An official website of the United States government
Abstract Recent advances in explainable artificial intelligence (XAI) methods show promise for understanding predictions made by machine learning (ML) models. XAI explains how the input features are relevant or important for the model predictions. We train linear regression (LR) and convolutional neural network (CNN) models to make 1-day predictions of sea ice velocity in the Arctic from inputs of present-day wind velocity and previous-day ice velocity and concentration. We apply XAI methods to the CNN and compare explanations to variance explained by LR. We confirm the feasibility of using a novel XAI method [i.e., global layerwise relevance propagation (LRP)] to understand ML model predictions of sea ice motion by comparing it to established techniques. We investigate a suite of linear, perturbation-based, and propagation-based XAI methods in both local and global forms. Outputs from different explainability methods are generally consistent in showing that wind speed is the input feature with the highest contribution to ML predictions of ice motion, and we discuss inconsistencies in the spatial variability of the explanations. Additionally, we show that the CNN relies on both linear and nonlinear relationships between the inputs and uses nonlocal information to make predictions. LRP shows that wind speed over land is highly relevant for predicting ice motion offshore. This provides a framework to show how knowledge of environmental variables (i.e., wind) on land could be useful for predicting other properties (i.e., sea ice velocity) elsewhere. Significance StatementExplainable artificial intelligence (XAI) is useful for understanding predictions made by machine learning models. Our research establishes trustability in a novel implementation of an explainable AI method known as layerwise relevance propagation for Earth science applications. To do this, we provide a comparative evaluation of a suite of explainable AI methods applied to machine learning models that make 1-day predictions of Arctic sea ice velocity. We use explainable AI outputs to understand how the input features are used by the machine learning to predict ice motion. Additionally, we show that a convolutional neural network uses nonlinear and nonlocal information in making its predictions. We take advantage of the nonlocality to investigate the extent to which knowledge of wind on land is useful for predicting sea ice velocity elsewhere.
more »
« less
Advancing Arctic Sea Ice Remote Sensing with AI and Deep Learning: Opportunities and Challenges
Revolutionary advances in artificial intelligence (AI) in the past decade have brought transformative innovation across science and engineering disciplines. In the field of Arctic science, we have witnessed an increasing trend in the adoption of AI, especially deep learning, to support the analysis of Arctic big data and facilitate new discoveries. In this paper, we provide a comprehensive review of the applications of deep learning in sea ice remote sensing domains, focusing on problems such as sea ice lead detection, thickness estimation, sea ice concentration and extent forecasting, motion detection, and sea ice type classification. In addition to discussing these applications, we also summarize technological advances that provide customized deep learning solutions, including new loss functions and learning strategies to better understand sea ice dynamics. To promote the growth of this exciting interdisciplinary field, we further explore several research areas where the Arctic sea ice community can benefit from cutting-edge AI technology. These areas include improving multimodal deep learning capabilities, enhancing model accuracy in measuring prediction uncertainty, better leveraging AI foundation models, and deepening integration with physics-based models. We hope that this paper can serve as a cornerstone in the progress of Arctic sea ice research using AI and inspire further advances in this field.
