Publications

2024

1. DiffESM: Conditional Emulation of Temperature and Precipitation in Earth System Models with 3D Diffusion Models

   Seth Bassetti, Brian Hutchinson, Claudia Tebaldi, and 1 more author

   Journal of Advances in Modeling Earth Systems, 2024

   Abs HTML PDF

   Earth System Models (ESMs) are essential tools for understanding the impact of human actions on Earth’s climate. One key application of these models is studying extreme weather events, such as heat waves or dry spells, which have significant socioeconomic and environmental consequences. However, the computational demands of running a sufficient number of simulations to analyze the risks are often prohibitive. In this paper we demonstrate that diffusion models – a class of generative deep learning models – can effectively emulate the spatio-temporal trends of ESMs under previously unseen climate scenarios, while only requiring a small fraction of the computational resources. We present a diffusion model that is conditioned on monthly averages of temperature or precipitation on a 96×96 global grid, and produces daily values that are both realistic and consistent with those averages. Our results show that the output from our diffusion model closely matches the spatio-temporal behavior of the ESM it emulates in terms of the frequency of phenomena such as heat waves, dry spells, or rainfall intensity.

2023

1. 

   DiffESM: Conditional Emulation of Earth System Models with Diffusion Models

   Seth Bassetti, Brian Hutchinson, Claudia Tebaldi, and 1 more author

   In ICLR 2023 Workshop on Tackling Climate Change with Machine Learning, 2023

   Abs HTML PDF

   Earth System Models (ESMs) are essential tools for understanding the impact of human actions on Earth’s climate. One key application of these models is studying extreme weather events, such as heat waves or dry spells, which have significant socioeconomic and environmental consequences. However, the computational de- mands of running a sufficient number of simulations to analyze the risks are often prohibitive. In this paper we demonstrate that diffusion models – a class of gen- erative deep learning models – can effectively emulate the spatio-temporal trends of ESMs under previously unseen climate scenarios, while only requiring a small fraction of the computational resources. We present a diffusion model that is con- ditioned on monthly averages of temperature or precipitation on a 96 × 96 global grid, and produces daily values that are both realistic and consistent with those av- erages. Our results show that the output from our diffusion model closely matches the spatio-temporal behavior of the ESM it emulates in terms of the frequency of phenomena such as heat waves, dry spells, or rainfall intensity.

2022

1. ICLR

   Conditional Emulation of Global Precipitation with Generative Adversarial Networks

   Brian Hutchinson, Alexis Ayala, Chris Drazic, and 7 more authors

   In AGU Fall Meeting Abstracts, 2022

   Abs

   Estimating risk of future impacts as driven by climate hazards, especially extremes, requires a robust characterization of these hazards’ statistics. Earth System Models enable research on the earth’s future climate under alternative plausible scenarios, but are extremely costly to run, providing only a limited number of scenarios/ensemble members. Earth System Model output emulators aim to supplement ESMs, allowing computationally efficient generation of new scenarios or realizations of internal variability, the latter being particularly important when studying extremes. In this paper we propose an approach to generating realistic time series of global daily precipitation fields requiring orders of magnitude less computation, using a conditional generative adversarial network (GAN) as an emulator of an Earth System Model (ESM). Specifically, we present a GAN that emulates daily precipitation output from a fully coupled ESM, conditioned on monthly mean values. The GAN is trained to produce spatio-temporal samples: 28 days of precipitation in a 92 × 144 regular grid discretizing the globe. We evaluate the generator by comparing generated and real distributions of precipitation metrics including average precipitation, average fraction of dry days, average dry spell length, and average precipitation above the 90th percentile, finding that the generated samples closely match those of real data, even when conditioned on climate scenarios never seen during training.