Saferain Logo   Rainwater Harvesting: An Alternative to the Roof Washer

Rainwater Harvesting






When rain falls on a roof it usually starts gently and then increases in intensity. This rainfall will wash adhering pollutants such as soot and bird manure into the storage cistern causing faecal contamination. With leaf debris present, "ponding" of stagnant water in gutters can occur making matters worse.

It is for this reason many authorities discourage or forbid the use of urban collected rainwater for human consumption. However, due to water shortages, rainwater collection for non-potable uses such as toilet flushing etc is beginning to be encouraged.

Where potable rainwater systems are allowed, the order in which components are installed are:

(a) roof;
(b) gutters (screened);
(c) in-line basic type filter (optional);
(d) roof-washer/first flush diverter;
(e) tank/cistern;
(f) UV or O3 sterilizer;
(g) micron filter.
In some cases disinfectant (usually chlorine) is also used.


For serious researchers on the topic "Rainwater Harvesting", how many times have you read "It is well known/generally accepted that the first volume of rain after a dry spell if allowed into the cistern can cause pollution." And so it can. For this reason roof washers were developed and manufactured with the premise that … "A rule of thumb is that 1 gallon of water is sufficient to trap for every 100 sq.ft. of foot area. "

"Studies have shown that the first mm of rain will remove most contaminants from a roof. " What studies? What's "most"? It would depend on what is on the roof. If the roof has salt deposits from being close to the sea then the above statements might well be true, but what about creosote deposited by a fireplace? What about dust and leaf debris that is furthermost from the downspout? If your house has just been crop dusted inadvertently or otherwise, do you think a gallon per 100 sq.ft. would be enough to stop the pollutants from entering the cistern?

These roof washers, whether they are first flush pits, holding tanks, settling tanks, roof washer boxes etc can only be described as containers. All containers work the same way; they collect the first volume of water and presume that the rest of the water going to the cistern will be free from any pollution. What can happen instead is that the contained water can mix with the cistern water causing the cistern water to become a biological hazard.

When I speak of container and ball diverters I refer to those units that isolate the first flush of water by means of a float ball. Instead of using a float ball why not use a sink ball? Instead of containing the initial water why not dump it either to the storm drain or to the cistern overflow? Instead of having a large container that's awkward to install and heavy, why not have one that's just 7 lbs when operating? Instead of having a device where you have to unscrew an access to clean out the sludge and debris deposited why not one which self-cleans for all debris under ½ inch? And finally instead of a device that diverts a fixed volume why not one which has adjustable automatic diversion?


In 1993 SafeRain was invented. SafeRain is a first flush water diverter which combines what I call a “differential volume” action with an extensible member. There have been many “differential volume” first flush diverters but none appear to have had much success. Briefly “differential volume” device intercepts a portion of the rain inside a body (sphere or cylinder) which is balanced by a counter-weight.

When the combined weight of the intercepted water and the valve body exceed the counter-weight then the body will move in one motion to seal a valve seat causing diversion sometimes called the “pilot” principle Instead of a counter-weight, SafeRain suspends the valve body (ball) from a shaft with a length of special water-proof elastic.

The shaft enables the ball position relative to the seat to be adjusted. This changes the flow-rate at which the ball closes the seat. SafeRain was the first to employ “an extendible member” (elastic). Above this shaft is a flat plate with holes in it (baffle plate). This attenuates the flow of incoming water into the ball. It also prevents premature closure when there is a recent surge of water.

Recently, there has been a copy of SafeRain released into the US market. This version uses a ball (looking like an exact replica of a SafeRain ball) attached to a shaft by a spring. A spring cannot be wound up so no adjustment is possible. Also no baffle plate directly above the ball is provided so that in the event of a sudden surge the ball is more likely to close prematurely.


When rain starts it will either go around the sides of the baffle or hit it (Figure 1). Some of this rain will go through the holes in the baffle and some of this will enter the ball. Typically about 0.0005 of the rain enters the ball and if this entry rate into the ball is faster than the drain rate then the ball will begin to descend eventually closing the valve seat. The water will then divert to the cistern (Figure 2).

Figure 1

Figure 2

When rain ceases the water in the unit will slowly drain out of the ball (Figure 3).
When it is empty the elastic will pull it back to its original position (Figure 4).

Figure 3

Figure 4

By winding the ball up or down the flow rate at which the ball closes can be adjusted. The standard model has a closing range between 5 litres/minute and 35 litres/minute. The "normal" setting is about 10 litres/minute - 15 litres/minute. Of course it can be closed or opened manually. The leakage rate for the ball when closed is about 50-100 ml/minute.

In drizzle conditions the unit will not work automatically. This is desirable since the roof would not be swept of adhering materials. A sudden rain-burst would not fill the ball immediately since the ball floats and the baffle plate attenuates the immediate impact of the incoming water. This enables the furthermost debris etc to reach the downspout before closure.

The above description is schematic only. There are other design features not mentioned that are included for smooth operation. The design can be adopted for inflow from any direction. The lifespan of the elastic is 7+ years. The pressure drop across the unit is minimal. This is important in tropical and monsoonal areas. SafeRain is not a leaf filter although it does have an internal ½ inch mesh stainless steel screen.

For urban use rain must fall firstly to clear the air of smog etc, and then go to clean the roof. The efficiency with which it does the former depends on the intensity while the latter depends on the momentum of the moving rainwater across a surface (roof). Ignoring roof pitch we can equate this with flow rate.

The use of small roof washer boxes should be discouraged. They are ideal breeding grounds for microorganisms. The use of large settling tanks, holding tanks or any other container that allows for suspended solids to settle, although an improvement on the small roof washer boxes, should be avoided. To say that "if you can see it, it's pollution" and "if you can't see it, it's not there", represents a triumph of hope over logic.

The realisation that the first flush of water needs to be separated from a cistern if that cistern is being used for consumption has been acknowledged since the 1920's. The technology of these first flush systems since then has never strayed from the container/volume idea. SafeRain has designed a flow-rate first flush diverter that has taken rainwater harvesting out of the 20th century and into the 21st.

Trent Church

October 4, 2001

Recurrence Interval/Rainfall Intensity
A Sensible Alternative to the "First Flush" as a Design Parameter
Authored by: Thomas R. Adams, P.E. and Robert A. Strong. Jr.
Vortechnics Inc., 1997 Vortechnics, Portland, Marine

Web link:


* Tested by CSIRO: Report No. HH1665. Available from: Hydraulics Laboratory, CSIRO Division
of Building, Construction and Engineering, Gate 2 Delhi Road, North Ryde, NSW 2113
©2001 The images and text contained within this site remain the property of SAFERAIN