A multiscale perspective for improving conservation of Conchos pupfish

Desert spring systems of the American southwest hold high local fish endemism and are ranked among the most threatened ecosystems in the world. The prioritization of conservation resources to protect species living within these arid landscapes requires knowledge of species abundance and distribution. The plight of Conchos pupfish (Cyprinodon eximius) is representative of freshwater fishes the world over, including population extirpations caused by human poisoning of streams and reservoir construction, to the extent that the species was once considered extinct in the USA. We developed a distance-sampling framework to monitor Conchos pupfish abundance and coupled this approach with species distribution modeling to guide conservation actions. Our multiscale approach included surveying abundances within 5-m transects at three reaches of the Devils River, where the last known USA populations persist. We combined this fine-scale analysis with species distribution modeling for stream segments across the range of the species in Mexico and USA. Modeling revealed Conchos pupfish abundance among transects was negatively correlated with current velocity and detection was negatively correlated with water depth. Estimated abundance at a reach where the species was previously reintroduced was greater than other reaches combined in November 2019, lowest in March 2021 when reach water levels were very low, then equivalent with other reaches by October 2021 after water returned to the reach. Modeled Conchos pupfish distribution illustrated a high probability of occurrence on the periphery of the species' overall range within Texas, USA and broadly across Chihuahua, Mexico, where proposed protected areas might benefit the species. Our study provides conservation guidance by establishing (1) baseline and trajectory values for abundance, (2) transect locations where abundances might be managed within existing protected areas, (3) reaches where high abundances could be used for future repatriation, and (4) stream segments where future surveys might be conducted to assess conservation opportunities.