Welcome to the Geo-Resolution 2021 poster session. Students from the region submitted posters on a wide range of research topics connected to geospatial sciences or using geospatial tools.
Traditionally, this would have been a standard poster session, but to accommodate for the pandemic Geo-Resolution is a hybrid event with a limited in-person gathering and an online audience. Therefore the in-person attendees can review the printed posters and the online audience can review the posters digitally on this webpage. To help bridge the gap the students provided a video or audio recording in which they talk about their work.
Some posters are traditional prints and other posters are digitally interactive. We have links to the digital copy or the interactive maps below.
Vote for Your To Three Posters
The top three posters will win $1,000, $750 and $500 respectively. Please submit your votes by Tuesday, September 14 before 2 p.m.
We hope that you enjoy reading about the interesting research of students in our geospatial ecosystem.
Geo-Resolution 2021 Student Posters
Poster 1: Sediment Pattern and Rate of Bathymetric Changes Due to Construction of Breakwater Extensions at Nowshahr Port
Authors: Farhad Sakhaee, Southern Illinois University Edwardsville; Mohamad Hossein, Southern
Illinois University Edwardsville; Rohan Benjankar, Southern Illinois University Edwardsville.
Abstract: "Erosion, scour and sedimentation are the most common phenomena which should be considered in the design of marine structures. Seas and oceans are dynamic environments, hence understanding their bathymetric changes results in more efficient design of marine structures and also enhances their durability. To achieve this goal one needs to know about currents and waves and the effects of their interactions. Also one needs to know about the rate of sedimentation and erosion in the field. On the other hand, to obtain a sufficient depth in order to facilitate marine traffic in harbors the rate of sediment deposition must be calculated and a plan for an effective annual dredging program should be developed. In addition, creating a calm situation inside the harbor for ships anchorage is needed. Breakwaters will fulfill this duty by damping the energy of waves through diffraction."
Poster Link: https://drive.google.com/file/d/1V8oSbLsKs82fLiPbu6Gk8LFo_2sgpQIl/view?usp=sharing
Video Link: https://drive.google.com/file/d/1bxvBrtHpGbp0vq2DSU1dDROFFH-Da3Ez/view?usp=sharing
Poster Two: CAFOs Story Map
Authors: Daniel O'Farrell, Saint Louis University; Melissa Vatterott, Missouri Coalition for
Abstract: “This ArcGIS Online Story Map presentation was created as an educational tool to visualize and learn about the phenomena of Concentrated Animal Feeding Operations (CAFOs) and their detrimental impact on the environment and public health.”
Poster Link: https://arcg.is/1W8rTf
Poster Three: Racial Bias in Contaminated Land Locations: Assessment of St. Louis Metro Area
Radioactive Sites and Brownfield Contaminated Land
Author: Alex Yentumi, Harris-Stowe State University.
Abstract: “Motivated by the desire to rehabilitate contaminated lands in a way that prioritizes social justice, the research goal was to conduct a case study of the St. Louis Metropolitan Area to see how populations of African-Americans were impacted by contaminated lands. Specifically, by considering how contaminated lands coincided with black populations, an assessment was made of how much contaminated lands are impacting blacks. Two types of contaminated lands were studied in particular: radioactive sites and brownfields. We used the Saint Louis Metro Area for study."
Poster Link: https://drive.google.com/file/d/1FpYDIvr8JElVsq988ODKTrOIsCQ2E39d/view?usp=sharing
Video Link: https://drive.google.com/file/d/164DkulXIJZ0kLVvUcCfXSzy-zpI3SPlA/view?usp=sharing
Poster Four: Mapping Coral Reefs Through Imagery
Author: Austin Blakeslee, Lindenwood University.
Abstract: “It is crucial to the ocean's ecosystems to assess the health of coral reefs, as they are being threatened through ocean acidification, pollution, over-harvesting, rising ocean temperatures among other stressors. Understanding how reefs are affected by environmental and human pressures and working with resource managers to identify how to sustain coral reef systems is essential for the future of the health of the oceans. Researchers need to determine how many healthy reef areas exist. 3D imaging through remote sensing would play an essential role in assessing coral reef health so that resources can be mitigated.”
Poster Link: https://storymaps.arcgis.com/stories/be842e9664ba4f5a9668c183b2857b6d
Poster Five: America's Electric Vehicle Charging Infrastructure in 2021
Author: Henning Lohse-Busch, Saint Louis University.
Abstract: "This study provides an assessment of the fast-charger infrastructure for long distance travel across America and a regional comparison to gas stations. Tesla has better coverage since they designed their supercharger network to enable their customers to drive from coast to coast and top to bottom. The SAE J172 network which is built and supported by several companies has some ‘holes’ in their coverage but has a denser network of L2 chargers in metropolitan areas. Most Electric Vehicles do have a ubiquitous charging network for overnight charging at home that does cover typical commuting needs. The electric utilities will have to plan for the increase in demand for more home charging for Electric Vehicles. There are 50 times more gas stations in America compared to fast chargers across all networks. These fast chargers dispense energy at 60 to 200 times slower energy rates (power). Even if Electric Vehicles require less energy compared to their gasoline counterparts to cover the same distance, an increase in Electric Vehicles in the market needs to be matched by an increase in DCFC stations."
