Early season invasives transform native arid and semi-arid biological communities decreasing biodiversity while increasing likelihood of catastrophic fires. Landscape and climate influence spatiotemporal variation in the establishment and spread of invasive annuals. The aggressive invasive annual cheatgrass (Bromus tectorum) has the capability to rapidly change landscape dynamics and the high plasticity of cheatgrass creates wide variability in growing conditions across space and time. National Parks are sensitive landscapes with unique management goals and cheatgrass has direct impacts on key management objectives related to invasive grass-fire cycles and biodiversity conservation.
Our aim was to develop a systematic model to characterize the status and environmental controls of cheatgrass invasions in the northern Colorado Plateaus National Park units. Detection of Early Season Invasives (DESI) is a tool developed to identify areas with sufficient cheatgrass cover to meet detection thresholds in Landsat imagery. Coupling those images with Hotspot spatial analysis tools (ArcGIS) we examined establishment patterns of cheatgrass and biophysical drivers that influence its persistence or absence. Our objectives were to 1) map the presence and persistence of cheatgrass in seven national park units across a 10-year period, and 2) identify the important biophysical parameters that define the establishment and persistence of cheatgrass.
Results/Conclusions
On average cheatgrass is found in enough abundance to meet detection thresholds in 6% of all park areas. With each park ranging from 3-20% coverage. Persistent populations make up on average 13% of total cheatgrass populations with 85% being identified as ephemeral. Biophysical factors that were used to discriminate against these categories of persistent hotspots, ephemeral, and absent areas of cheatgrass included elevation, clay and sand soil content, distance to park boundaries and high visitation areas, plant available water, Standardized Precipitation-Evapotransportation Index (SPEI), antecedent fall precipitation, average winter temperatures, and spring precipitation. Using stepwise discriminant analysis the largest contributors in defining these categories with all parks combined were elevation and plant available water. When parks were categorized with other like-parks each park group had different characteristics that were more heavily weighted than others. For example, in colder parks winter temperature became much more heavily weighted in the discriminant functions while in the hotter parks percent clay was a top weighted factor. We conclude that cheatgrass populations can be modelled across space and time and across environmental gradients remotely. The scale of analysis can also be adjusted to meet specific management objectives in a spatially and temporally meaningful way.