Abstract: Species distribution models (SDMs) are important for quantifying species-landscape relationships. However, they often rely upon presence-only occurrence data, which alone is insufficient to infer the absolute abundance of species. Methods for jointly analyzing presence-only and presence-absence occurrence data may be of value to land management and conservation efforts that require estimates of plant species absolute abundance. We have investigated a new spatially-inhomogeneous Poisson point process model to ascertain its ability to quantify the influence of elevation, slope, and aspect on the abundance of plants in the Organ Mountains of New Mexico. Both mesic (Pinus spp., Cheilanthes spp., Juniperus spp., and Selaginella spp.) and xeric (Larrea tridentata, Opuntia spp., and Yucca baccata) species were included to capture a range of life histories. Presence-only data were aggregated from the Global Biodiversity Information Facility, iNaturalist, and herbaria metadata. Presence-absence data were obtained from the Assessment Inventory and Monitoring (AIM) dataset. From the point-process model, we inferred the posterior distribution for density of occurrence for each species across the entire mountain landscape and compared inferred distributions to data withheld from the analysis as a check. We expected that (1) elevation would powerfully predict species abundance, (2) mesic and xeric species would be distributed at higher and lower elevations respectively, and (3) the influence of slope and aspect to vary across taxa. Model estimates converged on biologically reasonable abundance values for every taxon. Abundance estimates for mesic species were greater at higher elevations and northeastern slopes. Those for xeric species were greater at lower elevations. Elevation and slope strongly influenced the probability of abundance for all taxa. East-west aspect influenced abundance estimates of Larrea tridentata, Cheilanthes spp, Juniperus spp. and Opuntia spp., whereas north-south aspect influenced abundance estimates of Pinus spp. Selaginella spp, and Yucca spp. This study introduces a new approach for integrating presence-only and presence-absence data to obtain estimates of plant species absolute abundance across the landscape, a quantity of increasing relevance to land managers and conservation biologists; for them, relative density or occurrence is not sufficient for planning or action. We have found our approach to be useful across a wide range of altitudes and of plant habitats from xeric to mesic. Additional development is required, but this represents a promising advance when inferences of plant abundance are needed.
