Assessing Flood Risks for Hog Farms with Satellite Imagery

Felicia Yan

Overview

North Carolina is the second largest hog producer in the US, and the state has historically faced issues with contaminants from hog waste lagoons, which are worsened by flooding events that cause them to be inundated or overflow. In this study, a method to improve upon the existing database of swine operations was explored by using NAIP imagery and object-based image classification to segment hog farms and their features. Between maximum likelihood, random trees, and support vector machine classifiers, the support vector machine classifier achieved the highest overall accuracy of about 90-92%. Flood risk was then assessed by analyzing six flood parameters including elevation, slope, land use/land cover, NDVI, topographic wetness index (TWI), and drainage density, as well as existing floodplain maps from the North Carolina Floodplain Mapping Program (NCFMP). This research presents a method of using image classification to help interpret flood risks for hog farms and provides information for local decision-making about mitigating the impacts of floods in the future.

My Story

I've always loved the environment and learning about environmental issues and their social impacts. After seeing news about North Carolina hog farms flooding after every hurricane, I was curious about why those events continued to happen and wanted to investigate the local environmental issue further. I couldn't go out with the pandemic, but I felt that there was still something I could do at home.

After using my dual enrollment email to acquire ArcGIS software for free (although I'm still not entirely sure I was allowed to), I used online videos and NASA ARSET tutorials to explore ways to work with satellite imagery, much of which is open access and free to download. I loved every part of conducting research, with many hours spent downloading terabytes of data, creating training samples by hand, and processing rasters.

After working for several months, I entered my local science fair and Student Academy of Science competitions, which were amazing experiences to present my research, and now I'm honored to be here as an AJAS Fellow!