Using drone imagery analysis in rare plant demographic studies

Abstract

For plant species of conservation concern, knowledge of changes in abundance through time is a minimum requirement for informed management. This information is usually acquired through on-the-ground monitoring, which entails counting individuals in defined areas over multiple years. Demographic studies, which involve tracking individual plants through time, are usually carried out at limited spatial scales and over shorter time frames than monitoring, but are more useful to management. In this study we explored the use of drone (UAV or unmanned aerial vehicle) imagery analysis as a tool for collecting demographic data for dwarf bear poppy (Arctomecon humilis), an endangered species restricted to gypsum outcrops in the northeastern Mojave Desert, USA. We obtained imagery at 15 m altitude during peak flowering at four populations in spring 2019. Each poppy plant in the imagery was georeferenced, measured and scored for flowering. To estimate reproductive output, we developed independent data sets relating plant diameter to flower number, then sampled to determine mean fruit set per flower and seeds per fruit. We used these relationships along with plant diameter and reproductive status for each plant in the drone imagery to estimate seed rain on an area basis across nine 0.6 ha demography plots at each population. This method enabled us to collect demographic data on >3,000 plants, including estimated production of ca. 3.7 million seeds, across >20 ha of habitat. We also analyzed imagery acquired in both 2018 and 2019 at two of the four populations and quantified recruitment, growth, and mortality of individual georeferenced plants. Our study is among the first to demonstrate the utility of drone imagery analysis in plant demographic studies. The method is most applicable for non-clonal perennial species with distinctive morphology that occur in habitats with low vegetative cover.