Executive Summary

The pressing need for sustainable freshwater supplies has increased the use of roof-based 
rainwater collection systems for potable applications. Although rainwater harvesting systems can
be simple and inexpensive to construct, various sources of contamination within the collection 
system can negatively affect water quality.  In addition to environmental factors (e.g., seasonal 
variations) that affect rainwater quality, harvested rainwater quality is affected by the roofing 
material on the catchment surface. The main objective of this research was to provide 
information to the rainwater harvesting community in Texas regarding the impact of roofing 
material on harvested rainwater quality.

In this study, five pilot-scale roofs (asphalt fiberglass shingle, Galvalume® metal, concrete tile,
cool, and green) and three full-scale residential roofs (two asphalt fiberglass shingle and one 
Galvalume® metal) were equipped with rainwater sampling devices to collect the “first flush” 
and water after the first flush. The harvested rainwater was collected from multiple rain events 
and analyzed for the following parameters: pH, conductivity, turbidity, total suspended solids
(TSS), total coliform (TC), fecal coliform (FC), nitrate, nitrite, total organic carbon (TOC), 
dissolved organic carbon (DOC), selected synthetic organic compounds, and selected metals. 
Generally, the first flush contained the highest concentrations of microbial and chemical 
contaminants in comparison to the subsequent collection tanks, indicating that the quality of 
harvested rainwater improved with roof flushing. However, the rainwater harvested after the first 
flush did contain some contaminants at concentrations above United States Environmental 
Protection Agency (USEPA) drinking water standards (i.e., turbidity, TC, FC, iron, and 
aluminum). This indicates that harvested rainwater must be treated prior to potable use.

Based on the pilot- and full-scale studies, none of the roofing materials emerged as clearly 
superior to the others in terms of the quality of the rainwater harvested after the first flush. From 
our limited data set, green roofs do not appear to be the best candidates for rainwater harvesting
for indoor domestic use if the harvested rainwater is disinfected with chlorine.  Although the 
rainwater harvested after the first flush from the green roof consistently had the lowest values of 
TSS, turbidity, nitrite, aluminum, iron, copper, and chromium, it also had the highest values of 
DOC; if disinfected by chlorination, the high DOC concentrations could lead to high 
concentrations of disinfection by-products.

While metal and tile roofs are commonly used for rainwater harvesting in developed countries, 
our limited data set suggests that asphalt fiberglass shingle and cool roofs also might be 
considered for this purpose given the quality of harvested rainwater that they produced; 
additional studies of asphalt fiberglass shingle and cool roofs are needed to provide a robust data 
set on harvested water quality.