High-Resolution Spatio-Temporal Wind with Deep Echo State Networks and Stochastic Partial Differential Equations

T0 review · 3 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A three-stage pipeline of support-point reduction, batch-updated deep echo state networks, and non-stationary SPDE interpolation yields more accurate two- and three-hour-ahead wind forecasts for Saudi Arabia than the closest rival, saving…

desk verdict Solid engineering extension undone by a test-set leak: hyperparameters and calibration are tuned on the 2016 test year, so the headline forecast gains and dollar savings are in-sample. read the letter →

arxiv 2412.07265 v1 pith:EBFOQIZC submitted 2024-12-10 stat.ML cs.LG

Kesen Wang , Minwoo Kim , Stefano Castruccio , Marc G. Genton This is my paper · ORCID

classification stat.MLcs.LG MSC 62M3062M1060G6068T07

keywords EchoStateNetworkStochasticPartialDifferentialEquationSupportPointsSpatio-TemporalForecastingWindEnergyDimensionReductionNonstationarySpatialProcessForecastCalibration

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to establish that a hybrid statistical-machine-learning pipeline can forecast country-scale wind fields at the short lead times grid operators need, in a country with almost no observational network. The pipeline reduces 53,333 simulated wind locations to 3,173 support points, models the dynamics of those knots with a deep echo state network that re-estimates its output weights every 75 hours, and reconstructs the full wind field with a non-stationary stochastic partial differential equation. Compared with the closest competing approach, the pipeline reports lower median squared forecast error at two- and three-hour leads, and essentially equal error at one hour, both at the knots and across all locations. Converting speed to power at 75 optimal inland farm sites, the annual absolute energy error drops by about $1.69 \times 10^7$ kWh, which the paper values at up to $1.18 million per year in avoided misprediction cost. If true, this offers a practical route to accurate wind-energy forecasting without dense in-situ observations.

What carries the argument

Three linked components carry the argument. Support points minimize the energy distance $$E(F,F_n)=\frac{2}{n}\sum_i E\|X-s_i\|_2 - \frac{1}{$n^{2}$}\sum_{i,j}\|s_i-s_j\|_2 - E\|X-X'\|_2$$ to select $n_{\mathrm{red}}=3{,}173$ knots that best represent the empirical distribution of the $53{,}333$ locations. A deep echo state network with sparse spike-and-slab random weight matrices and quadratic output states models the temporal dynamics at the knots; the batch update re-estimates the ridge-regression output weights every $b=75$ hours. A non-stationary SPDE, $$(\$kappa^{2}$(s)-\$\Delta$)^{\$\alpha$/2}(\tau(s)Y(s))=\mathcal{W}(s),$$ with $\log\kappa(s)$ and $\log\tau(s)$ expanded in Fourier basis functions, is discretized by finite elements into a Gaussian Markov random field, giving sparse-precision interpolation from the knots to the whole domain. Calibration then shrinks the SPDE covariance toward the empirical covariance with a parameter $\delta$ to make prediction intervals reach nominal coverage.

What would settle it

Compare B-ESN and ESN forecasts against independent hourly wind measurements or a reanalysis product at dozens of stations not used in the ten-station validation; if the two-hour, all-location median MSPE advantage (0.133 vs 0.141) does not persist after correcting for WRF bias, the central claim fails.

Watch

Extended reading notes

Core claim

On its own terms, the paper establishes that the proposed B-ESN pipeline, using support-point reduction to 3,173 knots, a deep echo state network with output weights re-estimated every $b=75$ hours, and interpolation through a non-stationary SPDE, produces more accurate and more stable wind speed forecasts at two- and three-hour leads than the closest competing ESN approach. At the knots the median MSPE falls from 0.125 to 0.115 for two-hour leads and from 0.173 to 0.157 for three-hour leads; across all 53,333 locations it falls from 0.141 to 0.133 and from 0.189 to 0.173. At the 75 optimal inland wind-farm sites, the annual absolute wind-energy difference drops from $2.770 \times 10^8$ kWh to $2.601 \times 10^8$ kWh, which the paper values at up to $\$1.183$ million per year in avoided misprediction cost.

Load-bearing premise

The forecast improvements are measured against WRF-simulated wind that was validated at only ten monitoring stations, so a bias in the simulator over unobserved regions, especially complex terrain and the Empty Quarter, would erase the reported gains in real-world operation.