Poster Link: https://arcg.is/0fu0CL
Video Link: (embedded in Story Map)
Poster Six: Dragonfly species with sexual ornaments are less vulnerable to land use changes, but not recent climate warming, across the American Midwest
Authors: Noah T. Leith, Saint Louis University; Michael P. Moore, Washington University in
St. Louis; Kim A. Medley, Washington University in St. Louis; Kasey D. Fowler-Finn,
Saint Louis University.
Abstract: “Understanding which factors determine extinction risk is a major concern for modern biologists. Theory predicts that sexual selection (competition for mates) can drive either extinction or adaptation following human-induced environmental change. However, previous empirical studies show conflicting results. One explanation for these discrepancies is that sexual selection and sexually selected traits may help organisms withstand some threats but render them more vulnerable to others. Leveraging >90,000 citizen science observations, we used a novel, spatially explicit analysis of dragonfly communities to test if sexual selection buffers species against some threats more than others. We found that dragonfly species with wing coloration—a proxy for more intense sexual selection—make up a larger proportion of local communities in areas with more impervious surface and greater agricultural land use. In contrast, we found no relationship between an area’s recent warming and the proportion of species that were ornamented. Overall, our work suggests that sexual selection may buffer species against extinction in changing environments, but these advantages may be threat specific.”
Poster Link: https://drive.google.com/file/d/1dUPoZyfFz2_5KUN7QdktkKxwWQ5weeJ2/view?usp=sharing
Video Link: https://drive.google.com/file/d/13Og6jT6nX4U60x9ENtLTrNXwheGCDf5t/view?usp=sharing
Poster Seven: Beyond the Latino Vote: A Geo-Spatial Visualization Tool to Analyze Socio-Spatial Variation in Latino Voting Patterns
Authors: Laura Brugger, Saint Louis University; Pedro Spindler-Ruiz, Saint Louis University.
Abstract: “Latino voters have often been portrayed as a monolithic group, even though studies have shown that Latino voting patterns are not uniform, and instead vary based on key socio-spatial characteristics such as education, class, region, and country of origin, among many others. While peer-reviewed studies have examined socio-spatial variation in Latino voting patterns at the national, state, and county level, a socio-demographic spatial analysis at the precinct level is yet to be seen. The absence of research at smaller areal units such as precincts can lead to masked differences, especially in large counties. This inaccurate portrayal of a Latino voting group presents challenges for protecting voting rights among a growing electorate. Through this project, we sought to gain a better understanding of variation in how Latinos voted in the 2020 Presidential Election through the creation of an open-source, interactive GIS visualization tool and by estimating socio-demographic characteristics at the precinct-level. We plan to use this tool to inform strategies that address the ecological fallacy and strengthen the Voting Rights Act for Latinos in the United States.”
Poster Link: https://arcg.is/1rvHOu
Video Link: (embedded in Story Map)
Poster Eight: An Experimental-Computational Investigation of Damaged UAV Wings- Detecting and Locating Damage Using Machine Learning
Author: Siddharth Chandra Shekar, Saint Louis University.
Abstract: “With a steady increase in the role that UAVs play in our day to day lives, their robustness and resilience becomes more important for safe operations. In-flight wing damage could potentially interfere with the operations of these small autonomous aircraft - the level of damage could vary and sometimes might not immediately render the aircraft unflyable in level flight, but the same damage could potentially lead to undesirable effects while the aircraft is making maneuvers. If the on-board flight controller does not identify such wing damage, it is likely for such sudden changes to be too much for the aircraft to safely recover from. In either condition (benign or unrecoverable), awareness of the extent of damage is critical in order to continue the mission or land safely. The work presented here is a follow-up to our prior work to investigate asymmetric wing damage using a combination of wind tunnel experiments; in this effort we discuss implementing machine learning techniques on the data from computational simulations and wind tunnel experiments to predict the potential areas and extent of damage on the wing surface.”
Poster Link: https://drive.google.com/file/d/1tZeBlfZWmPc3glklFODQklxs8va376Hz/view?usp=sharing
Video Link: https://drive.google.com/file/d/1FOSbzALzgOdv90s3pilJD_aTZNR94X2v/view?usp=sharing
Poster Nine: Public Colleges and COVID-19 Vaccination Rates
Author: Alyssa Coleman, Saint Louis University.
Abstract: "Introduction: As the US attempts a return to normalcy, many schools and businesses are requiring COVID-19 vaccination. The vaccination requirement and the availability of vaccine appointments at academic institutions may be a major contributor to vaccination rates. Research Questions: Is there an association between the number of academic institutions within a county’s 30-mile radius and that county’s COVID-19 vaccination rate. Additionally, is there a difference in change in COVID-19 vaccination rates between May 2021 and July 2021 between states with a high density of 4-year colleges and universities and states with lower density."
