The San Pedro River of southeast Arizona is a north-flowing tributary of the Gila River. The area of the drainage basin upstream of the 40-km-long study reach is about 3,200 km². This study traces the historical evolution of the San Pedro River channel—specifically, the deepening, widening, and sediment deposition that have occurred since 1900—and it aims to evaluate the causes of channel widening and deepening, the rate of widening, and the present stability of the channel.

Alluvium of the river valley consists of upper Holocene pre- and postentrenchment deposits. The pre-entrenchment alluvium, which forms the principal terrace of the inner valley, accumulated between about A.D. 1450 and 1900 in a relatively sluggish, low-energy fluvial system with extensive marshy reaches and high water table. In contrast, postentrenchment alluvium, which forms the terrace, floodplain, and channel of the San Pedro River, was deposited in a relatively high-energy, entrenched, and meandering fluvial system.

The river flowed in a shallow, narrow channel on the surface of the unentrenched valley before 1890. A series of large floods, perhaps beginning as early as 1881, eventually led to entrenchment of the channel between 1890 and 1908. This deepening placed the channel 1 to 10 m below the former floodplain. The channel has widened substantially since entrenchment through lateral migration and expansion of entrenched meanders; its present size is 5.7 times greater than before entrenchment. The rate of channel expansion, however, has decreased since about 1955, coincident with a decrease of peak-flood discharge. Channel area increased at 0.1 km² yr⁻¹ from entrenchment until 1955; since then the area increased at only 0.02 km² yr⁻¹, suggesting that the channel has stabilized and that further widening will probably be minor under present conditions of land use, discharge, and climate.

The reduction of peak-flow rates was related partly to increased channel sinuosity and to development of floodplains and riparian woodlands. The increased sinuosity produced a reservoir effect that attenuated flood waves, and the development of flood-plains enabled flood waters to spread laterally, thereby increasing transmission losses. In addition, flow rates were probably affected by improved land use and changes of rainfall intensity and short-term rainfall patterns, which reduced runoff and decreased the time necessary for channel stabilization. Livestock grazing decreased steadily after the turn of the century, and numerous stock ponds and small water-retention structures were constructed in tributaries. The cumulative effect of these structures probably reduced peak-flow rates. Short-term rainfall patterns of the wet season (June 15–October 15) have probably changed from annual alternation of above- and below-average rainfall to a biennial or longer pattern. Moreover, frequency of low-intensity rainfall (daily rainfall less than about 1.27 cm) was consistently above average for the decade 1957–1967. These factors probably improved conditions for growth and establishment of vegetation both in and outside of the channel.

The causes of the large floods that resulted in entrenchment are poorly understood, although climate and land use were key factors. Floods followed closely the rapid settlement of the area brought about by mining activity in the late 1870s; population rose from a few hundred to 6,000 in less than 5 yr. Extensive wood cutting for mine timber and fuel, suppression of wildfire, and reintroduction of large cattle herds undoubtedly exacerbated entrenchment. Flood-producing wet-season rainfall in the Southwest, however, was unusually heavy before, during, and shortly after entrenchment.