Editorial extensions

If this is right

For energy grid operations, two- and three-hour-ahead wind speed forecasts over Saudi Arabia would be both more accurate and more stable in time than those from the prior ESN approach, while one-hour-ahead accuracy remains essentially unchanged.
At the 75 optimal inland wind-farm sites, the annual absolute energy error drops from $2.770 \times 10^8$ kWh to $2.601 \times 10^8$ kWh, corresponding to up to $\$1.183$ million in annual savings at Saudi electricity prices.
The calibrated covariance $\hat{\Sigma}^*(\hat{\delta})$ brings marginal interval coverage close to nominal levels, with a median coverage of 0.949 for a nominal 0.95 interval at one-hour lead, whereas using either the SPDE covariance alone or the empirical covariance alone misses by a wide margin.
Operationally, keeping the batch size $b$ above about 70 hours keeps the periodic weight re-estimation inside a one-hour window, so the accuracy gain does not break the hourly forecast cycle.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Because the WRF benchmark was validated against only ten stations, the pipeline's real-world edge is untested in precisely the regions where the two models differ most; re-running the comparison against denser tower networks or reanalysis would reveal whether the reported margins survive.
The architecture is a transferable recipe: any country with one high-resolution simulation and sparse observations could use energy-distance knots, reservoir computing, and SPDE reconstruction, so the value may extend beyond Saudi Arabia.
The largest gains appear only at two- and three-hour leads, which suggests the $b=75$ batch refresh, not the spatial interpolation, may be the active ingredient; an ablation that turns batch updating off while keeping SPDE interpolation would isolate the mechanism.
The economic figure assumes a flat $\$70$/MWh electricity price; with a real-time energy market in Saudi Arabia, the value of reduced forecast error could be larger or smaller depending on when errors occur.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Quick take: this is a solid engineering paper whose central empirical claim does not survive its own Section 4.2. The pipeline—support points for dimension reduction, a deep ESN with a new batch-update rule, and non-stationary SPDE interpolation—is a genuine extension of Huang et al. (2022) and McDermott & Wikle (2017), and the supplement's simulation studies give reasonable component-level support. Code and data are provided, which helps.

The problem is the validation protocol. Section 4.2 says the ESN hyperparameter vector θ is chosen as the minimizer of MSPE on the 2016 testing set, and Table 1 then reports those same 2016 MSPE numbers as evidence of forecast skill. The batch window b=75 is picked from the MSPE-versus-b curve in Section 4.4, and the shrinkage δ in Section 5.2 is chosen so that the 2016 coverage is near nominal. So every headline number, including the $1.18 million annual saving, is in-sample. The comparison against ESN is similarly contaminated if ESN's hyperparameters were selected under the same protocol, as the text implies. This voids the paper's main claim as presented. The fix is straightforward—a proper train/validation/test split—but it has to be done.

There are smaller issues too. The harmonic periods in Section 3.1 are listed as 4380, 2920, 1460, 8, and 4 hours for what the text describes as one-year, half-year, one-day, 12-hour, and 8-hour cycles; those numbers don't match the stated periods. The SPDE smoothness is fixed at ν=0.5 to keep the GMRF property even though the estimated ν is about 0.3; the conclusion admits this, which is honest, but it means the spatial model is knowingly misspecified. And the evaluation treats WRF-simulated wind as ground truth with only ten monitoring stations for validation—a real limitation, though the authors are transparent about treating the simulation as a surrogate.

Where the paper earns its keep: the batch-update idea is new, the simulations are sensible, and the writing is clear. The methodological flaw is serious but fixable, and the components are worth preserving. My recommendation is to send this to peer review anyway—a good referee can push for a clean split and the corrected periods, and the paper would be publishable after major revision. It should not be accepted in its current form.

Referee Report

3 major / 6 minor

Summary. The paper proposes a spatio-temporal model for wind speed over Saudi Arabia consisting of three stages: spatial dimension reduction via support points, temporal forecasting with a deep Bayesian (batch-updated) Echo State Network, and spatial reconstruction with a non-stationary SPDE-based Matérn model. The method is applied to WRF-simulated hourly wind speed from 2013–2016 over 53,333 locations, and the authors report improved 2- and 3-hour lead forecasts compared with ESN, GRU, LSTM, VAR, and persistence, as well as annual energy savings up to about $1.18 million against the closest competitor. The paper also contains simulation studies on spatial reduction and temporal forecasting, and it makes code and data available.