Poster Link: https://storymaps.arcgis.com/stories/30c9abf55acf48479dd08940cec87ef5 Video Link: https://drive.google.com/file/d/1TY5OtSDS49OOvHZmb-RLX0jEXpggSF6C/view?usp=sharing
Poster 10: Sentiment and Emotions of Tweets Regarding COVID-19 Vaccination from a Midwestern Urban Region
Authors: Bryce Takenaka, Saint Louis University; Stephen Scroggins, Saint Louis University;
Shreya Nagendra, Saint Louis University; Enbal Shacham, Saint Louis University.
Abstract: "Background: In the US, unprecedented attempts to achieve wide scale COVID-19 vaccination distribution have been met with a charge of false and misleading information. Misinformation has the potential to adversely affect vaccine uptake and social media is a hospitable environment for misinformation to rapidly spread. Objective: The purpose of this study is to investigate the sentiments and emotions of Twitter data on COVID-19 vaccines within the St. Louis region, since the FDA approval of the vaccines."
Poster Link: https://drive.google.com/file/d/1xU8e_8BxFaO3as6DX_w6Hn0J68U9aL_S/view?usp=sharing
Video Link: https://drive.google.com/file/d/14WXluKTDaou8E9OnERoYGbj_QhtM-87L/view?usp=sharing
Poster 11: St. Louis EcoUrban Assessment
Authors: Julia Feller, Washington University in St. Louis; Rebecca Weaver, The Nature Conservancy
Missouri; Brandon Cox, The Nature Conservancy Missouri.
Abstract: “The St. Louis EcoUrban Assessment is a tool by The Nature Conservancy in Missouri (TNC) that both analyzes environmental hazards in the St. Louis region and lifts up the work of community networks, place-based organizations, and other anchor institutions working to ameliorate those risks. An ArcGIS StoryMap tells a narrative of environmental challenges, the communities most burdened by them, and the projects by community organizations that can serve as models for use of the tool in the future. An interactive map component accompanying the StoryMap allows users to choose specific environmental factors and community assets to map and view.”
Poster Link: https://storymaps.arcgis.com/stories/1fa2d6b75d8740cd82dd0583bf44833a
Video Link: https://drive.google.com/file/d/1maVpaSsofSfQIVvdP2HCQ-b4myUU_EZl/view?usp=sharing
Poster 12: Helping Achieve Maritime's Mission: Deep Dive into Japanese Electronic Navigational Charts
Author: Lauren Wratchford, Saint Louis University.
Abstract: “The purpose of this project is to evaluate Japanese Electronic Navigational Charts (ENC) against Raster Navigational Charts (RNC) and NGA Digital Nautical Charts (DNC) to aid NGA in its transition from DNC to ENC production. NGA is transitioning from a military product (DNC) to a commercial, internationally standardized product (ENC) in order to have a greater advantage on the world stage and to produce more accurate, timely products for DoD and other customers. The Maritime Safety Office’s move to ENC supports NGA’s greater mission to develop a more data-centric production environment. As Japan’s charts and sources are historically reliable and of high quality, it was predicted that Japanese ENC could be swiftly and easily incorporated into NGA products. Though it was predicted that Japanese ENC could be easily integrated into NGA’s Safety of Navigation products, a majority of the Japanese ENC have different extents and scales from NGA’s existing DNC. Generally, as product scale becomes smaller, there are more discrepancies between products. Consequently, the level of effort required for ENC incorporation becomes significantly larger. While the project is ongoing, it can be concluded that most of the Japanese ENC reflects current and accurate Safety of Navigation data and should be incorporated into NGA’s holdings.”
Poster Link: https://drive.google.com/file/d/1vkYuJOGgtuqGsCkVPjD2mlVBuYmnvTzQ/view?usp=sharing
Video Link: https://drive.google.com/file/d/1HGscN5pqUJD5-qMKhfkOlIHD2hLr1zCD/view?usp=sharing
Poster 13: Geospatial Analysis of Change to Kenya’s Agro-Climatic Zones (1980-2020)
Authors: Ted Lawrence, Saint Louis University; Justin Vilbig, Saint Louis University; Enbal
Shacham, Saint Louis University; Vasit Sagan, Saint Louis University.
Abstract: “Climate change is a grand challenge for humanity as the global mean temperature in 2017 reached 1.0°C above pre-industrial levels. Changes to temperature and precipitation patterns are especially challenging for developing countries where livelihoods are agriculturally dependent. In Kenya, agricultural activities are organized according to agro-climatic zones (ACZs) based on average annual temperature and precipitation estimates that Kenya’s Ministry of Agriculture established in 1980. Since the ACZs creation however, Kenya’s climate has also experienced changes, such as a 1.0°C increase in temperature. Therefore, we used geospatial analytics to examine how the ACZs have changed over 40 years and developed a current map of their spatial arrangement. Our research questions were: (1) how have average annual temperature and precipitation changed within each ACZ; and 2) how have the spatial patterns and distributions of the ACZs changed? To investigate these questions, we used georeferenced climate data from 1980 and 2020 to map and analyze changes to temperature and precipitation patterns, and related them to changes in ACZs. We show that Kenya has on average become warmer and dryer, and how ACZs with higher temperatures and lower precipitation have spatially expanded and shifted geographically. Our study can inform policies related to food security and agricultural development under climate change and can help to guide actions related to UN sustainable development goals in Kenya.”
