Executive Summary: Fisheries data compilation efforts for this project fell within two large watersheds in Arizona; the Verde River watershed (Desert LCC) and the Little Colorado River watershed (Southern Rockies LCC). We divided the project into two phases; 1) data compilation for the Arizona Game and Fish Fisheries Information Systems (FINS) and 2) a demonstration of FINS through model development and species distribution data. During phase 1, we compiled, cleaned, assigned National Hydrography Dataset (NHD) reach codes to historical data for 113,230 fish records in the Verde River watershed and 43,828 fish records from the Little Colorado River watershed. These records were standardized to meet the Arizona Game and Fish Department FINS standards. From these records we developed three types of geospatial data layers: 1) individual species, 2) reach information, and 3) fish collection records.
In phase 2, we demonstrated the utility of FINS by employing a multivariate adaptive regression splines method to model species presence/absence throughout the entire Verde River watershed. Catchments boundaries were used to represent the area of land contributing to individual stream segments as defined by the NHD+ database, and represent an ecologically relevant scale to quantify environmental factors likely to influence local species occurrence. Environmental and physical attributes associated with each catchment was used to predict presence/absences of species across the watershed. Distribution models were created for 12 native fish species and 18 non-native fish. To test the models accuracy, we electro-fished, seined and/or set hoop nets at 58 sites that had not been previously surveyed. The models ability to predict presence/absence was good. With the exception of rainbow trout, the model predicted species presence/absence in greater than 86% of the previously unsurveyed location.
This product was co-funded by multiple Landscape Conservation Cooperatives: Desert LCC and the Southern Rockies LCC.
