![Sample our Geography journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5937/TOC-Subject-Sample-2021-Geography.jpg"})
Abstract
Shifts in vegetation composition and cover are the result of processes acting at different levels such as landscapes, hill slopes, or plant interspaces. Analytical approaches designed for discrete objects which are based upon the inherently hierarchical nature of complex systems are well suited to research applications conducted across spatial scales. We quantified spatial and temporal vegetation dynamics over 71 years at three spatial scales, landscape, plot, and patch, in a Chihuahuan Desert ecosystem in southern New Mexico, USA, using object-based analysis. We analyzed time series aerial photography from 1937 to 2008 to include automated image analysis at the landscape scale and manual delineation of shrub image objects at the patch scale. We sought to identify patch mechanisms associated with changes in shrub patch density and percent cover by characterizing structural changes in individual shrub patches from one image to the next in the time series. The classification scheme captured colonization by new shrub patches, growth or decline in patch area, and patch stability (i.e., change in size of less than 15%). Patch growth was categorized as growth by coalescence with neighboring patches or canopy expansion. Similarly, patch decline was distinguished as either loss of patch area due to canopy dieback or fragmentation of conglomerate patches. Interpretations of change in patch density based solely on shrub colonization and mortality can be too simplistic. Increases in patch density can result from an influx of new patches or fragmentation of patches into its constituent patches; conversely, decreases in density may be due to mortality of patches or coalescence of existing patches. We demonstrate that patches grew in size at the beginning of the study in conjunction with increases in shrub cover (0.5% in 1937 to 11% in 1960) and patch density increased during the initial encroachment phase of shrub proliferation (4 patches ha−1 in 1937 to 80 patches ha−1 in 1960). Shrub cover remained stable at 7% from 1967 to 1989 and over this period, patch dynamics were broadly characterized by growth and persistence of patch area with roughly equal proportions of mortality and colonization. Shrub cover increased linearly from 8% in 1989 to 14% in 2008, approaching a projected maximum shrub cover of 18% based on mean annual precipitation (MAP) of 230 mm. Patch fate over this period constituted growth and persistence of shrub patch area whereas appearance of new patches remained relatively stable. Shrub patch dynamics were nonlinear and variable over time. We documented the transition from grass- to shrub-dominated states with patch dynamics signifying a shifting mosaic in which shrub patch establishment, growth, and mortality wax and wane. Monitoring patch dynamics will become increasingly important in actively managed ecosystems as an important indicator of impending shifts in ecosystem structure and function.
Acknowledgments
A. Moreno's dedication and hard work in digitizing and categorizing shrub patches is appreciated and recognized. C. Steele, B. Nolen, and S. Schrader shared insights and suggestions regarding geoprocessing challenges that arose during the course of this project. P. Gronemeyer assisted with assessing image classification accuracy. K. Havstad and three anonymous reviewers provided constructive criticisms that improved the quality of this work. This research was funded by the US Department of Agriculture, Agricultural Research Service, and the National Science Foundation Long-Term Ecological Research Program, Jornada Basin LTER V: Landscape Linkages in Arid and Semiarid Ecosystems.