Presentation Abstracts – Turfgrass Trends for Changing Times

Dr. Daniel O’Brien, Dr. Mike Kenna, and Dr. Jordan Booth – United States Golf Association Green Section Research

“Emerging Technologies for Sustainable Turfgrass Management “

John Hurst, The Toro Company, Sr. Director Center for Technology, Research, and Innovation and Robotics

• “Pesticide Use in Turfgrass Systems: Balancing Efficacy, Safety, and Public Perception” – Dr. Travis Gannon, Professor, Pesticide Fate in Turfgrass Systems

Maintaining high-quality turfgrass requires consistent control of weeds, insects, and diseases, yet turf managers must achieve this under increasing labor, budgetary, and time constraints. Reliable pesticide performance depends on selecting the appropriate product, rate, timing, and placement—decisions informed by pest biology, site conditions, equipment calibration, and environmental factors. These decisions are further complicated by the rise of herbicide resistance. At the same time, limited development of new active ingredients and lengthy, expensive registration and reregistration processes—often exceeding a decade—restrict the availability of novel tools, underscoring the importance of preserving existing chemistries. Beyond management and regulatory challenges, public perceptions also influence pesticide use. Conversations around pesticide safety frequently center on emotionally charged terms such as “toxic,” “cancer,” or “water contamination,” which may obscure scientific context and blur the distinction between hazard and exposure. Clear communication about why pesticides are used and how risks are minimized through PPE, timing, placement, and strict label adherence is essential for maintaining public trust and for highlighting the environmental safeguards already in place. Sustainable pesticide use in turfgrass systems ultimately depends on responsible management practices. Emphasizing careful application, environmental stewardship, resistance management, and thorough record-keeping helps ensure that pesticides remain safe and effective tools. By adopting informed management strategies supported by training, professional development, and research-based guidance, turf managers are better equipped to maintain turf quality, meet regulatory expectations, and communicate effectively with stakeholders. This integrated approach aligns efficacy, safety, and public confidence, supporting long-term success amid growing scrutiny and evolving challenges.

• “Pixels, Patterns, and Perception: Automating Turfgrass Evaluations with UAVs and Artificial Intelligence” – Rob Austin, Research & Extension Specialist, Turfgrass Phenomics

The evaluation of turfgrass is a cornerstone of the turfgrass industry, serving a diverse range of stakeholders. Breeders rely on turf quality assessments to select and advance cultivars and experimental lines, while seed companies use these evaluations to promote and market their products. Similarly, recreation and parks departments reference turf quality data to support procurement decisions, and both landscape professionals and homeowners consider such information when purchasing seed or sod. Given its widespread impact, it is paramount that turfgrass evaluation methods are repeatable, consistent, and standardized across the industry. However, turfgrass evaluation poses significant challenges due to its inherently subjective nature. Aesthetic and functional attributes, including color, density, leaf texture, uniformity, and overall quality, are typically assessed using visual ratings and are subject to the judgment of the evaluator. The National Turfgrass Evaluation Program (NTEP) has established widely accepted protocols for these assessments, and its visual rating system is the prevailing standard in the United States and internationally. Nevertheless, the reliance on human observation introduces potential for bias and inconsistency. Studies have demonstrated that even trained evaluators may produce variable results, making comparative analysis difficult. However, recent technological advances in the fields of Unmanned Aerial Vehicles (UAVs) and Artificial Intelligence (AI) are providing new opportunities for more objective turfgrass evaluation. This presentation explores the use of UAV-acquired imagery in predicting turfgrass quality ratings and shares findings that highlight the potential of integrating remote sensing and artificial intelligence into standard turfgrass assessment practices.

• “Improved Resilience and Increased Sustainability: 15 Years of Sustained NIFA funding in Turfgrass Breeding” – Dr. Susana Milla-Lewis, Professor, Turfgrass Breeding and Genetics

• “Measuring Organic Matter at The Speed of Light” – Payton Perkinson, PhD student, Turfgrass Management

