{"id":{"repo_id":"unr","oai_identifier":"oai:scholarwolf.unr.edu:11714/11574"},"canonical_url":"https://search.dev.ndltd.org/etd/unr/oai:scholarwolf.unr.edu:11714/11574","repository":{"repo_id":"unr","name":"University of Nevada - Reno","base_url":"https://scholarwolf.unr.edu/server/oai/request"},"display":{"title":"Three-dimensional wildland fuel mapping with remote sensing and machine learning","abstract":"The 21st century has seen a global increase in fire size and severity, including the western United States. Historically loaded forests, changing precipitation patterns, and lengthening fire seasons place a significant strain on land management and firefighting capacity. Fire behavior is a fine scaled process affected by climate, weather, topography, and fuel. The amount and arrangement of fuels are of particular interest because they can be directly controlled by managers to manage present-day fire risk and influence future response to fire. Existing fuel maps have a coarse resolution that limits their utility to understand fire behavior. In this dissertation, I address this research gap by using airborne and terrestrial LiDAR (ALS, TLS; respectively) and machine learning to make fine scaled predictions of fuels beneath the canopy. In my first chapter I use a gradient boosting regressor to boost the signal of ALS at the surface floor by using the ALS signal itself as features. I demonstrate that this method does a better job of detecting changes in surface fuels after two forest fires compared to ALS alone. In my second chapter I use a generative deep learning model to reduce fine-scale occlusion in ALS point clouds, leading to better measurement of vegetation beneath the canopy. In my third chapter, I compare this method with ALS to examine changes in fuel loading before and after a forest fire. I also examine how these changes influence management by comparing differences in common fuel metrics. I found that the use of machine learning and deep learning can significantly improve the capabilities of ALS beneath the forest canopy.","abstract_html":"The 21st century has seen a global increase in fire size and severity, including the western United States. Historically loaded forests, changing precipitation patterns, and lengthening fire seasons place a significant strain on land management and firefighting capacity. Fire behavior is a fine scaled process affected by climate, weather, topography, and fuel. The amount and arrangement of fuels are of particular interest because they can be directly controlled by managers to manage present-day fire risk and influence future response to fire. Existing fuel maps have a coarse resolution that limits their utility to understand fire behavior. In this dissertation, I address this research gap by using airborne and terrestrial LiDAR (ALS, TLS; respectively) and machine learning to make fine scaled predictions of fuels beneath the canopy. In my first chapter I use a gradient boosting regressor to boost the signal of ALS at the surface floor by using the ALS signal itself as features. I demonstrate that this method does a better job of detecting changes in surface fuels after two forest fires compared to ALS alone. In my second chapter I use a generative deep learning model to reduce fine-scale occlusion in ALS point clouds, leading to better measurement of vegetation beneath the canopy. In my third chapter, I compare this method with ALS to examine changes in fuel loading before and after a forest fire. I also examine how these changes influence management by comparing differences in common fuel metrics. I found that the use of machine learning and deep learning can significantly improve the capabilities of ALS beneath the forest canopy.","abstract_has_math":false,"creators":["Hartsook, Theodore Elliott"],"institution":null,"degree_name":null,"degree_level":"Doctorate Degree","degree_discipline":null,"degree_department":null,"school":null,"contributors":[],"advisors":["Greenberg, Jonathan"],"committee_chairs":[],"committee_members":["Bisbing, Sarah","Hanan, Erin","Nussear, Kenneth","Tavakkoli, Alireza"],"year":2025,"date_issued":"2025","date_published":"2025","updated_at":"2026-07-27T21:46:30Z","subjects":["deep learning","fire","forest","lidar","machine learning","remote sensing"],"languages":["en_US","English"],"rights":[],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"https://scholarwolf.unr.edu/handle/11714/11574","outbound_label":"Repository record","outbound_source":"dc:identifier.uri"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Greenberg, Jonathan"]},{"key":"dc:contributor.committeemember","label":"Committee Member","values":["Bisbing, Sarah","Hanan, Erin","Nussear, Kenneth","Tavakkoli, Alireza"]},{"key":"dc:creator","label":"Author","values":["Hartsook, Theodore Elliott"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2025-09-08T18:42:44Z"]},{"key":"dc:date.available","label":"Dc Date