Swedish Joint Project for Next-Generation Modeling of Cloud Feedbacks and Climate Change using AI

Project leader: Dr V. Phillips, University of Lund, Sweden

Co-Investigator: M. Ohlsson, Department of Astronomy and Theoretical Physics,

Lunds Universitet, Lund, Sweden

Co-Investigators: T. Landelius, SMHI, Norrkoping, Sweden

Stakeholders: S. Montin, Energiforsk, Sweden; M. Billstein, Vattenfall, Sweden

Overview

The project is entitled “Next-Generation Modeling of Cloud Feedbacks and Climate Change using AI: Implications for Alternative Energy”. It is led by Lund University and was funded in Autumn 2020 by Sweden’s Innovation Agency (Vinnova). The project will last about 3 years (November 2021-2024).

This joint project involves four institutions in Sweden:

· SMHI, in Norrkoping;

· Energiforsk;

· Vattenfall;

· Department of Physical Geography and Ecosystem Science, Lund University, in Sweden.

The aim is to innovate a new technology for prediction of climate change. Since clouds control the climate sensitivity to any forcing, the coarse representations of clouds in a global model will be replaced by our detailed cloud scheme accelerated by AI.

This advance will enable cloud-radiation feedbacks, which dominate the climate sensitivity to greenhouse gas emissions, to be assessed more reliably. Implications for long-term policy decisions about renewable energy sources in Sweden will be assessed.

Conventional climate models must treat clouds and related physical processes statistically since they are too small spatially to resolve on the numerical grid. From Sherwood et al. (2014).

AI approach with deep learning

The plan is to apply a neural network to predict cloud properties from large-scale conditions resolved by the global model. Any neural network consists of layers of neurons, each receiving a signal and sending another signal to other neurons in the next layer. The strength of the connections adjusts during learning. A neural network can be trained with specimen datasets to predict an output from given inputs. Once trained, the outputs are predicted by the combination of input values.

In the project, detailed simulation of an observed storm will provide the initial training dataset. The neural network, once trained, will be applied in every grid-box of the global model. The type of neural network to be applied is a ‘deep learning’ convolutional scheme that involves both time and space with a memory of past cloud conditions.

In short, this AI scheme will be applied in every gridbox of the global model to treat the unresolved small-scale processes such as clouds and radiation.

Neural networks involve layers of neurons, each being connected to other neurons in the next layer.

Cloud modelling

The aerosol-cloud model (AC) predicts the microphysical, dynamical and radiative properties of clouds. AC includes at least 7 chemical species of aerosol: primary biological aerosols, non-biological insoluble organics, soluble organics, sea-salt, ammonium sulphate, mineral dust and soot. Interstitial and immersed/embedded components of each aerosol species are predicted in cloud and precipitation. The model predicts active IN and cloud condensation nuclei (CCN) concentrations from the chemistry, sizes and loading of aerosols.

An observed case of a storm is simulated by AC and after verifying accuracy of the simulation, the dataset is used to train a ‘deep learning’ form of neural network.

A typical meteorological research aircraft for sampling clouds in field campaigns such as MC3E.

Aircraft measurements

Observed cases of clouds shown above will be used to validate the AC model. Aircraft flew through clouds measuring ice and drop concentrations while radar and ground-based measurements. National Center for Atmospheric Research in USA will help to provide the data from past field campaigns:

· MC3E (funded by DoE/NASA): Warmer-based clouds in precipitating deep convective

systems were sampled by aircraft during May 2011 over Oklahoma;

Comparison of the 3D detailed simulation (not the AI scheme) of a storm observed in MC3E with coincident aircraft observations, for the ice concentration. Generally, ice concentrations are the most challenging quantity to predict correctly, as many processes of ice initiation must be represented. From simulations described by Patade et al. (2022).

Current Progress in Project

A postdoctoral scholar began work on the project in late 2021. Another two scholars will join the project in late 2023 for inclusion of storm electrification aspects in the AI scheme.

The current scholar has created the AI scheme, and in Year 1 applied it ‘off-line’ to simulate the MC3E case in comparison with a detailed simulation by a cloud model. The 3D cloud simulation (see diagram above) provided the data for training the AI scheme.

In Year 2, the AI scheme has been implemented in a global model, namely the Community Atmospheric Model (CAM) from NCAR in USA. Technical obstacles have been overcome, such as how to do a mixed-language compilation of the global model code (‘FORTRAN’) to enable a simulation with the AI scheme (‘python’). Training of the AI scheme with a global dataset is now being done to enable a present-day climate simulation.

Moreover, two papers by Phillips (2022, 2023) have been published in a peer-reviewed journal about new theories for microphysical quasi-equilibrium in clouds and the balance between warm and cold precipitation. These theories, though not involving AI directly, have informed our choice of input variables for the neural network. In these, and in some other papers also published over the last year or so, support by Vinnova has been acknowledged.

Stakeholder meetings occurred in September 2022, May 2023, and November 2023, where results for the performance of the AI scheme in the MC3E case and in the global model were communicated. The stakeholders are business managers in the renewable energy industry.

Papers relevant to the project published so far include a study of time-dependence of ice initiation by Waman et al. (2023) and a global modeling study of the impact from secondary ice processes by Zhao et al. (2023). Moreover, a paper will appear in Nature Communications about the mechanisms for precipitation in cloud ensembles worldwide, by Gupta et al. (2023).

Bibliography

Gupta, A. K., D. Waman, A. Deshmukh, A. Jadav, S. Patade, V. T. J. Phillips, A. Bansemer, J. A. Martins and F. Goncalves, 2023: The microphysics of the warm-rain and ice crystal processes of precipitation in simulated continental convective storms. Commun Earth Environ. (Nature portfolio), 4, 226

Patade, S., Waman, D., Deshmukh, A., Gupta, A. K., Jadav, A., Phillips, V. T. J., Bansemer, A., Carlin, J., and A. Ryzhkov, 2022: The influence of multiple groups of biological ice-nucleating particles on microphysical properties of mixed-phase clouds observed during MC3E. Atmos. Chem. Phys., 22, 12055–12075

Phillips, V. T. J., 2022: Theory of in-cloud activation of aerosols and microphysical quasi-equilibrium in a deep updraft. J. Atmos. Sci., 79, 1865–1886

Phillips, V. T. J., 2023: A theory for the balance between warm rain and ice crystal processes of precipitation in mixed-phase clouds. J. Atmos. Sci., in press

Waman, D., A. Deshmukh, A. Jadav, S. Patade, M. Gautam, V. T. J. Phillips, A. Bansemer, and J. Jakobsson, 2023: Effects from time dependence of ice nucleus activity for contrasting cloud types. J. Atmos. Sci., 80, 20132039

Zhao, X., X. Liu, G. McFarquhar, S. Patade, V. T. J. Phillips, Y. Shi, and M. Zhang, 2023: Important ice processes are missed by climate models in Southern Ocean mixed-phase clouds: bridging SOCRATES observations to model developments. J. Geophys. Res. Atmos., 128(4), e2022JD037513