There is great interest in managing organic matter in turfgrass, particularly on golf course putting greens. Traditional analysis techniques have required extensive laboratory preparation of samples, resulting in a long interval before results are known. The objective of this study was to determine if a portable near-infrared (NIR) spectrometer could accurately predict organic matter on sand-based root zones with minimal sample preparation. One hundred 2.5-cm diameter soil cores were taken from sand-based putting greens containing varying levels of organic matter. Each core was separated into 3 depths: 0-2 cm, 2-4 cm, and 4-6 cm, and verdure was not removed. Each sample was processed and scanned in a sequential manner: (1) samples collected from the field and immediately ground with a mortar and pestle; (2) samples dried at 105°C for 48 hours then scanned again; (3) samples ground finer using a ball mill and scanned for a final time. Each sample was scanned three times for each of the three preparation methods. After scanning was completed, samples were subjected to loss on ignition (LOI) at 440°C for two hours. Each scan was assigned its respective LOI values, and a model was developed for each preparation method. The ball milled samples had the highest prediction accuracy, followed by the dried and ground with a mortar and pestle samples, and the fresh ground samples had the lowest prediction accuracy. However, fresh ground samples had relatively successful prediction accuracy, suggesting the analysis technique could be used for rapid analysis of organic matter content on sand-based putting greens.

• “Exploring genetic diversity for seed-related traits in zoysiagrass” – Balihar Kaur, PhD student, Turfgrass Breeding and Genetics

Zoysiagrass is one of the most economically important turfgrass species in the United States due to its moderate drought and shade tolerances, tolerance to weed encroachment, and reduced maintenance needs. Seeded cultivars are preferred by the industry because of their economic as well as storage and transport advantages. However, traits such as seedhead density, seed set and uniform seed maturity have not been fully elucidated in the species, and problems like poor seed fill and high levels of seed dormancy pose a challenge in the successful establishment of zoysiagrass stands. The objective of this study is to investigate genetic variation in seed-related traits among a collection of 268 diverse zoysiagrass accessions. In addition, genome-wide association studies will be conducted to uncover genomic regions/candidate genes controlling these traits of interest. Data on various reproductive traits like flowering time, inflorescence abundance, seedhead color, seedhead morphology, seed morphology, total seed weight, overall seed yield, percent seed fill and germination potential is being collected from field trials in North Carolina and Texas. Preliminary results from year 1 revealed differences in seedhead color, staggered flowering times,  variability in seedhead abundance from sparse to prolific, and wide range in seedhead length. Additionally, germination assays using multiple dormancy-breaking treatments identified genotype and treatment specific differences in germination behavior. These results highlight the potential for selecting high-yielding and low-dormancy genotypes to enhance the development seeded zoysiagrass cultivars.

• “Behavioral Differences and Management Responses of Root-Knot Nematodes Affecting Turfgrasses in North Carolina” – Jack Mascarenhas, PhD student, Turfgrass Pathology

While historically plant-parasitic nematodes were overlooked as a major threat in turfgrass, it has recently become an increasing concern in turfgrass management. Root-knot nematodes (Meloidogyne spp.) in particular have gained attention in green pest management, as they rank among the top 10 most diagnosed pests in bermudagrass and bentgrass samples received by the Turf Diagnostic Lab. Root-knot nematodes are a common pest in multiple cropping systems but can often be overlooked in turf due to the smaller gall size on turf roots compared to agronomic crops. However, root-knot nematode feeding on turf results in compromised root systems that are more susceptible to secondary pathogens and less efficient at water and nutrient uptake. To date, there are nine species of root-knot nematode known to parasitize turfgrass. In North Carolina, M. graminis, M. incognita, M. marylandi, and M. naasi are the predominant root-knot nematode species. However, little is known about the biological differences between root-knot nematode species or their management responses in turfgrass. There are also limited molecular tools tailored to identify different species of root-knot nematode in turfgrass. Previous in vitro studies have found that nematicide sensitivity can vary among different Meloidogyne spp., so further insight into the behavioral and sensitivity differences among root-knot nematodes could help optimize management strategies in turfgrass. The objectives of this research are to identify root-knot nematode (RKN) species present in turfgrass samples submitted to the NC Turfgrass Diagnostics Lab over a three-year survey using J2 extractions and species-specific primers, assess juvenile sensitivity to key nematicides through in vitro assays across a range of concentrations, and determine population thresholds and temperature-dependent infection timing by inoculating creeping bentgrass and ultradwarf bermudagrass with varying RKN densities. Additionally, field studies at golf course sites with high RKN pressure will evaluate the efficacy and optimal timing of nematicide programs, including multiple active ingredients and application intervals, to enhance management strategies for RKN in turfgrass systems. Through this research, we aim to gain a clearer understanding of the biology of Meloidogyne spp. affecting turf and to determine if species-specific differences should be considered when formulating management strategies.

