JRJamin Kurtis Rader

Jamin Kurtis Rader

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Jamin Kurtis Rader has a total of 15 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Attribution of the record-high 2023 SST using a deep-learning framework

    by Jamin K Rader, Charlotte J Connolly, M A Fernandez, Emily M Gordon

    The global-mean sea surface temperature (SST) reached a record high in 2023, exceeding the 2016 record by 0.14 °C. This unprecedented change in global-mean SST has major implications for our understanding of internal variability and the forced response in our changing climate. In this work, we use neural networks trained on simulated climate data to separate the contributions of internal variability and the forced response within observations. Performing attribution reveals that internal variability was responsible for +0.07 °C of the 2023 global mean SST, due to anomalously warm conditions in the Pacific, Atlantic, and Indian Ocean basins. Furthermore, these results provide a line of evidence for accelerated forced warming in recent years. Continued monitoring of the climate will be critical for understanding the drivers behind this unprecedented SST record.

  • Open Access

    Internal climate variability modulates decadal changes in ocean anthropogenic carbon storage

    by Holly C Olivarez, Nicole S Lovenduski, Elizabeth Maroon, Jens D Müller, Amanda R Fay, Kristen M Krumhardt, Michael N Levy, Keith Lindsay, Galen A McKinley, Jamin K Rader

    The ocean removes man-made (anthropogenic) carbon from the atmosphere and thereby mitigates climate change. Observations from global hydrographic surveys reveal the spatial and temporal evolution of the ocean inventory of anthropogenic carbon and suggest substantial decadal variability in historical storage rates. Here, we use a 100-member ensemble of an Earth system model to investigate the influence of external forcing and internal climate variability on historical changes in ocean anthropogenic carbon storage over 1994 to 2014. Our findings reveal that the externally forced, decadal changes in storage are largest in the Atlantic (2–4 mmol m−3 decade−1) and positive nearly everywhere. Internal climate variability modulates regional ocean anthropogenic carbon storage trends by up to 10 mmol m−3 decade−1. The influence of internal climate variability on decadal storage changes is most prominent at depths of ∼300 m and at the edges of the subtropical gyres. Internal variability in anthropogenic carbon in the extratropics has high spectral power on decadal to multi-decadal timescales, indicating that the approximately decadal repetitions of hydrographic surveys may produce storage change estimates that are heavily influenced by internal climate variability.

  • Open Access

    Optimizing Seasonal‐To‐Decadal Analog Forecasts With a Learned Spatially‐Weighted Mask

    by Jamin K. Rader, Elizabeth A. Barnes

    Seasonal‐to‐decadal climate prediction is crucial for decision‐making in a number of industries, but forecasts on these timescales have limited skill. Here, we develop a data‐driven method for selecting optimal analogs for seasonal‐to‐decadal analog forecasting. Using an interpretable neural network, we learn a spatially‐weighted mask that quantifies how important each grid point is for determining whether two climate states will evolve similarly. We show that analogs selected using this weighted mask provide more skillful forecasts than analogs that are selected using traditional spatially‐uniform methods. This method is tested on two prediction problems using the Max Planck Institute for Meteorology Grand Ensemble: multi‐year prediction of North Atlantic sea surface temperatures, and seasonal prediction of El Niño Southern Oscillation. This work demonstrates a methodical approach to selecting analogs that may be useful for improving seasonal‐to‐decadal forecasts and understanding their sources of skill.


  • Open Access

    Optimizing Seasonal-to-Decadal Analog Forecasts with a Learned Spatially-Weighted Mask

    by Jamin Kurtis Rader, Elizabeth A. Barnes

    Seasonal-to-decadal climate prediction is crucial for decision-making in a number of industries, but forecasts on these timescales have limited skill. Here, we develop a data-driven method for selecting optimal analogs for seasonal-to-decadal analog forecasting. Using an interpretable neural network, we learn a spatially-weighted mask that quantifies how important each grid point is for determining whether two climate states will evolve similarly. We show that analogs selected using this weighted mask provide more skillful forecasts than analogs that are selected using traditional spatially-uniform methods. This method is tested on two prediction problems within a perfect model framework using the Max Planck Institute for Meteorology Grand Ensemble: multi-year prediction of North Atlantic sea surface temperatures, and seasonal prediction of El Niño Southern Oscillation. This work demonstrates a methodical approach to selecting analogs that may be useful for improving seasonal-to-decadal forecasts and understanding their sources of skill.


