Carly Eakin, Department of Wildlife, Fisheries, and Conservation Biology, University of Maine, Orono, ME 04469; firstname.lastname@example.org
Aram Calhoun Dr., Department of Wildlife, Fisheries, and Conservation Biology, University of Maine, Orono, ME 04469; email@example.com
Malcolm Hunter Dr., Department of Wildlife, Fisheries, and Conservation Biology, University of Maine, Orono, ME 04469; firstname.lastname@example.org
Vernal pool-breeding amphibians are threatened by habitat degradation associated with urbanization. Land use change near pools can alter conditions that influence larval development and, via carry-over effects, adult survival and reproductive success. Better understanding amphibian responses in urbanizing landscapes may help guide conservation of these amphibians in developing landscapes. In 2014-16 we surveyed pool characteristics and wood frog (Lithobates sylvaticus) reproductive effort, larval development, and larval abundance at 35 pools across the urbanization gradient near Bangor, Maine, USA. Additionally, we used GIS to quantify impervious, forest, and water cover within 1,000 m of pools. We used regression tree analyses (RTA) to identify the importance of landscape and pool characteristics in explaining variation in larval condition and developmental rate (both adjusted for developmental stage and Julian day) and survival. We created partial dependence plots to examine the influence of important predictor variables. RTA explained 47% of the variation of larval development rate, with higher than expected development rates associated with pools with > 5% impervious surface within 300 m and pools with short hydroperiods. Tree cover within 1,000 m of pools was an important for partitioning pools; i.e., pools with ≤ and > 52% tree cover within 1,000 m had higher and lower than expected development rates, respectively. RTA explained 17% of the variation of relative survival and identified hydroperiod length as positively associated with relative survival, with the highest survival for pools drying mid-July or after. RTA was not effective in explaining variance of larval condition. Our results suggest that land cover change, as measured by impervious surface and tree cover at large scales (300–1,000 m), could increase larval developmental rate. Faster developing wood frog larvae have lower survival and fecundity as adults, thus land conversion at these scales could be detrimental their populations. Low relative survival at pools with shorter hydroperiods indicates the sensitivity of wood frog larvae to the duration of inundation in pools without predatory fish. Because of the high variability within these data and the relatively low-levels of urbanization near Bangor, we caution that these emerging patterns are likely context dependent. However, these results support the need to maintain natural hydrology and terrestrial life-zones near pools.