Streamflow Connectivity Under a Changing Climate with Implications for Endangered Desert Fishes in the Verde River, Arizona

We simulate climate change impacts to streamflow in a desert river system to quantify potential changes in longitudinal connectivity. Under a changing climate, desert river systems may experience increases in streamflow intermittency and potential regime shifts from perennial to intermittent or ephemeral flow, which can cause temporary or permanent fragmentation of aquatic habitats. The Verde River in central Arizona, USA (14,229 km² catchment area) is composed of a naturally perennial mainstem with several contributing tributaries of both perennial and intermittent flow. We chose this particular river system because it supports 12 species of native fish, many of which depend on perennial streamflow and all of which are either state or federally listed as endangered, threatened, or species of concern. In addition, hydrologic alteration has remained low relative to other major rivers in the region.

Streamflow is simulated with the Soil Water Assessment Tool (SWAT), a continuous, distributed precipitation-runoff model that uses spatially explicit elevation, soil, and land cover data coupled with daily precipitation and temperature data. We calibrate and validate the model using 2 10-year daily discharge records (1988-1998 and 1999-2009) from 5 USGS-operated stream gauges (2 located along the mainstem and 3 located on tributaries). Both 10-year periods are hydrologically similar and representative of current normal regional climate conditions. We apply forecasted changes in monthly temperature and precipitation to the 10-year validation period to simulate streamflow response to climate change. Projected values are averages taken from an ensemble of 11 Global Circulation Models (GCM) under the A2 scenario for 2070-2099, which forecasts a 3-6^oC increase in monthly temperatures and a 30-50% decrease in precipitation for some months. We partition the Verde River catchment into sub-basins such that model output is generated every 2 kilometers of channel length to identify potential interruptions in longitudinal connectivity.

Expected flow regime changes include increased frequency in no-flow days in already intermittent reaches in the upper basin and the introduction of no-flow days in some perennial reaches that currently experience reduced baseflow from diversions. We couple model output with long-term fish distributional data to forecast how climate-induced changes in longitudinal connectivity may affect the population persistence of native and non-native fishes within the basin. These data will be used to identify stream reaches that may be vulnerable to increased flow intermittency and potential habitat fragmentation under a changing climate.