Abstract Detail

Biogeography

Soifer, Lydia [1], Peroni, Patricia [1], Thawley, Chris [1].

The effects of spatial and temporal scales on distribution models of Cypripedium acaule in North Carolina.

Species distribution models (SDMs) are widely used to inform conservation management plans. However, modeling results may be influenced by the spatial scale of environmental predictors and the temporal and spatial ranges of occurrence records used in the model. Orchids are a common target of plant conservation efforts. As habitat specialists with distribution records that are often patchy in space and time, their distributions may be challenging to assess with commonly used presence-only distribution modeling techniques, such as Maxent. In this study we investigated 1) the influence of spatial distribution and temporal range of occurrence records from different data sources (publicly available records and herbarium records) on predicted distributions and 2) the ability of relatively large-scale environmental variables to accurately represent the distribution of Cypripedium acaule (the pink lady’s slipper orchid) in North Carolina. We compared the predicted distribution of different models using niche equivalency tests and evaluated model accuracy by calculating true-presence and false-absence rates using populations identified in the field that were not included in the models. We also determined the extent to which macro-scale soil properties and canopy cover represent the same properties at the micro-site level. The model made from publicly available records performed the best based on all measures and identified a significantly different distribution from the model made with herbarium records. We found that spatial and temporal sampling biases impacted model outputs, particularly in cases in which large areas of North Carolina where not sampled and in which older herbarium records may be mismatched with current environmental conditions. Macro- and micro- scale measurements of the soil environment and canopy cover significantly differed, suggesting that SDMs may not capture environmental heterogeneity influential to the distribution of C. acaule at the microsite level. Given the differences in predicted distributions between models and in measurements of soil properties and canopy cover between macro- and micro-scale environmental factors, the possible effects of spatial and temporal sampling bias and scale of environmental variables on predicted distributions should be considered prior to using SDM results for conservation planning of habitat specialists such as orchids.

1 - Davidson College, Biology, Box 7118, Davidson, NC, 28035, USA

Keywords:
Orchidaceae
Maxent
spatial scale.

Presentation Type: Oral Paper
Session: BIOG1, Biogeography I
Location: Virtual/Virtual
Date: Tuesday, July 28th, 2020
Time: 10:15 AM
Number: BIOG1002
Abstract ID:211
Candidate for Awards:None