Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Žiga Zaplotnik has a total of 49 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Forecast‐Error Diagnostics in Neural Weather Models

    by Uroš Perkan, Gregor Skok, Žiga Zaplotnik

    Deep learning (DL) weather prediction models offer notable advantages over traditional physics‐based models, including auto‐differentiability and low computational cost, enabling detailed diagnostics of forecast errors. Using our convolutional encoder–decoder model, ConvCastNet, we relax selected subdomains of the forecast fields systematically towards “true” weather states (ECMWF ERA5 reanalysis) and monitor the forecast skill gain in other regions. Our results show that a medium‐range midlatitude forecast improves substantially when the stratosphere and boundary layer are relaxed, while relaxation of the tropical atmosphere has a limited effect. This underscores the need for a more accurate representation of the stratosphere and the planetary boundary layer to improve medium‐range weather predictability. Additionally, we investigate the relationship between the forecast‐error sensitivity to initial conditions and relaxation experiments. By utilising auto‐differentiability, we identify overlapping regions of large error sensitivity and strong forecast skill improvement from relaxation. Average midlatitude error sensitivity to initial conditions shows negligible influence from the Tropics, corroborating the results of the tropical relaxation experiments. The error sensitivity shows a physically consistent influence of upstream weather dynamics and sea‐surface temperatures on forecast accuracy. The latter also highlights the importance of representing the atmosphere–ocean coupling accurately in numerical weather prediction models. This combined approach could provide valuable heuristics for diagnosing neural model errors and guiding targeted model improvements.


  • Open Access

    Comment on egusphere-2026-77

    by Marco Stefanelli, Žiga Zaplotnik, Gregor Skok

    In three-dimensional variational data assimilation (3DVar) for numerical weather prediction (NWP), the observation operator H plays a central role by mapping model state variables to an observation equivalent. For weather radar, however, specifying H is particularly challenging: reflectivity is a nonlinear, microphysics-dependent diagnostic quantity that only indirectly relates to the model’s prognostic variables, making traditional parameterised radar operators complex, regime-dependent and difficult to tune. In this study, we propose a neural-network (NN)-based observation operator for radar reflectivity and apply it within a 3DVar data assimilation (DA) framework. Using five years (2019–2023) of radar reflectivity data from the Lisca radar and 4.4 km-resolution short-range forecasts from ALADIN model over Slovenia, we train a convolutional encoder–decoder neural network to map model temperature, humidity, horizontal wind components and surface pressure fields to radar reflectivity. Across independent test cases spanning clear-sky, stratiform and convective regimes, the NN-based operator accurately reproduces the spatial structure and intensity of observed reflectivity, relying primarily on the model state in the vicinity of the observation point. In the extreme precipitation case, which caused widespread floods in Slovenia on August 4, 2023, assimilating the full radar disc reduces the domain-averaged reflectivity root-mean-square error (RMSE) from 5.99 dBZ to 3.47 dBZ and improves the alignment between the analysed and observed convective bands. Embedded within 3DVar, the Jacobian of the NN observation operator allows radar reflectivity observations to inform model state variables, producing corresponding analysis increments. The proposed NN radar observation operator offers a flexible alternative to traditional parameterised radar operators for improving convective-storm forecasts.

  • Open Access

    Comment on egusphere-2026-77

    by Marco Stefanelli, Žiga Zaplotnik, Gregor Skok

    In three-dimensional variational data assimilation (3DVar) for numerical weather prediction (NWP), the observation operator H plays a central role by mapping model state variables to an observation equivalent. For weather radar, however, specifying H is particularly challenging: reflectivity is a nonlinear, microphysics-dependent diagnostic quantity that only indirectly relates to the model’s prognostic variables, making traditional parameterised radar operators complex, regime-dependent and difficult to tune. In this study, we propose a neural-network (NN)-based observation operator for radar reflectivity and apply it within a 3DVar data assimilation (DA) framework. Using five years (2019–2023) of radar reflectivity data from the Lisca radar and 4.4 km-resolution short-range forecasts from ALADIN model over Slovenia, we train a convolutional encoder–decoder neural network to map model temperature, humidity, horizontal wind components and surface pressure fields to radar reflectivity. Across independent test cases spanning clear-sky, stratiform and convective regimes, the NN-based operator accurately reproduces the spatial structure and intensity of observed reflectivity, relying primarily on the model state in the vicinity of the observation point. In the extreme precipitation case, which caused widespread floods in Slovenia on August 4, 2023, assimilating the full radar disc reduces the domain-averaged reflectivity root-mean-square error (RMSE) from 5.99 dBZ to 3.47 dBZ and improves the alignment between the analysed and observed convective bands. Embedded within 3DVar, the Jacobian of the NN observation operator allows radar reflectivity observations to inform model state variables, producing corresponding analysis increments. The proposed NN radar observation operator offers a flexible alternative to traditional parameterised radar operators for improving convective-storm forecasts.

