Species richness and diversity are two metrics of biodiversity that may serve as useful indicators of ecosystem condition, quality, or health. Ideally, their response to a variety of natural drivers and anthropogenic stressors in the ecosystem of interest would be clearly understood, thereby providing a means to separate trends in ecosystem condition caused by management or stressors from fluctuations caused by natural variability.  This understanding is particularly important for adaptive management and ecosystem monitoring programs, in which experimental control of various factors is not possible and distinguishing signal from noise may be difficult.  Several networks of the U.S. National Park Service’s Vital Signs monitoring program are using plant species richness and diversity as measures of ecosystem health in Great Plains grasslands, but the relative sensitivity of these metrics to natural, annual fluctuations in weather vs. management actions (grazing, fire) or stressors (N deposition) has not been determined.  Therefore, we used the AICc model selection technique with six existing datasets from Great Plains grasslands to assess the relative importance of multiple weather models for explaining interannual variability in these metrics, and a literature review to compare the relative magnitude of temporal variability within experiments to the magnitude of treatment effects.

Results/Conclusions In our analyses of existing long- and short-term datasets, the relationships between plant species richness or diversity and various weather models (precipitation and temperature by season) were highly variable among datasets, among experimental treatments or vegetation types within datasets, and among richness and diversity metrics.  Across datasets, temperature models tended to have a stronger relationship to native richness than did precipitation models, but the reverse was true for exotic richness. Native richness and diversity shared a significant weather model in slightly less than half of the cases investigated, but exotic richness and diversity shared no significant models.  The strength of the relationship between richness and weather models was generally greater than between diversity and weather for both natives and exotics.  Over all the published studies we could find, the average ratio of maximum temporal variation to maximum treatment effect was 0.74 and 0.69 for richness and diversity, respectively. Clearly, fluctuations in precipitation and temperature contribute a substantial amount of “noise” to temporal patterns in Great Plains grassland plant species richness and diversity, but the impact of these fluctuations is not clear and is apparently highly situation-specific.