CANCELLED - Primary productivity and the richness of ant assemblages in arid and semiarid ecosystems of Sonora, Mexico

Background/Question/Methods

The ant fauna of central Sonora, Mexico was surveyed with the objective of evaluating the relationships among net primary productivity and ant abundance and richness. Ant assemblages were sampled in both native habitats and in nonnative grass pastures, which increasingly dominate the landscape of central Sonora. Sampling was conducted across a longitudinal precipitation gradient 200 km in length running from the Sea of Cortez in the west to the foothills of the Sierra Madre Occidental in the east. Temperature varies little across this gradient, but as precipitation increases from west to east, the density of vegetation in both native habitats and exotic grass pastures increases dramatically. Net primary  productivity was estimated through Landsat imagery with the Normalized Difference Vegetation Index (NDVI) at a scale of 30 m², and ants were sampled with pitfall traps. Other aspects of vegetation were also quantified. Ant abundance was estimated through species occurrences, a species occurrence being the occurrence of one or more individual ants of a given species within a single pitfall trap.

Results/Conclusions

Over 44,000 individual ants and over 4,500 species occurrences were collected. The number of individual ants per site ranged from 300 to 5,865, and the number of species occurrences from 63 to 220. The mean number of species occurrences was greater in native habitats than in nonnative grass pastures. 72 species and 22 genera were recorded across all sites, and site richness ranged from 13 to 31 species. Mean NDVI value of study sites ranged from 0.218 to 0.892, indicating a substantial increase in primary productivity across the gradient. Mean NDVI values explained only small proportion of variation in ant species richness across sites and were not related ant abundance. In addition, ant abundance proved to be unrelated to ant richness. The best predictors of ant richness were tree cover and perennial plant richness. Given the wide range of NDVI values across study sites, it is somewhat surprising that NDVI values were not a better predictor of ant species richness. These findings suggest that energy availability may be a less important determinant of the richness of Sonoran ant assemblages than other factors.