Significance. If the reported results were valid, the paper would offer a scalable pipeline for high-dimensional spatio-temporal forecasting that combines dimension reduction, recurrent neural networks, and SPDE-based interpolation, with a clear application to wind-energy planning in a region with scarce observational infrastructure. The batch-update mechanism for the ESN output layer is a simple and plausible enhancement, and the support-points approach to knot selection is theoretically motivated. The authors are transparent in providing code and data and in comparing against several baselines. However, the central empirical comparison is compromised by test-set leakage in hyperparameter selection and calibration, as detailed below, so the forecast improvements and monetary savings are not trustworthy as presented.

major comments (3)

[Section 4.2 / Table 1] The hyperparameter vector θ is chosen by minimizing the mean squared error on the 2016 testing year (Section 4.2: 'The hyper-parameter vector θ is then chosen as the minimizer of the mean squared error on the testing data.'), and the same 2016 data are then used to compute the MSPEs reported in Table 1 and the energy savings in Table 3. Consequently, the claimed improvements of B-ESN over ESN at two- and three-hour leads are not out-of-sample results; they are artifacts of fitting the model to the test period. The batch size b=75 is likewise selected from the MSPE-versus-b tradeoff in Section 4.4, which appears to be evaluated on the same test year. The central forecast comparison must be redone with all hyperparameters and design choices (θ, b, δ) selected using only data up to 2015, with 2016 reserved exclusively for evaluation.
[Section 4.6 / Table 2] The calibration parameter δ in equation (8) is chosen to make the prediction intervals for the spatial mean achieve the expected coverage on the test period (Section 4.6: 'The parameter δ ∈ (0,1) is chosen to make prediction intervals for the spatial mean as close as possible to the expected coverage.'), and Table 2 then reports that the calibrated intervals achieve coverage close to the nominal level. This is circular: the empirical coverage of the constructed intervals is fitted by construction, not a validation of the calibration procedure. The coverage comparison should be performed on data not used to select δ, for example by a within-training cross-validation scheme or by holding out a separate period.
[Section 3.4 / Section 4.5 / Conclusion] There is an internal inconsistency about when the SPDE solution is a Gaussian Markov random field. Section 3.4 states that 'if ν is an integer, the discrete solution of the SPDE (5) is a Gaussian Markov random field,' Section 4.5 says 'If ν is an integer or half integer, Z is a GMRF,' and the Conclusion says 'we fixed the smoothness parameter ν = 0.5 to ensure the Gaussian Markov property.' Since ν=0.5 is not an integer (and for d=2, α=ν+1=1.5 is not an integer), the claim that ν=0.5 ensures the GMRF property is not supported by the stated condition. Moreover, the estimated ν≈0.3 is disregarded without a sensitivity analysis. The authors should either correct the GMRF condition, justify ν=0.5 by evaluating alternative smoothness values, or use the rational SPDE approximation (Bolin & Kirchner, 2020) that they cite.

minor comments (6)

[Section 2] The evaluation treats WRF-simulated fields as ground truth, and the simulation is validated at only ten monitoring stations (Section 2, citing Giani et al. 2020). The paper should state explicitly that all forecast errors and energy savings are relative to the simulation, not to observations, and that the ten stations may not cover complex terrain such as the Empty Quarter.
[Table 1] The 'All locations' part of Table 1 includes interpolation error; reporting the SPDE interpolation-only error as a separate row would help separate the temporal and spatial contributions to the B-ESN gains.
[Section 5.3] The energy savings are computed under a single flat electricity price ($70/MWh) and without uncertainty quantification; the conclusion acknowledges this, but the abstract's claim of '$1.18 million' savings should be accompanied by a sensitivity range or qualified as an upper-bound estimate.
[Figure 2c] The reversed x-axis in Figure 2c is confusing; consider plotting b on a conventional increasing axis with the label indicating that computation time grows as b decreases.
[Section 3.4] Typo: 'Matén' should be 'Matérn'.
[Section 5.1] Typo: 'MPSE' should be 'MSPE' in the sentence about Table 1.

Circularity Check

3 steps flagged · score 6.0 of 10

Test-year leakage in hyperparameter, batch-size, and calibration selection invalidates the central out-of-sample forecast comparison.

fitted input called prediction [Section 4.2 (Temporal Model inference), evaluated in Table 1 and Section 5.1]
"we perform cross-validation by using wind speed data from 2013 to 2015 as a training set and 2016 as a testing set. The hyper-parameter vector θ is then chosen as the minimizer of the mean squared error on the testing data."

