HURRI-GAN: A Novel Approach for Hurricane Bias-Correction Beyond Gauge Stations using Generative Adversarial Networks
Authors: Noujoud Nadera, Hadi Majed, Stefanos Giaremis, Rola El Osta, Clint Dawson, Carola Kaiser, Hartmut Kaiser
Organizations: Center of Computation and Technology, Louisiana State University, Baton Rouge, 70803 LA, US · Saint-Joseph University of Beirut, Beirut, 1104, Lebanon · Department of Physics, Aristotle University of Thessaloniki, Thessaloniki , 54124, Greece · Center for Interdisciplinary Research and Innovation, Aristotle University of Thessaloniki, Thessaloniki, 57001, Greece · Oden Institute for Computational Engineering and Sciences, The University of Texas at Austin, Austin 78712, TX, USA · Department of Computer Science, Louisiana State University, Baton Rouge, 70803 LA, US
The coastal regions of the eastern and southern United States are impacted by severe storm events, leading to significant loss of life and properties. Accurately forecasting storm surge and wind impacts from hurricanes is essential for mitigating some of the impacts, e.g., timely preparation of evacuations and other countermeasures. Physical simulation models like the ADCIRC hydrodynamics model, which run on high-performance computing resources, are sophisticated tools that produce increasingly accurate forecasts as the resolution of the computational meshes improves. However, a major drawback of these models is the significant time required to generate results at very high resolutions, which may not meet the near real-time demands of emergency responders. The presented work introduces HURRI-GAN, a novel AI-driven approach that augments the results produced by physical simulation models using time series generative adversarial networks (TimeGAN) to compensate for systemic errors of the physical model, thus reducing the necessary mesh size and runtime without loss in forecasting accuracy. We present first results in extrapolating model bias corrections for the spatial regions beyond the positions of the water level gauge stations. The presented results show that our methodology can accurately generate bias corrections at target locations spatially beyond gauge stations locations. The model's performance, as indicated by low root mean squared error (RMSE) values, highlights its capability to generate accurate extrapolated data. Applying the corrections generated by HURRI-GAN on the ADCIRC modeled water levels resulted in improving the overall prediction on the majority of the testing gauge stations.
Figures & tables
Figure 1: Overview of the methodology framework: (A) Data Pre-processing phase includes offset extraction using Eq. 2.1 , data cleaning, and normalization. (B) Modeling phase involves the structure of the TimeGAN components and model evaluation using regression metrics. (C) Output phase includes the final application of the pre-trained model for bias extrapolation. The generated offsets are then used to correct the forecasted data using Eq. 2.2 .
Hurricane
Category
No. of stations
No. of hourly offsets
Ian (2022)
H5
250
26250
Harvey (2017)
H4
247
25935
Ida (2021)
H4
264
18216
Idalia (2023)
H4
304
31920
Matthew (2016)
H4
236
24780
Hermine (2016)
H1
259
27195
Table 1: Hurricanes considered in this study, their category based on the Saffir-Simpson hurricane scale [ 46 ] , the number of station and the total amount of hourly offset data in each.
Figure 2: Station clustering for Hurricane Harvey (2017). This map illustrates the clustering of stations based on their coordinates. A total of 24 clusters were identified, with each color representing a different cluster.
Embedder, Recovery, Generator
Layers
No. of Neurons
Activation
Output Shape
GRU
256
–
5*256
GRU
256
–
5*256
GRU
256
–
5*256
GRU
256
–
5*256
GRU
256
–
5*256
Table 2: Schematic overview of the structure of the HURRI-GAN model (GRU: Gate Recurring Unit, Bi-GRU: Bidirectional Gated Recurrent Unit).
Hurricane
MSE
RMSE
MAE
Ian (2022)
0.129
0.359
0.241
Harvey (2017)
0.121
0.347
0.214
Ida (2021)
0.116
0.34
0.197
Idalia (2023)
0.124
0.352
0.22
Matthew (2016)
0.066
0.257
0.169
Hermine (2016)
0.085
0.292
0.177
Table 3: Metrics in feet of the real vs ML-predicted offsets generated via our previous LSTM-based model [ 30 ] for each of the considered hurricanes.
Error Metric
RMSE
Parameters
No. of Layers
No. of neurons
No. of Epochs
(ft.)
[4, 5]
[128, 256]
[2000, 3000, 4000]
Ian(2022)
Min
0.654
5
256
4000
Max
0.724
4
128
4000
Mean - Std
0.669
5
256
3000
Mean + Std
0.707
4
256
2000
Table 4: Hyperparameter tuning results for the HURRI-GAN model. The table presents the RMSE values (in feet) along with the optimal hyperparameters for each hurricane, including the number of network layers, number of neurons, and number of epochs. The values for each of the aforementioned parameters considered in this analysis are listed in brackets. The number of layers is analyzed for Generator, Embedder and Recovery. The supervisor is one layer less than the other components, while the discriminator architecture always consists of two layers.
Figure 3: RMSE values for the testing stations across different hurricanes. The map for each hurricane shows the locations of the testing stations along with their corresponding RMSE values (in feet), indicated by the color gradient. The considered hurricanes are Harvey (2017), Hermine (2016), Ian (2022), Ida (2021), Idalia (2023), and Matthew (2016) and the considered agencies are NOAA, USACE, USGS, TCOON, and PRSN. The hurricane paths are outlined in red.
