Remote sensing of phenology: Towards the comprehensive indicators of plant community dynamics from species to regional scales

Abstract

Abstract Remote sensing of vegetation phenology has long been used to characterize ecosystem functions and responses to climate at spatial and temporal scales unfeasible to field surveys. However, the potential of remote sensing to elucidate mechanistic drivers of phenology and the underlying plant community processes at such scales remains under‐discussed. This review synthesizes possibilities to advance this knowledge using multi‐temporal remote sensing and discusses remaining challenges and progress in instruments and analytical tools. Recent evidence indicates that, besides documenting vegetation seasonality and responses to climate, remote sensing of phenology can help meet emerging needs for indicators of plant diversity, vegetation structure and ecosystem change. Responses of phenological metrics to stressors over large, heterogeneous regions may provide clues on ecological resilience manifested in asynchronies, recovery of vegetation cycles and stable microrefugia. At the same time, important barriers persist in relation to choosing among phenological estimation methods and paradigms, characterizing phenological events beyond changes in photosynthetically active biomass, and mechanistic interpretation of phenological patterns. Synthesis. Increasing temporal frequency of products, opportunities for multi‐sensor data fusion, and advances in historically less available hyperspectral, active microwave and lidar instruments promise to help navigate these barriers and enable more comprehensive assessments of seasonality. Progress in customizable local platforms such as unoccupied aerial vehicles and phenocams may further enrich ground‐level understanding of phenology and validate satellite‐based assessments. However, remote sensing analyses alone are insufficient for mechanistic interpretation of phenology, which can be challenged by artefacts in remote sensing data and sensitivity of estimated metrics to landscape structure and spatial resolution of the inputs. Robust and informative phenological assessments call for rigorous collaborations with field ecological studies, strategic selection of ancillary environmental and geographic data, and wider adoption of causal inference approaches to address these needs and support novel explorations in plant ecology.