Quantifying phenological changes and invasions of Rosa bracteata cover on Southeast Texas rangelands

Invasive plants can dominate rangelands by displacing native species critical to ecosystem functioning. Quantifying the amount and configuration of invasives at very-fine spatial resolutions enables efficient monitoring and control. Advances in drone technology for “on-demand” imagery provide new opportunities to assess changes in vegetation cover. The objectives of this study are to (1) assess seasonal differences in quantifying the cover configuration of Rosa bracteata and (2) analyze temporal changes in R. bracteata cover between two years to quantify spread. We collected quarterly drone imagery on a ranch in southeastern Texas (29°55'53.5"N 95°50'54.0"W) from December 2023 to June 2025. We classified R. bracteata using an object-based image classification and computed patch metrics including area (AREA), shape index (SHAPE), and Euclidean nearest neighbor distance (ENN). We evaluated seasonal effects due to plant phenology on image classifications by comparing different seasons within the same year (2024 and 2025) using Kolmogorov-smirnov (P< 0.05) tests. To quantify cover change, we compared classified R. bracteata imagery from March 2024 to March 2025. The average overall accuracy for all image classifications was 90.7%, while the average producer’s and user’s accuracy for R. bracteata were 90.4% and 90.2% respectively. In 2024, AREA, SHAPE, and ENN showed statistically similar values for December, March and July, and statistically different values for September. In 2025, AREA, SHAPE, and ENN values were statistically similar for February, March, and June. This suggests that time of year may affect the quantification of R. bracteata in late growing seasons. Overall, the net change in R. bracteata cover for one year was a gain of 5.9%. Given this level of rapid spread, developing strategies to quantify the configuration and spread of invasives at very-fine scales can improve our ability to selectively treat patches and monitor growth to improve ecosystem health.