How Can We Improve the Accuracy and Timeliness of Extreme Weather Forecasts to Help Reduce the Losses and Damages?| Report by CBCGDF-SSBSP (3)

At present, human society is suffering from various extreme weather such as rainstorms and floods, high temperature heat waves, strong winds and dust. In such a severe situation, what can ordinary people do? What do they need to know? How to further enhance the practice and understanding of environmental protection and green development among the general public?

Professor YUE Xiaoguang, Executive Secretary General of the China Biodiversity Conservation and Green Development Foundation South-South Biodiversity Science Project(CBCGDF-SSBSP) and European University of Cyprus, sorted out some relevant questions and interviewed several members of the International Engineering and Technology Association (IETI), hoping that they can give some easy-to-understand answers to this. The following is the relevant interview content:

Q: How can we improve the accuracy and timeliness of extreme weather forecasts to help reduce the losses and damage?

Rita Yi Man Li, Secretary General of IETI and Professor of Hong Kong Shue Yan University: Improving the accuracy and timeliness of extreme weather forecasts is crucial to reducing losses. Here are some recent advances in this field:

1) Generative AI for ensemble forecasting: Google Research researchers introduced SEEDS (Scalable Ensemble Envelope Diffusion Sampler), a generative AI model that can efficiently generate ensembles of weather forecasts using diffusion models. SEEDS significantly reduces computational costs and is able to better describe rare or extreme weather events. In addition, Google DeepMind's GraphCast model can predict weather conditions 10 days in advance with higher accuracy and speed.

2) Deep learning technology: Deep learning models have become a promising tool for weather forecasting, including extreme event prediction. Researchers are using these techniques to improve forecast accuracy and enhance our understanding of weather dynamics. For example, one study proposed using deep generative models to improve the accuracy and resolution of precipitation forecasts, which is critical for predicting extreme rainfall events.

3) Probabilistic Forecasts: Weather is an inherently random phenomenon, and deterministic models have limitations due to the chaotic nature of the atmosphere (the “butterfly effect”). Probabilistic forecasts, such as those generated by SEEDS, provide estimates of uncertainty, enabling better decision making.

Original article:https://mp.weixin.qq.com/s/lahDfHm5L2YK66suZ0yAFw

Translator:Daisy

Checked by Sara

Editor: Daisy

Contact: v10@cbcgdf.org; +8617319454776

Contribution

Do you know? We rely on crowd-funding and donations. You have the opportunity to help an international movement to advance biodiversity conservation. Donate TODAY to power up the movement to make it a better world for all life.

Donation(501C3)Paypal: intl@wbag.org 

https://www.paypal.com/donate/?hosted_button_id=2EYYJJZ8CGPLE