A gaging station has been operated by the U.S. Geological Survey at Lees Ferry, Arizona, since May 8, 1921. In March 1963, Glen Canyon Dam was closed 15.5 miles upstream, cutting off the upstream sediment supply and regulating the discharge of the Colorado River at Lees Ferry for the first time in history. To evaluate the pre-dam variability in the hydrology of the Colorado River, and to determine the effect of the operation of Glen Canyon Dam on the downstream hydrology of the river, a continuous record of the instantaneous discharge of the river at Lees Ferry was constructed and analyzed for the entire period of record between May 8, 1921, and September 30, 2000. This effort involved retrieval from the Federal Records Centers and then synthesis of all the raw historical data collected by the U.S. Geological Survey at Lees Ferry. As part of this process, the peak discharges of the two largest historical floods at Lees Ferry, the 1884 and 1921 floods, were reanalyzed and recomputed. This reanalysis indicates that the peak discharge of the 1884 flood was 210,000±30,000 cubic feet per second (ft³/s), and the peak discharge of the 1921 flood was 170,000±20,000 ft³/s. These values are indistinguishable from the peak discharges of these floods originally estimated or published by the U.S. Geological Survey, but are substantially less than the currently accepted peak discharges of these floods. The entire continuous record of instantaneous discharge of the Colorado River at Lees Ferry can now be requested from the U.S. Geological Survey Grand Canyon Monitoring and Research Center, Flagstaff, Arizona, and is also available electronically at http://www.gcmrc.gov. This record is perhaps the longest (almost 80 years) high-resolution (mostly 15- to 30-minute precision) times series of river discharge available. Analyses of these data, therefore, provide an unparalleled characterization of both the natural variability in the discharge of a river and the effects of dam operations on a river.

Following the construction and quality-control checks of the continuous record of instantaneous discharge, analyses of flow duration, sub-daily flow variability, and flood frequency were conducted on the pre- and post-dam parts of the record. These analyses indicate that although the discharge of the Colorado River varied substantially prior to the closure of Glen Canyon Dam in 1963, operation of the dam has caused changes in discharge that are more extreme than the pre-dam natural variability. Operation of the dam has eliminated flood flows and base flows, and thereby has effectively "flattened" the annual hydrograph. Prior to closure of the dam, the discharge of the Colorado River at Lees Ferry was lower than 7,980 ft³/s half of the time. Discharges lower than about 9,000 ft³/s were important for the seasonal accumulation and storage of sand in the pre-dam river downstream from Lees Ferry. The current operating plan for Glen Canyon Dam no longer allows sustained discharges lower than 8,000 ft³/s to be released. Thus, closure of the dam has not only cut off the upstream supply of sediment, but operation of the dam has also largely eliminated discharges during which sand could be demonstrated to accumulate in the river. In addition to radically changing the hydrology of the river, operation of the dam for hydroelectric-power generation has introduced large daily fluctuations in discharge. During the pre-dam era, the median daily range in discharge was only 542 ft³/s, although daily ranges in discharge exceeding 20,000 ft³/s were observed during the summer thunderstorm season. Relative to the pre-dam period of record, dam operations have increased the daily range in discharge during all but 0.1 percent of all days. The post-dam median daily range in discharge, 8,580 ft³/s, exceeds the pre-dam median discharge of 7,980 ft³/s. Operation of the dam has also radically changed the frequency of floods on the Colorado River at Lees Ferry. The frequency of floods with peak discharges larger than about 29,000 ft³/s has greatly decreased, while the frequency of smaller floods, with peak discharges between 18,500 and 29,000 ft³/s, has increased substantially. Operation of the dam has greatly extended the duration of smaller floods; for example, each of the four longest periods of sustained flows in excess of 18,500 ft³/s occurred after closure of the dam.