| Abstrakt: |
Surge‐type glaciers are present in many cold environments in the world. These glaciers experience a dramatic increase in velocity over short time periods, the surge, followed by an extended period of slow movement, the quiescence. This study aims at understanding why only few glaciers exhibit a transient behavior. We develop a machine learning framework to classify surge‐type glaciers, based on their location, exposure, geometry, climatic mass balance and runoff. We apply this approach to the Svalbard archipelago, a region with a relatively homogeneous climate. We compare the performance of logistic regression, random forest, and extreme gradient boosting (XGBoost) machine learning models that we apply to a newly combined database of glaciers in Svalbard. Based on the most accurate model, XGBoost, we compute surge probabilities along glacier centerlines and quantify the relative importance of several controlling features. Results show that the surface and bed slopes, ice thickness, glacier width, climatic mass balance, and runoff along glacier centerlines are the most significant features explaining surge probability for glaciers in Svalbard. A thicker and wider glacier with a low surface slope has a higher probability to be classified as surge‐type, which is in good agreement with the existing theories of surging. Finally, we build a probability map of surge‐type glaciers in Svalbard. The framework shows robustness on classifying surge‐type glaciers that were not previously classified as such in existing inventories but have been observed surging. Our methodology could be extended to classify surge‐type glaciers in other areas of the world. Plain Language Summary: 1% of the glaciers in the world exhibit intermittent phases of accelerated motion, called surge. These accelerations are not fully understood and we do not know why only few glaciers experience such behavior. Surging glaciers may lead to dramatic advances over rivers and damming up lakes that are then prone to a sudden and possibly catastrophic drainage. The Svalbard archipelago, located in the high Arctic, hosts more than one hundred surging glaciers. By analyzing statistically several data‐sets, we calculate the probability for every glacier to experience surge events. Our results show that specific combinations of surface and bed slopes, glacier width and ice thickness control glacier surge probability. To a smaller extent climatic parameters such as the mass a glacier may lose or gain during the year and the amount of melt water available also contribute to the surge probability. These findings are in good agreement with existing theories explaining surge dynamics. We produce the first probabilistic map of surging for all the glaciers in Svalbard. Our classification highlights some glaciers that were not classified as surging glaciers in glacier inventories but have been observed surging, confirming the robustness of our framework. Our method is applicable to other world regions. Key Points: We establish a machine learning framework to evaluate the probability of glacier surgeWe build a combined database of glaciers in Svalbard that contains thirteen featuresWe compute the first map of glacier surge probability in Svalbard and we quantify the relative importance of relevant features [ABSTRACT FROM AUTHOR] |
Plný text