Recent floods on the Wichita River at Wichita Falls, Texas, have reached higher stages compared to historical floods of similar magnitude discharges. The U.S. Geological Survey (USGS) has operated streamflow-gaging station 07312500 Wichita River at Wichita Falls, Tex., since 1938 and flood measurements near the location of the present gage were first made in 1900. Floods recorded in 2007 and 2008 at this gaging station, including the record flood of June 30, 2007, reached higher stages compared to historical floods before 1972 of similar peak discharges. For flood measurements made at stages of more than 18 feet, peak stages were about 1 to 3 feet higher compared to peak stages of similar peak discharges measured before 1972. Flood measurements made at stages of more than 18 feet also indicate a decrease in the measured mean velocity from about 3.5 to about 2.0 feet per second from 1941 to 2008. The increase in stage and decrease in streamflow velocity for similar magnitude floods indicates channel conveyance has decreased over time. A study to investigate the causes of reduced channel conveyance in the Wichita River reach from Loop 11 downstream to River Road in Wichita Falls was done by the USGS in cooperation with the City of Wichita Falls. Historical photographs indicate substantial growth of riparian vegetation downstream from Loop 11 between 1950 and 2009. Aerial photographs taken between 1950 and 2008 also indicate an increase in riparian vegetation. Twenty-five channel cross sections were surveyed by the USGS in this reach in 2009. These cross sections were located at bridge crossings or collocated with channel cross sections previously surveyed in 1986 for use in a floodplain mapping study by the Federal Emergency Management Agency. Four channel cross sections 3,400 to 11,900 feet downstream from Martin Luther King Jr. Boulevard indicate narrowing of the channel. The remaining channel cross sections surveyed in 2009 by the USGS compared favorably with cross sections surveyed in 1986 for the Federal Emergency Management Agency, with no substantial differences noted. Comparison of channel cross sections surveyed in 2009 to those from historic bridge plans indicate no change in cross section has occurred at most of the bridges from Loop 11 downstream to River Road in Wichita Falls, except for obstructions noted at the Scott Avenue bridge and Martin Luther King Jr. bridge. Although obstructions in the channel at these bridges only partially block flow, they could also be contributing to reduced channel conveyance. Step-backwater profiles were used by the USGS to verify channel roughness. The main channel roughness coefficients (Manning's n values) from 2009 surveys were virtually unchanged from those used in a 1991 hydraulic model done for the Federal Emergency Management Agency. The average overbank roughness coefficient (Manning's n value) was 0.15, more than double the value of 0.06 used in the 1991 hydraulic model. Increased overbank vegetation has resulted in higher stages conveying the same amount of discharge, particularly for discharges more than 4,000 cubic feet per second.