Poster Link: https://drive.google.com/file/d/1jkdgw1aGCXi5xffW5fKdrf2Pa_ovaaXF/view?usp=sharing
Video Link: https://drive.google.com/file/d/1AUraEZZMmQlGqJrvjaCA0wiNPFy1ha6t/view?usp=sharing
Poster 14: The Development of a UAV Platform for Intelligent Flight Control
Authors: Henry Wright, Saint Louis University; Srikanth Gururajan, Saint Louis University.
Abstract: "A new era of aviation is upon us. With the onset of unmanned aerial vehicles (UAVs),
the nature of flight is poised to change forever. It is predicted that the total value
of business and labor across all industries that could be improved with UAVs is upwards
of $127.3 billion The increase in the demand for UAVs will naturally lead to increased
autonomy for reasons of both safety and cost, but we will not be able to fully reap
the benefits of these autonomous UAVs until they are able to fly safely in urban or
cluttered environments, under a wide range of conditions. Research in intelligent
adaptive flight controllers to address this is active and strides have been made,
but the robustness of these flight controllers in autonomous vehicles experiencing
unmodeled nonlinearities including system failures, especially under non-ideal atmospheric
conditions, remains to be determined. In the case of autonomous UAVs, these failures
can be random and possibly unidentifiable. Consequently, the most promising method
of control would be for the system to learn the best allocation of its resources,
following a failure. This current research project focuses on bridging this gap through
the extensive development and testing of an autonomous UAV in these unmodeled conditions.
The UAV, modeled after a T-38, is currently being manufactured and the intelligent
flight controller is beginning to be developed. Flight testing has not yet occurred,
but it is expected that this flight controller aided by machine learning will outperform
traditional adaptive control techniques.”
Poster Link: https://drive.google.com/file/d/1qG8UVc7nAAMy2UCyEEPRGMpyJ-NdaBx1/view?usp=sharing
Video Link: https://drive.google.com/file/d/15p3LIAh4hO1XW1QnqKhl6ALX_iPikcH9/view?usp=sharing
Poster 15: Diversity, Desirability, and Community: How Parents Assess Residential and Educational Fit Through Spatially Conscribed Social Networks
Author: Heather Sweeney, University of Missouri.
Abstract: “The decoupling of housing and school choice through a saturated school choice marketplace facilitated by educational policies over recent decades has resulted in the reification of residential and educational segregation. If we aim to reduce such segregation and work toward greater educational access, we must better understand what shapes parents' processes of residential and school choice, and how these processes are interrelated. Through a close examination of how mothers assess neighborhood and school fit, this basic qualitative asked: How do urban, middle-class mothers assess neighborhood and school fit within an open enrollment environment? The core data collection included interviews with 9 St. Louis mothers who interacted with private, charter, and public school options as they made their school choice decision in an open enrollment environment. Each of these families has school-aged children and currently resides within one of three neighborhoods on the south side of St. Louis that has experienced a steady increase in housing and commercial development. While navigating the housing decision-making process and the school choice process the mothers constructed a perception of desirability, diversity, and community to assess fit. Perceptions which resulted from a belief or experience with the ability to gain community membership. To inform and strengthen the qualitative analysis, census demographic data as well as St. Louis metropolitan area public school enrollment data were placed in a Geographic Information System (GIS) analysis.”
Poster Link: https://drive.google.com/file/d/1SBaOVuw8p6FyiZeW9bwne-fMKqHQXLtj/view?usp=sharing
Video Link: https://drive.google.com/file/d/1rcePTxboNm3GBtd2hpM8cnld5ZQEGlH_/view?usp=sharing
Poster 16: The Evolution of GIS Applications in Monitoring Illinois Mine Subsidence
Authors: Nick Milner, Southern Illinois University Edwardsville; Logan Pelo, Southern Illinois
Abstract: “Since the first underground mining operation in 1848, subsurface coal has been extracted from over 1,177,000 acres of land in Illinois. Seventeen percent of all underground mines in Illinois underlie urban areas. Therefore, this presentation will demonstrate the progress made in developing products in two key areas of mine subsidence mitigation. The first is the creation of a digitized and georeferenced database of historical mine maps. The second is a comprehensive surveying program to monitor mine subsidence. The original mine maps were sourced from various local government institutions and were scanned, catalogued, and then georeferenced to topographic quadrangles. Many were divided into smaller sections for scanning purposes, resulting in a total database of 10,000+ records with over 9,000 having been completely georeferenced. In parallel to the database creation, surveying of subsidence events has occurred for over 30 years. Since the inception of the surveying program, approximately 80 sites in Southern Illinois and roughly 35 sites in Springfield have been routinely surveyed (at varying frequencies). The application of GIS technologies benefits both the historic mine map georeferencing and surveying program. ERDAS Imagine software was integral in georeferencing mine maps to the surface of the earth, with most maps being spatially accurate to within 30-50ft of their mapped locations. GIS technologies has enabled the creation of a spatial database within ArcGIS Online to increase the management capabilities and efficiency for the entire survey program. Ultimately, the use of GIS in the subsidence program increases the capability to inform and serve the public.”