  • Open Access

    Detection of Forced Change Within Combined Climate Fields Using Explainable Neural Networks

    by Jamin K. Rader, Elizabeth A. Barnes, Imme Ebert‐Uphoff, Chuck Anderson

    Assessing forced climate change requires the extraction of the forced signal from the background of climate noise. Traditionally, tools for extracting forced climate change signals have focused on one atmospheric variable at a time, however, using multiple variables can reduce noise and allow for easier detection of the forced response. Following previous work, we train artificial neural networks to predict the year of single‐ and multi‐variable maps from forced climate model simulations. To perform this task, the neural networks learn patterns that allow them to discriminate between maps from different years—that is, the neural networks learn the patterns of the forced signal amidst the shroud of internal variability and climate model disagreement. When presented with combined input fields (multiple seasons, variables, or both), the neural networks are able to detect the signal of forced change earlier than when given single fields alone by utilizing complex, nonlinear relationships between multiple variables and seasons. We use layer‐wise relevance propagation, a neural network explainability tool, to identify the multivariate patterns learned by the neural networks that serve as reliable indicators of the forced response. These “indicator patterns” vary in time and between climate models, providing a template for investigating inter‐model differences in the time evolution of the forced response. This work demonstrates how neural networks and their explainability tools can be harnessed to identify patterns of the forced signal within combined fields.


  • Open Access

    Detection of forced change within combined climate fields using explainable neural networks

    by Jamin Kurtis Rader, Elizabeth A. Barnes, Imme Ebert-Uphoff, Chuck Anderson

    Assessing forced climate change requires the extraction of the forced signal from the background of climate noise. Traditionally, tools for extracting forced climate change signals have focused on one atmospheric variable at a time, however, using multiple variables can reduce noise and allow for easier detection of the forced response. Following previous work, we train artificial neural networks to predict the year of single- and multi-variable maps from forced climate model simulations. To perform this task, the neural networks learn patterns that allow them to discriminate between maps from different years—that is, the neural networks learn the patterns of the forced signal amidst the shroud of internal variability and climate model disagreement. When presented with combined input fields (multiple seasons, variables, or both), the neural networks are able to detect the signal of forced change earlier than when given single fields alone by utilizing complex, nonlinear relationships between multiple variables and seasons. We use layer-wise relevance propagation, a neural network explainability tool, to identify the multivariate patterns learned by the neural networks that serve as reliable indicators of the forced response. These “indicator patterns” vary in time and between climate models, providing a template for investigating inter-model differences in the time evolution of the forced response. This work demonstrates how neural networks and their explainability tools can be harnessed to identify patterns of the forced signal within combined fields.

  • Open Access

    Detection of forced change within combined climate fields using explainable neural networks

    by Jamin Kurtis Rader, Elizabeth A. Barnes, Imme Ebert-Uphoff, Chuck Anderson

    Assessing forced climate change requires the extraction of the forced signal from the background of climate noise. Traditionally, tools for extracting forced climate change signals have focused on one atmospheric variable at a time, however, using multiple variables can reduce noise and allow for easier detection of the forced response. Following previous work, we train artificial neural networks to predict the year of single- and multi-variable maps from forced climate model simulations. To perform this task, the neural networks learn patterns that allow them to discriminate between maps from different years—that is, the neural networks learn the patterns of the forced signal amidst the shroud of internal variability and climate model disagreement. When presented with combined input fields (multiple seasons, variables, or both), the neural networks are able to detect the signal of forced change earlier than when given single fields alone by utilizing complex, nonlinear relationships between multiple variables and seasons. We use layer-wise relevance propagation, a neural network explainability tool, to identify the multivariate patterns learned by the neural networks that serve as reliable indicators of the forced response. These “indicator patterns” vary in time and between climate models, providing a template for investigating inter-model differences in the time evolution of the forced response. This work demonstrates how neural networks and their explainability tools can be harnessed to identify patterns of the forced signal within combined fields.