How an 1887 Sonoran earthquake dried up Southern Arizona wells and springs, fed others

For Tucson residents, the 1887 Sonoran earthquake was not only a distant disaster. It changed where water could be found across Southern Arizona, drying springs and wells in some places while forcing groundwater closer to the surface in others. The effects reached San Xavier, downtown Tucson, Fort Lowell, the Rillito River and the San Pedro River. Some changes lasted for decades, and a Tucson author argues that the earthquake may even have influenced the later decline of the Santa Cruz River, although that possibility remains disputed.

The earthquake struck near Bavispe, Sonora, and registered an estimated 7.4 on the Richter scale. Although its epicenter was in northern Mexico, Tucson and much of Southern Arizona were close enough to experience major ground movement and changes beneath the surface. Mary Elizabeth Reynolds documents those effects in her new book, “The Quake that Drained the Desert,” which examines how the earthquake altered regional water supplies and how residents, farmers and military officials responded.

At San Xavier Mission, the earthquake shut down the key spring that supplied water to the community. A different spring appeared upstream, however, allowing Tohono O’odham farmers to quickly construct a reservoir and return to growing crops. In central Tucson, El Ojito, a spring that had provided water to nearby residents since 1770, also dried up. The spring served a recreation area known as Carrillo Spring, making its disappearance both a practical and social loss for people living in the growing settlement.

Fort Lowell, on Tucson’s east side, experienced a major decline in its well water. Before the earthquake, the Army could reach water about 30 feet below ground. Afterward, the water level fell to roughly 100 feet. Water stored in the fort’s tower also began to taste brackish. The Army ultimately abandoned Fort Lowell primarily because the Apache Wars had ended, but the worsening difficulty of drawing water from the aquifer may have contributed to that decision. The nearby Rillito River also stopped flowing throughout the year west of Fort Lowell, close to the area where Craycroft Road now crosses the riverbed. The earthquake was considered one cause, along with groundwater use by farms and cattle operations.

The San Pedro River showed a sharply different pattern. A section near Charleston, the former mining town north of Tombstone, temporarily dried up. Near St. David, farther downstream, the earthquake pushed artesian groundwater close to the surface. That newly accessible water supported orchards growing pecans, walnuts and other crops. The San Pedro also surged in the St. David area, flushing stagnant pools where mosquitoes had multiplied and been associated with cases of malaria. Reports from the period described other changes, including sudden rises and falls in wells, stronger spring flows, new water appearing at the surface, and water and mud being forced from cracks in the ground.

The varied results reflected the region’s complicated underground geology. Springs occur where groundwater reaches fractures or encounters an impermeable layer that redirects it toward the surface. Strong shaking can shift those layers, closing an existing route in one location while opening a new pathway somewhere else. Julio Betancourt, a retired U.S. Geological Survey scientist and longtime student of the Santa Cruz River, said liquefaction may compress waterlogged sediments and squeeze water out of an aquifer. He also pointed to riverbed downcutting that began around 1890, which lowered the channel and may have helped lower the water table.

That possibility is especially significant for the Santa Cruz, which once carried water year round in portions of Tucson. By the 1940s, the river through the city generally flowed only during and after major storms, a condition that continues today except where Tucson and Pima County agencies release treated sewage effluent. Reynolds presents a geological scenario in which the earthquake may have altered the aquifer beneath the river enough to contribute to its decline. Betancourt considers that possible, but Robert Webb, another retired USGS hydrologist who has extensively studied the river, strongly disagrees. Webb says there is no evidence the earthquake affected the Santa Cruz and argues that excessive groundwater pumping is the well documented explanation for the loss of its perennial reaches.

The two experts also differ over how consistently the river flowed in the past. Webb says the Santa Cruz was never continuously flowing from its headwaters north of the Mexican border to the Gila River, except during the historic floods of October 1983. He says year round water historically occurred mainly near the river’s junction with the Rillito River and Cañada del Oro Wash, and near the base of “A” Mountain west of downtown. Betancourt’s view leaves more room for the earthquake to have influenced later changes, particularly because the riverbed began cutting downward soon after the quake.

The earthquake’s effects fit a broader pattern seen after powerful inland earthquakes. Reynolds compares the Sonoran event with Idaho’s 1983 Borah Peak earthquake, which also occurred within a tectonic plate in the Basin and Range province. After that quake, researchers found that about half of the measured wells in Idaho, Montana and Wyoming either permanently dropped or became too muddy for drinking. Old Faithful at Yellowstone, more than 200 miles away, also shifted from erupting every 59 to 70 minutes to intervals of 77 to 78 minutes four months later. For Tucson and Southern Arizona, the 1887 quake remains a reminder that groundwater can change dramatically even when the shaking occurs far away, and that the region’s water history was shaped by both natural forces and human pumping.