Poster Link: https://drive.google.com/file/d/1qh_6Nve0iy3zbwB3g1xa8_vwYLAT0LaL/view?usp=sharing
Video Link: https://drive.google.com/file/d/11-6AH2VI_yrnoNbDmmquTGWJQLIVzm3N/view?usp=sharing
Poster 17: Corn Yield Prediction Using UAV-based Hyperspectral and LiDAR Data Fusion and Automated Machine Learning
Authors: Kamila Dilmurat, Saint Louis University; Vasit Sagan, Saint Louis University; Maitiniyazi
(Mason) Maimaitijiang, Saint Louis University.
Abstract: “Predicting crop grain yield before harvest is decisive for grain policy-making and food security; Moreover, yield prediction at fine-scale is significant in precision agriculture and high-throughput plant phenotyping. The rapid advancement of Unmanned Aerial vehicles (UAV) and sensor technologies allowed remote sensing data acquisition with high spatial, spectral and temporal resolution at a low cost. This poster aims to illustrate the capacity of the UAV-based high-resolution hyperspectral, LiDAR dataset and their fusion in predicting corn grain yield within an automated machine learning framework. To this end, high-resolution UAV-based hyperspectral and LiDAR data were collected throughout the growing season of 2020 over a cornfield near Urbana Champaign, Illinois, USA. Canopy spectral and texture features were extracted from hyperspectral imagery, and canopy structure features were derived from LiDAR point clouds. Canopy spectral, texture and structure elements, and their combination, were used to predict corn yield. The contribution of hyperspectral and LiDAR data fusion to improving overall prediction accuracy and the potential of data fusion in advancing spatial transferability and robustness of yield prediction models were examined.”
Poster Link: https://drive.google.com/file/d/1N4BxvJV9pBG9R3t_UwPRimPsELGoh3F4/view?usp=sharing
Video Link: https://drive.google.com/file/d/1i8H2C9ogh8YwQ7r7LuqeP0oUo4wKqIdT/view?usp=sharing
Poster 18: The Double Vulnerability of COVID-19: Mapping New PrototypesAuthors: Ruaa Al Juboori, Saint Louis University; Divya S. Subramaniam, Saint Louis University; Travis Loux, Saint Louis University; Chris Prener, Saint Louis University; Ness Sandoval, Saint Louis University.
Abstract: "Background: COVID-19 health outcomes varied within regions and communities, prompting questions about which populations and regions are at higher risk for unfavorable outcomes. Therefore, this study aimed to create new prototypes to explain the association between location, racial/ethnic majority, and COVID -19 Case Fatality Rate (CFR) and vaccination.”
Poster Link: https://storymaps.arcgis.com/stories/1b8e82ae40d047ecaaef5ecb6ab9eff8
Video Link: https://drive.google.com/file/d/1i89sdpiM9B4jEyDmVexVGC5LmPabnH9h/view?usp=sharing
Poster 19: Take Me Out to the Ballgame: COVID-19 Prevalence and Mobility Surrounding Sporting Events
Authors: Stephen Scroggins, Saint Louis University; Enbal Shacham, Saint Louis University.
Abstract: "As evidence continues to point to “super-spreader” events being responsible for COVID-19 outbreaks, understanding mobility surrounding these events becomes integral to prevention and reduction of mobility and mortality. The purpose of this study was to identify geographic patterns of movement among visitors to the 2021 opening weekend games of the St. Louis Cardinals. Smart-device GPS data was used to identify attendees among the first three games of the St. Louis Cardinals during the first week of April 2021 (n=1,925). ZIP code level business location data and COVID-19 case-rates were then integrated into aggregated mobility data. Clustering analysis was used to determine significant hot-spot ZIP codes where visitors reside and what types of additional locations were visited after the games. Increases in COVID-19 rates were found to be significantly associated, not with attendee resident locations, but commercial locations where event visitors went after the game. Over 1/3 of the sample was determined to have gone to at least one additional commercial location after the game, 50% of which were determined to be bars/restaurants. Post-event mobility was found to be associated with COVI-19 rate increases. This added layer of risk not only places more burden on location visitors, but employees of these locations too. With little-to-no state preventative measures enacted in Missouri, spatial risk and prediction of COVID-19 incidence have been difficult to describe. These results highlight the complex relationship between geography, mobility, and infectious disease transmission and should be leveraged to develop more effective prevention policy.”