• “Data-driven fertilization management of ultradwarf hybrid bermudagrass putting greens” – Nirmal Timilsina, PhD student, Turfgrass Management

Precision fertilization management requires an accurate understanding of putting greens growth rate for achieving sustainable fertilizer use. This study uses a machine learning framework to build a growth rate prediction model for short-term clipping yield of ultradwarf bermudagrass. Field experiments were conducted in 2023 and 2024 on two sand-based putting greens planted with ‘G12’ and ‘Champion’ bermudagrass at North Ridge Country Club and the NC State Turfgrass Field Lab, respectively, in Raleigh, NC. On the ‘G12’ bermudagrass green, the combined treatments of four nitrogen rates (0, 4.9, 9.7, and 19.5 kg/ha every other week], two Trinexapac-ethyl rates (TE, 0 and 0.034 kg/ha every 170 growing degree days), and two mowing heights (low and high) were applied. On the ‘Champion’ bermudagrass green, only the same four N rates were evaluated. Foot traffic was applied weekly at 1000 rounds with golf shoes. Clippings, vegetative indices [normalized difference red edge index (NDRE), normalized difference vegetation index (NDVI)], and soil volumetric moisture content at 7.6 cm depth were collected at least three times per week on both greens. To predict clipping yield, machine learning models were built in Python (Scikit-learn package) using accessible predictor variables such as daily weather, management practices (soil water content, historical N and TE applications, and foot traffic), vegetation indices (NDVI and NDRE), and post-application time (days after N application and growing degree days after TE application). Extreme gradient boosting (XGBoost) performed the best and accurately predicted daily ultradwarf bermudagrass clipping yield on both greens, with R² = 0.89 for the ‘G12’ green and R² = 0.88 for the ‘Champion’ green. The machine-learning-based growth rate prediction model developed in this study could help turfgrass managers allocate resources and prioritize management practices more effectively.

• “Optimizing preemergence herbicides in turfgrass by understanding influential factors” – Robby Andrews, MS student, Pesticide Fate

Effective preemergence (PRE) herbicide programs in turfgrass systems require the alignment of many influential factors, such as weed emergence, application timing, effective dose required, and herbicide persistence. To refine PRE programs for annual bluegrass, crabgrass, and goosegrass using prodiamine, indaziflam, and simazine, four complementary studies were developed, each targeting a key component of PRE performance. First, ongoing multi-site field experiments are characterizing seedling emergence patterns under bare ground and turf cover while continuously monitoring soil temperatures and moisture. Second, field studies are quantifying PRE persistence behaviour in spring and fall seasons as influenced by early versus late application timings, and increased soil moisture via a simulated saturated period. Preliminary residues quantified by HPLC-DAD-MS and fitted to first-order kinetic models to estimate field half-lives in the 0-2 inch soil layer reveal average persistence across seasons under normal field conditions to be 44, 49, and 35 days for prodiamine, indaziflam, and simazine, respectively. Third, concurrently and following the same design as persistence experiments, efficacy studies are being conducted in areas with an established history of annual bluegrass, crabgrass, and goosegrass to evaluate the effects and implications of application timing, soil moisture, and PRE herbicide on resulting control. Efficacy evaluations of annual bluegrass and crabgrass reveal prodiamine as a strong control choice, as well as indaziflam although more strongly influenced by application timing. Simazine remains a strong short-term option, benefited by both its PRE and POST control characteristics. Fourth, greenhouse dose-response bioassays are used to determine critical doses (e.g. EC50, EC90) required for each herbicide-weed combination. Collectively, these studies will integrate weed emergence patterns, soil residue persistence, and dose requirements to define application windows that maintain effective concentrations in the critical weed seed germination zone while minimizing repeat applications, thereby improving the reliability of existing PRE tools and ensuring more consistent, predictable herbicide performance.

