Remote mapping of foodscapes using sUAS and a low cost BG-NIR sensor

Abstract

The assessment of landscape condition for large herbivores, also known as foodscapes, is fast gaining interest in conservation and landscape management programs worldwide. Although traditional approaches are now being replaced by satellite imagery, several technical issues still need to be addressed before full standardization of remote sensing methods for these purposes. We present a low-cost method, based on the use of a modified blue/green/near-infrared (BG-NIR) camera housed on a small-Unmanned Aircraft System (sUAS), to create foodscapes for a generalist Mediterranean ungulate: the Iberian Ibex (Capra pyrenaica) in Northeast Spain. Faecal cuticle micro-histological analyses were used to assess the dietary preferences of ibexes and then individuals of the most common plant species (n = 19) were georeferenced to use as test samples. Because of the seasonal pattern in vegetation activity, based on the NDVI (Smooth term _Month = 21.5, p-value < .01, R² = 43%, from a GAM), images were recorded in winter and spring to represent contrasting vegetation phenology using two flight heights above ground level (30 and 60 m). Additionally, the range of image pixel sizes was 3.5-30 cm with the smallest pixel size representing the highest resolution. Boosted Trees were used to classify plant taxa based on spectral reflectance and create a foodscape of the study area. The number of target species, the sampling season, the height of flight and the image resolution were analysed to determine the accuracy of mapping the foodscape. The highest classification error (70.66%) was present when classifying all plant species using a 30 cm pixel size from acquisitions at 30 m height. The lowest error (18.7%), however, was present when predicting plants preferred by ibexes, at 3.5 cm pixel size acquired at 60 m height. This methodology can help to successfully monitor food availability and seasonality and to identify individual species.

Keywords: Capra pyrenaica; Food resources monitoring; Remote sensing; Vegetation assessment; sUAS.