The starting point is COSMO-REA6, a regional reanalysis for the European CORDEX domain developed by the Hans-Ertel-Centre for Weather Research together with the universities of Bonn and Cologne and the German Meteorological Service. Reanalyses blend numerical weather models with historical observations to produce a physically consistent record of past atmospheric conditions, and COSMO-REA6 does so on a grid of roughly six kilometers with forty vertical levels. But even at this resolution, the model’s deterministic gust diagnostic struggles with the chaotic, localized nature of wind gusts, which are produced by turbulent eddies deflecting fast-moving air parcels down to the surface or by convective downdrafts spreading horizontally when they strike the ground. These mechanisms, first formalized in diagnostics by researchers such as Brasseur and Nakamura, are deterministic by design and carry no measure of uncertainty, which is precisely what forecasters and emergency managers need most.

Ertz and Friederichs attack this problem with the tools of extreme value statistics, the branch of mathematics built specifically for rare, high-impact events. Because a gust measurement represents the highest three-second average wind speed within an hour, the theoretically appropriate model for daily maxima is the generalized extreme value distribution. After fitting single-station models and finding that the shape parameter showed no clear signal or leaned slightly negative, the authors deliberately fixed it at zero, yielding a Gumbel distribution. This choice avoids imposing an artificial upper bound on gust speeds, a property of the Weibull-type distribution that would be physically implausible for storm-driven extremes, while keeping parameter estimates stable.

The real innovation lies in how the model handles space. At the top level of the hierarchy, gust observations follow a Gumbel distribution whose location and scale parameters vary in both space and time. These parameters are expressed as linear regressions on predictor variables drawn from COSMO-REA6, chiefly the maximum ten-meter wind speed and the mean horizontal wind, plus a covariate capturing the difference between the station altitude and the model’s representation of that altitude. Crucially, the regression coefficients themselves are not treated as fixed numbers. Instead, each coefficient is a realization of a two-dimensional Gaussian random field with a constant mean and an isotropic Matérn covariance function, meaning that stations close to one another tend to share similar gust behavior. This latent-process structure allows information to be pooled across the entire network of 109 synoptic stations, stabilizing estimates at data-sparse locations and enabling kriging-based interpolation to any unobserved point in the domain.

Perhaps the most elegant trick in the paper is a correction to the distance metric itself. Mountain stations such as the Zugspitze or the Brocken can sit within a few kilometers of valley towns like Braunlage, yet their gust climates are worlds apart. Under a purely geographic distance measure, the model would wrongly treat these stations as near neighbors, contaminating the estimated spatial dependency structure for the whole region. The authors solve this by adding a scaled elevation difference to the great-circle distance between stations, effectively pushing mountain and valley locations apart in the covariance calculation. The scaling factor is treated as an unknown parameter with a prior centered on the ratio of buoyancy to Coriolis forces, a value of roughly one hundred derived from quasigeostrophic scale analysis. The result is that mountaintop observations can remain in the training data without distorting the picture elsewhere, and the estimated correlation lengths become larger and more stable.

Training such a model is no small computational feat. The authors used the Stan probabilistic programming language via its Python interface, running the No-U-Turn Sampler, an adaptive form of Hamiltonian Monte Carlo, for chains of 1500 iterations with 500 discarded as burn-in. The full spatial model required about an hour of sampling time, and the subsequent interpolation of parameter fields across a 121-by-161 grid cell domain over Germany took roughly three hours on an ordinary desktop computer, using an iterative scheme that simulates horizontal slices of the country one after another, each conditioned on its southern neighbor. The authors emphasize that the entire pipeline is reproducible on standard hardware and that the code and preprocessed data are openly available.

To find out whether all this statistical machinery actually pays off, the team subjected the model to a demanding verification regime. Data from odd-numbered years between 2001 and 2018 trained the model, while even-numbered years served for evaluation, and a leave-one-out cross-validation ensured that the station being predicted was always excluded from training. Performance was judged with proper scoring rules: the Brier score for probabilities of exceeding warning thresholds of 14, 18, 25, 29 and 33 meters per second, and the quantile score for predictive quantiles ranging from the 75th to the 99.9th percentile. Both scores were decomposed into miscalibration, resolution and uncertainty components using modern isotonic-regression-based estimators, allowing the authors to see not just whether the model was skillful but why.

