A technique for estimating the magnitude and frequency of floods in the Houston, Texas, metropolitan area was developed by use of a multiple-regression flood-frequency analysis of flow data from unregulated streams in the area. A regression model, relating flood-peak discharge to concurrent rainfall and antecedent soil moisture conditions, was used to simulate 67-year records of annual peak discharges. Flood-frequency characteristics were determined for the simulated annual peaks and for the observed annual peaks at each of 22 gaging stations. Drainage area, bank-full channel conveyance, and percentage of urban development were used as independent variables; and weighted flood-frequency discharges were used as dependent variables in the multiple regression analysis.
Relationships applicable to unregulated streams were developed for predicting floods with recurrence intervals of 2, 5, 10, 25, 50, 100, and 500 years. Drainage basins ranged in area from 1.33 to 182 square miles. The percentage of urban development in these basins ranges from 37 to 98.9 percent.
The relationships indicate that as a basin changes from a completely natural state to one of complete urbanization, the magnitude of a 2-year peak discharge is increased by a factor of 4.2, the magnitude of a 50-year peak is increased by a factor of 4.9, and the magnitude of a 100-year peak is increased by a factor of 4. 9.