Poster Link: https://drive.google.com/file/d/1Mr8zQCkxUM8CBd8cmWFvw0RDwQtY-xvj/view?usp=sharing
Video Link: https://drive.google.com/file/d/1l-CIGCIlEY0TJRmGRtCUyrGjqf_xC6Jq/view?usp=sharing
Poster 20: Mapping Field-scale Nitrogen Use Efficiency of Maize using Remote Sensing and Machine Learning
Authors: Sourav Bhadra, Saint Louis University; Vasit Sagan, Saint Louis University; Stephen
Patrick Moose, University of Illinois Urbana-Champaign.
Abstract: “Efficient utilization of Nitrogen fertilizer is an important feature for Maize in terms of sustainable crop management and breeding operations. This study explores the use of Unmanned Aerial Vehicle (UAV) based remote sensing techniques and machine learning to estimate Nitrogen Utilization Efficiency (NUE) of Maize non-destructively and rapidly. An experimental site was set up with 66 maize genotypes from which 368 sample plots were selected to manually measure NUE after harvesting. During the reproductive stage, hyperspectral data was collected using an UAV. After processing the hyperspectral data cubes, four commonly used machine learning regression algorithms (i.e., partial least squares, random forest, support vector machine and deep neural network) were developed to estimate NUE from plot-level hyperspectral reflectance spectra. Results suggest that the deep neural network outperformed the three other machine learning models (R2=0.83). Additionally, a NUE map was achieved from the deep neural network model to locate the genotypes with high/low NUE. The spatial distribution NUE can be very helpful to crop breeders and farmers for identifying more sustainable maize genotypes.”
Poster Link: https://drive.google.com/file/d/1FlfCdprCTLf16330LxJT9uPGMxHnNgmn/view?usp=sharing
Video Link: https://drive.google.com/file/d/1pyCyCCI-Qp0qbDZ7KxyCzBg7BoCtuttq/view?usp=sharing
Poster 21: Spatiotemporal Assessment of Harmful Algae Bloom in Illinois inland Waters
Authors: Supria Sarkar, Saint Louis University; Ruopu Li, Southern Illinois University Carbondale.
Abstract: “Harmful Algae Blooms (HABs) in inland water bodies (e.g., lakes and ponds) pose a serious threat to human health and the natural ecosystem. Therefore, it is imperative to assess HABs and their potential triggering factors over broader spatiotemporal scales in order to keep the natural ecosystem clean. This study explores the spatiotemporal pattern of HABs across the inland water bodies in Illinois. The assessment was conducted by using Chlorophyll-a (Chl-a) concentration in water samples collected from lakes in Illinois. The dataset was collected by Illinois Environmental Protection Agency (IEPA) from 1998 to 2018. Seven environmental and water quality zones were selected for spatial analyses. Additionally, the temporal patterns were assessed using time-series decomposition of monthly Chl-a concentration datasets. Results indicate that: a) the Collinsville region in southwestern part of Illinois exhibited higher mean concentration of Chl-a in its lakes than any other regions for the last two decades; and b) the lakes with increasing trends in monthly mean Chl-a concentrations were also focused in the southwestern region. The findings of this study will help policy makers to drive resources towards vulnerable lakes that are prone to HABs outbreaks.”
Poster Link: https://drive.google.com/file/d/1yWgsH74wUwzbxk1SF1qG5Famdmr__H1s/view?usp=sharing
Video Link: https://drive.google.com/file/d/11BVizrSIJE0AWwxPXS8QCNbBubzfZyu-/view?usp=sharing
Poster 22: A Novel Ground Truth Extraction Method to Identify Optimal Date/Band Combinations of Sentinel 2 Imagery for Use in a Machine Learning Environment for Land Use/Cover Assessment in Whiteside County, Illinois
Authors: Sara Chamberlin, Southern Illinois University Edwardsville; Stefanie Pryor, Southern
Illinois University Edwardsville; Clayton Pearson, Southern Illinois University Edwardsville.
Abstract: “Recently, there has been significant interest in assessing land use change across the US. Land use assessment is vital to understanding the inventory of natural resources, which are tied to the country’s economic and environmental viability/sustainability. At the same time, there has been a substantial change in satellite technology and mapping algorithms for characterizing annual land use/cover and monitoring temporal changes. This study developed a novel approach for generating ground truth data in Whiteside County, Illinois for the 2019 growing season. The ground truth data were used for training a supervised machine learning algorithm using Sentinel 2 data. The USDA’s Cropland Data Layer was used along with GIS analysis and remote sensing techniques to extract high confidence level pixels of corn, soybean, alfalfa, fallow, forest, and grass. Ranked Stepwise analysis and a Random Forest machine learning algorithm were used to determine image dates and spectral band combinations that are best suited for separating the different land use/covers. The Stepwise algorithm allowed for the identification and ranking of the most influential date/band combinations that optimized the Random Forest model’s accuracy. The results of this study indicated that only 10 of the band/dates were needed to achieve a classification accuracy over 97%. Nearly half of the date/band combinations were mid-to-late season middle infrared, with the others consisting of mid-to late-season red, red edge, and near infrared bands. This novel approach for ground truth extraction, coupled with evolving satellite technology and machine learning algorithms will enable higher quality land classifications in the future.”