The 2016 year is used to select θ by minimizing MSPE, and the very same year's MSPEs are reported in Table 1 as evidence that B-ESN beats ESN at two- and three-hour leads. A model selected on a dataset cannot subsequently be used to demonstrate out-of-sample skill on that dataset; the lower MSPE is the fitted objective itself, not an independent forecast comparison.
fitted input called prediction [Section 4.4 (Computational Sensitivity); batch size b=75 used in Section 5.1/Table 1]
"although more frequent updates of the coefficient matrix are beneficial for the prediction accuracy of long-range forecasts, as indicated by various slopes of the MSPE curves for the three lead hour forecasts shown in Figure 2c... our B-ESN model relying on support points, batch update (b = 75) and SPDE for interpolation results in improved forecasts for two- and three-hour lead prediction horizon in terms of both accuracy and stability."

The batch size b=75 used in Table 1 is selected after inspecting MSPE-versus-b curves, apparently for the same 2016 evaluation period, and the resulting MSPEs are then reported as predictive skill of the B-ESN configuration. The forecast gain attributed to the batch update is therefore partly an artifact of selecting b on the test-year MSPE curve rather than a genuine out-of-sample comparison.

1 more flagged steps

fitted input called prediction [Section 4.6 (Forecast Calibration) and Section 5.2 (Table 2)]
"The parameter δ ∈ (0, 1) is chosen to make prediction intervals for the spatial mean as close as possible to the expected coverage. ... Finally, among all choices of ˆδ(∆), we choose the one which gives the most accurate marginal coverage: the median (across all locations in the domain) empirical coverage of a 95% needs to be as close as possible to the nominal 95% value."

δ is fitted on the 2016 forecast residuals so that empirical coverage is close to 0.95, and Table 2 then reports 95% coverages of 0.949, 0.946, and 0.940 as evidence that the calibrated covariance performs well. The agreement with the nominal level is imposed by the selection criterion on the same data, so the coverage table validates the fitting target rather than the uncertainty model independently.

full rationale

The paper's strongest claims—improved two- and three-hour wind forecasts over ESN and annual savings of $1.18 million—rest on Table 1 and Section 5.1, but the MSPEs in Table 1 are computed on the same 2016 testing year that Section 4.2 explicitly uses to select the ESN hyperparameter vector θ by minimizing test MSPE. Batch size b=75 is also chosen from MSPE-versus-b curves in Section 4.4 before being presented as part of the winning B-ESN configuration, and the calibration parameter δ in Section 4.6/5.2 is selected on 2016 coverage before Table 2 reports that coverage as successful validation. These are cases where a fitted quantity is presented as a prediction or validation result, not independent evidence. The method itself has non-circular components: the support-point and SPDE machinery are evaluated in properly split synthetic simulations, and citations to Huang et al. (2022), Song et al. (2024), and Bonas & Castruccio (2023) are provenance rather than load-bearing self-citation chains. However, as presented, the central application comparison is compromised by test-set leakage in model selection, so the headline forecast and savings results do not support the paper's central claim out-of-sample.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

No new physical entities or latent constructions are introduced; all model components are existing statistical or numerical constructs.

free parameters (6)

Number of support points nred = 3173
Chosen by hand to match the number of knots in Huang et al. (2022), not by data-driven selection.
Batch window b = 75
Selected as a tradeoff between forecast accuracy and computation time (Section 4.4); the search is not described as an independent validation.
ESN hyperparameters theta (nh,D, m, nuD, lambda, etaWD, etaWinD, piWD, piWinD, alpha) = 2500, 1, 0.9, 0.15, 0.05, 0.01, 0.1, 0.01, 1 (Table S4)
Chosen as the minimizer of the MSE on the 2016 testing data (Section 4.2), so they are fitted to the evaluation period.
SPDE smoothness nu = 0.5
Fixed by hand in the Conclusion to ensure the Gaussian Markov property, although the paper states integer nu is required and the estimated value is near 0.3.
Shrinkage parameter delta = Table S3, varies by lead and region (e.g., 0.36 to 0)
Fitted to make coverage of prediction intervals match the nominal level on the 2016 testing period (Section 4.6); the reported coverage is therefore achieved by construction.
Wind shear coefficient alpha(s) = Estimated per pixel via linear regression using Crippa et al. (2021)
Borrowed from prior work and refitted to the same data; used in the energy conversion, so the savings estimate depends on it.

