ArcticNAP

Arctic climate change: Exploring the Natural Aerosol baseline for improved model Predictions

Started
November 1, 2021
Status
Completed
Share this project

Abstract

The Arctic is warming two to three times faster than the global average. Arctic amplification has repercussions for global climate, northern hemispheric weather and local livelihoods. Current models have difficulties to simulate Arctic change, making future scenarios uncertain. To improve model skill the factors contributing to Arctic amplification need to be better represented. We focus on the role of aerosols. Aerosols can interact with solar radiation directly, as well as indirectly through their incluence on cloud radiative properties. Considerable effort has gone into describing the role of anthropogenic aerosols in Arctic climate change. However, there has been much less focus on natural aerosols, which represents a large part of overall model uncertainty. This is partly because they are emitted and processed via complex pathways involving several environmental compartments, which are difficult to represent in models. To complicate things further, the natural state of the Arctic is changing rapidly thereby changing aerosol processes.

We propose to use an unprecedented combination of in situ measurements, satellite-based observations and numerical weather prediction data to reveal the processes controlling natural Arctic aerosols and driving climate-relevant aerosol properties. Aerosol data are available from eight Arctic observatories covering up to 30 years, and from two high Arctic drift expeditions. By developing a latent variable model, which accommodates the different data types, time series with gaps, and domain knowledge for constraints, we explore relationships between environmental variables, such as sea ice extent, chlorophyll-a concentrations, or meteorological conditions, and aerosol concentrations. After quantifying these relationships, we compare the output from five Earth System models (ESM) against the data driven model results, to identify where ESMs can be improved. This project is an important step in this research field, as the response of the Arctic environment to future climate change (including atmospheric and oceanic conditions) remains an open research question and provides an avenue for continuing research.

People

Collaborators

SDSC Team:
Eliza Harris
Michele Volpi
William Aeberhard

PI | Partners:

EPFL, Extreme Environments Research Laboratory:

  • Dr. Jakob Boyd Pernov
  • Prof. Julia Schmale

More info

description

Motivation

The natural production of aerosols in the Arctic plays a likely important role in climate change, although this role is not well understood, particularly in a changing climate. The main goal of the project is to construct models which can identify key environmental drivers of the key aerosol species methanesulfonic acid (MSA), and potentially describe processes that may evolve with climate change.

Proposed Approach / Solution

We combine many data sources in an unprecedented way, including in situ measurements of MSA (see Figure 1), remote sensing, and outputs from numerical weather prediction models. By pre-processing climatic and environmental features as residence-time weighted averages based on a particle dispersion model (FLEXPART) we can model and predict MSA concentrations in space-time. We propose two models: a random forest model and an additive model tailored for the task at hand. Both models have similar predictive performance in terms of MSA concentration forecasts, but they yield different inference as far as feature selection is concerned. In parallel to this work, using gradient-boosted trees with selected key features (Figure 2) we have identified the main source regions for MSA and project a seasonal shift over the next 50 years (Figure 3).

Impact

Our results, including the climatic and environmental drivers we identified in a fully data-driven way, can in principle directly inform global climate models in their capacity to represent aerosols in the Arctic. This can improve climate predictions at a global scale and thus have a substantial impact in certain medium- and long-term projections.

