| Summary: | Spatial analyses of marine mammal stranding data can be used to identify stranding patterns including those associated with Unusual Mortality Events (UMEs). Despite the power of the tool, it has rarely been used to improve surveillance and monitoring. Using ArcGIS and SaTScan, we analyzed 12 years of stranding data from San Juan County, Washington to better understand patterns of carcass deposition. We plotted locations for 631 dead marine mammals that stranded from 2002-2014 and aggregated stranding events into 1000m segments of shoreline. Stranding “hot spots” included beach segments that had significantly higher carcass deposition according to the Getis-Ord Gi* statistic in ArcGIS or segments that were encompassed by significant spatial clusters using the discrete Poisson model in SatTScan. There was a greater frequency of pinniped strandings (84%) compared to cetacean strandings (16%), and there was 65% overlap between cetacean and pinniped hot spots. Hot spots were characterized for the locally abundant harbor seal (Phoca vitulina). We identified 52 hot spots in ArcGIS and 62 hot spots in SaTScan with 81% consistency between the two methods. High proximity to public pedestrian access points was a strong determining factor for significant hot spots, likely due to increased reporting effort in those areas. In addition to this potential reporting bias, beach segments with longer fetch distances and nearly level or gentle slopes were more likely to be hot spots and could prove to be ideal locations to survey to increase routine collection of quality pinniped carcasses for postmortem analysis and disease surveillance. In the case of UMEs or oil spills, targeting these locations could maximize carcass recovery with limited response effort. Other stranding networks in the region could use this methodology to identify local marine mammal hot spots and improve stranding response, especially during times of high expected mortality and low response resources.
|
|---|