[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-120219-en":3,"doc-seo-120219-105":29,"detail-sidebar-cat-0-en-105":89},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":20,"language":21,"language_code":22,"site_id":23,"html_lang":22,"table_of_contents":24,"faqs":25,"seo_title":26,"seo_description":14,"update_tm":27,"read_time":28},120219,7971461740886,"Theodore","https://ap-avatar.wpscdn.com/davatar_3d24733baf745e90a7e4bdd5f77d97b2",8,"Research & Report","Machine Learning in Analyzing Atmospheric Gravity Waves","A LightGBM-based machine learning workflow improves the identification of clear-sky image windows from the ANtarctic Gravity Wave Instrument Network (ANGWIN) all-sky imager (ASI) database. The Halley model filters “clean” windows while reducing obstructions such as clouds, aurora, bright moonlight, and fog that otherwise distort power spectrum graphs. Iterative retraining with misclassified samples and expanded training data yields higher selectivity, lower analysis cost, and faster turnaround. Validation shows strong accuracy, with specific misreporting issues addressed across Halley Model 1.0–3.0.","Machine Learning in Analyzing Atmospheric Gravity Waves  \nAnastasia Brown  \nUtah State University Physics Department and Center for Atmospheric and Space Sciences Kenneth Zia  \nUtah State University Physics Department and Center for Atmospheric and Space Sciences P.-Dominique Pautet  \nUtah State University Physics Department and Center for Atmospheric and Space Sciences Yucheng Zhao  \nUtah State University Physics Department and Center for Atmospheric and Space Sciences Max Haehnel  \nUtah State University Physics Department and Center for Atmospheric and Space Sciences Michel J. Taylor  \nUtah State University Physics Department and Center for Atmospheric and Space Sciences  \nIntroduction  \nObjective: Improve Halley LightGBM Model to better identify images of clear skies.  \n• The ANtarctic Gravity Wave Instrument Network (ANGWIN) is an international collaboration aimed at investigating the upper atmosphere dynamics over a continent-size region, using a network of all-sky imagers (Fig. 1) .  \n• ANGWIN network began collecting All-Sky-Imager (ASI) data in 2012 (Table. 1) . The ASI data is then shifted through to find windows of “clean”data (no clouds, aurora, or moonlight, Fig. 2) .  \n• Once found, the “clean” image windows are processed, and power spectrum graphs are made to identify wave activity. In the middle of winter, each station collecting ASI data can easily produce well over 8,000 images a night.  \n• To streamline the identification of “clean” windows in the extensive database of all sky-imager data obtained since 2012, we have developed a machine learning algorithm that sorts “clean”(marked as 0) images from“obscured”(marked as 1) images (Zia 2022) .  \n• Inspired by the use of machine learning to quickly sort through large Themis Aurora data sets with a reported 96% accuracy (Clausen et al., 2018) .  \n• Already, a Light Gradient Boosted Machine (LightGBM) model is used to sort through two station’s data (Fig. 3 and 4) .  \n• The model for the Halley reports a 99.2% accuracy, but when validating the data sorted from the Halley station, it was discovered that 62.3% of the “clean” windows identified were misreported (Fig. 5) .  \nMethods  \nChecked “clean” widows identified by the Halley model. The model predominantly misidentified images with moon glare and/or fog as “clean”. Removed ambiguous and misleading examples from training set.  \nBuffed its weak areas by adding themisidentified “obscured” windows into the training set. Increased the number of frames in the training set.  \n• Halley Model 2.0 was tested ageists August 2012: 98.7% reported accuracy.  \n• The process was repeated again to create Halley Model 3.0, and that model was tested against July 2012: 98.8% reported accuracy (Fig. 6) .  \nResults and Conclusion  \nBoth models yielded power spectrum graphs more consistent with “clean” windows, and both models were more selective in the “clean” windows they identified.  \nThe use of machine learning in ASI cleaning removes the bottle neck created by the large data set. The second iteration of the LightGBM Halley model proved to an improvement over the first and third model. A more accurate model results a lower cost of analysis and a faster turnout.  \nHalley Model 1.0: of the days tested only 37.7% of “clean”windows identified were labeled correctly. 99.2% accuracy. Halley Model 2.0: of the days tested 76.9% of “clean”windows identified were correctly labeled. 98.7% accuracy.  \nHalley Model 3.0 found 23 “clean” windows in the month of July 2012, but the “clean” windows were heavily obscured. The model appeared to be overfitted.  \nReferences:  \nClausen, L. B. N., & Nickisch, H. (2018) . Automatic classification of auroral images from the Oslo Auroral THEMIS (OATH) data set using machine learning. Journal of Geophysical Research: Space  \nPhysics, 123, 5640– 5647 . [https://doi.org/10.1029/2018JA025274](https://doi.org/10.1029/2018JA025274)  \nFig. 3: Light Gradient-Boosting Machine (LightGBM) models quickly sort “clean” an","cbCaifCman6Fk1Ze","https://ap.wps.com/l/cbCaifCman6Fk1Ze","pdf",1280329,1,"English","en",105,"# Introduction\n## Objective\n# Methods\n# Results and Conclusion\n# References and Figures","[{\"question\":\"What is the main objective of this work on ANGWIN all-sky imager data?\",\"answer\":\"Improve the Halley LightGBM model to better identify clear-sky (“clean”) image windows from ASI observations, enabling more reliable wave-activity power spectrum graphs.\"},{\"question\":\"How does the method handle misclassified “clean” windows?\",\"answer\":\"It checks Halley model outputs, removes ambiguous misleading examples from the training set, then retrains by adding misidentified “obscured” windows into the training set while increasing the number of training frames.\"},{\"question\":\"What performance change is reported across Halley model iterations?\",\"answer\":\"Halley Model 1.0 and later versions show high reported accuracies (e.g., 99.2% and 98.7%), and the iterative process makes the models more selective in the “clean” windows they identify, while Model 3.0 shows signs of overfitting during July 2012 evaluation.\"}]","Machine Learning in Analyzing Atmospheric Gravity Waves | PDF",1785728788,3,{"code":4,"msg":30,"data":31},"ok",{"site_id":23,"language":22,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":84,"head_meta":86,"extra_data":88,"updated_unix":27},"machine-learning-in-analyzing-atmospheric-gravity-waves","",{"@graph":35,"@context":83},[36,52,66],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,49],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":28},"https://docshare.wps.com/document/research-report/",{"item":50,"name":13,"@type":42,"position":51},"https://docshare.wps.com/document/machine-learning-in-analyzing-atmospheric-gravity-waves/120219/",4,{"url":50,"name":13,"@type":53,"author":54,"headline":13,"publisher":56,"fileFormat":59,"inLanguage":22,"description":14,"dateModified":60,"datePublished":60,"encodingFormat":59,"isAccessibleForFree":61,"interactionStatistic":62},"DigitalDocument",{"name":9,"@type":55},"Person",{"url":40,"name":57,"@type":58},"DocShare","Organization","application/pdf","2026-08-03",true,{"@type":63,"interactionType":64,"userInteractionCount":20},"InteractionCounter",{"@type":65},"ViewAction",{"@type":67,"mainEntity":68},"FAQPage",[69,75,79],{"name":70,"@type":71,"acceptedAnswer":72},"What is the main objective of this work on ANGWIN all-sky imager data?","Question",{"text":73,"@type":74},"Improve the Halley LightGBM model to better identify clear-sky (“clean”) image windows from ASI observations, enabling more reliable wave-activity power spectrum graphs.","Answer",{"name":76,"@type":71,"acceptedAnswer":77},"How does the method handle misclassified “clean” windows?",{"text":78,"@type":74},"It checks Halley model outputs, removes ambiguous misleading examples from the training set, then retrains by adding misidentified “obscured” windows into the training set while increasing the number of training frames.",{"name":80,"@type":71,"acceptedAnswer":81},"What performance change is reported across Halley model iterations?",{"text":82,"@type":74},"Halley Model 1.0 and later versions show high reported accuracies (e.g., 99.2% and 98.7%), and the iterative process makes the models more selective in the “clean” windows they identify, while Model 3.0 shows signs of overfitting during July 2012 evaluation.","https://schema.org",{"og:url":50,"og:type":85,"og:title":13,"og:site_name":57,"og:description":14},"article",{"robots":87,"canonical":50},"index,follow",{"doc_id":7,"site_id":23},{"code":4,"msg":5,"data":90},[91,95,99,103,108,113,118,121,126,129,133],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":92,"show_sort_weight":93,"slug":94},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":96,"show_sort_weight":97,"slug":98},"Literature",80,"literature",{"id":51,"doc_module":4,"doc_module_name":45,"category_name":100,"show_sort_weight":101,"slug":102},"Exam",70,"exam",{"id":104,"doc_module":4,"doc_module_name":45,"category_name":105,"show_sort_weight":106,"slug":107},5,"Comic",60,"comic",{"id":109,"doc_module":4,"doc_module_name":45,"category_name":110,"show_sort_weight":111,"slug":112},6,"Technology",50,"technology",{"id":114,"doc_module":4,"doc_module_name":45,"category_name":115,"show_sort_weight":116,"slug":117},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":119,"slug":120},30,"research-report",{"id":122,"doc_module":4,"doc_module_name":45,"category_name":123,"show_sort_weight":124,"slug":125},9,"Religion & Spirituality",20,"religion-spirituality",{"id":124,"doc_module":4,"doc_module_name":45,"category_name":127,"show_sort_weight":124,"slug":128},"World Cup","world-cup",{"id":130,"doc_module":4,"doc_module_name":45,"category_name":131,"show_sort_weight":130,"slug":132},10,"Lifestyle","lifestyle",{"id":134,"doc_module":4,"doc_module_name":45,"category_name":135,"show_sort_weight":104,"slug":136},19,"General","general"]