assumptions (5)

domain assumption The WRF simulation at 6 km resolution with the MYJ scheme is an adequate proxy for true wind over Saudi Arabia, and forecast skill against the simulation transfers to reality.
The model is built and evaluated entirely on simulated WRF fields (Section 2). Validation is against only 10 monitoring stations, so unobserved regions may have systematic biases.
standard math The discrete solution of the SPDE (5) is a Gaussian Markov random field when nu is an integer (Lindgren et al., 2011).
Invoked in Section 3.4 to justify sparse precision matrices and fast inference; the application then fixes nu=0.5, which contradicts this condition.
domain assumption The out-of-sample evaluation is valid, i.e., the test year 2016 is not used for model selection.
This assumption is violated in Section 4.2, where hyperparameters are chosen to minimize testing MSE.
domain assumption The residual process after trend removal is independent in time with unit variance, and the ESN residuals are Gaussian with a covariance captured by the calibrated shrinkage covariance.
Used to define the likelihood and calibration; the Gaussianity is only partially checked (Figure S3).
standard math The support points selected by energy distance are representative of the full spatial distribution for interpolation (Song et al., 2024).
Borrowed from the cited theoretical result; relied on to justify the dimension reduction in Section 3.2.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Modeling High-Resolution Spatio-Temporal Wind with Deep Echo State Networks and Stochastic Partial Differential Equations." pith.science (2026). https://pith.science/paper/EBFOQIZC

@misc{pith2026241207265,
  author       = {Pith},
  title        = {Pith review of: Modeling High-Resolution Spatio-Temporal Wind with Deep Echo State Networks and Stochastic Partial Differential Equations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EBFOQIZC}},
  note         = {Machine review of arXiv:2412.07265}
}

read the original abstract

In the past decades, clean and renewable energy has gained increasing attention due to a global effort on carbon footprint reduction. In particular, Saudi Arabia is gradually shifting its energy portfolio from an exclusive use of oil to a reliance on renewable energy, and, in particular, wind. Modeling wind for assessing potential energy output in a country as large, geographically diverse and understudied as Saudi Arabia is a challenge which implies highly non-linear dynamic structures in both space and time. To address this, we propose a spatio-temporal model whose spatial information is first reduced via an energy distance-based approach and then its dynamical behavior is informed by a sparse and stochastic recurrent neural network (Echo State Network). Finally, the full spatial data is reconstructed by means of a non-stationary stochastic partial differential equation-based approach. Our model can capture the fine scale wind structure and produce more accurate forecasts of both wind speed and energy in lead times of interest for energy grid management and save annually as much as one million dollar against the closest competitive model.

Figures

Figures reproduced from arXiv: 2412.07265 by the authors.

**Figure 2.** Computation time for inference and forecasting (three-hour ahead) of the ESN model ( [PITH_FULL_IMAGE:figures/full_fig_p014_2.png] view at source ↗

**Figure 3.** Location-wise relative forecasting errors between ESN and B-ESN for up to three-hour lead [PITH_FULL_IMAGE:figures/full_fig_p019_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

58 extracted references · 56 canonical work pages

[1]

& Gasim, A

Aldubyan, M. & Gasim, A. (2021). Energy price reform in saudi arabia: Modeling the economic and environmental impacts and understanding the demand response. Energy Policy , 148, 111941

work page 2021
[2]

M., Trouv \'e , E., Nicolas, J.-M., & L \^e , T

Atto, A. M., Trouv \'e , E., Nicolas, J.-M., & L \^e , T. T. (2016). Wavelet operators and multiplicative observation models—application to sar image time-series analysis. IEEE Transactions on Geoscience and Remote Sensing , 54(11), 6606--6624

work page 2016
[3]

& Kirchner, K

Bolin, D. & Kirchner, K. (2020). The rational SPDE approach for Gaussian random fields with general smoothness. Journal of Computational and Graphical Statistics , 29(2), 274--285

work page 2020
[4]

& Simas, A

Bolin, D. & Simas, A. B. (2023). rSPDE: Rational Approximations of Fractional Stochastic Partial Differential Equations . R package version 2.3.3

work page 2023
[5]

Bollerslev, T. (1986). Generalized autoregressive conditional heteroskedasticity. Journal of Econometrics , 31(3), 307--327

work page 1986
[6]

& Castruccio, S

Bonas, M. & Castruccio, S. (2023). Calibration of spatio-temporal forecasts from citizen science urban air pollution data with sparse recurrent neural networks. Annals of Applied Statistics , 17(3), 1820--1840

work page 2023
[7]

C., Boone, E., Alamri, F., Hari, B., Kavila, I., Simmons, S., Jarvis, S., Burr, W., Pagendam, D., Chang, W., & Castruccio, S

