Dear Colleagues,

Wind forecasting can be carried out utilizing a number of methods. Wind forecasts can come from numerical prediction models where the dynamical and thermodynamical variables are solved from a set of coupled partial differential equations based on the principles of geophysical fluid dynamics and thermodynamics or empirical models using statistics or machine learning. Additionally, wind forecasts can be derived by combining numerical weather forecasts and statistical or machine learning models.

Wind diagnostics may employ methods to characterize the temporal and spatial variabilities using methods such as Fourier analysis. In addition, machine learning methods such as self-organizing maps have been used to identify different wind regimes.

Many countries are adopting renewable energy, such as wind energy, in order to reduce the consumption of fossil fuels to combat pollution and climate change. However, accurate prediction of wind speed/power is essential in power grid integration due to the intermittent nature of wind power. Wind speed forecasts from numerical weather prediction models have been used to provide wind power forecasts by using the power curves or diagnostic relationship between wind speed and wind power. Machine learning models have also been used with numerical weather forecasts to provide wind power forecasts.

Manuscripts on all aspects of wind and wind power forecasting and diagnostics are welcome for this Special Issue.

Dr. William Cheng
Guest Editor

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Atmosphere is an international peer-reviewed open access monthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2400 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

Authors: Xiaohui Li¹, Xinhai Han², Jingsong Yang^1,2,3∗, Jiuke Wang⁴, Guoqi Han⁵

1 State Key Laboratory of Satellite Ocean Environment Dynamics, Second Institute of

Oceanography, Ministry of Natural Resources, Hangzhou 310012, China

2 School of Oceanography, Shanghai Jiao Tong University, Shanghai 200030, China

3 Southern Marine Science and Engineering Guangdong Laboratory (Zhuhai), Zhuhai 519082,

China

4 National Marine Environmental Forecasting Center, Beijing 100081, China

5 Fisheries and Oceans Canada, Institute of Ocean Sciences, Sidney, BC, Canada, V8L 4B2,

Canada

Abstract: An improved spatiotemporal sequence model for wind speed prediction is proposed based on attention mechanisms and generative adversarial network (GAN). The model adopts ECMWF wind field forecast data for training and can predict wind fields for 12 or 24 hours forecasts at intervals of 3 hours. The Attention Spatiotemporal Predictive GAN (AST-GAN) model will utilize both temporal and spatial features for wind speed forecasting by incorporating the Long Short-Term Memory (LSTM) Network for capturing spatiotemporal dependencies and the attention mechanism for helping the model weigh the significance of different spatial features. Furthermore, the GAN component allows for probabilistic forecasting, providing a measure of uncertainty in the predictions. The proposed AST-GAN model will undergo comprehensive evaluations of the spatiotemporal uncertainty of the wind speed, to determine the efficacy of the model in providing accurate and reliable predictions. The results of these evaluations will provide valuable insights into the strengths and limitations of the proposed model.