  • Open Access

    A Unified Neural Background‐Error Covariance Model for Midlatitude and Tropical Atmospheric Data Assimilation

    by Boštjan Melinc, Uroš Perkan, Žiga Zaplotnik

    Estimating and modeling background‐error covariances remains a core challenge in variational data assimilation (DA). Operational systems typically approximate these covariances by transformations that separate geostrophically balanced components from unbalanced inertia‐gravity modes—an approach well‐suited for the midlatitudes but less applicable in the tropics, where different physical balances prevail. This study estimates background‐error covariances in a reduced‐dimension latent space learned by a neural‐network autoencoder (AE). The AE was trained using 40 years of ERA5 reanalysis data, enabling it to capture flow‐dependent atmospheric balances from a diverse set of weather states. We demonstrate that performing DA in the latent space yields analysis increments that preserve multivariate horizontal and vertical physical balances in both tropical and midlatitude atmosphere. Assimilating a single 500 hPa geopotential height observation in the midlatitudes produces increments consistent with geostrophic and thermal wind balance, while assimilating a total column water vapor observation with a positive departure in the nearly‐saturated tropical atmosphere generates an increment resembling the tropical response to (latent) heat‐induced perturbations. The resulting increments are localized and flow‐dependent, and shaped by orography and land‐sea contrasts. Forecasts initialized from these analyses exhibit realistic weather evolution, including the excitation of an eastward‐propagating Kelvin wave in the tropics. Finally, we explore the transition from using synthetic ensembles and a climatology‐based background error covariance matrix to an operational ensemble of data assimilations. Despite significant compression‐induced variance loss in some variables, latent‐space assimilation produces balanced, flow‐dependent increments—highlighting its potential for ensemble‐based latent‐space 4D‐Var.


  • Open Access

    Forecast Error Diagnostics in Neural Weather Models Using Gridpoint Relaxation

    by Uroš Perkan, Žiga Zaplotnik, Gregor Skok

    In physics-based numerical weather prediction models, underlying physical laws and numerical discretisation schemes offer some interpretability of the simulated processes and point to potential sources of errors. In contrast, data-driven deep-learning (DL) models lack explicit physics-based interpretation of their predictions. However, the low computational cost and auto-differentiable implementation allow for fast and extensive diagnostics of the origins of forecast errors using well-established explainable AI methods and traditional diagnostics tools. Here, we combine saliency maps and gridpoint relaxation to perform a multivariate regional analysis of the sources of forecast errors in global DL weather forecasting.We developed ConvCastNet, a DL global weather prediction model based on depth-wise separable convolutional neural networks. The model predicts 6 atmospheric variables at 13 pressure levels and includes 2 additional single-level prognostic and 7 constant input fields. The forecast is computed on a 3-degree lon-lat grid and uses 12-hour autoregressive time stepping to roll out the forecast. ConvCastNet achieves significant success in predicting geopotential at the 500 hPa pressure level, with 8.5 days of useful forecast (based on an anomaly correlation coefficient greater than 0.6), which makes it a suitable tool for diagnostics of the origin of the forecast error.We use ConvCastNet to systematically nudge subdomains of the forecast fields for 1) planetary boundary layer, 2) stratosphere, and 3) tropics towards a "true" weather state (reanalysis) and monitor the forecast skill improvements beyond selected subdomains. Our results show that an 8-day mid-latitude weather forecast improves significantly with relaxation in the stratosphere, whereas relaxation in the tropics has no significant impact on mid-latitude. This highlights the importance of accurately representing the stratosphere for medium-range weather prediction and the limited impact of tropical variability on mid-latitude forecasts.Additionally, we investigate the relationship between model error sensitivity to initial conditions and relaxation experiments. By utilising the model's auto-differentiability, we analyse saliency maps, i.e. the gradients of the forecast errors with respect to input fields, to identify overlapping regions of large error sensitivity and high impact of relaxation to the truth. We find the model sensitivity largely consistent with physics-based expectations, with local errors being sensitive to the upstream dynamics and varying sensitivity to different variables and pressure levels. We believe that this combined approach provides valuable heuristics for diagnosing neural model errors and guiding targeted model improvements.