Figure 4: RMSE distribution of the extrapolated offsets with the HURRI-GAN model in the testing stations by hurricane and agency. The considered hurricanes are Harvey (2017), Hermine (2016), Ian (2022), Ida (2021), Idalia (2023), and Matthew (2016) and the considered agencies are NOAA, USACE, USGS, TCOON, and PRSN.
Figure 5: Comparison of observed ( blue ), modeled ( orange ) and HURRI-GAN-corrected modeled ( green ) for: (a) Hermine (2016, category H1), (b) Harvey (2017, category H4), and (c) Ian (2022, category H5). The left side displays the location of two testing gauge stations (marked by colored disks) and the hurricane path (in red); while the right side shows the evaluation of the regression performance for these gauge stations. Evaluation statistics in each plot represent the performance of regression between modeled and observed water levels (without AI) and HURRI-GAN bias corrected and observed water levels (with AI).
Figure 6: Computational time (in seconds) required for the HURRI-GAN model to sequentially extrapolate bias corrections (i.e., inference time), as a function of the total amount of corrections.
Storm surge forecasting remains a critical challenge in mitigating the impacts of tropical cyclones on coastal regions, particularly given recent trends of rapid intensification and increasing nearshore storm activity. Traditional high fidelity numerical models such as ADCIRC, while robust, are often hindered by inevitable uncertainties arising from various sources. To address these challenges, this study introduces StormNet, a spatio-temporal graph neural network (GNN) designed for bias correction of storm surge forecasts. StormNet integrates graph convolutional (GCN) and graph attention (GAT) mechanisms with long short-term memory (LSTM) components to capture complex spatial and temporal dependencies among water-level gauge stations. The model was trained using historical hurricane data from the U.S. Gulf Coast and evaluated on Hurricane Idalia (2023). Results demonstrate that StormNet can effectively reduce the root mean square error (RMSE) in water-level predictions by more than 70% for 48-hour forecasts and above 50% for 72-hour forecasts, as well as outperform a sequential LSTM baseline, particularly for longer prediction horizons. The model also exhibits low training time, enhancing its applicability in real-time operational forecasting systems. Overall, StormNet provides a computationally efficient and physically meaningful framework for improving storm surge prediction accuracy and reliability during extreme weather events.
Noujoud Nader, Stefanos Giaremis, Clint Dawson +3
Center for Computation and Technology, Louisiana State University, Baton Rouge, 70803, LA, US. · Department of Physics, Aristotle University of Thessaloniki, Thessaloniki, 54124, Greece. · Oden Institute for Computational Engineering and Sciences, The University of Texas at Austin, Austin, 78712, TX, US.
Predicting storm surge induced by landfalling tropical cyclones is crucial for flood mitigation and coastal risk management. Traditionally, physics-based numerical models simulate storm surge by solving the Navier--Stokes equations using numerical methods, but these simulations are computationally expensive. Generative models are promising for storm surge emulation because they can generate diverse realizations rather than producing a single deterministic prediction. However, their use for storm surge emulation remains largely unexplored. In this paper, we leverage diffusion models for storm surge surrogate modeling, combining a baseline prediction stage with conditional generation to provide a more interpretable modeling framework. We develop SurgeGen, a two-stage generative framework for generating storm surge scenarios conditioned on hypothetical storms with parameters defined in a continuous space. First, a baseline model produces a coarse estimate of the storm surge height. This estimate then conditions a diffusion model, which generates refined storm surge scenarios that better capture spatial patterns and variability. We demonstrate that our approach can generate realistic and diverse storm surge scenarios under conditions both within and outside the training distribution.
Shunan Zheng, John J. Hasenbein
Operations Research and Industrial Engineering University of Texas at Austin Austin, TX, USA
Emergency managers need to know where floodwater is, how deep it is, and how it will change over the coming hours across an entire river basin. During a flood, however, real-time measurements come from only a handful of stream gauges, and high-resolution hydrodynamic models are too costly to rerun each time new data arrive or to run as large ensembles. We present C-STRIDE, an observation-driven AI digital twin that turns short records from a few stream gauges, together with terrain and rainfall, into basin-wide maps of water depth and extends these predictions up to a day ahead. It is trained on simulations from a calibrated two-dimensional hydrodynamic model and needs no separate data-assimilation step. In the Des Plaines River basin near Chicago, six gauges inform predictions over 4.2 million 30-m grid cells. Terrain improves the predictions most, rainfall keeps errors from growing over longer horizons, and together they reduce errors by about 40% compared with gauge records alone. When future rainfall is known, errors remain near 15% one day ahead, compared with nearly 40% without rainfall. Given real instead of simulated gauge records, the model shifts its predictions toward the observed hydrographs at three of six gauges without retraining, and it runs about 150 times faster than the hydrodynamic model. These results show how sparse gauges, terrain, and rainfall can be combined into fast, continuously updated flood predictions, a step toward operational flood digital twins that still requires testing with real-time data and rainfall forecasts.
Yanjie Tong, Phillip Si, Yuan Qiu +1
School of Computational Science and Engineering, Georgia Institute of Technology, 756 West Peachtree Street Northwest, Atlanta, 30308, GA, USA