Physical Sciences and Mathematics Commons^™

Harmful Algal Bloom Monitoring In South Florida Using Water And Aerosol Payloads On Uncrewed Aircraft, Kayla D. Taylor, Justin Fratto, Avinash Muthu Krishnan, David Zink, Blake Neal, Jose Cabrera, Emel Sen-Kilic, Marc Compere, Kevin Adkins

Student Works

Harmful algal blooms (HABs) threaten aquatic ecosystems, public health, and tourism and fishing industries throughout the United States, including Florida. Because HABs can develop and spread rapidly, geographically representative sampling efforts can be very difficult and labor-intensive, and traditional monitoring methods, such as satellite-based remote sensing, often lack the spatiotemporal resolution needed to characterize bloom dynamics. This project, funded by the Environmental Protection Agency (Grant No. 05D36825), aims to develop an uncrewed aircraft (UA)-based water and aerosol sampling approach to improve HAB monitoring and support prediction and mitigation efforts. One year into our funding period, we have integrated a multi-depth, …

A Deep Learning-Based Approach For Mapping Shrubs In Arctic Tundra From Very High-Resolution Imagery, Darko Radakovic, Mark Chopping, Bruce Cook, Aparna S. Varde, Stefan A. Robila

Department of Earth and Environmental Studies Faculty Scholarship and Creative Works

Arctic shrub expansion is accelerating due to climate warming, but accurately mapping it at fine spatial scales remains challenging due to spectral overlap between shrubs and surrounding vegetation, heterogeneous tundra landscapes, and limited availability of very high-resolution imagery. This study presents a novel mapping and validation approach that integrates very high-resolution (< 1 m) QuickBird (QB) and WorldView (WV) satellite imagery across 12 North Slope sites, using deep learning architectures to map changes in shrub cover from 2002 to 2022. Deep learning models, including Convolutional Neural Network (CNN), ResNet, VGG, U-Net, and Vision Transformer, were trained on 4,100 3-channel pansharpened multispectral images (200×200 m). Two traditional machine learning models (Random Forest and Support Vector Machine) served as a control group. Leave-one-site-out cross-validation showed that all models performed well (AUC-ROC of 0.78–0.86, F1 at validation of 0.33-0.46) and had a moderate to strong agreement with field reference data (R2 = 0.76–0.87). The pretrained VGG and ResNet models achieved the strongest field-reference alignments for shrub cover prediction, while the difference with traditional models was small. Temporal trend analyses for July and August demonstrate a cover change of 0.06–0.45% yr-1 with unweighted mean area slopes and 0.01–0.92% yr-1 inverse-variance weighting. Growth rates were found to depend on the analytical design. Furthermore, the limiting factor in mapping shrub cover was not model complexity but the quality of the training data labels (annotation method explained ∼26% of the error variance). Per-scene cover error was seen across architectures (ρ = 0.67). Comparison with NASA Goddard's LiDAR, Hyperspectral and Thermal (G-LiHT)-derived Canopy Height Models revealed a weak-to-moderate relationship (R2 = 0.04–0.77), suggesting horizontal cover and vertical canopy structure capture complementary aspects of shrub abundance. These results demonstrate the potential of deep learning to map decadal shrub expansion at sub-meter resolution.

Remote Sensing Integrated Modeling Of Fall Armyworm Parasitoid Interactions In Africa Using Fractional Differential Equations, Marian Adan, Steeven Belvinos Affognon, Henri E.Z. Tonnang, Klaus Greve, Christian Borgemeister, Georg Goergen

All Peer-Reviewed Publications

The invasion of fall armyworm (Spodoptera frugiperda; FAW) into Africa poses a major threat to agricultural systems, particularly maize production, necessitating effective biological control strategies. This study presents an ecological modelling framework that integrates temperature data with fractional differential equations (FDEs) to investigate interactions between FAW and its parasitoid natural enemies. By incorporating temperature-driven effects on life-history traits and ecological memory dynamics, the model captures key features of host–parasitoid interactions under variable environmental conditions. Equilibrium states corresponding to FAW extinction, parasitoid extinction, and coexistence were analytically derived, and sensitivity analysis of the FAW reproduction number highlighted its central role in …

Dbssnet: Dual-Branch Spectral-Spatial Network With Data-Driven And Knowledge-Guided Band Selection For Uav Hyperspectral Wheat Rust Detection, Subin Kim

All Graduate Theses and Dissertations, Fall 2023 to Present

Wheat rust is a serious plant disease that can reduce crop yield and quality. In practice, the disease is often noticed only after visible symptoms appear, when some damage may already be difficult to reverse. This thesis studies whether drone-based imaging can help detect wheat rust earlier and more reliably in field environments. 

