UWEC CERCA 2025

Variable Rate Agriculture (VRA) is a data-driven approach aimed at reducing the environmental impact of commercial farming. It leverages machine learning models (MLMs) to enhance crop yield predictions more rapidly than traditional soil analyses. However, MLMs require large datasets, and agricultural data is often limited. Cross-validation (CV) techniques help improve model generalization by testing model performance on reserved subsets of data, even with limited data. This study used a three-year dataset on hybrid wheat, covering pre-planting, crop growth, and yield mass from Minnesota. Four machine learning models—linear regression, random forest, XGBoost, and feed-forward neural networks (FFNN)—were developed to link pre-growing conditions with yield outcomes. Two CV methods, Random CV and Spatial Grid CV, were applied to compare model performance, assessing overfitting using the coefficient of determination (R²) and Root Mean Squared Error (RMSE). Feature selection was performed to pinpoint critical spectral indices impacting model output. Findings indicated that Random CV generally outperformed Spatial Grid CV across both full and reduced feature sets. While linear regression suffered from feature selection limitations, FFNN showed occasional improvement. Overall, Random CV proved more effective, especially with a diverse dataset, enhancing model reliability in VRA applications.

Transportation is a fundamental human activity, and cycling offers significant potential to improve community health, reduce carbon emissions, and lower transportation costs both for individuals and the community at large. To foster widespread adoption, sufficient cycling infrastructure is essential, as safety issues present a major barrier to ridership. Understanding cyclists' perceptions of safety, location, and the built environment is crucial for enhancing safety and increasing participation. This poster presents findings from a study on spatial perceptions of cycling safety in Marburg, Germany. Through a web-based survey, respondents identified two major clusters of unsafe areas. In general, unsafe locations were identified as such due to inadequate infrastructure and potential conflicts with motor vehicles. The study reveals that while women and men share similar levels of confidence in cycling, safety concerns have a greater influence on where women choose to ride. The insights from this study are valuable for identifying specific areas in Marburg that require improvement, but they also offer guidance for urban planners seeking to design safer and more equitable transportation systems.

Our research is focused on the geomorphic history of the Niobrara River, located in northern Nebraska. This river exhibits noteworthy landforms, including a large number of terraces. The goal is to determine what event(s) may have taken place to cause these formations. Through analyzing the terraces, it’s apparent that it has drastically aggraded and degraded to form a complex series of terraces. The Niobrara is located just north of the Nebraska Sand Hills, which is a crucial aspect of the landscape that contributes to the river’s geologic history. To understand what happened, we mapped all the terraces within the Niobrara Scenic River Corridor. We utilized ArcGIS Pro to digitize each terrace, then calculated their average height above the river. Once completed, we classified the terrace heights into categories to reveal geographic patterns. We hypothesize that there was a river piracy event that caused the Niobrara to reroute towards the Keya Paha River, a tributary to the Niobrara, and abandon the ancestral course that is now the Elkhorn River. A possible explanation includes the transport of sediment from the Sand Hills to create a sand dam on the river, which would have formed an overflowing lake, spilling over the dam boundaries.

Two Holocaust-era trenches sit between the Abava River and a local road near Sabile, Latvia. The mass murder of Sabile Jews took place on August 6,1941. The intended execution of Sabile’s Roma population never occurred leaving an open trench. The goal of the project is to locate the filled in trench containing 240 Jewish citizens and compare it to the open Roma trench. To find the Jewish trench, an 8x9m grid was laid down and ground penetrating radar (GPR) lines were collected every 0.25m using a Sensors & Software pulseEKKO Pro system with 500MHz antennae. The GPR data was then geometrically corrected using a Topcon laser leveler. Results show horizontal, to sub-horizontal semi-continuous reflection patterns with erosional truncations from 2-4m across the grid. These truncations were filled with concave upward reflection patterns. In comparison to the open Roma trench, it is hypothesized that the Jewish trench lays partially within the imaged lines and is roughly 2-4m wide and 5m deep. The study, for the first time, has been able to compare two Holocaust trenches (one open and one filled in). The results can be used by others to compare mass execution sites worldwide.

During the early stages of the Holocaust in 1941, over 5,470 people were shot by bullets within the city limits of Liepāja, Latvia. Near the Liepāja Lighthouse, a film recorded by Reinhard Wiener captures the scene of active killings while a 1945 Soviet Commissions Map illustrates four execution trenches. The map and film present geographic reference points that suggest trench locations near the Lighthouse. The research aims to build upon prior research to map the continuation of execution trenches. The primary method, ground penetrating radar (GPR), a non-invasive technique that transmits high-frequency electromagnetic pulses into the subsurface. Seven grids were laid out using survey tapes with a 0.25m line spacing and ranging from 4.5-29m in length and 12-33m in width. All seven grids used an antennae frequency of 500MHz with an odometer wheel triggered step size of 0.02m. The data was processed using Sensors and Software pulseEKKO Pro V6 software. Pairs of erosional truncation reflections are interpreted as trench outlines within local subhorizontal semicontinuous coastal aeolian sediments. Two possible trenches extend in grids 1 and 5. The project builds an understanding of how to detect mass burial sites worldwide and help the Liepāja Jewish community identify memorial locations for execution trenches.

