Rigorous analysis of rainwater harvesting system design can improve reliability and water quality

May 2012

PROBLEM

Rainwater harvesting systems that collect and convey rainwater from roofs to storage tanks are often the best or only source of water for many communities in the developing world. A common problem with these systems is that dust and pollutants that accumulate on a roof during dry periods are swept into the storage tank along with the rainwater. While some systems divert the “first flush” to a separate storage vessel, questions remain about the quantity that must be flushed in order to maintain water quality and the impact of the first flush on stored water volume and community water needs.

APPROACH

Senior lecturer Peter Shanahan and graduate student Kelly Doyle S.M. ’08 performed a rigorous analysis of three existing rainwater harvesting systems in Bisate, Rwanda, and determined how incorporation of first-flush diversion would impact the systems’ efficacy. Using a two-year record of daily rainfall for Bisate and a discontinuous 18-year record of daily rainfall for nearby Musanze (1977-1992 and 2002-2005), Doyle determined the probability of rain on any day for each month, assigned a rainfall depth to each day and scaled values to achieve the correct monthly total adjusted to match the two-year record for Bisate. Shanahan and Doyle then computed the storage-reliability-yield of the rainwater harvesting systems, defining reliability as days per year on which the community’s water demand was met. They designed field studies that used six alternative first-flush strategies, varying the quantity of water to be diverted (from 0 to 2 millimeters) and the number of days between rains when pollutants would collect on a roof, then tested water quality in the storage tanks under each first-flush strategy.

FINDINGS

The researchers found that diverting the first flush reduced reliability by a maximum of 8 percent of the total rainfall for an area, which is a relatively small cost for ensuring water quality. In the case of Bisate, diverting the first millimeter of rainwater after three or more days with little or no rainfall flushed enough pollutants to maintain good water quality in the storage tank while reducing the number of days the system failed to meet demand by only four to seven days per year (a 4 percent decrease in reliability). The first millimeter of runoff resulted in a 50 percent reduction in turbidity and a decrease in E. coli to safe levels. The flushed water can be used for irrigation or other non-consumption purposes.

IMPACT

This work provides a quantitative basis for the design of rainwater harvesting systems in the developing world, where the United Nations Development Programme estimates there are roughly 1 billion people who lack clean drinking water. In many communities, women and children make daily trips to distant, polluted pools to collect water, which hampers their ability to work or attend school. Thus, the development of rigorous methods for the design and operation of rainwater harvesting systems and the installation of such systems can have an enormous impact on the present — and future — quality of life for millions of people.

MORE

The results of this work appeared in a paper by Kelly Doyle and Peter Shanahan in Volume 2 Issue 1 of the Journal of Water, Sanitation and Hygiene for Development. Doyle now works at Princeton Hydro, an environmental engineering firm in Princeton, N.J. MIT research specialist Jean Pierre Nshimyimana S.M. ’10 was the environmental scientist who helped design and build the Bisate rainwater harvesting system and collected some of the field data for this research prior to his graduate work at MIT.

Rainwater falling on the corrugated metal roofs of select buildings in Bisate flows into gutters, through a series of pipes and into storage tanks. Three rainwater harvesting systems — at the health clinic, the primary school and the gorilla trackers’ house (Bisate is adjacent to the Virunga Forest, home to the lowland mountain gorillas once studied by Dian Fossey and now studied by the Karisoke Research Center) — collect water in 20 storage tanks with a combined capacity of 106 thousand liters. Photo / Daria Cresti, M.Eng. ’07

Download this issue as a pdf.