In winter 2011–12 and summer 2012, the U.S. Geological Survey (USGS), in cooperation with the U.S. Army Corps of Engineers, Albuquerque District and the U.S. Fish and Wildlife Service New Mexico Fish and Wildlife Conservation Office in Albuquerque, New Mexico, evaluated the physical characteristics and fish assemblage composition of available mesohabitats over a range of streamflows at 15 sites on the Middle Rio Grande in New Mexico. The fish assemblage of the Middle Rio Grande includes several minnow species adapted to hydrologically variable but seasonably predictable rivers, including theHybognathus amarus (Rio Grande silvery minnow), a federally listed endangered species. Gaining a better understanding of habitat usage by the Rio Grande silvery minnow was the impetus for studying physical characteristics and fish assemblages in the Middle Rio Grande during different streamflow conditions. Data were collected at all 15 sites during winter 2011–12 (moderate streamflow), and a subset was collected at the 13 most downstream sites in summer 2012 (low streamflow). Sites were grouped into four river reaches separated by diversion dams listed in downstream order (names of the diversion dams are followed by short names of the sites nearest each dam in parentheses, listed in downstream order): (1) Cochiti (Peña Blanca), (2) Angostura (Bernalillo, La Orilla, Barelas, Los Padillas), (3) Isleta (Los Lunas I, Los Lunas II, Abeytas, La Joya, Rio Salado), and (4) San Acacia (Lemitar, Arroyo del Tajo, San Pedro, Bosque del Apache I, and Bosque del Apache II). Stream habitat was mapped in the field by using a geographic information system in conjunction with a Global Positioning System. Fish assemblage composition was determined during both streamflow regimes, and fish were collected by seining in each mesohabitat where physical characteristic data (depth, velocity, dominant substrate type and size, and percent embeddedness) and water-quality properties (temperature, dissolved oxygen, specific conductance, and pH; during summer 2012 only) were measured.

Nineteen species of fish were collected among the 15 sites and four reaches over both sampling periods; 10 of these 19 species are introduced. Fish-species richness (total number of fish species collected at each site during each sampling event) among sites that were sampled during both sampling periods ranged from 6 at Rio Salado to 12 at La Orilla. Fish were most abundant at the Lemitar site (1,786 individuals) and least abundant at the San Pedro site (275 individuals). The native Cyprinella lutrensis (red shiner) was the most abundant species collected among all of the sites, accounting for about 42 percent of fish collected. Fish-species richness and catch per unit effort (CPUE) were higher (or equivalent) at all sites during summer 2012 compared to winter 2011–12.

The relations between fish assemblage composition (that is, total abundance, which refers to the number of individuals of each species that were collected) and selected environmental variables (physical characteristic data collected at the mesohabitat scale [depth, velocity, and substrate particle size], and mesohabitat types) were explored by using canonical correspondence analysis. Environmental variables explained 8 percent (p=0.48) of the variability in the Middle Rio Grande fish assemblage during winter 2011–12, and Rio Grande silvery minnow were weakly associated with sand substrates, relatively moderate velocities (qualitative descriptors are derived from synthetic gradients extracted from CCAs), and relatively shallow depths. Environmental variables explained 14 percent (p < 0.01) of the variability in the Middle Rio Grande fish assemblage during summer 2012, when Rio Grande silvery minnow were associated with run mesohabitats, relatively high velocities, sand substrates, and relatively moderate depths.

The mean fish-species richness was greater in summer 2012 than in winter 2011–12 for each mesohabitat type, and the overall fish-species richness across all mesohabitat types was 0.62 during winter 2011–12, compared to 1.49 during summer 2012. The highest mean CPUE during winter 2011–12 was in isolated pools (54.3 fish per 100 square meters [m²]), whereas the lowest was in flats (18.9 fish per 100 m²). Ranges in CPUE were higher in summer 2012 relative to winter 2011–12 in each mesohabitat type sampled. As in winter 2011–12, the highest mean CPUE during summer 2012 was in isolated pools (233 fish per 100 m²), whereas the lowest was in flats (29.6 fish per 100 m²). Overall mean CPUE per mesohabitat across all mesohabitat types was 29.1 fish per 100 m² during winter 2011–12 compared to 85.3 fish per 100 m² during summer 2012.

Four species of minnows (red shiner, Rio Grande silvery minnow, Pimephales promelas [fathead minnow], and Platygobio gracilis[flathead chub]) were selected to compare preferred mesohabitat characteristics because all are small-bodied minnows and because more than 200 individuals of each of these species were collected. Red shiner were collected across the largest range of depths in both winter 2011–12 (0.02–4.31 feet [ft]) and summer 2012 (0.05–3.4 ft), as well as the largest range of velocities (0.0–4.31 feet per second [ft/s]) during winter 2011–12 among the four minnow species of interest. Rio Grande silvery minnow occurred in the narrowest range of depths (0.30–2.1 ft) during summer 2012, as well as the narrowest range of velocities in both winter 2011–12 (0.0–3.18 ft/s) and summer 2012 (0.02–1.51 ft/s).

Water-quality properties were only collected during summer 2012, when low-streamflow conditions existed and water-quality properties were thought to be potentially most limiting to aquatic life. Area-weighted mean water temperatures tended to be higher at the sites that were sampled in August 2012 (25.57 degrees Celsius [°C]) compared to June 2012 (24.61 °C). The highest area-weighted mean water temperature at a given site (29.03 °C) was measured at the Lemitar site on August 7, 2012, coincident with the lowest measured discharge (4.13 cubic feet per second [ft³/s]). Area-weighted mean dissolved oxygen concentrations tended to be lower in August (7.46 milligrams per liter [mg/L]) compared to June (8.33 mg/L). The highest area-weighted mean dissolved oxygen concentration (9.13 mg/L) was measured at the Lemitar site on August 7, 2012, and the lowest area-weighted mean dissolved oxygen concentration (6.23 mg/L) was measured at the Los Padillas site on August 10, 2012. Area-weighted specific conductance in the sites upstream from La Joya did not exceed 400 microsiemens per centimeter (μS/cm) at 25 °C, whereas the area-weighted mean specific conductance at La Joya (837 μs/cm at 25 °C), Rio Salado (857 μs/cm at 25 °C), and Lemitar (1,300 μs/cm at 25 °C) were all well above the average of the area-weighted means for the 10 remaining sites (433 μs/cm at 25 °C). Lower area-weighted mean pH values were measured at the 3 sites in and near Albuquerque (La Orilla, Barelas, and Los Padillas—7.98, 8.08, and 7.81, respectively) compared to any of the 10 remaining sites, which had an overall mean pH of 8.44.