Bonas, M., Datta, A., Wikle, K. C., Boone, E., Alamri, F., Hari, B., Kavila, I., Simmons, S., Jarvis, S., Burr, W., Pagendam, D., Chang, W., & Castruccio, S. (2024a). Assessing predictability of environmental time series with statistical and machine learning models. Environmetrics . in press

work page 2024
[8]

H., & Castruccio, S

Bonas, M., Richter, D. H., & Castruccio, S. (2024b). A physics-informed, deep double reservoir network for forecasting boundary layer velocity. Journal of the American Statistical Association - Applications and Case Studies . in press

work page 2024

Show all 58 references

[9]

Bonas, M., Wikle, C., & Castruccio, S. (2024c). Calibrated forecasts of quasi-periodic climate processes with deep echo state networks and penalized quantile regression. Environmetrics , 35(1), e2833

2024
[10]

Castruccio, S., Ombao, H., & Genton, M. G. (2018). A scalable multi-resolution spatio-temporal model for brain activation and connectivity in fmri data. Biometrics , 74(3), 823--833

2018
[11]

Chen, W., Castruccio, S., & Genton, M. G. (2021). Assessing the risk of disruption of wind turbine operations in saudi arabia using bayesian spatial extremes. Extremes , 24, 267--292

2021
[12]

Chung, J., Gulcehre, C., Cho, K., & Bengio, Y. (2014). Empirical evaluation of gated recurrent neural networks on sequence modeling. arXiv preprint arXiv:1412.3555

2014 arXiv
[13]

G., & Castruccio, S

Crippa, P., Alifa, M., Bolster, D., Genton, M. G., & Castruccio, S. (2021). A temporal model for vertical extrapolation of wind speed and wind energy assessment. Applied Energy , 301, 117378

2021
[14]

C., Alexander, C

Dowell, D. C., Alexander, C. R., James, E. P., Weygandt, S. S., Benjamin, S. G., Manikin, G. S., Blake, B. T., Brown, J. M., Olson, J. B., Hu, M., Smirnova, T. G., Ladwig, T., Kenyon, J. S., Ahmadov, R., Turner, D. D., Duda, J. D., & Alcott, T. I. (2022). The high-resolution r...

2022
[15]

Doya, K. et al. (1992). Bifurcations in the learning of recurrent neural networks 3. Learning (RTRL) , 3, 17

1992
[16]

Espeholt, L., Agrawal, S., Sønderby, C., Kumar, M., Heek, J., Bromberg, C., Gazen, C., Carven, R., Andrychowicz, M., Hickey, J., Bell, A., & Kalchbrenner, N. (2022). Deep learning for twelve hour precipitation forecasts. Nature Communications , 13, 5145

2022
[17]

Friedman, J., Hastie, T., & Tibshirani, R. (2008). Sparse inverse covariance estimation with the graphical lasso. Biostatistics , 9(3), 432--441

2008
[18]

G., Castruccio, S., & Crippa, P

Giani, P., Tagle, F., Genton, M. G., Castruccio, S., & Crippa, P. (2020). Closing the gap between wind energy targets and implementation for emerging countries. Applied Energy , 269, 115085

2020
[19]

G., & Aldrich, E

Gneiting, T., Larson, K., Westrick, K., Genton, M. G., & Aldrich, E. (2006). Calibrated probabilistic forecasting at the stateline wind energy center: The regime-switching space--time method. Journal of the American Statistical Association , 101(475), 968--979

2006
[20]

J., Raible, C

G \'o mez-Navarro, J. J., Raible, C. C., & Dierer, S. (2015). Sensitivity of the wrf model to pbl parametrisations and nesting techniques: evaluation of wind storms over complex terrain. Geoscientific Model Development , 8(10), 3349--3363

2015
[21]

Goodfellow, I., Bengio, Y., & Courville, A. (2016). Deep Learning . MIT press

2016
[22]

Gualtieri, G. (2019). A comprehensive review on wind resource extrapolation models applied in wind energy. Renewable and Sustainable Energy Reviews , 102, 215--233

2019
[23]

Hastie, T. J. (2017). Generalized additive models. In Statistical models in S (pp.\ 249--307). Routledge

2017
[24]

Hering, A. S. & Genton, M. G. (2010). Powering up with space-time wind forecasting. Journal of the American Statistical Association , 105(489), 92--104

2010
[25]

& Schmidhuber, J

Hochreiter, S. & Schmidhuber, J. (1997). Long short-term memory. Neural Computation , 9(8), 1735--1780