Unlike an ordinary color photograph, a hyperspectral image records reflected light at many narrow wavelengths. These measurements can reveal useful information about plant condition, but they are also high dimensional, noisy, and difficult to analyze when only a limited number of labeled field samples are available. To …

Vision Transformers And Convolutional Neural Networks For Land Use Scene Classification, Arun D. Kulkarni

Computer Science Faculty Publications and Presentations

Land use scene classification (LUSC) from remote sensing imagery plays a critical role in environmental monitoring, urban planning, and sustainable resource management. In recent years, deep learning methods have significantly advanced the state-of-the-art, with Convolutional Neural Networks (CNNs) dominating the field because of their strong ability to capture local spatial features. However, the emergence of Vision Transformers (ViTs) has introduced a new paradigm that models long-range dependencies through self attention mechanisms, potentially enabling improved global context understanding. This study presents a comparative assessment of Vision Transformers and CNN-based architectures for remote sensing land use scene classification. Representative CNN models, such …

Quantifying Terrain Controls On Satellite-Based Snow Water Equivalent Estimation: A Spatially Explicit Machine Learning Approach, Brant Giovannetti

Geography and the Environment: Graduate Student Capstones

Terrain variables are widely incorporated into machine learning Snow Water Equivalent (SWE) models but are rarely evaluated for their independent contribution relative to spectral predictors. Using a four-tier stepwise Random Forest framework with Harmonized Landsat Sentinel-2 imagery and Airborne Snow Observatory LiDAR ground truth, this study isolates the contribution of elevation, slope, northness, and eastness across Peak and Ablation snowpack regimes in the East Taylor River Watershed, Colorado. During peak snowpack, adding terrain improved R² by 0.214, with elevation alone accounting for 42.8% of model importance. During ablation, full-dataset terrain gains were modest, increasing R² by only 0.036. However, when …

Multimodal Machine Learning For Soil Burn Severity Mapping Across California Wildfires, Sanjana Checker

Master's Theses

Accurate mapping of soil burn severity (SBS) is critical for post-fire watershed management, erosion risk assessment, and ecological recovery planning, yet traditional field-based approaches remain costly, time-intensive, and spatially limited. This thesis presents a machine learning pipeline for wall-to-wall SBS classification across California wildfires using multi-sensor satellite imagery, terrain derivatives, and bioclimatic covariates. Field-collected SBS observations (n = 2,180) from 52 wildfires occur- ring between 2013 and 2025, sourced from the U.S. Forest Service and CAL FIRE, were used to train and evaluate multiple classification architectures within a Google Earth Engine and Google Cloud-based prediction framework. After upsampling the unburned …

Spatiotemporal Pah Patterns In Size-Fractionated Particles (Pm>10-Pm0.1) From Northern Thailand Biomass Burning Via Sentinel-2, Phakphum Paluang, Watinee Thavorntam, Sarawut Sangkham, Phuchiwan Suriyawong, Hisam Samae, Thaneeya Chetiyanukornkul, Masami Furuuchi, Worradorn Phairuang

Research outputs 2022 to 2026

Biomass burning, particularly from forest fires and crop residue burning during the dry season, is a major source of particulate pollution across many Asian countries. However, accurately identifying these emissions remains challenging due to uncertainties in burned area estimation and the limited availability of country-specific emission factors. This study quantified the spatiotemporal distribution of emissions from biomass burning using satellite imagery. Burned areas were classified using a random forest (RF) algorithm implemented on the Google Colaboratory (Colab) platform. The RF model showed strong performance, with a kappa coefficient of 0.85 and an average accuracy of 0.81. Emission estimates for the …

Backscattering By Particles Of Sizes <0.2 Μm In Seawater, Xiaodong Zhang, Yuanheng Xiong

Faculty Publications

We measured the angular scattering functions at 517 nm in different size fractions, including bulk and very small particles (VSP for diameters <0.2 μm), in a variety of waters ranging from coasts to open oceans and from surface to depths as deep as 3,000 m. The measurements revealed a strong covariation in backscattering between VSP and bulk particle populations. The fractional contribution by VSP decreases from 60% in clear oceanic water where bulk particle backscattering coefficient bbp is on the order of 0.001 m⁻¹ to <2% in turbid coastal water where bbp ∼0.1 m⁻¹. An empirical model is developed (r² = 0.89) to estimate bbp for VSP from the bulk bbp, which can now be routinely obtained from satellite observations.

