How many ant workers are in a typical square meter of forest floor?
Individual miniWinkler samples typically contain 50-150 workers. The figure shows box-plots of the number of workers per sample, for the three sampling periods (one outlier value of 657 in 2014 not plotted). There were clearly fewer workers per sample in 2010 compared to 2008, but abundances were higher again in 2014.
Has community species richness changed over the sampling period?
Data were analyzed with EstimateS 9.0. In making data matrices, clearly arboreal elements of the fauna and pest ants in the Osa Biodiversity Center buildings were deleted: Azteca, Camponotus sericeiventris, Crematogaster erecta, Monomorium floricola, Paratrechina longicornis, Tapinoma melanocephalum, and Tetramorium bicarinatum.
Species accumulation curves were calculated for each year's sampling. For each level of sampling effort, sample coverage was estimated using the formula in Chao and Jost (2012). For a given level of sampling, sample coverage is the proportion of individuals in the entire community that are in the species in the sample. For all three sampling periods, sample coverage was over 0.95. To compare richness across sampling periods, species accumulation curves were rarefied to a common coverage of 0.95. The figure shows estimated richness with 95% confidence intervals (EstimateS output) at coverage 0.95.
There is a trend toward a richness decline, but the confidence intervals are broadly overlapping.
Chao, A., and L. Jost. 2012. Coverage-based rarefaction and extrapolation: standardizing samples by completeness rather than size. Ecology 93:2533-2547.
Are there changes in community composition over time?
To investigate this I calculated the Incidence-based Chao-Sorensen Estimate of community similarity (Chao et al. 2000) between pairs of years. This index varies from 0-1, with 0 being no species in common and 1 being complete overlap. Chao's estimate adjusts for undersampling effects in which species are present in a community but appear absent due to insufficient sampling. If the Osa community is changing, the overlap measure comparing the first and later sampling periods should decrease over time. The estimated overlap measure between 2008 and 2010 was 0.964; between 2008 and 2014 it was also 0.964. These results suggest no change over time in community composition.
However, individual species can show large changes in abundance. Several species showed strong declines between 2008 and 2010, but all of these showed rebounds to higher abundances in 2014. These included Pheidole multispina, Octostruma balzani, Solenopsis JTL-003, Hypoponera trigona, Strumigenys gundlachi, and Solenopsis brevicornis. In contrast, Crematogaster tenuicula and Solenopsis azteca showed large increases between 2008 and 2010 and then declined from 2010 to 2014.
Chao, A., W.-H. Hwang, Y.-C. Chen, and C.-Y. Kuo. 2000. Estimating the number of shared species in two communities. Statistica Sinica 10:227-246.
How large might the total ant community be?
The species accumulation curve for the combined data set (all years) shows that the curve is still climbing. The species richness estimate (ICE) is also gradually climbing, suggesting that there is not enough information to estimate community richness. However, we can interpret the final ICE value (176 species) as a minimum and confidently predict that there are at least that many species in the litter ant community.
Distinctive species
The Osa Biodiversity Center forest is home to a high diversity of Neotropical ants. Some of them are new to science and have not been named yet. One distinctive new species, Pheidole carinote, was first discovered in the 2008 sample and was officially published as a new species in 2009. It is one of the most abundant species in the samples. Pheidole carinote is locally abundant but so far is only known from the Osa Peninsula