| This dissertation covers an intensive study of the landscape phenology of Wisconsin's temperate mixed forest. It endeavors to connect conventional plant phenology study back to its ecological complexity (from gardens/trees to the forest) and to compare field-observed phenology with remotely sensed phenology for regional to global monitoring/forecasting applications (from the forest to biomes). A new research perspective: landscape phenology (LP) is proposed in this dissertation. LP is defined as an approach to seasonal vegetation dynamics that integrates spatial patterns and temporal processes within heterogeneous environment across multiple scales.;High density in situ observations, remote sensing data, and spatio-temporal analysis are employed for understanding patterns and processes within the complexity of seasonal landscape dynamics. In particular, bi-daily spring forest phenologies of multiple tree/shrub species and understory plants were observed using field protocols or digital photography; high-frequency micrometeorological measurements were used in tandem with LiDAR-based microtopography/canopy heights as well as soil condition data, to characterize microenvironments; and high-resolution, multi-temporal satellite images were employed to facilitate plant community delineation and landscape scaling. A hierarchical upscaling approach is introduced, aiming to integrate in situ phenological observations with the remotely sensed phenological measures.;Primary results from this work include: a detailed account of spatio-temporal variations of spring plant phenology and their environmental drivers within a typical seasonal forest; thermal time (accumulated growing degree hours) driven linear phenological models for six forest species; a landscape-level phenological progression regime driven by antecedent weather fluctuations; a conceptual landscape phenology model that assigns phenological behaviors to levels of population, community, and ecosystem patch; and a nested hierarchical upscaling approach that successfully validates a satellite-based phenological index algorithm used for deriving NASA MODIS land cover dynamic products. |
