Forb variability over two decades in a Chihuahuan Desert grassland assemblage: local dynamics in a regional context

Background/Question/Methods

While shrub and graminoid abundance often dominate the discussion of desert vegetation community dynamics, annual and perennial forbs are a diverse and often prevalent component of these communities depending on climatic conditions and other drivers.  There is a particular interest in how understanding forb dynamics under natural conditions can inform the restoration of degraded desert grasslands systems.   Accordingly, we analyzed forb frequency and abundance from ten datasets on the Sevilleta National Wildlife Refuge (SEV LTER) that span two decades.  These data come from desert grasslands under a relatively natural fire regime and without large herbivores. A forb’s abundance was represented by mean frequency or cover standardized by the standard deviation of all forbs in a single dataset over all years it was detected.  A forb’s prevalence was equal to its standardized frequency of occurrence over all years.   To provide context, we then compared the species richness from the monitoring datasets to a regional species pool represented by suite of 241 relevé plots collected from protected areas across the northern Chihuahuan Desert.   

Results/Conclusions

The frequency analysis yield four class of prevalence and abundance:  1) above average prevalence-above average abundance, 2) above average prevalence-below average abundance, 3) below average prevalence- above average abundance, and 4) below average prevalence- below average abundance.  Among 128 different species, only 11 (8.6%) were detected every year.  On average, only 35% of forbs could be found in any given year.  More importantly, species composition changed greatly with turnover rates averaging 65% annually.   In an analysis of frequency distributions, forb diversity was dominated by species below average in both prevalence year to year and abundance when present.  Only 18 species were above average with respect to prevalence/abundance, and 82 are below/below.   Of the 18 species, only five were highly prevalent and abundant (>1 S.D.).  While informative about for this particular set of species, the species set represented only 13% of the forb species pool found in the regional dataset (977 species from sandy-plains desert grasslands).  Accordingly, the results suggest that additional region-wide analysis of other long-term datasets is warranted to further aid restoration practices.