• “Automated Seedhead Counting in Zoysiagrass: Early Insights from AI-Based Image Analysis” – Stefano Fratton, PhD student, Turfgrass Breeding and Genetics

One of the biggest bottlenecks in breeding for improved seed yields in zoysiagrass is the accurate evaluation of seedhead abundance. Seedheads are extremely small and variable in morphology, density and maturity, which makes them difficult to measure. As manual counting is slow, subjective, and impractical at scale, breeders are forced to rely on simple categorical scores instead of true quantitative data. To address these challenges, we aim to develop a high-throughput phenotyping (HTP) pipeline that incorporates artificial intelligence (AI) and image analysis to generate fast, objective, and scalable evaluation of seedhead abundance. Because seedheads are easily lost in background noise, a preliminary pilot study was conducted in a controlled greenhouse environment with fixed lighting and uniform backgrounds. RGB images captured from multiple angles were used to train a YOLO object-detection model capable of identifying and counting seedheads. Under these controlled conditions, automated counts aligned well with manual counts (R² ≈ 0.8–0.9), with further improvements in accuracy when images were cropped and multi-view inputs were used. These results demonstrate that AI-based phenotyping can provide consistent, quantitative measurements of seedheads, enabling the evaluation of larger populations more efficiently and supporting breeding efforts to improve seed yield in zoysiagrass.

• “Targeted Weed Control in Turfgrass Using Precision Spot Spraying Technology” – Dr. Navdeep Godara, Assistant Professor, Turfgrass Weed Science

Ecorobotix’s ARA research sprayer utilizes artificial intelligence and machine learning to distinguish problematic weeds from desirable turfgrass. The research sprayer can target weeds at 1.2-inch spray resolution and is expected to be commercially launched in 2026. However, additional research is needed to quantify herbicide savings, and to evaluate weed control efficacy and turfgrass safety when herbicides are applied via targeted versus broadcast treatment. A field experiment was conducted in fall of 2025 to evaluate false-green kyllinga control, bermudagrass safety, and spray volume reduction of herbicide admixtures using targeted versus broadcast applications with ARA spraying system. The study was arranged as a two-factor factorial (application method × herbicide) in randomized complete block design with four replications. Treatments included nontreated control, glyphosate (415 g ai ha-1), pelargonic acid (16 kg ha-1), pyrimisulfan (68 g ha-1), a tank mix of sulfentrazone (280 g ha-1) and pyridate (350 g ha-1), a tank mix of bentazon (1066 g ha-1), halosulfuron (70 g ha-1), and sulfentrazone (190 g ha-1). Application methods include broadcast or targeted herbicide treatment using ARA research sprayer, calibrated to deliver 400 L ha-1 of spray volume at 4.5 km ha-1. At trial initiation, false-green kyllinga cover ranged from 14 to 19% in bermudagrass, making it ideal site for evaluating targeted applications. Spray volume delivered was dependent on the application method (P <0.0001) but was not affected by herbicide treatment or its interaction with application method. Targeted applications reduced the spray volume by 77% compared to the broadcast method. Except for pelargonic acid, all herbicides controlled false-green kyllinga more than 87% at 8 weeks after treatment (WAT), regardless of application method. Bermudagrass cover at 4 WAT was influenced by the herbicide × application method interaction (P = 0.0029). Broadcast application of glyphosate reduced turf cover to 43%, but bermudagrass cover was greater than 73% following other treatments at 4 WAT. Targeted applications of glyphosate minimized collateral damage to bermudagrass compared to the broadcast treatment. False-green control in bermudagrass was herbicide-dependent and unaffected by application method; however, targeted applications reduced the risk of turfgrass injury from nonselective herbicides. Based on other research trials conducted on dallisgrass, smooth crabgrass, and broadleaf weeds, targeted applications reduced spray volume by ~53% to 95% compared to the broadcast method, which is primarily driven by weed density. Future research will focus on identifying weed density thresholds for adopting targeted spraying technology.

• “Developing and using genomic tools for St Augustinegrass” – Dr. Joe Gage, Assistant Professor, Crop Genetics and Genomics

• “Verification or disproving the empirical data for sand replacement post aerification in USGA sand based putting greens” – Dr. Roch Gaussoin, Professor Emeritus, University of Nebraska-Lincoln

• “Detection and distribution of single nucleotide polymorphisms associated with SDHI-insensitivity in Clarireedia spp.” – Dylan Dean, PhD student, Turfgrass Pathology