Spectral Classification Of Diverse Lithologies Using Multi- And Hyperspectral Satellite Data: A Comparative Study, Önder Gürsoy, Emre Özelkan, Rutkay Atun, Ayşe Betül Çalişkan, Ahmet Efe

Turkish Journal of Earth Sciences

Accurate lithological mapping requires selecting the appropriate remote sensing data and classification methods. This study evaluates the performance of four satellite datasets—Landsat 8 OLI, Sentinel-2A, ASTER, and Hyperion EO-1—using three spectral classification techniques: Matched Filtering (MF), Spectral Angle Mapper (SAM), and Spectral Information Divergence (SID). The study area is located between the Zara and Koyulhisar districts in eastern Türkiye and comprises diverse lithological units. A total of 49 rock samples collected in the field were used for validation. The results indicate that MF consistently outperformed the other methods, achieving the highest accuracy with Landsat 8 (Kappa = 94.2%). ASTER data …

Hyperspectral Image Classification Using Novel 1-D And 2-D Deep Neural Networks, Özlem Polat, Zümray Dokur, Tamer Ölmez

Turkish Journal of Earth Sciences

Hyperspectral image (HSI) classification is of critical importance in many fields including agriculture, geology, environmental monitoring, and urban planning. In recent years, many researchers have utilized deep neural networks (DNNs), known for their high performance in the classification of HSIs. When 2-D/3-D convolutional neural networks are used in HSI classification, filters are applied using input patches typically larger than 11 × 11. This allows spectral and spatial features to be evaluated together. However, this combination creates several problems. Because HSIs have low spatial resolution, they often do not contain strong texture details. Furthermore, features with little relevance to classification make …

Assessing Vegetation Health Following The 2014 Carlton Complex Fire Using Remote Sensing, Rachael Sv Pentico

2026 Symposium

Climate change, combined with decades of altered fire regimes and fire suppression, has increased fuel accumulation across western U.S. forests, contributing to more frequent, larger, and more intense wildfires. Remote sensing offers an effective approach for analyzing these large-scale fire events and their ecological impacts, as satellite imagery enables affordable, accessible, and reproducible monitoring of vegetation change across broad spatial and temporal scales. This study examined the 2014 Carlton Complex Fire in the Methow Valley, Washington. Burning 265,108 acres, it was the largest wildfire in Washington State history at the time and caused extensive agricultural and forest damage. Post-fire vegetation …

Evapotranspiration Everywhere, All The Time: Towards A Unified View From Earth Observation, Joshua B. Fisher, Martha C. Anderson, Diego G. Miralles, Kanishka Mallick, Paul C. Stoy, Youngryel Ryu, Wim G. M. Bastiaanssen

Biology, Chemistry, and Environmental Sciences Faculty Articles and Research

Scientists want to know everything, everywhere, and all the time. This is particularly true in Earth science, where we seek to understand processes that span from the molecular to the planetary scale in how the world works, how it affects us, and how we impact it—especially the water cycle. Evapotranspiration (ET) was the last component to be measured in closing the water cycle: for decades, closing the water budget meant adding up all the measurable components, then inferring ET as the residual. Early measurements relied on water loss from pans and weighing lysimeters, followed by sensors inserted into plants to …

A Remote Sensing Approach To Identifying Zones Of Instability In Mine Tailings Impoundments Using Multi-Sensor Satellite Data, Arden K. Pierce

Honors Theses

Mine tailings impoundments are structures built to contain the chemical and sedimentary byproducts generated by ore extraction activities. In this common method of mining waste storage, the tailings can be pumped into containment ponds and held behind earthen embankments to prevent the release of hazardous substances into the surrounding environment. Monitoring environmental factors that influence the structural integrity of these impoundments is essential for identifying areas vulnerable to instability and failure. Soil moisture is a factor that can affect the stability of mine tailings impoundment structures. This study evaluates the relationship between satellite-derived soil moisture and surface deformation at mine …

