Can Machines Learn to Predict Weather? Using Deep Learning to Predict Gridded 500‐hPa Geopotential Height From Historical Weather Data

Abstract

We develop elementary weather prediction models using deep convolutional neural networks (CNNs) trained on past weather data to forecast one or two fundamental meteorological fields on a Northern Hemisphere grid with no explicit knowledge about physical processes. At forecast lead times up to 3 days, CNNs trained to predict only 500‐hPa geopotential height easily outperform persistence, climatology, and the dynamics‐based barotropic vorticity model, but do not beat an operational full‐physics weather prediction model. These CNNs are capable of forecasting significant changes in the intensity of weather systems, which is notable because this is beyond the capability of the fundamental dynamical equation that relies solely on 500‐hPa data, the barotropic vorticity equation. Modest improvements to the CNN forecasts can be made by adding 700‐ to 300‐hPa thickness to the input data. Our best performing CNN does a good job of capturing the climatology and annual variability of 500‐hPa heights and is capable of forecasting realistic atmospheric states at lead times of 14 days. Although our simple models do not perform better than an operational weather model, machine learning warrants further exploration as a weather forecasting tool; in particular, the potential efficiency of CNNs might make them attractive for ensemble forecasting.

Optimization of Deep Learning Precipitation Models Using Categorical Binary Metrics

“…It is increasingly common in meteorology to use machine learning approaches for identifying patterns in the atmosphere using large amounts of historical data (Dueben & Bauer, ; Scher & Messori, ; Ukkonen & Mäkelä, ; Weyn et al, ). This approach, of extracting the underlying physical relationships in the atmosphere from data, opens an opportunity to explore new algorithms that optimize the output based on different verification metrics.…”

Deep spatial transformers for autoregressive data-driven forecasting of geophysical turbulence

“…Such work has been done in the past for atmospheric dynamics and, more generally, for chaotic dynamical systems [1,2,3,4], including ocean dynamics [5]. Another approach involves using the resolved field of a complex chaotic and turbulent flow directly in a data-driven model to be trained on and subsequently evolved forward in time as has been shown in [6,7,8,9]. The assumption here, is that the data-driven model learns the effects of the unresolved or under-resolved sub-grid scale processes directly on the resolved field from a long time history of the fields in the training dataset.…”

Improving data-driven global weather prediction using deep convolutional neural networks on a cubed sphere

“…Using convolutional neural networks (CNNs), [37] and [39] trained an algorithm on simulations from a simplified GCM that significantly outperformed baseline metrics and effectively captured the simplified-GCM dynamics. [43], hereafter WDC19, trained CNNs similar to those of [37] and [39] with over 20 years of historical reanalysis data to produce forecasts of 500 hPa height and 300-700 hPa thickness over the northern hemisphere. Their best CNN formulation was able to outperform a climatological benchmark for root-mean-squared error (RMSE) in the 500 hPa height field out to about 5 days of forecast lead time.…”