• Corpus ID: 235829738

Extreme Precipitation Seasonal Forecast Using a Transformer Neural Network

@article{Civitarese2021ExtremePS,
  title={Extreme Precipitation Seasonal Forecast Using a Transformer Neural Network},
  author={Daniel Civitarese and Daniel Szwarcman and Bianca Zadrozny and Campbell D. Watson},
  journal={ArXiv},
  year={2021},
  volume={abs/2107.06846}
}
An impact of climate change is the increase in frequency and intensity of extreme precipitation events. However, confidently predicting the likelihood of extreme precipitation at seasonal scales remains an outstanding challenge. Here, we present an approach to forecasting the quantiles of the maximum daily precipitation in each week up to six months ahead using the temporal fusion transformer (TFT) model. Through experiments in two regions, we compare TFT predictions with those of two baselines… 

Figures and Tables from this paper

Extreme precipitation forecasting using attention augmented convolutions

Extreme precipitation wreaks havoc throughout the world, causing billions of dollars in damage and uprooting communities, ecosystems, and economies. Accurate extreme precipitation prediction allows

Intercomparison of deep learning architectures for the 1 prediction of precipitation fields

TLDR
A comprehensive evaluation of a set of deep learning architectures to realistically simulate precipitation, including heavy precipitation events and extreme events over the European domain and finds that a simplified version of the original U-Net with a single encoder-decoder level achieves similar skill scores as deeper versions for predicting precipitation extremes, reducing overall complexity and computing resources.

Feature scaling and attention convolutions for extreme precipitation prediction

  • Computer Science
  • 2022
TLDR
This work presents an in-depth exploration of the use of deep neural networks for the prediction of short-term severe precipitations, and determined that feature scaling is crucial in the context of severe precipitation prediction.

References

SHOWING 1-10 OF 15 REFERENCES

Sub‐seasonal to seasonal prediction of rainfall extremes in Australia

Seasonal climate prediction to date has largely focussed on probabilistic forecasts for above‐ and below‐average conditions in climate means. Here, we examine the possibility of making sub‐seasonal

A data-driven multi-model ensemble for deterministic and probabilistic precipitation forecasting at seasonal scale

Seasonal precipitation forecasting is valuable for regional water management and agricultural food security. Current numerical models have large uncertainty in model structure, parameterization and

Application of a Hybrid Statistical–Dynamical System to Seasonal Prediction of North American Temperature and Precipitation

Recent research demonstrates that dynamical models sometimes fail to represent observed teleconnection patterns associated with predictable modes of climate variability. As a result, model forecast

The sub-seasonal to seasonal prediction project (S2S) and the prediction of extreme events

The sub-seasonal to seasonal prediction project (S2S) is a 5-year project, established in 2013 by the World Weather Research Program (WWRP) and the World Climate Research Program (WCRP). This paper

ENSO and Extreme Rainfall Events in South America

Abstract The influence of the opposite phases of ENSO on the frequency of extreme rainfall events over South America is analyzed for each month of the ENSO cycle on the basis of a large set of daily

The climate hazards infrared precipitation with stations—a new environmental record for monitoring extremes

TLDR
The Variable Infiltration Capacity model, a novel blending procedure incorporating the spatial correlation structure of CCD-estimates to assign interpolation weights, is presented and it is shown that CHIRPS can support effective hydrologic forecasts and trend analyses in southeastern Ethiopia.

ENSO Normals: A New U.S. Climate Normals Product Conditioned by ENSO Phase and Intensity and Accounting for Secular Trends

Climate normals are traditionally calculated every decade as the average values over a period of time, often 30 years. Such an approach assumes a stationary climate, with several alternatives

S2S reboot: An argument for greater inclusion of machine learning in subseasonal to seasonal forecasts

TLDR
It is suggested that S2S prediction will be most beneficial to the public by incorporating mixed or a hybrid of dynamical forecasts and updated statistical techniques such as machine learning.