Drone-Derived Vegetation Responses to a Rainfall Pulse in a Semi-Arid Rangeland

In semi-arid rangelands, rainfall pulses drive vegetation growth. Understanding vegetation responses to rainfall is important for evaluating forage availability and ecosystem resilience. Traditional field methods provide valuable measurements but are often limited in capturing fine-scale changes at the landscape level. This study evaluates very high-resolution drone imagery to quantify herbaceous and woody vegetation responses following a rainfall event in South Texas, USA. Drone surveys were conducted two weeks before rainfall (March 12, 2025) and two and six weeks after the event (April 12, 2025 and May 12, 2025), generating approximately 5 cm orthomosaics and canopy height models to estimate vegetation volume across the study area. We collected field data to identify native and invasive grass communities. From this, we randomly selected 150 sample areas (3 m diameter) for each vegetation cover type from the imagery. We used the nonparametric Kruskal-Wallis test (alpha = 0.05) to evaluate differences in mean volume among sampling periods and the Kolmogorov-Smirnov test for pairwise comparisons among distributions of functional groups. Invasive and native vegetation did not differ significantly among sampling periods, whereas woody vegetation increased significantly from March (10,776,330 cm³) to April (14,568,020 cm³) and remained elevated through May (P < 0.05). Pairwise comparisons showed that woody vegetation differed significantly between March and April and between March and May (15,129,680 cm³; P < 0.05), while no significant differences were observed between April and May or among sampling periods for herbaceous vegetation. The greater structural response observed in woody vegetation is consistent with previous research suggesting that growing-season rainfall, combined with access to deeper soil moisture through deep root systems, may support greater volume growth than in shallow-rooted herbaceous species. By integrating drone-derived imagery and traditional field methods, this approach captures fine-scale vegetation dynamics at the landscape level.