Search for: All records

The Pleistocene Epoch was characterized by extensive glacier systems in numerous mountain ranges around the world. Mapping glacial landforms and deposits over many decades of prior work has afforded reconstructions of mountain glaciers, chiefly during the last Pleistocene glaciation and subsequent deglaciation. The availability of high-resolution satellite imagery, digital terrain models, and numerical chronologies of glacial deposits and landforms provides opportunities for mapping paleoglacier outlines and reconstructing ice thickness and volume during specific periods across glaciated regions at different spatial scales. Most paleoglacier reconstructions require outlines corresponding to a specific valley and terminus. However, various formats of digital paleoglacier outlines exist in the literature, some of which encompass entire glacier complexes or ice caps without differentiating between individual valleys and outlet glaciers. Also, unlike inventories of present-day glaciers such as the Randolph Glacier Inventory, digitized paleoglacier outlines lack standardized attributes. In this study, we developed an ArcGIS toolbox to subdivide paleoglacier outlines into individual polygons constrained within watershed boundaries (drainage basins) and to derive a consistent set of attributes related to the geometry, topography, and ice thickness of paleoglaciers. We demonstrate the applications of this toolbox in glaciated mountain areas in Costa Rica, the western U.S., and the central Tibetan Plateau. Although some manual adjustments are still necessary, this toolbox provides an efficient means to standardize the format and derive attributes for paleoglacier outlines. Our proposed framework and newly developed ArcGIS toolbox for standardizing paleoglacier outline formats and attributes improve the value, accuracy, and utility of paleoglacier mapping and paleoclimate reconstruction, and facilitate consistency and comparability among model simulations of glacier and climate changes from the past to present and into the future. 

Vision Language models (VLMs) have transformed Generative AI by enabling systems to interpret and respond to multi-modal data in real-time. While advancements in edge computing have made it possible to deploy smaller Large Language Models (LLMs) on smartphones and laptops, deploying competent VLMs on edge devices remains challenging due to their high computational demands. Furthermore, cloud-only deployments fail to utilize the evolving processing capabilities at the edge and limit responsiveness. This paper introduces a distributed architecture for VLMs that addresses these limitations by partitioning model components between edge devices and central servers. In this setup, vision components run on edge devices for immediate processing, while language generation of the VLM is handled by a centralized server, resulting in up to 33% improvement in throughput over traditional cloud-only solutions. Moreover, our approach enhances the computational efficiency of off-the-shelf VLM models without the need for model compression techniques. This work demonstrates the scalability and efficiency of a hybrid architecture for VLM deployment and contributes to the discussion on how distributed approaches can improve VLM performance. Index Terms—vision-language models (VLMs), edge computing, distributed computing, inference optimization, edge-cloud collaboration.