Scaling effects on landscape metrics in alpine meadow on the central Qinghai-Tibetan Plateau

Small-scale patchiness is common in alpine meadow on the Qinghai-Tibetan Plateau (QTP) and significantly affects ecological processes. The unmanned aerial vehicle (UAV) provides an effective way to monitor the small-scale patchiness at different spatial scales. However, there is a lack of studies concerning the scaling effects on landscape metrics derived from UAV data, which are essential for observation scales and landscape metrics selection. Here, we first analyzed the responses of landscape metrics and their relationships with controlling environmental factors to the changing of spatial scales. We then discussed the ideal observation scales and landscape metrics for alpine meadow on the QTP. The results showed that: 1) contiguity, edge density, shape, fractal dimension, patch density (PD), perimeter-area ratio (PAR) and connectance (CONNECT) metrics exhibited a predictable (linear or power law) response to the changes of spatial scales; 2) soil water content and temperature (SWC and ST), soil organic carbon (SOC) and air temperature (TEMP) were significantly correlated with percentage of landscape index (PL), largest patch index (LPI), aggregation (AI), average area of patches (PA), CONNECT, PAR, and PD (P < 0.05); 3) the correlations between SWC, ST, SOC, TEMP and most landscape metrics presented as an approximate parabola and peaked at the flight heights of 40–60 m. Our findings indicated that alpine meadow landscape on the QTP could be accurately quantified using UAV photogrammetry with PL, LPI, AI, CONNECT, PAR and PD metrics, and the optimal observation extents are 70 m × 52.5 m ~ 105 m × 78.75 m, with the grain sizes ranged from 1.75 cm to 2.19 cm.