The verdict is a qualified but meaningful success for the spatial approach. Against a spatially constant baseline model, which itself already achieved skill scores of 10 to 45 percent relative to local climatology, the spatial Bayesian model delivered median improvements of 2 to 5 percent for prediction quantiles, with the largest gains at the extreme upper quantiles that matter most for warnings. When the evaluation was restricted to the top quartile of forecast gust events, the spatial model’s advantage grew even more pronounced, and for the 99.9th percentile it nearly matched the skill of a model trained locally at each station, the theoretical ceiling of the approach. The score decomposition revealed that the gains came primarily from better resolution rather than better calibration, meaning the spatial model discriminates more effectively between windy and calm situations. For threshold exceedance probabilities the improvement was marginal, though the spatial model showed a slight edge at the highest 33 meters per second threshold.

The study is candid about its limits. At sheltered inland locations with weak mean winds, exemplified by Garmisch-Partenkirchen, even the locally trained model has no skill against climatology, because the reanalysis predictors simply lack explanatory power in complex, low-wind terrain. The spatial interpolation can also introduce bias where a station’s gust character differs sharply from its surroundings. Yet the overall message is compelling: by letting the data itself determine how gust behavior varies across the landscape, and by treating mountains as the statistical islands they truly are, the Bonn team has produced a proof of concept that spatially aware extreme value modeling can extract more reliable gust information from reanalysis archives. With extensions already envisioned for ensemble forecasts, diurnal and annual cycles, and land-sea contrasts, the framework could soon help weather services issue wind warnings that are not only more accurate but accompanied by the probabilities that decision-makers increasingly demand.

Subject of Research: Spatial Bayesian hierarchical extreme value post-processing of wind gusts from the COSMO-REA6 reanalysis over Germany

Article Title: Post-processing of wind gusts from COSMO-REA6 with a spatial Bayesian hierarchical extreme value model

Article References: Ertz, P., & Friederichs, P. (2025). Post-processing of wind gusts from COSMO-REA6 with a spatial Bayesian hierarchical extreme value model. Advances in Statistical Climatology, Meteorology and Oceanography, 11(2), 229-256. https://doi.org/10.5194/ascmo-11-229-2025

Image Credits: AI Generated

DOI: 10.5194/ascmo-11-229-2025

Keywords: wind gusts, extreme value statistics, Bayesian hierarchical model, COSMO-REA6, reanalysis, post-processing, spatial statistics, Gaussian random fields, Gumbel distribution, forecast verification, Germany, meteorology

A team at the Montréal Neurological Institute of McGill University has now published a detailed account of how they have managed this complexity, both inside their own laboratory and for the wider community. Writing in Nature Protocols, Zhen-Qi Liu, Vincent Bazinet and colleagues, led by Bratislav Misic, describe netneurotools, an open-source Python toolkit that has been continuously developed and maintained by the laboratory’s trainees since its inception. The paper is both a practical protocol for carrying out network neuroscience analyses and a manifesto for a different way of building scientific software, one in which the informal, ad hoc scripts that every laboratory accumulates are treated as a legitimate, shareable scientific resource.

The philosophy behind the toolkit is disarmingly simple. The authors describe netneurotools as the Swiss army knife of the laboratory: a collection of functions and routines that the group uses constantly but that belong to no established pipeline or package. Where large neuroimaging platforms excel at well-defined tasks such as preprocessing functional magnetic resonance imaging or reconstructing diffusion data, they are not designed to interoperate with one another. The gaps between them, the authors argue, are precisely where trainees are forced to improvise isolated heuristics and workarounds. netneurotools formalizes those improvisations, turning scattered personal scripts into documented, tested, reusable code that anyone can pick up.

Technically, the toolkit is built on the familiar foundations of the scientific Python ecosystem, drawing on array programming libraries such as NumPy, the algorithms of SciPy, machine learning utilities from scikit-learn, graph structures from NetworkX, and file-reading capabilities from nibabel and nilearn. It extends these foundations with capabilities that are specific to network neuroscience. These include utilities for handling cortical surface meshes and transforming data between the many parcellation schemes that fragment the field, from volumetric atlases to multi-resolution cortical subdivisions. Because a brain map computed on one parcellation cannot be directly compared with a map on another, robust surface-based resampling and interpolation are among the most valuable functions the package provides, sparing researchers from the error-prone manual conversions that have long been a rite of passage in the field.