Poster Link: https://drive.google.com/file/d/1mhuH-raNWEpRvN5fSHPJG7xx6k62zOHs/view?usp=sharing
Video Link: https://drive.google.com/file/d/18poZHVMlKY2NfugwkFu4gauSxcRyvb6G/view?usp=sharing
Poster 23: Predicting Terrorism in Europe with Remote Sensing, Machine Learning, and Spatial Statistics
Author: Caleb Buffa, Saint Louis University.
Abstract: “This study predicts terrorism in Europe at a previously unexplored spatial scale. Dependent variables consisted of satellite imagery and socio-environmental data. Five machine learning models were evaluated over a binary classification problem, the presence or absence of historical attacks within hexagonal-grid cells of 25 square kilometers. Four spatial statistics methods, Join Counts, Moran I’s, exploratory regression, and spatiotemporal correlation were conducted to assess the validity of results and improve our inferential understanding of spatial processes among terror attacks. This analysis resulted in a Random Forest model that is 99% accurate in predicting terrorism at a spatial resolution of approximately 5 kilometers. The results were validated by robust F1 and average precision scores, 96 and 97% respectively. Additional statistical analysis revealed the variance of nighttime lights and population density to be strongly and negatively correlated with the presence of terrorism per grid cell. This work concludes that remote sensing, machine learning, and spatial techniques are not only valuable, but important methods for providing deeper insights into terrorist activity and behavior."
Poster Link: https://arcg.is/9DS4L
Video Link: https://drive.google.com/file/d/1EinO0p07RMM5NNkEtBrFckanqfhFDJhW/view?usp=sharing
Poster 24: A Small Computer Cluster for Computation onboard an Unmanned Aerial System
Authors: Mukund Chidambaram, Lindbergh High School; Srikanth Gururajan, Saint Louis University.
Abstract: "This poster presents the design, capabilities, and future potential of a small computer cluster for use onboard an Unmanned Aerial System (UAS). This Raspberry Pi single board computer-based cluster is composed of a Raspberry Pi 3B+ controller and 4 Raspberry Pi Zero nodes connected via an interface board, named Cluster Hat. The cluster is designed with computing geospatial processes onboard an aircraft in mind. The nodes communicate with the controller via the Hat, while parallel processing is done via a software known as MPI4PY, a message passing interface which allows users to designate tasks to be done by specific nodes. Performance benchmark tests are currently being performed on the cluster and involve a prime number computation test. Additionally, the CPU temperature readings are also recorded every 0.2 seconds to establish the heat profile. Since the vast majority of geospatial processes are heavy on computation, this approach allows for more efficient task completion as the cluster handles jobs while in flight. When fully implemented, the cluster will be able to read data from a flight controller, and execute a trained neural network, built on Tensorflow and Keras, on said data to predict future values and designate faults in a sensor.”
Poster Link: https://drive.google.com/file/d/1yY3n88hBXjDbP0TxEQn8JM9wCxMyasqG/view?usp=sharing
Video Link: https://drive.google.com/file/d/19yCbgHrf8AteOHIVsujzQ6lSZFgppDM_/view?usp=sharing
Poster 25: Using Geostatistical Analysis and Machine Learning to Improve Tropospheric Correction for InSAR
Authors: Ngo Hi Kenny Yue, Missouri University of Science and Technology; Jeremy Maurer,
Missouri University of Science and Technology.
Abstract: "Interferometric synthetic aperture radar (InSAR) is a powerful tool to detect deformation and movement of the Earth’s surface. Many applications are using InSAR to study earthquakes, subsidence, and surface deformation on Earth. However, the tropospheric delay results in significant contamination when measuring small deformation signals with large spatial and temporal scales. Because of this issue, many recent studies have investigated the use of various corrections for tropospheric noise in InSAR. These can be broken into approaches using models, such as topography-correlated delay and some auxiliary data. An example of the latter approach includes using delays available from Global Navigation Satellite Systems (GNSS), interpolating to InSAR pixels to estimate, and removing the delay by computing or removing delays using global atmospheric models (GAMs). These both require interpolation, either from GNSS stations or GAM model nodes to the InSAR pixels. To date, studies have used linear interpolation and kriging or Gaussian regression. However, these methods have failed to accurately remove tropospheric noise from interferograms at short wavelengths, which are most difficult for global weather models and most important for identifying target localized processes. This study investigates using statistical analysis and machine learning approaches to improve the tropospheric corrections based on GAMs. We develop spatial correlation models for weather model parameters (Temperature, Pressure, and Water vapour) and interferogram delays in regions without tectonic signals. We use 100+ Sentinel-1 interferograms available from the ARIA project during 2019-2020 in the northeastern US as training images and determine spatial correlation characteristics. This method is to avoid having interferograms with non-tropospheric signals. We explore the use of spatial correlation and Deep Neural Networks (DNNs) to improve the interpolation scheme and downscale the GAMs to InSAR resolution."