Figure 1: Locations and MSA seasonal cycle for each station. (a) Map of the four Arctic stations used in this study. Stations are indicated with a red star. Map background is from Natural Earth. (b) Seasonal cycle of MSA at Alert (red), Gruvebadet (blue), Thule (cyan), Utqiaġvik/Barrow (magenta), and all stations combined (Pan-Arctic) (black). The thick lines represent the median and the shading represents the interquartile range for each month over the period 2010-2017 for Alert, Gruvebadet, and Thule. For Utqiaġvik/Barrow, the period is 2008-2014. Figure from Boyd Pernov et al. (2024a).
Figure 2: Overall importance of each variable for the prediction of mean MSA concentrations in different regions of the Arctic. The bars represent the median of the absolute SHAP value while the black lines represent the interquartile range. See Boyd Pernov et al. (2024a) for details.
Figure 3: Projected changes in the seasonal cycle of Pan-Arctic MSA. Changes in the distribution of projected Pan-Arctic MSA values for each decadal interval is displayed in (a). The middle line represents the median while the upper and lower limits of the box represent the 25th/75th percentiles. The total relative change in MSA expressed as a percent for each decadal interval compared to year 0 is displayed in (b). The vertical lines for each bar represent the upper/lower confidence intervals of the Theil-Sen slope used for extrapolating the trends in ERA5. See Boyd Pernov et al. (2024a) for details.

Gallery

Annexe

Publications

  • Pernov, J. B., Harris, E., Volpi, M., Baumgartner, T., Hohermuth, B., Henne, S., Aeberhard, W. H., Becagli, S., Quinn, P. K., Traversi, R., Upchurch, L. M., and Schmale, J. (2024a). Pan-Arctic Methanesulfonic Acid Aerosol: Source regions, atmospheric drivers, and future projections. Accepted in Climate and Atmospheric Science. Pan-Arctic Methanesulfonic Acid Aerosol: Source regions, atmospheric drivers, and future projections
  • Pernov, J. B., Harris, E., Volpi, M., Baumgartner, T., Hohermuth, B., Henne, S., Aeberhard, W. H., Becagli, S., Quinn, P. K., Traversi, R., Upchurch, L., and Schmale, J. (2024b). Pan-Arctic Methanesulfonic Acid Aerosol: Source regions, atmospheric drivers, and future projections. The European Aerosol Conference 2024.

Additional resources

Bibliography

  1. Pörtner, H.-O. et al. (2019). IPCC Special Report on the Ocean and Cryosphere in a Changing Climate. Report Home
  2. Schmale, J., Zieger, P., and Ekman, A. M. L. (2021). Aerosols in current and future Arctic climate. Nature Climate Change 11, 95-105. https://doi.org/10.1038/s41558-020-00969-5

Publications

Related Pages

More projects

ML-L3DNDT

Completed
Robust and scalable Machine Learning algorithms for Laue 3-Dimensional Neutron Diffraction Tomography
Big Science Data

BioDetect

Completed
Deep Learning for Biodiversity Detection and Classification
Energy, Climate & Environment

IRMA

In Progress
Interpretable and Robust Machine Learning for Mobility Analysis
No items found.

FLBI

In Progress
Feature Learning for Bayesian Inference
No items found.

News

Latest news

Smartair | An active learning algorithm for real-time acquisition and regression of flow field data
May 1, 2024

Smartair | An active learning algorithm for real-time acquisition and regression of flow field data

Smartair | An active learning algorithm for real-time acquisition and regression of flow field data

We’ve developed a smart solution for wind tunnel testing that learns as it works, providing accurate results faster. It provides an accurate mean flow field and turbulence field reconstruction while shortening the sampling time.
The Promise of AI in Pharmaceutical Manufacturing
April 22, 2024

The Promise of AI in Pharmaceutical Manufacturing

The Promise of AI in Pharmaceutical Manufacturing

Innovation in pharmaceutical manufacturing raises key questions: How will AI change our operations? What does this mean for the skills of our workforce? How will it reshape our collaborative efforts? And crucially, how can we fully leverage these changes?
Efficient and scalable graph generation through iterative local expansion
March 20, 2024

Efficient and scalable graph generation through iterative local expansion

Efficient and scalable graph generation through iterative local expansion

Have you ever considered the complexity of generating large-scale, intricate graphs akin to those that represent the vast relational structures of our world? Our research introduces a pioneering approach to graph generation that tackles the scalability and complexity of creating such expansive, real-world graphs.

Contact us

Let’s talk Data Science

Do you need our services or expertise?
Contact us for your next Data Science project!