  • Open Access

    A neural network-based observation operator for weather radar data assimilation

    by Marco Stefanelli, Ziga Zaplotnik, Gregor Skok

    Forecasting convective storms remains a significant challenge in Numerical Weather Prediction (NWP). Data Assimilation (DA) plays a crucial role by improving the initial conditions for forecasts through the integration of observational data with previous model outputs (background). Among observational platforms, weather radar is particularly valuable due to its high spatial and temporal resolution, offering detailed information for storm monitoring. Therefore, assimilating radar data into Numerical Weather Prediction models has the potential to substantially enhance the accuracy of storm forecasts. However, studies, such as Fabry and Meunier (2020), have shown that short-term precipitation forecasts produced through radar-based extrapolation methods (nowcasting) often outperform model-based forecasts with Data Assimilation. This is largely because radar primarily provides information on precipitation structures and intensities within the storm-affected region but lacks direct insights into broader environmental conditions like temperature, wind fields, and humidity. These atmospheric variables, both within and outside the storm system, are critical for accurate storm evolution forecasts. A promising approach to address this limitation is the application of machine learning (ML) to develop a more advanced observation operator for Data Assimilation. Specifically, an encoder-decoder neural network can be trained to relate Numerical Weather Prediction model variables (such as temperature, wind components, and relative humidity) to corresponding radar reflectivity observations. This ML-based observation operator captures complex non-linear relationships between model variables and radar data observations, while its Jacobian can propagate radar-derived information to other atmospheric variables in the Data Assimilation system, potentially leading to improved representation of storm dynamics and enhancing convective storm forecasting capabilities.


  • Open Access

    Variability and trends of the Pacific Walker circulation indices in multiple reanalyses

    by Katarina Kosovelj, Matic Pikovnik, Žiga Zaplotnik

    The Pacific Walker Circulation (PWC) is an important part of the global atmospheric circulation. Its strength is changing constantly due to internal variability, such as ENSO, IPO, and multi-decadal climate variability, as well as the forced response to increasing greenhouse gas concentrations. Despite the observational studies showing strengthening of the PWC, modelling studies predominantly predict its weakening with the reason for this discrepancy still unknown. Generally, results of different studies are difficult to compare, due to different datasets, different time intervals and different indices used. However, any change in PWC strength may have vastly different effects on the global climate, which underlines the importance of proper simulation of its trends in the changing climate. The main aim of our study is to assess the consistency of different PWC indices across reanalyses. We compared nine different PWC indices between six different reanalyses (ERA5, ERA-Interim, JRA-55, MERRA-2, NCEP and NOAA 20CRv3), both over their respective time spans and over the common period from 1981 to 2011. The comparison was performed for the trends of different lengths, as well as correlations between indices for each reanalysis and between reanalyses for each index. On average, time series of indices in reanalyses verify well compared to pure observation-based indices. They are able to capture the strongest El Niño years and distinguish between El Niño and La Niña states. As expected, the spread of indices is larger in the surface-based reanalyses than in the full atmospheric reanalyses, as the the lack of observations in the free atmosphere in the former make them less constrained by observations. The spread is generally the largest during the strongest El Niño events. Correlations between the indices are stronger for the indices based on physically related processes, such as indices based on surface pressure and surface wind or 500-hPa vertical velocity and upper tropospheric humidity. Indices based on surface variables are more strongly correlated between reanalyses than indices based on data in the upper-atmosphere. Trends of different lengths show mixed results, depending on the reanalyses and indices. The 20-year trends ending at about 2020 are mostly negative but not statistically significant. Ultimately, this raises a key question: Are the traditional PWC indices still suitable for capturing changes in today’s rapidly changing global climate?

  • Open Access

    Towards higher resolution data assimilation in ECMWF IFS

    by Ziga Zaplotnik, Josef Schroettle, Jorge Bandeiras, Benoit Vanniere, Emiliano Orlandi, Michael Maier-Gerber, Elias Holm, Massimo Bonavita

