Predicting Population Factors

We used two factors, inbreeding and isolation, to represent the effects of habitat loss and fragmentation on the PCC populations (Table 3.7 in Chapter 3). To predict future inbreeding, we developed the following relationship based on the history of development in Bay County to predict how future fragmentation and habitat loss may affect the genetic health of the known populations. We use a very simplified relationship to describe the genetic drift process with the understanding that drift may happen in much different ways (e.g., Sewall-Wright process; Futuyma 1998, p 303). However, these relationships require better understanding of the genetic and demographic processes operating in each population to be useful. We used historical aerial imagery and parcel data from Bay County to identify the approximate year that each of the four isolated populations in the western portion of the range were isolated and calculated the approximate time to 2016 when genetic samples were collected (Table 5.1). For example, Airport North was isolated between 1953 and 1964, so the site has been isolated for approximately 56 years. We then used the FIS values from Duncan et al. (2017) to calculate a genetic drift value per 2-year generation for each population in the western portion of the range, assuming Star Ave population represented normal for PCC. We used the mean value (0.0101) to predict drift as future development isolated populations. When a population was predicted to move into the moderate or high category of developed and unsuitable habitat (>33% of area supporting that population in those two landcover types), then we began the drift calculation.

Table 5.1. History of isolation and inbreeding of Panama City crayfish populations in the developed western portion of the species range.
Approximate isolation year
Approximate generations isolated
Inbreeding coefficient
Drift per generation
Airport North 1960 28 0.214 0.01
Airport South 1975 21 0.344 0.0073
Shriners 1982 17 0.359 0.0079
Talkington 1992 12 0.31 0.0153

To predict future isolation, we relied on the work of Duncan et al. (2017). They predicted the least cost path (LCP) between the centroids of all populations for the Status Quo scenario. Briefly, they updated the ‘developed’ classification within the lc6 resistance layer that was most supported from the models, such that we made two new resistance layers that reflected the current and future predicted distribution of urbanization based on SERAP for each time point in the future. They then determined new current and future predicted least cost distances between all of our sites given the new resistance layers. Finally, they used the new current least cost distances to obtain parameters that were then used to predict future genetic distances given the new future predicted least cost distances. Two sites, Airport North and Industrial were not included in this analysis. We estimated the LCP by hand as described in Chapter 3 and, because each site was in the low condition, we kept these sites in the low condition for all scenarios. Because of the computational intensity and time constraints, we used the status quo scenario as our prediction for all three scenarios. However, we realize it would represent the minimum change expected in LCP distances between populations and the Moderate and High Development scenarios may show more change in this population metric.