Scalable Roof Polygon Extraction And Geometric Characterization From Remotely-Sensed Data For Snow Load Assessment, Jashon Newlun

All Graduate Theses and Dissertations, Fall 2023 to Present

Heavy snow accumulation on rooftops is a serious structural risk in cold climates, and understanding how much snow builds up on different types of roofs is essential for safe building design. Currently, most data on roof snow loads comes from small, labor-intensive field surveys that cover only a handful of buildings at a time. This results in far too few measurements of buildings to draw confident conclusions about how snow behaves across communities. This thesis develops and demonstrates a new automated approach for measuring roof snow accumulation and extracting key building characteristics across thousands of buildings at once using airborne …

Urban Heat Mitigation In Newark, New Jersey Using Remote Sensing And Spatial Analysis, Ama Akowa Rhule

Theses, Dissertations and Culminating Projects

This study examines urban heat patterns in Newark, New Jersey, using remote sensing and spatial analytics to support targeted mitigation strategies. Land Surface Temperature (LST) was derived from a Landsat 8 Operational Land Imager/Thermal Infrared Sensor (OLI/TIRS) scene acquired on July 28, 2025, representing a summer snapshot of surface thermal conditions. While urban heat island effects are typically considered long-term climatic phenomena, this study uses a single-date observation to capture spatial variability in temperature across the city. Land cover variables, including tree canopy and impervious surface fractions, were obtained from the National Land Cover Database (NLCD), and demographic variables, including …

Post-Wildfire Erosion And Biogeochemistry: Integrating Aerial Imagery And Soil Testing To Assess Landscape Recovery, Justin A. Allred

All Graduate Theses and Dissertations, Fall 2023 to Present

After a wildfire, land managers are required to monitor large tracts of land with limited time and budgets. Traditionally, tracking landscape recovery is a time intensive process. This research explored the effectiveness of using drones (UAVs), strategic soil sampling, and computer modeling as additional tools for land managers in order to monitor the recovery more efficiently.

By using drones to collect aerial imagery and using machine learning, plant regrowth was monitored in back-to-back years. The machine learning model performed well at telling broad groups apart (trees vs grass) but it struggled to identify differences between plant species. Because many plants …

Passive Microwave Remote Sensing Of Flash Drought Impacts On Vegetation, Quinton R. Deppert

School of Natural Resources: Dissertations, Theses, and Student Research

In 2002, Dr. Mark Svoboda characterized a new form of drought known as flash drought. Flash drought was defined as a rapid decline in vegetation health caused by severe heat and drought. In recent years, attempts to quantify the impacts of flash drought via precipitation, soil moisture, evapotranspiration, and temperature indicators have proliferated. What has rarely been quantified is what the rapid decline in vegetation health amid flash drought looks like through remote sensing. This is because vegetation health indices like the Normalized Difference Vegetation Index (NDVI) are derived from the visible and infrared regions of the electromagnetic spectrum and …

Integrating Multi-Source Data With Machine Learning Techniques To Upscale Wetland Carbon Dioxide Fluxes, Abdullah Sulaiman Abdullah Al Fazari

Electronic Theses and Dissertations 2020 - Present

Accurate quantification of atmospheric carbon dioxide (CO₂) fluxes in wetland ecosystems is essential for understanding their role in both regional and global carbon dynamics, particularly in the context of climate change. However, the spatial and temporal heterogeneity of wetlands presents major challenges for developing reliable upscaling models. This research developed and validated a comprehensive framework to upscale CO₂ fluxes across the Everglades National Park (ENP) and Big Cypress National Preserve (BCNP) in South Florida through the integration of multi-source datasets, including AmeriFlux eddy covariance (EC) tower measurements, NASA’s BlueFlux airborne CO₂ data, and multispectral satellite imagery from Landsat 8 OLI …

Developing A Multi-Band Gnss Remote Sensing Station To Measure Precipitable Water Vapor For Flash Flood Nowcasting, Madison Gleydura

