Comparing Spatial Covariance to Common Field Vegetation Metrics

Le résumé fourni par la source

Spatial covariance is a metric used to delineate grassland cores and track state transitions driven by woody encroachment in rangelands, and it does so by calculating the degree of spatial coexistence between functional groups like grasses and trees at different scales. Spatial covariance is being used in biome-scale frameworks aimed at combatting woody plant encroachment and supporting grassland conservation, but little is known about how spatial covariance relates to on-the-ground vegetation metrics. Here, we begin to address this knowledge gap by comparing tree-grass spatial covariance (0.81 ha scale) to a variety of established, field vegetation structure metrics. Generalized additive models revealed that relationships between spatial covariance and field-based vegetation metrics were weak to moderate (deviance explained ranging from 10% to 47%). Tree density had the strongest relationship with spatial covariance (deviance explained = 47%): tree density nonlinearly decreased to near-zero when spatial covariance became positive. Average visual obstruction and visual percent grass cover had the weakest relationships with spatial covariance (deviance explained = 13% and 10%, respectively). The stronger tie between tree density and spatial covariance and the very weak link between dominant vegetation cover aligns with the intended usage of spatial covariance in rangeland management, which is to locate treeless grassland cores (i.e., spatial covariance values near zero) and prioritize them for defending “core” rangeland habitats. This also aligns with current applications of spatial covariance in grassland conservation as a metric of early warning, variation, intactness, and boundary strength, not of central tendencies or dominance. Additional testing at different scales and in different ecosystem contexts will help to further understand how spatial covariance relates to other field vegetation metrics.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.