Ochoa-Tocachi, B.F., Bardales, J.D., Antiporta, J., Pérez, K., Acosta, L., Mao, F., Zulkafli, Z., Gil-Ríos, J., Angulo, O., Grainger, S., Gammie, G., De Bièvre, B. and Buytaert, W. (2019). “Potential contributions of pre-Inca infiltration infrastructure to Andean water security.” Nature Sustainability, 2(7), 584-593. https://doi.org/10.1038/s41893-019-0307-1
A study of the amunas of Huamantanga, canals about 1,400 years old that divert wet-season stream water onto hillslopes, where it infiltrates and feeds springs further down. Dye tracing found that infiltrated water “is retained for an average of 45 d before resurfacing”. The authors then model the system scaled up to Lima’s source areas, where it could delay about 99 million m³ of streamflow a year and raise dry-season flow by 7.5% on average. The field tracing is measured; the basin result is a scenario, and it moves water in time with no gain in annual volume. The wiki cites it as the strongest evidence that retention can shift water into the dry season, and as the clearest case of a timing gain kept apart from a volume gain.