1997
[26]

Huang, H., Castruccio, S., & Genton, M. G. (2022). Forecasting high-frequency spatio-temporal wind power with dimensionally reduced echo state networks. Journal of the Royal Statistical Society: Series C (Applied Statistics) , 71(2), 449--466

2022
[27]

Ingebrigtsen, R., Lindgren, F., & Steinsland, I. (2014). Spatial models with explanatory variables in the dependence structure. Spatial Statistics , 8, 20--38

2014
[28]

echo state

Jaeger, H. (2001). The “echo state” approach to analysing and training recurrent neural networks-with an erratum note. Bonn, Germany: German National Research Center for Information Technology GMD Technical Report , 148(34), 13

2001
[29]

Jaeger, H. (2007). Echo state network. Scholarpedia , 2(9), 2330

2007
[30]

Jani \'c , Z. I. (2001). Nonsingular implementation of the mellor-yamada level 2.5 scheme in the ncep meso model. Office note (National Centers for Environmental Prediction (U.S.)) , 437

2001
[31]

Kingma, D. P. & Ba, J. (2014). Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980

2014 arXiv
[32]

& Rue, H

Lindgren, F. & Rue, H. (2015). Bayesian spatial modelling with r-inla. Journal of Statistical Software , 63, 1--25

2015
[33]

Lindgren, F., Rue, H., & Lindstr \"o m, J. (2011). An explicit link between gaussian fields and gaussian markov random fields: the stochastic partial differential equation approach. Journal of the Royal Statistical Society: Series B (Statistical Methodology) , 73(4), 423--498

2011
[34]

Lorenz, E. N. (1996). Predictability: A problem partly solved. In Proc. Seminar on predictability , volume 1: Reading

1996
[35]

Luko s evi c ius, M. (2012). A practical guide to applying echo state networks. Neural Networks: Tricks of the Trade: Second Edition , (pp.\ 659--686)

2012
[36]

L \"u tkepohl, H. (2013). Vector autoregressive models. In Handbook of Research Methods and Applications in Empirical Macroeconomics (pp.\ 139--164). Edward Elgar Publishing

2013
[37]

Maass, W., Natschl \"a ger, T., & Markram, H. (2002). Real-time computing without stable states: A new framework for neural computation based on perturbations. Neural Computation , 14(11), 2531--2560

2002
[38]

& Joseph, V

Mak, S. & Joseph, V. R. (2018). Support points. The Annals of Statistics , 46(6A), 2562--2592

2018
[39]

McDermott, P. L. & Wikle, C. K. (2017). An ensemble quadratic echo state network for non-linear spatio-temporal forecasting. Stat , 6(1), 315--330

2017
[40]

McDermott, P. L. & Wikle, C. K. (2019a). Bayesian recurrent neural network models for forecasting and quantifying uncertainty in spatial-temporal data. Entropy , 21(2), 184

2019
[41]

McDermott, P. L. & Wikle, C. K. (2019b). Deep echo state networks with uncertainty quantification for spatio-temporal forecasting. Environmetrics , 30(3), e2553

2019
[42]

Mellor, G. L. & Yamada, T. (1982). Development of a turbulence closure model for geophysical fluid problems. Reviews of Geophysics , 20(4), 851--875

1982
[43]

& Nuno, G

Nakov, A. & Nuno, G. (2013). Saudi arabia and the oil market. The Economic Journal , 123(573), 1333--1362

2013
[44]

Nychka, D., Bandyopadhyay, S., Hammerling, D., Lindgren, F., & Sain, S. (2015). A multiresolution gaussian process model for the analysis of large spatial datasets. Journal of Computational and Graphical Statistics , 24(2), 579--599

2015
[45]

Nychka, D., Furrer, R., Paige, J., & Sain, S. (2021). fields: Tools for spatial data. R package version 16.3

2021
[46]

& Held, L

Rue, H. & Held, L. (2005). Gaussian Markov Random Fields: Theory and Applications . CRC press

2005
[47]

G., & S rbye, S

Simpson, D., Rue, H., Riebler, A., Martins, T. G., & S rbye, S. H. (2017). Penalising Model Component Complexity: A Principled, Practical Approach to Constructing Priors . Statistical Science , 32(1), 1--28

2017
[48]

Skamarock, W., Klemp, J., Dudhia, J., Gill, D., Barker, D., Duda, M., Huang, X., Wang, W., & Powers, J. (2008). A description of the advanced research wrf version 3 (ncar tech. note ncar/tn-475+ str, 113 pp). Boulder, CO: National Center for Atmospheric Research