    Enhancing the spatial resolution of global analyses and forecasts enables weather prediction systems to more accurately capture rapidly evolving mesoscale and convective-scale atmospheric phenomena. The Integrated Forecast System (IFS) of ECMWF consists of an Earth-system model coupled with an incremental 4D-Var data assimilation (DA) system, which has recently been experimentally upgraded to higher spatial resolution.This upgrade involved higher resolution (4.4 km) 4D-Var trajectories and higher resolution (20 km) 4D-Var minimizations using the tangent linear model (TLM) and its adjoint (ADM). This configuration requires less observation thinning and resulted in greater use of high-resolution observations, such as those from multispectral imagers aboard Himawari and Meteosat satellites. Shorter time steps in TLM and ADM further enabled the use of shorter observation time slots (400 seconds), allowing for more accurate comparison between observations and their model equivalents.We demonstrate that this setup leads to significant improvements in the accuracy of initial conditions and enhances tropospheric forecast skill, extending medium-range predictability by 6 to 12 hours. We argue that the improved forecast skill can be attributed to a better representation of large-scale features in the stratosphere. Higher resolution DA and forecasts also lead to improved forecasting of extreme precipitation events, as well as the track and intensity of tropical cyclones. Furthermore, the improved initial conditions significantly increase the predictability machine-learning-based forecasts.The benefits of the higher resolution DA system are demonstrated using the case of Tropical Cyclone Otis, which made landfall as a Category 5 cyclone, but was predicted by most global and regional NWP models to peak only at tropical storm intensity – a major forecasting bust. By employing higher resolution 4D-Var DA, we were able to replicate the observed rapid intensification of TC Otis.

  • Open Access

    Exploring a neural network-based background-error covariance model

    by Boštjan Melinc, Uroš Perkan, Žiga Zaplotnik

    Data assimilation of atmospheric observations traditionally relies on variational and Kalman filter methods. Several ideas of merging data assimilation with machine learning have emerged in recent years. One possibility would be to perform variational data assimilation in the latent space of an autoencoder (AE). In our approach, we define and minimise the three-dimensional variational (3D-Var) data assimilation cost function there to determine the analysis that optimally fuses simulated observations and the encoded 24-hour neural network forecast (background), accounting for their errors. Similar to several previous studies, the latent space has Gaussian properties, which are favourable for variational data assimilation, and the climatology-based background-error covariance (B) matrix measured and represented in the latent space turns out to be quasi-diagonal, which, together with the cost function manipulation, leads to a vast computation process speed-up. The main focus of this presentation is to study the physical feasibility of the analysis increments that we obtain using a multilevel multivariate AE and the U-net type neural network forward model (NNfwd) in this kind of framework. The analysis increments after observing geopotential in the midlatitudes give a multivariate geostrophic pattern, obey the thermal wind balance, and resemble some distinctive local features, such as orography and land-sea distribution. Also, the magnitude of the analysis increment at the observation location and its standard deviation closely match their theoretical values, and the impact of the observation reasonably fades with distance in both horizontal and vertical directions. On the other hand, assimilating increased moisture in the tropics leads to increments that align with the physics of the emerging tropical convective cloud systems despite the clouds and precipitation not being explicitly forecasted by the AE and NNfwd. Also, the increments after assimilating the observations in the tropics seem to distinguish between the local horizontal length scales. Finally, we study the plausibility of the model response after running the forecast with NNfwd using the analysis as the initial condition.

  • Open Access

    The Changes of the Northern Hadley Cell Strength in Reanalyses and Radiosonde Observations

    by Matic Pikovnik, Žiga Zaplotnik

    This study investigates the mean meridional wind patterns and their long-term trends within the Northern Hadley Cell (NHC) from 1980 to 2022, using reanalysis datasets and radiosonde observations. The Hadley circulation, a key component of Earth’s atmospheric circulation, is vital in redistributing heat and momentum between the equator and mid-latitudes. Understanding changes in its strength and structure is essential for interpreting current climate dynamics and improving projections of future climate change.Our analysis reveals a consistent discrepancy between reanalyses and radiosonde measurements in representing the vertical structure of the NHC. Specifically, reanalyses tend to underestimate the mean poleward flow in the upper troposphere compared to radiosonde data while accurately capturing equatorward flow in the lower troposphere. Despite climate model projections generally indicating a weakening trend of the NHC under global warming, our examination of radiosonde data shows no statistically significant long-term trend. This absence of a clear signal in observational data introduces additional uncertainty into future projections of the Hadley circulation's response to climate change.In contrast, reanalysis datasets, including ERA5, suggest a strengthening of the NHC over recent decades, mainly driven by an intensification of the upper-tropospheric poleward branch. To assess the reliability of this trend in reanalyses, we analysed the analysis increments from ERA5, which represent the corrections applied during data assimilation. Our findings indicate that the observed strengthening trend is not an artefact introduced by the data assimilation process. Instead, the analysis increments tend to nudge the first-guess model state, which underestimates the NHC strength, toward stronger meridional circulation, aligning more closely with assimilated observations.These results highlight the importance of using multiple data sources when assessing long-term atmospheric trends and caution against relying solely on either reanalyses or climate model projections. The divergence between observational and reanalysis-based trends underscores ongoing uncertainties in the behaviour of large-scale circulation systems under climate change and calls for further investigation into the causes of these discrepancies.