Dollar spot is an important disease of both warm and cool-season turfgrasses, and it is associated with multiple causal species within the fungal genus Clarireedia (formerly Sclerotinia homoeocarpa). Various fungicides control Clarireedia spp., but this historical usage imposed selection pressures for fungicide-insensitive populations to develop. Boscalid was the first modern succinate dehydrogenase inhibitor (SDHI) used specifically to control Clarireedia spp.; however, boscalid-insensitive populations of Clarireedia spp. were documented by the late 2010s. SDHI insensitivity is thought to be caused by various single nucleotide polymorphisms (SNPs) within the SdhB, SdhC, and SdhD genes. The current distribution of SDHI-insensitive populations of Clarireedia spp is unknown, and this study aims to identify the extent of SNPs associated with SDHI resistance in a large population of Clarireedia spp. To accomplish this, DNA was extracted from Clarireedia isolates collected from various states, and PCR was performed using primers to amplify SdhB, SdhC, and SdhD. The amplified PCR products were then Sanger sequenced to determine the presence of SNPs. Our data documented multiple mutations across the Sdh genes for these isolates. The presence of these mutations in the sequenced population of dollar spot isolates highlights the importance of regular SNP testing so that growers and superintendents can be offered better management guidance.

• “Demonstrating the Feasibility of Genomic Prediction in St. Augustinegrass” – Jaswinder Kaur, PhD student, Turfgrass Genomics

St. Augustinegrass [Stenotaphrum secundatum (Walt.) Kuntze] is a widely used warm-season turfgrass valued for its shade tolerance, dense canopy, and vigorous growth. Traditional breeding relies on evaluating nurseries of about 1,000 genotypes across three diverse locations in North Carolina, but only the top 50 to 70 advances to multi-location replicated trials, meaning that less than 10 percent of the initial candidates ever progress beyond the first stage. The early nursery evaluations are the most resource-intensive, yet the large population size restricts the ability to measure important agronomic traits with sufficient detail and accuracy. Genomic prediction offers a way to streamline this bottleneck by predicting nursery line performance from genetic data alone, enabling a substantial reduction in the number of genotypes planted in early nursery trials. By eliminating a large proportion of low-performing lines before they ever reach the field, breeders can allocate space and resources to fewer, higher-potential genotypes, allowing for larger plots, more detailed data collection, and ultimately better selections. In this study, we assessed the feasibility of genomic selection using 155 diverse St. Augustinegrass accessions genotyped using whole-genome sequencing and evaluated for six traits: turfgrass quality, winterkill, leaf length, leaf width, internode length, and internode width. Genomic Best Linear Unbiased Prediction (GBLUP) models were used to estimate genomic breeding values for each trait. Tenfold cross-validation indicated the highest predictive ability for winterkill (0.74), while moderate accuracies were observed for the other traits, ranging from 0.29 to 0.43. These findings suggest that genomic prediction is particularly valuable for early-stage nursery selection, which represents the most resource-demanding phase of cultivar development in St. Augustinegrass. Overall, the results demonstrate the feasibility of incorporating genomic selection into St. Augustinegrass breeding programs and lay the groundwork for improving prediction models and expanding their use in future breeding efforts.

• “Rethinking Annual Bluegrass Weevil Monitoring: Enhancing Detection with Computer Vision and Drones” – Gram Grant, MS student, Turfgrass Entomology

The annual bluegrass weevil (Listronotus maculicollis Kirby; ABW) is a damaging insect pest of annual bluegrass (Poa annua) and creeping bentgrass (Agrostis stolonifera) in western North Carolina. Current monitoring strategies sample adult populations through the weekly application of soap flushes to fixed areas. This research sought to improve ABW monitoring by increasing sampling area and determining the spatial distribution of populations. An unmanned aerial system was used to collect near-infrared images of ABW on the turfgrass surface. These images were used to train a YOLOv8s-p2 computer vision model for ABW detection, automating counts of ABW from each image. The trained model achieved precision and recall values of t 0.9770 and 0.9620, indicating high detection accuracy. Following the development of a trained model ten golf course fairways were imaged, capturing photos every 4.5 meters. Images were processed using the trained computer vision model to obtain spatially referenced counts of adult ABW. Heatmaps were generated for each fairway, allowing for the spatial distribution of AWB populations to be visualized. These results demonstrate the potential of unmanned aerial system scouting and computer vision models to monitor surface-active insect pest populations across large areas.

