The Barker and Addicks Reservoirs, located about 16 miles west of Houston, Texas, provide flood-detention storage for storm runoff. Of interest are the water-quality characteristics in the study area and changes in water quality during detention. Study area sampling sites were selected upstream along Buffalo Bayou for Barker Reservoir and on Bear Creek and Langham Creek for Addicks Reservoir, within the reservoirs, near the reservoir outflows, and below the confluence of each reservoir outflow at the streamflow station Buffalo Bayou near Addicks. Flow data were available at all sites except in the reservoirs. Analyses of samples collected during both low flow and storm runoff show that, in general, the waters of the study area were low in mineralization, but the aesthetics of the water was a problem.

The inorganic constituents, trace metals, and pesticides rarely exceeded maximum contaminant levels recommended by the U.S. Environmental Protection Agency for public supply using 1976 and 1977 criteria for primary and secondary standards. All species of nutrients, except ammonia nitrogen and phosphorus, almost always were below the recommended maximum contaminant levels. Phosphorus almost always exceeded these levels.

Aesthetic problems are evident. Large values of suspended solids, turbidity, and color were common. Small dissolved-oxygen values commonly occurred in the reservoirs. Possible bacterial problems are indicated because coliformbacteria densities exceeded recommended levels in about 25 percent of the samples.

The effects of the reservoirs on the water-quality characteristics of storm runoff were analyzed using three approaches. The first approach was a comparison of the discharge-weighted average values of nine selected constituents at each streamflow-gaging station during four storms. Reservoir effects on the quality of runoff detained 1 to 4 days in the two reservoirs were inconsistent. However, the reservoirs consistently had an effect on the water quality of runoff that was detained the longest (more than 8 days). Biochemical oxygen demand, suspended solids, turbidity, color, total nitrogen, and total organic carbon discharge-weighted average values were consistently smaller after flowing through the reservoirs. Dissolved solids and total phosphorus values were consistently larger after flowing through the reservoirs.

The second approach was an analysis of the means of the discharge-weighted average values computed for the four hydrologic events using the Student t-test. Statistical results Indicate that reservoir detention significantly reduced suspended solids (the mean decreased from 178 milligrams per liter at the Inflows to 105 milligrams per liter at the outflows) and turbidity (the mean decreased from 119 nephelometrlc turbidity units at the inflows to 66 nephelometric turbidity units at the outflows).

The third approach was a comparison at each site of the mean, maximum, and minimum values computed for seven constituents that did not correlate with discharge. These constituents or properties of water were temperature, pH, dissolved oxygen, dissolved oxygen percent saturation, total-coliform bacteria, fecal-conform bacteria, and fecal-streptococci bacteria. The only consistent water-quality changes observed were with the three bacteria groups, which were decreased by flood-water detention.