Network analysis itself forms a second major pillar. The toolkit implements routines for generating group-representative structural brain networks using distance-dependent consensus thresholding, an approach designed to respect the fact that anatomical connection probability falls with physical distance in the brain. It provides algorithms for randomizing weighted networks while preserving key topological properties, a crucial step in any null-model-based analysis, including a simulated annealing method developed by the same group for rigorously controlling network structure. It also implements a library of network communication models, which ask how signals could theoretically travel along the wiring of the brain, from classical shortest-path routing inspired by the Floyd, Roy and Warshall algorithms to navigation strategies and diffusion-style models that better capture the biology of neural signaling.

Statistical machinery rounds out the package. Network neuroscience increasingly relies on spatial statistics, because brain measures are arranged in space and neighboring regions are not independent. netneurotools includes implementations of spatial autocorrelation measures such as Moran’s I and Geary’s C, along with bivariate extensions that quantify spatial associations between two brain maps. It offers null models that preserve the spatial autocorrelation of data before statistical testing, a safeguard against the inflated significance that naive permutation schemes can produce. Dominance analysis, a technique from psychology for assessing the relative importance of correlated predictors in regression, is also available, addressing a common challenge when multiple brain properties compete to explain a neural phenomenon.

The protocol paper walks readers through complete workflows that chain these functions together to answer neurobiologically meaningful questions. Example analyses include relating brain network organization to microarchitectural features such as receptor distributions and cell-type composition, examining how strongly the brain’s structural wiring constrains its functional dynamics across different imaging modalities, and generating spatially informed null models for testing whether an observed pattern of structure-function coupling is unusual. Workflow diagrams in the paper show how data flow from raw parcellated imaging outputs, through the toolkit’s conversion, modeling and statistical layers, to interpretable figures, giving trainees a template they can adapt to their own projects rather than a black box they must trust blindly.

Beyond its technical content, the article makes a cultural argument that is likely to resonate far beyond one laboratory. The authors position netneurotools as a necessary counterweight to out-of-the-box software packages, arguing that smaller, ad hoc functions deserve recognition as real scientific contributions. By opening a window into the inner workings of a laboratory, the toolkit invites a new kind of discourse among research groups, one in which the unglamorous glue code that actually holds a project together is shared, critiqued and improved collectively. The package has been open to contributions from neuroscientists across the globe since its inception, and its development by trainees reflects a deliberate pedagogical choice: writing and maintaining shared infrastructure is itself a form of scientific training.

The timing of this publication is significant. A recent assessment of open-source neuroscience software described the field’s dependence on volunteer-maintained tools as precarious, and the proliferation of analysis pipelines has made reproducibility a persistent concern. By documenting their toolkit in a peer-reviewed protocols journal, the Misic laboratory is making a case that sustainability in computational neuroscience depends not only on large, polished platforms but also on transparent, community-maintained collections of mid-sized tools that bridge the gaps between them. The approach aligns with the FAIR principles for research software, which call for software to be findable, accessible, interoperable and reusable.

For a field whose data keep multiplying in modality and scale, the message is practical and quietly radical at once. The connectome may be the most complicated object ever mapped, but the daily work of studying it is made of thousands of small, concrete operations: converting a file, resampling a surface, rewiring a network, testing a spatial statistic. netneurotools gathers those operations into one open, living toolbox, and in doing so suggests that the health of network neuroscience may depend as much on how generously its practitioners share their everyday tools as on any single breakthrough analysis.

Subject of Research: An open-source, trainee-developed Python toolkit for network neuroscience analysis and brain imaging data integration

Article Title: netneurotools: a trainee-oriented approach to network neuroscience

Article References: Liu, Z.-Q., Bazinet, V., Hansen, J. Y., Milisav, F., Luppi, A. I., Ceballos, E. G., Farahani, A., Suarez, L. E., Shafiei, G., Markello, R. D., & Misic, B. (2026). netneurotools: a trainee-oriented approach to network neuroscience. Nature Protocols. https://doi.org/10.1038/s41596-026-01446-7

Image Credits: AI Generated

DOI: 10.1038/s41596-026-01446-7

Keywords: netneurotools, network neuroscience, Python toolkit, brain imaging, connectomics, open-source software, brain networks, spatial statistics, null models, FAIR principles, McGill University, Nature Protocols