Turnover and species reordering during shrub encroachment in the northern Chihuahuan Desert

Background/Question/Methods

Shrub encroachment is a widespread phenomenon leading to significant changes in community structure and ecosystem functioning. The replacement of C4 grasses by C3 shrubs and small trees can lead to dramatic changes in vegetation structure, but it is unclear if these changes reflect wholesale replacement of grassland communities by shrubland or if shrub encroachment primarily reflects reordering of species already present in an area. In addition, the degree to which these changes in community structure differ among taxa remains uncertain. In the northern Chihuahuan Desert, blue grama is being replaced by black grama and black grama grassland is being invaded by creosotebush. We used long-term species composition data from plants (21 years), small mammals (32 years), and native bees communities (14 years) from these three communities at the Sevilleta National Wildlife Refuge (SNWR) to assess the degree to which shrub encroachment reflected reordering, as well as the degree to which reordering differed among taxa. For comparisons within taxa, non-metric multidimensional scaling (NMDS), perMANOVA and dispersion of centroid analyses were based on species presence-absence and abundance data. Presence-absence data quantify turnover whereas abundance data can be used to measure community reordering.

Results/Conclusions

Based on perMANOVA, plant species composition differed significantly between blue grama-, black grama-, and creosotebush-dominated communities. This occurred based on abundance data, abundance data minus the site dominants, and presence-absence data. However, multivariate distances among centroids declined when comparing quantitative and qualitative analyses. Overall, 30 species were unique to blue grama grassland, 12 species to black grama grassland and 23 species to creosotebush shrubland. However, 53 of 155 species were shared among the three vegetation types reflecting a high level of species reordering compared to turnover. Small mammal communities showed strong differences between grassland and shrubland communities based on abundances, and differences persisted with presence-absence data, but considerable overlap occurred among the two grassland communities. Seven of 13 species occurred in all three sites and there were no unique species in either grassland. Bee communities in blue grama grassland were distinct from those of the other two sites, but considerable overlap occurred between black grama and creosotebush communities. A total of 149 species (out of 326) were shared among all three sites. Our results demonstrate that community reordering is a significant component of grassland to shrubland state transitions and that the degree of dissimilarity between states varies among taxa.