The Jewish Cemetery in Vikija, Lithuania has been abandoned since the Nazi and Soviet occupation of Lithuania. The Karnofsky family were a prominent Jewish family who moved to New Orleans in 1900 where they met Louis Armstrong and his family. Honoring this relationship inspired the restoration of the cemetery. Geographic research was conducted using ground penetrating radar (GPR) to image disturbances in the subsurface to locate cemetery features. GPR transmits electromagnetic pulses into the ground and reflects off changes in the natural sediments and anthropogenic features and data is displayed on a digital video logger (DVL). A pulseEKKO Pro 500 MHz antennae GPR unit with 0.25m spacing between lines, and a Topcon RL-H4C self-leveling laser and receiver are used to geometrically correct the GPR data. Between the road and the grassy edge of the cemetery three grids and one line were recorded, Grid 1 was 13.5x22m , Grid 2 was 5x12m, Grid 3 was 10x10m, and a 34m line. The reflections are interpreted to be boundaries of a mourning room and/or ritual purification building. Research on this site will continue to locate boundaries to allow for restoration of the site.

Ten kilometers outside Riga, a city located on the northern coast of Latvia, stands a public park that may seem mundane, but beneath the surface lies a dark past. Today the site is known as Mazjumpramuizha Park, but it was once the location of the Jungfernhof concentration camp during WWII. while there is one interpretive sign that acknowledges this history, it does not tell the full story and even locals remain unaware of what happened here. This silenced history is the focus of my research project, which aims to locate the prisoner barracks that once stood on the site and have been described by survivor testimonies. To accomplish this, I worked alongside a team commissioned by The Jews of Latvia Museum, utilizing ground penetrating radar— a non-invasive scanning technology that transmits electromagnetic pulses into the subsurface—to reveal what lies beneath. In conjunction with a Topcon-RL-H4C self-leveling laser to determine ground elevation, our team unearthed evidence of the camp’s barracks, including barbwire and the building’s rock foundation. For survivors, our findings provide evidence to help bring closure and launch memorial efforts via the creation of a steel monument inscribed with the victims’ names known as the Locker of Memory.

Rural residents have lower levels of access to health care and health services, and health insurance coverage rates are lower as well; as a result, the health of rural residents is poorer compared to their urban counterparts. The lack of healthcare providers, longer distances to healthcare facilities, and lower economic status for rural populations place extra burdens on both providers and patients to meet and receive needed care. The COVID-19 pandemic not only highlighted these disparities but also spurred new forms of care delivery. Telehealth, through synchronous and asynchronous remote appointments, provided healthcare providers with opportunities to connect with patients outside of physical office visits. The rise of Telehealth not only increased options for patients but also protected all parties from the spread of the COVID-19 (Hirko et al. 2020). Yet, throughout much of the United States rural broadband options are limited or nonexistent.This project is an extension of a previous collaboration with the Mayo Clinic (please see “A Spatial Analysis of Cellular Singal Strength in Western Wisconsin”). It has been determined that a strength of less than -115 dBm will not produce a viewable telehealth appointment. This poster will discuss one of the outcomes of this project (1) kriging analyses and a probability map of cellar strength for Verizon and AT&T near Menomonie, WI. Such analysis will provide Mayo Clinic with a better understanding of where telehealth opportunities are possible in rural areas for patients who lack internet access.

Dunn County Potter’s Field contains over 100 unmarked graves of persons from the County Poor House and the County Asylum (both now demolished) in Menomonie, Wisconsin. Community interest in locating the graves prompted non-invasive scans of Potter’s Field using ground penetrating radar (GPR). GPR transmits electromagnetic (EM) pulses into the ground, which reflect off subsurface changes before being received and recorded digitally. A 16x50m grid was collected along the site’s western boundary with a Sensors and Software pulseEKKO Pro GPR unit with 500MHz antennae. GPR traces were triggered every 0.02m via odometer wheel, and lines were collected with a spacing of 0.25m. Alternating strong and weak reflections are repeated in a sub-parallel, semicontinuous pattern throughout the grid. The strong reflections are often underlain by steep-angled hyperbolics. The reflections were identified 0.6–1.5m below the surface, each measuring 1.0m by 2.25m, and spaced 1.0m apart. Reflections were interpreted as two probable rows of individual graves. In Fall 2024, 42 crosses were placed at each probable grave, however further scanning is recommended to locate remaining lost graves. The GPR work at Dunn County Potter’s Field is a model for other cemeteries; providing means of locating and restoring honor upon unmarked graves worldwide.

During the Holocaust from the 15-17 of December 1941, ~2,749 Jewish women and children were massacred behind a foredune along the coastline of Šķēde, Latvia. The execution trench has not been located, despite numerous photos of the killings. The research aims to identify the location of the trench using shallow, non-invasive subsurface imaging tools to respect Jewish Halacha burial laws, avoiding disruption of the deceased. Based on earlier research including current and past aerial imagery, witness testimonies, and historical photos of the killings, a 15x30m (NW-SE) GPR grid was laid down within the foredune. Using a Sensors & Software pulseEKKO Pro GPR system, lines were collected every 0.25m with 500MHz antennae and a step size of 0.02m, which allowed subsurface depths of ~4m (0.10m/ns velocity). The subsurface reflections dip at a 30-34° angle, with erosional truncation occurring at ~17-21m and extending ~3.1m below the surface; these dipping layers are interpreted as a probable trench ~5m wide throughout the grid. Based on witness testimonies, the trench is hypothesized to be ~100m long (S-N). The findings have been shared with the local Jewish community for memorialization, and the research provides a framework for identifying unmarked graves worldwide.