Site-specific abiotic factors in the response of native prairies to habitat restoration actions
In response to invasions by introduced cool-season grasses across units of the National Wildlife Refuge System in the northern Great Plains, the U.S. Fish and Wildlife Service and the U.S. Geological Survey implemented the Native Prairie Adaptive Management (NPAM) program in 2009. NPAM provides real-time, condition-specific decision support to refuge managers while reducing uncertainty in how prairies respond to management actions. NPAM is a broad-scale application, spanning 19 Service field stations across four states and two grassland biomes: the eastern tallgrass prairie and the western mixed grass prairie. Development of initial models and decision alternatives specific to grassland type lacked information on potentially key biotic or abiotic attributes that could affect prairie response to treatment. Using data gathered over five cycles of management and monitoring, we now ask (1) is the response (percent cover by native grass and forb species) dependent on site-specific abiotic factors, (2) do these factors interact with treatments applied, and (3) what has been the adjusted 5-year trend in the response? We analyzed data in the mixed-grass range of the program, where more units occurred and where a full complement of abiotic predictor variables was available.
Response at the management unit level depended on climatic position of the unit: percent cover of native grasses and forbs was negatively related to average annual precipitation, and magnitude of the relationship was greater at colder sites. The transect-level response was negatively related to a metric derived from the site’s Ecological Site Description (USDA Natural Resources Conservation Service), which range management experts believe characterizes vulnerability of the site to invasion by nonnative plant species and the potential of the site to recover from invasion via management. For some treatments (prescribed fire, burn-graze combination), efficacy varied with prior-year precipitation. After adjustment for abiotic factors, management effects, and the prior-year amount of native cover, percent cover of native grasses and forbs on units in the mixed-grass range has increased after initiation of NPAM. These results suggest that NPAM is meeting its objective of improving conditions of native grasslands through learning-focused decision support. That this trend is evident in five years is encouraging in light of the decades-long history of invasion. The results also suggest ways in which the predictive models that form the core of NPAM can be enhanced to deliver decision support at greater resolution of site condition.