• “Exploring alternatives to quartz-based sands in golf course management” – David Toscano, MS student, Turfgrass Management

Quartz-based topdressing sands are essential for maintaining firm, fast, and smooth putting surfaces on golf greens, yet their finite availability, rising costs, and high carbon footprint from mining present multiple sustainability challenges. The project aims to evaluate the potential of basalt as an alternative topdressing material. Basalt offers several potential advantages including a darker color (may assists faster spring green up), lower cost, lower carbon footprint of production, and potential CO₂ sequestration. A two-year field study is underway on creeping bentgrass and ultradwarf bermudagrass greens in Raleigh, NC. Treatments include a quartz sand and a basalt sand, each applied at two rates: (1) high rate (USGA method), requiring approximately 26 ft³ and 36 ft³ per 1,000 ft² annually for bermudagrass and creeping bentgrass, respectively, and (2) low rate (PACE Turf Method), in which monthly topdressing amounts are calculated as turfgrass growth potential × 300 lbs per 1,000 ft². Topdressing is applied every two weeks. Preliminary results from June and July 2025 indicate that sand type and topdressing rate did not significantly affect surface firmness or soil volumetric water content for both greens. High rate (USGA method) resulted in significantly greater normalized difference red edge (NDRE) values compared with low rate (PACE method), regardless of whether quartz or basalt was used on both greens. Sand type did not significantly influence NDRE within the same rate. On creeping bentgrass, quartz sand at the lower rate produced significantly lower ball roll distances (BRDs) than all other treatments. On bermudagrass, quartz sand applied at a higher rate resulted in significantly lower BRDs than basalt sand applied at a lower rate. suggesting basalt may enhance BRDs under reduced topdressing amount. Total organic matter and percent green cover will be evaluated to have a complete understanding on how sand types and rates affect both greens’ performance.

• “Metabolomic and Microbiome Responses to Progressive Drought Stress in Drought-Tolerant and Drought-Sensitive Bermudagrass Cultivars” – Sayada Momotaz Akther, PhD candidate, Soil Microbiology

Plant–microbe interactions in the rhizosphere play a critical role in determining plant responses to environmental stress, yet the integrated roles of root metabolomics, exudation, and microbiome dynamics remain poorly understood in turfgrass systems. Here, we combined untargeted metabolomics and shotgun metagenomics to investigate the effects of host genotype and drought stress on root metabolites, exudates, and rhizosphere microbial communities in two contrasting bermudagrass cultivars: drought-tolerant TifTuf and drought-susceptible Tifway. Genotype emerged as the dominant factor shaping microbial community composition, with TifTuf selectively enriching the beneficial bacterium Massilia putida. In contrast, drought stress exerted a stronger influence on root exudate chemistry, triggering the release of specific metabolites that reprogrammed microbial function without major community restructuring. Differentially abundant microbial genes under drought included TccC toxin complex proteins and Type VI secretion system components, indicative of intensified microbial competition and colonization strategies. Functional enrichment also highlighted universal stress proteins and polyamine transporters, underscoring the mechanisms of microbial resilience. Together, these findings demonstrate that bermudagrass adapts to water deficit not by overhauling its microbiome, but by fine-tuning microbial functional activity through targeted metabolite signaling. This study provides new insights into plant–microbiome communication and highlights candidate metabolites and microbial partners that could be leveraged to enhance drought resilience in turfgrass.

• “Exploring Spectral Responses of Turfgrass Seedheads Using UAV-Based Pika L Hyperspectral Imaging” – Caiwang Zheng, postdoctoral researcher, Translational Plant Phenomics

Hyperspectral characterization of turfgrass seedheads offers an opportunity to quantify reproductive development through fine-scale spectral variation. In this study, we used UAV-based Pika L hyperspectral imagery acquired across nine dates during the 2025 growing season to investigate how canopy spectral responses relate to seedhead traits, including seedhead abundance, color transitions, and maturity progression. The Pika L sensor provided continuous reflectance information from 400-1,000 nm, enabling detailed examination of wavelength-specific and index-based patterns during seedhead emergence and senescence. Ground-measured seedhead metrics, particularly the seedhead count rating, which represents visual density scores, are being used as reference variables for ongoing correlation analysis. Building on these spectral-rating relationships, we aim to develop hyperspectral-based classification models designed specifically to distinguish seedhead count rating classes. This ongoing work seeks to identify the most informative spectral regions and features for rating-based discrimination and to establish a foundation for automated assessment of turfgrass reproductive development using UAV hyperspectral data.