Rain seldom falls on the desert lowlands of coastal Peru, so people in the area have always depended on the water that flows down from the Andes during the rainy season

But streams in this part of the world come and go quickly, so indigenous people built a system of canals and ponds to channel excess rainwater and create groundwater. Now a group of researchers says that a scaled-up version could help improve Peru’s water management.Ancient engineers (not aliens)1,400 years ago, Chavin and Wari indigenous communities on the slopes of the Andes Mountains dug systems of stone-lined and earthen canals to channel excess rainwater from streams to areas where the ground could soak up more of the water. From there, the water gradually trickled through sediment and cracks in the rock until it reached springs downslope. “Water is stored in the soils and travels much slower beneath the surface than it would as overland flow,” Boris Ochoa-Tocachi, a civil engineer at Imperial College London, told Ars Technica. Water that would otherwise have been lost to flooding feeds springs that remain active even into the dry season.Today, most of these once-widespread canals—called amunas in the Quechua language—lie abandoned or clogged. But in a few rural communities, like Huamantanga in the central Andes, people have used and maintained parts of the ancient amunas for centuries. Eleven of the original canals still operate, feeding 65 active springs and 14 small ponds.Recently, with help from local non-governmental organizations, people at Huamantanga started enhancing these pre-Incan systems. In most cases, they used concrete to make the upstream section of the canal more watertight so that more water reaches the permeable ground downslope. And when Ochoa-Tocachi and his colleagues injected tracer dyes into the canals and checked to see how much—and how quickly—the dye emerged at springs downstream, it turned out that the amunas system still works remarkably well.In some places, it took just two weeks for the dye to make its way through the system and bubble up in a spring; elsewhere, it stayed underground for up to five months before emerging. On average, water took about 45 days to flow through the canals and underground channels to reach springs. The springs were still flowing about a month after the end of the dry season, and the team calculated that the amunas caught about half of the water discharged by streams above Huamantanga.
A diagram of how the amunas divert water from flooded streams to groundwater and springs, where it emerges after weeks or months underground.
Ochoa-Tocachi et al., Nat. Sustain., 2019.

This part of the amunas system is called a diversion canal; it routes water away from the stream toward areas where it can soak into the ground.
Sam Grainger, Imperial College London, 2015.

Canals like this divert water during the wet season to areas with more permeable ground. Water is stored in the soils and becomes available during the dry season.
Musuq Briceño, CONDESAN, 2012.

Diversion canals channel water into earth-bottomed infiltration canals like this one, where water can begin to soak into the ground on its way to a pond or basin.
Musuq Briceño, CONDESAN, 2012.

A water pond as part of the pre-Inca infiltration system during the wet season. Ponds like this store water when it emerges downstream after it was infiltrated in the highlands. Water in the pond is used by the local community for agriculture and livestock grazing, and water can also soak further into the soil.
Sam Grainger, Imperial College London, 2015.

A water pond as part of the pre-Inca infiltration system during the dry season. Ponds like this store water when it emerges downstream after it was infiltrated in the highlands. Water in the pond is used by the local community for agriculture and livestock grazing, and water can also soak further into the soil.
Junior Gil-Ríos, CONDESAN, 2014.

The community of Huamantanga in the central Andes, where the pre-Inca infiltration system is located. The city of Lima would be located downstream in the horizon background.
Junior Gil-Ríos, CONDESAN, 2014.

The Andean highlands in the Chillon river basin where Huamantanga is located. The city of Lima would be located downstream in the horizon background.
Sam Grainger, Imperial College London, 2015.
Previous SlideNext Slide

It’s not terribly surprising that the system works; after all, other people around the world have devised similar ways of managing water supplies. Medieval Arab rulers in Spain built systems called careo from the 9th to 15th centuries, and structures called paar collect rainwater after it percolates through sandy soil in Rajasthan, India. Across North Africa and the Middle East, sloping tunnels called qanats carry groundwater from within hills to wells dug on the lower slopes.“The challenge is scalability”Canal systems like the one at Huamantanga were built to support a relatively small community; today, about 1,000 people live in the town, where they rely on the water to support livestock and farming. But Ochoa-Tocachi and his colleagues suggest that a larger-scale version of the ancient design could be useful for the teeming city of Lima, which relies on a complex infrastructure and still struggles to provide enough water to its 12 million residents during the dry season. (Even with about 330 million cubic meters of storage capacity in the form of reservoirs and dams, Lima still needs about 43 million cubic meters more water to meet its demand during the dry season.)“The challenge is scalability to estimate regional impacts for a much larger population in the lowlands and urban centers,” Ochoa-Tocachi told Ars. His team ran a computer simulation that took the data from Huamantanga and applied it to the Rimac River basin, which waters the Lima. The team found that the systems could boost water flow through the Rimac River by an average of 7.5 percent during the city’s dry season. The impact would be most noticeable in the first couple dry months—around a 33 percent increase—and then taper off to less than one percent by the end of the season.“We observed that the water volumes will follow a curve that peaks after two months and is sustained even eight months later,” explained Ochoa-Tocachi. “The mean residence time of water inside the soils is 45 days, which means that if a unit of water is infiltrated in day 0, half of this unit would come out within the first 45 days, and the other half will come out after 45 days, but for as long as 10 months after the initial infiltration. This means that water will emerge unevenly (more water during the early months and less water during the later months).”That’s definitely not a magic bullet (or magic water gun), but Ochoa-Tocachi and his colleagues claim it’s enough to be a useful part of Lima’s overall water supply system, perhaps by relieving some of the demand on reservoirs, dams, and other infrastructure. If everything works as well as the simulation predicts, a large enough amunas system could reroute about 99 million cubic meters of water a year down the river to Lima. “This could allow serving a higher water demand with the same infrastructure,” they wrote, adding that it could also help buffer against short droughts during the wet season.Officials not ready to jump inSo far, there has been no official move from the Peruvian government or Lima’s water utility. But the benefits of a canal system resemble those targeted by a 2014 law in Peru, which offers incentives for ecosystem-services projects—efforts that use the natural ecosystem to help provide food and water, control climate, and control pests and diseases. The law emphasizes “integration of indigenous and scientific knowledge,” wrote Ochoa-Tocachi and his colleagues, and mentions rehabilitation of traditional water harvesting infrastructure like the amunas.But that doesn’t mean municipal or federal agencies in Peru are ready to jump in. “Although, as mentioned before, local NGOs have become interested in restoring some of the canals, Lima’s water utility SEDAPAL is still more reluctant to invest in this type of infrastructure,” Ochoa-Tocachi said, adding that in part, that’s because of a lack of concrete (pun 100 percent intended) data. The latest study provides some of that data, but more is needed to answer detailed questions about how well the amunas would work in different areas of Peru’s desert coast or how well they could serve a large city like Lima.

default-poup