Research & Projects
Pest climate suitability modeling, Center for Integrated Pest Management. Jan 2021 - present
Model pest establishment risk and create maps to support the Cooperative Agricultural Pest Survey.
Spatio-temporal network model for forecasting invasive species spread, USDA Animal and Plant Health Inspection Service. Mar 2020 - present
Mathematical mode of plant pest migration coupling international trade with drivers of biological invasions. Developed mathematics equations, parameter calibration, and sensitivity analysis.
GitHub repository: https://github.com/ncsu-landscape-dynamics/PoPS-Global
Contaminated consignment simulation to support risk-based inspection design, USDA Animal and Plant Health Inspection Service. Apr 2019 - Dec 2020
Created an open-source Python package designed to measure inspection outcomes under various cargo contamination scenarios to support recommendations for border inspection protocols.
GitHub repository: https://github.com/ncsu-landscape-dynamics/popsborder
Crop stress mapping with unmanned aerial systems (UAS), NC State. May 2018 - Jun 2020
Monitored nutrient and disease stress in tobacco, soybeans, and grain sorghum with UAS. Used statistics to understand relationship between crop canopy structure and crop stress over time. Related blog post: https://cnr.ncsu.edu/geospatial/news/2018/07/02/taking-flight-with-drone-research/
GitHub repository: https://github.com/kellynm/canopy_structure
Bayesian state-space model of peak fall leaf color, Near-term Ecological Forecasting Initiative Summer Course, Boston University, (course group project). Jun 2020
Created a dynamic state-space model using observations and weather variables to predict peak color.
GitHub repository: https://github.com/LucienFitzpatrick/Fitzpatrick_NEFI_2020
Oblique imagery to address distortion in unmanned aerial systems data, NC State. Jan 2018 - Aug 2019
Conducted experiments to measure the impact of using oblique UAV camera angles on resulting digital surface model accuracy and crop canopy reconstruction.
Estimating the extent of beaver-created wetlands in North Carolina, NC State. Feb 2019 - May 2019
Used machine learning (random forest) with lidar data and multispectral imagery to identify beaver dam locations and their associated wetlands in a NC watershed.
GitHub repository: https://github.com/kellynm/beaverDams
Data mining and predictive analytics for cut flower pest inspections, USDA Animal and Plant Health Inspection Service. Nov 2018 - Mar 2019
Mined five years of imported cut flower inspection data to do predictive modeling using neural networks and random forest. Developed Dash visualization and analytics dashboard.
GitHub repository: https://github.com/kellynm/cutflower_model
Automated crop biomass estimation using unmanned aerial systems, NC State. Jan 2018 - Dec 2018
Designed and tested an approach to estimate grain sorghum biomass from UAS imagery and lidar data. Wrote a Python ArcMap tool to automate UAS data processing and create crop biomass report. Analyzed relationship between crop vigor over time and topographic parameters of field.
GitHub repository: https://github.com/kellynm/relative-biomass-maps
Watermelon and tomato trials at CEFS farm, Center for Environmental Farming Systems, NC A&T University, Goldsboro, NC. Apr 2013 - Sept 2013
Collected field trial data for horticulture research projects at the Small Farm Research Unit. Collaborated with NCSU and NC A&T faculty and graduate students to maintain research plots.
Gender dimensions of integrated pest management in Sub-Sahara Africa, Center for International Research, Education, and Development, Virginia Tech. Jul 2008 - Jun 2010
Field research in Kampala, Uganda consisting of surveys, interviews, GPS data collection, and focus group discussions. Statistical analysis to understand relationship between gender, mobility, and IPM.
Graduate Research Assistant, Center for International Research, Education, and Development, Virginia Tech. Feb 2008 - Dec 2008
Prepared gender-focused literature reviews, maps, and reports for the Integrated Pest Management Collaborative Research Support Program.