Poster Link: https://drive.google.com/file/d/14QxZ5fRTCj4IbdZmFxX0wShASaZ35sHV/view?usp=sharing
Video Link: https://drive.google.com/file/d/1wfm9cDwWKeOXg4_fkfFunUq02U7_PDHM/view?usp=sharing
Poster 26: Deer Ridge Park Map
Author: Gianna Mitchell, Lindenwood University.
Abstract: "The research granted by Lake St. Louis, where they were in need for a nature trailhead map. This opportunity allowed the knowledge that had been acquired through numerous Geographic information classes and applied them in order to create this needed map. The opportunity presented many learning opportunities and problem solving as we were the first ones from our university to complete a GIS internship during a pandemic.”
Poster Link: https://storymaps.arcgis.com/stories/c88ad5fbf816420c879838b2ac475bb5
Poster 27: We Have the Watch, They Have the Time
Author: Caleb Buffa, Saint Louis University.
Abstract: “Terrorism poses a threat to societies in a manner that can be described by a derivation of an old adage, "States have the watch, terrorists have the time". Preventative and predictive tools become imperative in a field in which time is on the terrorists' side, and they only need to get a plan correct once while states must get it right every time. To that end, much research has focused on the political and sociological perspectives to understand how and why terrorism occurs. However, less has been done to understand terrorism's geographic component at scales larger than regional levels.”
Poster Link: https://arcg.is/9DS4L
Poster 28: CWS: A Geospatial Tool for Mapping Water SystemsAuthors: Andy Florian Irakoze, Saint Louis University; Flavio Esposito, Saint Louis University; Lee Voth-Gaeddert; Andrew Schranck. Abstract: "In low- and middle-income countries, providing access to clean water is critical as water, sanitation, and hygiene (WASH) have direct connections to disease transmission. In rural settings specifically, water is primarily delivered by community water systems (CWSs) which may service households, schools, and health clinics via a piped network run by gravity or water pumps. However, millions of CWSs across the world function sub-optimally due to insufficient data on operating conditions. CWSMap, an Information and Communications Technology (ICT) app, is being developed and prototyped with a user-centered, design approach to empower local communities to improve the management of their water resources in a way that requires minimal technical training and is easily transferable to other users. CWSMap provides users a simplified method to map and analyze their rural water system using a smartphone or tablet. Targeted users for this technology include rural water technicians, water engineers, and even untrained personnel such as volunteer water committees or community leaders. In higher income settings, advanced technologies are deployed by highly trained engineers to develop system maps and analyses. Water system mapping is a multi-million dollar industry currently accessible only to wealthier water utilities, however, internet-connected, handheld devices are often accessible in low- and middle-income countries, presenting a prime opportunity to leverage ICT for CWS improvements." Poster Link: https://drive.google.com/file/d/1xq6gawQ6T49mQHjP-wV2ZLycda7Q1P-S/view?usp=sharing Video Link: https://drive.google.com/file/d/1u_J-2kZQ7nHbvtwPK3with1zFSM0Plwo/view?usp=sharing
Poster 29: Don’t Go There: A Geospatial mHealth App for Gambling DisorderAuthors: Roberto Coral, Saint Louis University; Andy Florian Irakoze, Saint Louis University; Flavio Esposito, Saint Louis University; Jeremiah Weinstock, Saint Louis University. Abstract: “Current efforts to address and alleviate the suffering and harms associated with gambling disorder (GD) predominantly involve in-person psychotherapy or static web-based interventions, while dynamic, proactive approaches to the treatment of GD are needed. Currently, several medical health (mHealth) apps for GD exist, but none are empirically validated and, even though they could be potentially useful, these interventions are static, reactive approaches and they are not based on the geo-fencing principle. Advances in smartphone technology now allow for a paradigm shift. Just-in-time adaptive interventions (JITAI) are technology-based dynamic interventions that proactively respond to time-varying information to provide intervention at critical moments. In this paper, a novel JITAI mHealth app for GD called “Don't Go There” (DGT) is presented. It capitalizes on smartphones' global positioning software (GPS) or other zero-permission embedded sensors to recognizes a user's location. When the patient turns off the GPS capabilities, we infer the geolocation solving a probabilistic route matching problem. The purpose of DGT is to construct a geofence around a gambler's favored gambling establishment, to discourage participation. We hypothesize that the deployment of this app could lead to a reduction in gambling behavior and problems and improve psychological functioning.”
Poster Link: https://drive.google.com/file/d/1RzWrNyHI5ExnAtP9XM47XY_Uoc51xKd3/view?usp=sharing