2008
[49]

Sobash, R. A. & Ahijevych, D. A. (2024). Evaluating machine learning–based probabilistic convective hazard forecasts using the hrrr: Quantifying hazard predictability and sensitivity to training choices. Weather and Forecasting , 39(10), 1399 -- 1415

2024
[50]

Song, Y., Dai, W., & Genton, M. G. (2024). Large-scale low-rank gaussian process prediction with support points. Journal of the American Statistical Association , (just-accepted), 1--21

2024
[51]

Stein, M. L. (1999). Interpolation of Spatial Data: Some Theory for Kriging . Springer Science & Business Media

1999
[52]

J., Rizzo, M

Sz \'e kely, G. J., Rizzo, M. L., et al. (2004). Testing for equal distributions in high dimension. InterStat , 5(16.10), 1249--1272

2004
[53]

Tagle, F., Castruccio, S., Crippa, P., & Genton, M. G. (2019). A non-gaussian spatio-temporal model for daily wind speeds based on a multi-variate skew-t distribution. Journal of Time Series Analysis , 40(3), 312--326

2019
[54]

Bureau of Labor Statistics (2024)

U.S. Bureau of Labor Statistics (2024). Energy prices. https://www.bls.gov/regions/midwest/data/averageenergyprices_selectedareas_table.htm. Accessed: 2024-05-15

2024
[55]

Whittle, P. (1954). On stationary processes in the plane. Biometrika , 41(3-4), 434--449

1954
[56]

Xie, Y., Li, C., Li, M., Liu, F., & Taukenova, M. (2022). An overview of deterministic and probabilistic forecasting methods of wind energy. iScience , 26(1), 105804

2022
[57]

G., & Castruccio, S

Zhang, J., Crippa, P., Genton, M. G., & Castruccio, S. (2021). Assessing the reliability of wind power operations under a changing climate with a non-gaussian bias correction. The Annals of Applied Statistics , 15(4), 1831--1849

2021
[58]

& Genton, M

Zhu, X. & Genton, M. G. (2012). Short-term wind speed forecasting for power system operations. International Statistical Review , 80(1), 2--23

2012

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.

[1] [1]

& Gasim, A

Aldubyan, M. & Gasim, A. (2021). Energy price reform in saudi arabia: Modeling the economic and environmental impacts and understanding the demand response. Energy Policy , 148, 111941

work page 2021

[2] [2]

M., Trouv \'e , E., Nicolas, J.-M., & L \^e , T

Atto, A. M., Trouv \'e , E., Nicolas, J.-M., & L \^e , T. T. (2016). Wavelet operators and multiplicative observation models—application to sar image time-series analysis. IEEE Transactions on Geoscience and Remote Sensing , 54(11), 6606--6624

work page 2016

[3] [3]

& Kirchner, K

Bolin, D. & Kirchner, K. (2020). The rational SPDE approach for Gaussian random fields with general smoothness. Journal of Computational and Graphical Statistics , 29(2), 274--285

work page 2020

[4] [4]

& Simas, A

Bolin, D. & Simas, A. B. (2023). rSPDE: Rational Approximations of Fractional Stochastic Partial Differential Equations . R package version 2.3.3

work page 2023

[5] [5]

Bollerslev, T. (1986). Generalized autoregressive conditional heteroskedasticity. Journal of Econometrics , 31(3), 307--327

work page 1986

[6] [6]

& Castruccio, S

Bonas, M. & Castruccio, S. (2023). Calibration of spatio-temporal forecasts from citizen science urban air pollution data with sparse recurrent neural networks. Annals of Applied Statistics , 17(3), 1820--1840

work page 2023

[7] [7]

C., Boone, E., Alamri, F., Hari, B., Kavila, I., Simmons, S., Jarvis, S., Burr, W., Pagendam, D., Chang, W., & Castruccio, S

Bonas, M., Datta, A., Wikle, K. C., Boone, E., Alamri, F., Hari, B., Kavila, I., Simmons, S., Jarvis, S., Burr, W., Pagendam, D., Chang, W., & Castruccio, S. (2024a). Assessing predictability of environmental time series with statistical and machine learning models. Environmetrics . in press

work page 2024

[8] [8]

H., & Castruccio, S

Bonas, M., Richter, D. H., & Castruccio, S. (2024b). A physics-informed, deep double reservoir network for forecasting boundary layer velocity. Journal of the American Statistical Association - Applications and Case Studies . in press

work page 2024