Modeling abundance and nest occupation of Pseudomyrmex termitarius as a function of detection factors and habitat structure in fragments of urban vegetation

Understanding the relationship between abundance and habitat variability is an underlying theme of ecological studies. Due to imperfect detection, robust estimates of population size require large efforts, especially for megadiverse taxa such as Formicidae, where most of the surveys rely on presence-absence or community data. Here, we used mixed occupation models to estimate the population abundance and local distribution of Pseudomyrmex termitarius (Formicidae: Pseudomyrmecinae) in nests when detection was imperfect. Our goal was to create a sampling protocol to estimate unbiased population parameters, taking into account several sources of heterogeneity that can affect the estimates (e.g., researcher detection variability, survey period, habitat variables). The advantage of this protocol is that it is time- and cost-effective for accessing common and easily identifiable ant species. We distributed 44 transects (2.5 × 20 m) with five subplots each in 2.33 ha of the vegetation remnants of the Cuiabá Campus of the Federal University of Mato Grosso (UFMT). In each subsample, we recorded the presence or absence of P. termitarius workers, as well as the count of other cooccurring ant species during two consecutive days and time periods (midday and 16:00). Canopy openness and soil compaction were recorded in each transect. Among all the sampling trials, we directly counted a maximum of 364 P. termitarius workers in 92 subplots. Both nest abundance and occupation estimates were directly related to soil compaction. The estimated probability of detection was 39% and was not affected by the period. On the other hand, despite P. termitarius being a noncryptic species in the study area, it had a stronger effect on detectability. Although P. termitarius is a subordinate species, the co-occurrence of other species does not explain its abundance and distribution in our study. Perhaps P. termitarius's preference for sites with more compact soils could favor it by filtering out or reducing the abundance of potential competitors. After taking into account soil compaction and researcher detection variability, we estimated that the species occurred in 70% of the study area with a density of approximately 0.09 nests/m². We expect that this protocol can be an additional tool to improve the knowledge of ants in general and, especially, the management of alien or economically relevant ant species.