Doctoral Dissertations and Master's Theses

Flash flood nowcasting in Central and Southern Appalachia is particularly challenging due to steep terrain, narrow valleys, highly localized rainfall patterns, and limited measurement coverage. Traditional remote sensing methods, such as Doppler radar and microwave radiometry, suffer from reduced resolution at long range and signal blockage by mountains. GNSS-meteorology offers an established alternative for measuring precipitable water vapor and is currently integrated into several numerical weather models. Recent research demonstrates that commercial-grade GNSS receivers can produce tropospheric products comparable to those from geodetic-grade equipment. The gaps in mountain coverage can be addressed by developing a low-cost, self-contained embedded system that …

Classification Of Land Cover In Sentinel-2 Imagery Using Machine Learning Models, Ehsan Ali Al-Zubaidi, Mohammed Ridha Hammoodi, Ahmed Naser Alzurfi

Al-Bahir

Remote sensing data of medium resolution are commonly used to classify land cover, and machine learning (ML) models have taken on a central aspect in the necessary data analysis. Ordinarily, land cover is coded on a pixel basis on the basis of Digital Number (DN) values, which in turn are computed across several spectral bands. This paper is concerned with land cover mapping in Mosul, Iraq, based on satellite images captured by Sentinel-2. Two platforms featuring unsupervised classification algorithms were used, Google Earth Engine and ArcMap, making it possible to use K-means and X-means in Google Earth Engine and ISO …

Changes In Land, Ocean, Atmospheric Parameters Associated With The 2025 Myanmar (Mw 7.7) Earthquake, Feng Jing, Akshansa Chauhan, Ashwani Raju, Ramesh P. Singh

Mathematics, Physics, and Computer Science Faculty Articles and Research

Multiple parameters associated with the land, atmosphere, and ocean were analyzed to study short-term and immediate pre-earthquake changes associated with the 28 March 2025 Myanmar earthquake (Mw 7.7). Anomalous clear-sky outgoing longwave radiation (ClrOLR) and trace gases (CH₄, CO, and O₃) were detected within two months prior to the mainshock. Vertical changes at different pressure levels suggest a possible underground source. High-temporal-resolution observations of the infrared brightness temperature and surface air pressure revealed short-lived fluctuations shortly before the earthquake, which may reflect localized stress adjustments and surface latent heat flux release during the final stage of earthquake preparation. In the …

Using Ensemble Disagreement To Stabilize Conformal Prediction Under Distribution Shift, Patrick D. Murphy

Master's Theses

Semantic segmentation of eelgrass from drone imagery is crucial for coastal habitat monitoring, restoration, and management, as these habitats continue to see rapid changes due to climate change and human influence. However, the reliability of generalizing a deployed classification model relies on both high-accuracy segmentation as well as robust uncertainty quantification that holds up when conditions change over years or locations. Conformal prediction (CP) is a method that converts a classifier's output into prediction sets with a guaranteed average coverage level for in-distribution data. However, the “vanilla” conformal score can often under-cover in hard or out-of-distribution (OOD) regions under drift. …

Unraveling Patch Size Effects In Vision Transformers: Adversarial Robustness In Hyperspectral Image Classification, Shashi Kiran Chandrappa, Sidike Paheding, Abel A. Reyes-Angulo

Michigan Tech Publications

Highlights: This work investigates the effect of spatial patch size on the classification accuracy and adversarial robustness of Vision Transformer-based architectures for hyperspectral image analysis. What are the main findings? Smaller patch sizes generally exhibit stronger adversarial robustness while maintaining comparable clean classification performance. Larger patch sizes tend to reduce robustness by increasing sensitivity to localized adversarial perturbations, with some dataset-dependent variations. What are the implications of the main findings? Spatial patch size is an important design consideration when applying Vision Transformers to hyperspectral image classification tasks. The findings provide practical guidance for informed patch-size selection in robust, deployment-aware transformer-based …

Mapping Urban Vegetation Changes Using Planetscope Imagery And Gis: A Case Study Of The 2024 Dubai Flood, Sumayya Almansoori, Xin Hong

All Works

In April 2024, unexpected heavy rainfall triggered one of the most severe flooding events in Dubai’s recent history, causing widespread concern for urban infrastructure and green spaces. This study evaluates the flood’s impact on urban vegetation in South Dubai using high-resolution PlanetScope satellite imagery and the Normalized Difference Vegetation Index (NDVI). Vegetation conditions before and after the flood (April 14 and April 18–19, 2024) were quantified and compared using NDVI analysis within QGIS to assess changes in vegetation health and coverage. Results indicate significant declines in vegetation health in areas dominated by intensive turf management, such as Damac Hills and …

