DASH

DAta Science-informed attribution of changes in the Hydrological cycle

Started
December 1, 2017
Status
In Progress
Share this post

Abstract

The main goal of the DASH (“DAta Science-informed attribution of changes in the Hydrologic cycle“) project is to improve the detection and attribution of climate change by incorporating data science and statistical learning methods. Detection and attribution (D&A) is a key methodology in climate science whose aim is to study the causal links between external drivers of the climate system and observed changes in climate variables (IPCC 2021). Improving our understanding of the factors that contribute to climate changes, such as anthropogenic emissions or natural variability, is of great relevance to society in order to anticipate and adapt to such changes, both on a global scale and regionally, for example in Switzerland. According to the IPCC standard terminology detection is the process of determining whether there is an observed change in a climate variable that cannot be explained by internal variability alone but without providing a reason for that change, while attribution is the process of providing a likely reason for the detected change. While it is well established that global-scale climate warming over the last decades and century is driven by human emissions of greenhouse gases and aerosols (e.g., see the most recent IPCC report: Sixth Assessment Report), important gaps of our understanding remain when it comes to climate changes on a regional scale (e.g., can warming or changes in the water cycle over Switzerland be attributed with high confidence to anthropogenic climate change?), or on short time scales (do extreme events, or daily patterns of climate variables, carry a signal of climate change?).

People

Scientists

SDSC Team:
PI | Partners

Seminar for Statistics:

  • Prof. Nicolai Meinshausen

More info

National Center for Atmospheric Research:

  • Dr. Angeline Pendergrass

More info

description

Goal:

Advance the detection and attribution (D&A) of climate change by incorporating advanced machine learning, statistical learning and data science methods.

Impact:

Advance the state-of-the-art on D&A techniques using data science, to finally guide climate model development based on the agreement or mismatch of models and observations, and improve the confidence in the detection and attribution statements.

Proposed approach:

Incorporate new statistical techniques for D&A using machine learning and causal inference to achieve robust predictions. We developed a novel detection and attribution methodology based on supervised statistical learning that uses explicitly the information contained in a target of interest to identify “fingerprints” of climate change (see Figure attached for a temperature fingerprint) that are representative of the target, such as anthropogenic or volcanic fingerprints. The methodology relies on a hypothesis testing framework and uses a statistical learning techniques inspired from causal inference that incorporates robustness constraints. These constraints help to increase the confidence in the detection and attribution statements. This methodology was used for both attribution of greenhouse gases and aerosols, and for detection to make it robust even under potentially large internal climate variability, a remaining long standing issue in climate science. In parallel work, we showed that a climate change signal can be detected from daily patterns of temperature and humidity at the global scale for every day since 2012. On a different line of work, we refined a traditional climate science method (“dynamical adjustment”) using data science, and we show that this allows to uncover the forced climate response as a residual from circulation-induced variability.

Gallery

Bibliography

  1. Hasselmann, K. F. On the signal-to-noise problem in atmospheric response studies. In Joint Conference of Royal Meteorological Society, American Meteorological Society, Deutsche Meteorologische Gesellschaft and the Royal Society (ed. Shaw, D. B.) 251–259 (Royal Meteorological Society, 1979)
  2. C. Deser, R. Knutti, S. Solomon, and A. S. Phillips. Communication of the role of natural variability in future North American climate. Nature Climate Change, 2, 775-779, DOI: https://doi.org/10.1038/nclimate1562 (2012).
  3. D. Rothenhäusler, N. Meinshausen, P. Bühlmann, J. Peters. Anchor regression: Heterogeneous data meet causality. Journal of the Royal Statistical Society: Series B (Statistical Methodology), Vol. 83 (2), pages 215-246, https://rss.onlinelibrary.wiley.com/doi/abs/10.1111/rssb.12398 (2021)

Publications

Related Pages

More projects

ML4FCC

In Progress
Machine Learning for the Future Circular Collider Design
Big Science Data

CLIMIS4AVAL

In Progress
Real-time cleansing of snow and weather data for operational avalanche forecasting
Energy, Climate & Environment

SEMIRAMIS

Completed
AI-augmented architectural design
Energy, Climate & Environment

4D-Brains

In Progress
Extracting activity from large 4D whole-brain image datasets
Biomedical Data Science

News

Latest news

Climate-smart agriculture in sub-Saharan Africa: optimizing nitrogen fertilization with data science
November 6, 2023

Climate-smart agriculture in sub-Saharan Africa: optimizing nitrogen fertilization with data science

Climate-smart agriculture in sub-Saharan Africa: optimizing nitrogen fertilization with data science

Food insecurity in sub-Saharan Africa is widespread, with crop yields much lower than in many developed regions. The project aims to use laser spectroscopy to measure fluxes and isotopic composition of N2O from maize and potato crops subjected to a range of fertilization levels.
Street2Vec | Self-supervised learning unveils change in urban housing from street-level images
October 31, 2023

Street2Vec | Self-supervised learning unveils change in urban housing from street-level images

Street2Vec | Self-supervised learning unveils change in urban housing from street-level images

It is difficult to effectively monitor and track progress in urban housing. We attempt to overcome these limitations by utilizing self-supervised learning with over 15 million street-level images taken between 2008 and 2021 to measure change in London.
DLBIRHOUI | Deep Learning Based Image Reconstruction for Hybrid Optoacoustic and Ultrasound Imaging
February 28, 2023

DLBIRHOUI | Deep Learning Based Image Reconstruction for Hybrid Optoacoustic and Ultrasound Imaging

DLBIRHOUI | Deep Learning Based Image Reconstruction for Hybrid Optoacoustic and Ultrasound Imaging

Optoacoustic imaging is a new, real-time feedback and non-invasive imaging tool with increasing application in clinical and pre-clinical settings. The DLBIRHOUI project tackles some of the major challenges in optoacoustic imaging to facilitate faster adoption of this technology for clinical use.

Contact us

Let’s talk Data Science

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