Multi-scale assessment of semi-arid vegetation communities: climate, disturbance, and environment as spatiotemporal drivers of phenology and composition

Abstract

Ecosystems processes and functions include hierarchical and complex drivers. Assessing drivers of variation at multiple scales therefore helps predict biotic responses and improves our overall understanding of ecosystems. For example, the seasonal cycle and duration of events, phenology, represents a foundational process sensitive to changes in climate, and has cascading impacts across the ecosystem. The long-term record and expansion of remote sensing techniques provides an opportunity to both assess phenological changes through time at broad spatial extents while also assessing variability at finer spatial scales. At regional extents, satellite-based measurement can provide key insights into community level shifts, while finer scaled techniques such as unpiloted aerial vehicles (UAVs), spectral sensors, and automated digital cameras (phenocams) can investigate pattern differences at centimeter scales (i.e., plant and functional groups). I analyzed the year to year and spatial variability of phenology and composition of rangeland systems over multiple spatial scales to explore interrelated aspects of ecosystem functions. I used the AVHRR satellite record of phenology to examine spatial and temporal variability in phenological drivers and to identify key drivers and differences between the phenology of communities, including the role of ecological memory, the legacy impact of prior climate over months to years. In addition, by employing UAVs, spectral sensors, and phenocams I investigated the pattern and influence of heterogeneity on the phenology of grasses and shrubs. Finally, I investigated the interaction of multiple disturbances on the relative proportions of vegetation functional groups within a community. Key findings include productivity tradeoffs, where higher annual temperature increased peak but decreased growing season long productivity; climate conditions from the prior season and up to four prior years influenced date and productivity phenological measures; near earth sensors can characterize phenological variation at the microsite level; and there is an interactive effect of fire and development disturbance on non-native annual grass expansion. The vegetation of U.S. rangelands is projected to have consequential impacts from climate change, especially summer drying, and these impacts can be better quantified by including antecedent conditions and incorporating microsite differences into predictive models.