Data for "Anuran Populations Exhibit Species-specific Responses to Precipitation Conditions"

Description

Flooding and drought events are increasing in frequency and severity across the Midwest of the USA. Increased variability of water on the landscape can impact wildlife populations, especially those that rely on water throughout their life cycle. We modeled anuran abundance for six species that range statewide in Iowa, USA using data from the Multiple Species Inventory and Monitoring program and the PRISM Group. Timed searches were conducted on at least five occasions between mid-April and mid-October each year from 2007-2022, with a total of 4485 surveys across 396 sites. We used Bayesian hierarchical N-mixture models to estimate adult abundance of each species while accounting for imperfect detection probability. We tested three hydric metrics (precipitation mean and variance, and drought index) during each of three temporal periods (current year, 1-year prior, 2-years prior) to determine the weather conditions that best explain variation in abundance for each species. Responses to hydric metrics were species-specific, and mean precipitation tended to be the strongest predictor for nearly all species. Regardless of the hydric metric, abundance was often best predicted by the hydric conditions one year prior. Boreal chorus frogs and American toads responded similarly to each hydric metric and time scale, and bullfrog and Blanchard’s cricket frogs showed similar patterns to each other. These emerging commonalities between species, suggest that conservation should focus on measures to support species assemblages based on shared ecological characteristics to maximize conservation benefit.

Steps to reproduce

We conducted standardized visual encounter surveys for anurans each year from April to October. We surveyed each property for four person-hours five times throughout the year. Surveys were conducted at least two weeks apart and with surveys encompassing spring, summer, and fall seasons to maintain independence while capturing the entire active season for frogs. On average 56 properties across the state of Iowa were surveyed each year. Two species, Cope’s gray treefrog (Hyla chrysoscelis) and eastern gray treefrog (Hyla versicolor) were aggregated as Hyla sp., as these species could not be confidently differentiated. We used ArcGIS Pro to extract monthly precipitation for each property location from historic PRISM precipitation and temperature data for the state of Iowa. This data is at a 4km resolution. We used program R and the package SPEI (v 1.8.1) to calculate mean monthly precipitation, precipitation variance, and Standardized Precipitation Evapotranspiration Index (SPEI) for each property-year combination. SPEI was calculated using the Thornthwaite equation for potential evapotranspiration. The three precipitation metrics were calculated for a 12-month time frame starting in November prior to the survey year to October to the survey year. We also calculated these metrics one year prior and two year prior to the survey year to be included as abundance covariates in our models. Additionally we calculated a monthly SPEI index for each property visit to be included as a detection covariate in our models. For a more detailed description of amphibian survey methodology, see the Iowa Department of Natural Resources manual on their Multiple Species Inventory and Monitoring (MSIM) program: https://www.iowadnr.gov/programs-services/wildlife-diversity-program/species-inventory-monitoring#msim-manual All models were run using package spAbundance in program R using N-mixture models. We analyzed the data in a stacked design with data organized by property-year combination. We included year as a random intercept on the abundance (state) part of the model.

Institutions

• Iowa State University

  Iowa, Ames

Categories

Licence