Available","values":["2025-09-08T18:42:44Z"]},{"key":"dc:date.issued","label":"Date","values":["2025"]},{"key":"dc:type","label":"Dc Type","values":["Dissertation"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Doctorate Degree"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["deep learning","fire","forest","lidar","machine learning","remote sensing"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["English"]},{"key":"dc:language.iso","label":"Language (ISO)","values":["en_US"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://scholarwolf.unr.edu/handle/11714/11574"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["The 21st century has seen a global increase in fire size and severity, including the western United States. Historically loaded forests, changing precipitation patterns, and lengthening fire seasons place a significant strain on land management and firefighting capacity. Fire behavior is a fine scaled process affected by climate, weather, topography, and fuel. The amount and arrangement of fuels are of particular interest because they can be directly controlled by managers to manage present-day fire risk and influence future response to fire. Existing fuel maps have a coarse resolution that limits their utility to understand fire behavior. In this dissertation, I address this research gap by using airborne and terrestrial LiDAR (ALS, TLS; respectively) and machine learning to make fine scaled predictions of fuels beneath the canopy. In my first chapter I use a gradient boosting regressor to boost the signal of ALS at the surface floor by using the ALS signal itself as features. I demonstrate that this method does a better job of detecting changes in surface fuels after two forest fires compared to ALS alone. In my second chapter I use a generative deep learning model to reduce fine-scale occlusion in ALS point clouds, leading to better measurement of vegetation beneath the canopy. In my third chapter, I compare this method with ALS to examine changes in fuel loading before and after a forest fire. I also examine how these changes influence management by comparing differences in common fuel metrics. I found that the use of machine learning and deep learning can significantly improve the capabilities of ALS beneath the forest canopy."]},{"key":"dc:format","label":"Dc Format","values":["PDF"]},{"key":"dc:title","label":"Title","values":["Three-dimensional wildland fuel mapping with remote sensing and machine learning"]}]}],"canonical_facts":{"dc:contributor.advisor":["Greenberg, Jonathan"],"dc:contributor.committeemember":["Bisbing, Sarah","Hanan, Erin","Nussear, Kenneth","Tavakkoli, Alireza"],"dc:creator":["Hartsook, Theodore Elliott"],"dc:date.accessioned":["2025-09-08T18:42:44Z"],"dc:date.available":["2025-09-08T18:42:44Z"],"dc:date.issued":["2025"],"dc:description.abstract":["The 21st century has seen a global increase in fire size and severity, including the western United States. Historically loaded forests, changing precipitation patterns, and lengthening fire seasons place a significant strain on land management and firefighting capacity. Fire behavior is a fine scaled process affected by climate, weather, topography, and fuel. The amount and arrangement of fuels are of particular interest because they can be directly controlled by managers to manage present-day fire risk and influence future response to fire. Existing fuel maps have a coarse resolution that limits their utility to understand fire behavior. In this dissertation, I address this research gap by using airborne and terrestrial LiDAR (ALS, TLS; respectively) and machine learning to make fine scaled predictions of fuels beneath the canopy. In my first chapter I use a gradient boosting regressor to boost the signal of ALS at the surface floor by using the ALS signal itself as features. I demonstrate that this method does a better job of detecting changes in surface fuels after two forest fires compared to ALS alone. In my second chapter I use a generative deep learning model to reduce fine-scale occlusion in ALS point clouds, leading to better measurement of vegetation beneath the canopy. In my third chapter, I compare this method with ALS to examine changes in fuel loading before and after a forest fire. I also examine how these changes influence management by comparing differences in common fuel metrics. I found that the use of machine learning and deep learning can significantly improve the capabilities of ALS beneath the forest canopy."],"dc:format":["PDF"],"dc:identifier.uri":["https://scholarwolf.unr.edu/handle/11714/11574"],"dc:language":["English"],"dc:language.iso":["en_US"],"dc:subject":["deep learning","fire","forest","lidar","machine learning","remote sensing"],"dc:title":["Three-dimensional wildland fuel mapping with remote sensing and machine learning"],"dc:type":["Dissertation"],"thesis:degree_level":["Doctorate Degree"]},"updated_at":"2026-07-27T21:46:30Z"}