Compilation Of A Nationwide River Image Dataset For Identifying River Channels And River Rapids Via Deep Learning, Nicholas Brimhall, Kelvyn K. Bladen, Thomas Kerby, Carl J. Legleiter, Cameron Swapp, Hannah Fluckiger, Julie Bahr, Makenna Roberts, Kaden Hart, Christina L. Stegman, Brennan L. Bean, Kevin R. Moon

Mathematics and Statistics Student Research and Class Projects

Remote sensing enables large-scale, image-based assessments of river dynamics, offering new opportunities for hydrological monitoring. We present a publicly available dataset consisting of 281,024 satellite and aerial images of U.S. rivers, constructed using an Application Programming Interface (API) and the U.S. Geological Survey’s National Hydrography Dataset. The dataset includes images, primary keys, and ancillary geospatial information. We use a manually labeled subset of the images to train models for detecting rapids, defined as areas where high velocity and turbulence lead to a wavy, rough, or even broken water surface visible in the imagery. To demonstrate the utility of this dataset, …

Land Cover Classification Using Optimized Imagery Resolution And Machine Learning Algorithms For A Long-Term Monitoring And Restoration Project, Jessica R. Suoja

Cal Poly Humboldt theses and projects

Ecosystem services and functions are prone to water resource exploitation resulting in cascading effects that include decrease of biodiversity and loss of riparian vegetation, a keystone habitat in desert riparian ecosystems. Mono Lake is a prime example of overexploitation of water resources leading to legal action and eventually a legally mandated long-term monitoring and restoration project. Monitoring and restoration projects can benefit from techniques such as remote sensing and machine learning algorithms to generate accurate land cover classification maps for calculating land cover change over time. However, the spatial resolution of remote sensing imagery and the machine learning algorithms chosen …

Mapping Coastal Vegetation To Evaluate Salt Marsh Decline From 2016 To 2025 In South Carolina, Usa, Zak Henry Bartholomew

Theses, Dissertations and Capstones

Accelerating sea-level rise and storm surge events pose a substantial threat to salt marsh ecosystems, which provide critical ecosystem services. Classification and mapping of coastal vegetation through remote sensing can identify marsh dieback events and provide spatial context for patterns of marsh vulnerability. We used a multi-sensor approach that incorporated LiDAR data, multispectral imagery, and field-derived cover estimates to classify Spartina alterniflora (smooth cordgrass) dominated marsh types on Marine Corps Recruit Depot Parris Island (MCRDPI), a sea island in South Carolina, USA. In summer 2025, surveys were conducted within marsh classes to collect training and validation data. We used random …

A Deep Learning Approach For Mapping Shrubs, Wet Tundra And Surface Water In Arctic Tundra With Very High Resolution Satellite Imagery, Darko Radakovic

Theses, Dissertations and Culminating Projects

Arctic shrub expansion threatens to accelerate permafrost thaw through complex feedbacks, yet whether shrubs primarily indicate or drive degradation remains unresolved. This dissertation integrates deep learning analysis of two decades of satellite imagery with LiDAR canopy structure and radar soil moisture data to reveal that shrubs play a dual role: young, expanding shrubs signal active permafrost thaw, while mature, tall shrubs stabilize underlying permafrost through insulation. By demonstrating that vertical canopy structure predicts thaw depth better than cover extent alone, this work establishes a scalable framework for monitoring permafrost vulnerability across the rapidly changing Arctic. Arctic shrub expansion is accelerating …

Using Remote Sensing Technology To Develop A Framework For Improving Hydrologic Models, Marissa Cook

Theses, Dissertations and Capstones

With increased storm intensity due to climate change and urbanization, flash flooding has become an increasingly significant issue globally and regionally. Although the factors influencing urban flash flooding are well-known, there is a growing need for technology to accurately and remotely predict the chance of a flash flood occurring from any given rain event to give people time to prepare. This study aims to use multispectral satellite imagery to provide a framework for improving near real-time flood predictions in an urban area of a high gradient, fourth order stream impacted by flooding. Specifically, we utilize satellite imagery to create the …