Using neural networks to predict weather patterns and trends - challenges and opportunities. (21st October 2022)

• Record Type: Journal Article
• Authors: Mathur, Sandeep
• Is Part Of: International journal of computer aided engineering and technology. Volume 17:Number 4(2022)
• Page Start: 388
• Page End: 396
• Publication Date: 2022-10-21
• Subjects: neural networks -- weather prediction -- prediction models; Computer-aided engineering -- Periodicals; 620.00285
• Languages: English
• ISSNs: 1757-2657

Abstract

Weather prediction systems today are a function of a multitude of variables, which result in complicated and convoluted systems of mathematical simulations and equations. Simulating along the lines of these systems requires enormous amounts of computational power and resources at the scale of supercomputers and super-scalar and massively parallelised architectures. There has been a recent attempt to avoid such costly and often inefficient calculations by making use of artificial neural networks (ANN) to predict future weather and climate patterns at large. This paper will explore such methods of varying complexity, and review their effectiveness as compared to numerical weather predictions (NWP) models.