more »
« less
- Award ID(s):
- 2230034
- PAR ID:
- 10579086
- Publisher / Repository:
- MDPI
- Date Published:
- Journal Name:
- Remote Sensing
- Volume:
- 16
- Issue:
- 20
- ISSN:
- 2072-4292
- Page Range / eLocation ID:
- 3764
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Scientific modeling is a vital educational practice that helps students apply scientific knowledge to real-world phenomena. Despite advances in AI, challenges in accurately assessing such models persist, primarily due to the complexity of cognitive constructs and data imbalances in educational settings. This study addresses these challenges by employing diverse analytic strategies, including the Synthetic Minority Over-sampling Technique (SMOTE), aimed at enhancing fairness and efficacy in automated scoring systems. We analyze the impact of these strategies through a robust methodology, utilizing a combination of tenfold cross-validation and independent testing phases to ensure the reliability of AI assessments. Our findings highlight the effectiveness of deep learning AI in mirroring human judgment, with improvements in accuracy, precision, recall, and F1 scores across varied model assessments. Specifically, the application of SMOTE significantly improved the scoring fairness for minority class instances, which are often underrepresented in educational datasets. This study also delves into the discrepancies between AI and human evaluations, particularly in interpreting creatively expressed student models, which reveals the areas where AI technologies require further enhancements to better align with human evaluative standards. This study lays a foundation for future research to explore advanced AI techniques and training strategies, thus promoting fair and supportive feedback mechanisms that enhance student learning and creativity. By advancing AI applications in science education, this research addresses essential challenges in the automated analysis of complex student responses and supports broader academic goals.more » « less
-
Important natural resources in the Arctic rely heavily on sea ice, making it important to forecast Arctic sea ice changes. Arctic sea ice forecasting often involves two connected tasks: sea ice concentration at each pixel and overall sea ice extent. Instead of having two separate models for two forecasting tasks, in this report, we study how to use multi-task learning techniques and leverage the connections between ice concentration and ice extent to improve accuracy for both prediction tasks. Because of the spatiotemporal nature of the data, we designed two novel multi-task learning models based on CNNs and ConvLSTMs, respectively. We also developed a custom loss function which trains the models to ignore land pixels when making predictions. Our experiments show our models can have better accuracies than separate models that predict sea ice extent and concentration separately, and that our accuracies are better than or comparable with results in the state-of-the-art studies.more » « less
-
Abstract Anthropogenic warming has led to an unprecedented year-round reduction in Arctic sea ice extent. This has far-reaching consequences for indigenous and local communities, polar ecosystems, and global climate, motivating the need for accurate seasonal sea ice forecasts. While physics-based dynamical models can successfully forecast sea ice concentration several weeks ahead, they struggle to outperform simple statistical benchmarks at longer lead times. We present a probabilistic, deep learning sea ice forecasting system, IceNet. The system has been trained on climate simulations and observational data to forecast the next 6 months of monthly-averaged sea ice concentration maps. We show that IceNet advances the range of accurate sea ice forecasts, outperforming a state-of-the-art dynamical model in seasonal forecasts of summer sea ice, particularly for extreme sea ice events. This step-change in sea ice forecasting ability brings us closer to conservation tools that mitigate risks associated with rapid sea ice loss.more » « less
-
Abstract Atmospheric reanalyses are widely used to estimate the past atmospheric near-surface state over sea ice. They provide boundary conditions for sea ice and ocean numerical simulations and relevant information for studying polar variability and anthropogenic climate change. Previous research revealed the existence of large near-surface temperature biases (mostly warm) over the Arctic sea ice in the current generation of atmospheric reanalyses, which is linked to a poor representation of the snow over the sea ice and the stably stratified boundary layer in the forecast models used to produce the reanalyses. These errors can compromise the employment of reanalysis products in support of polar research. Here, we train a fully connected neural network that learns from remote sensing infrared temperature observations to correct the existing generation of uncoupled atmospheric reanalyses (ERA5, JRA-55) based on a set of sea ice and atmospheric predictors, which are themselves reanalysis products. The advantages of the proposed correction scheme over previous calibration attempts are the consideration of the synoptic weather and cloud state, compatibility of the predictors with the mechanism responsible for the bias, and a self-emerging seasonality and multidecadal trend consistent with the declining sea ice state in the Arctic. The correction leads on average to a 27% temperature bias reduction for ERA5 and 7% for JRA-55 if compared to independent in situ observations from the MOSAiC campaign (respectively, 32% and 10% under clear-sky conditions). These improvements can be beneficial for forced sea ice and ocean simulations, which rely on reanalyses surface fields as boundary conditions. Significance Statement This study illustrates a novel method based on machine learning for reducing the systematic surface temperature errors that characterize multiple atmospheric reanalyses in sea ice–covered regions of the Arctic under clear-sky conditions. The correction applied to the temperature field is consistent with the local weather and the sea ice and snow conditions, meaning that it responds to seasonal changes in sea ice cover as well as to its long-term decline due to global warming. The corrected reanalysis temperature can be employed to support polar research activities, and in particular to better simulate the evolution of the interacting sea ice and ocean system within numerical models.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript
- Journal Article:
- https://doi.org/10.3390/rs16203764
