As featured in Waterline Spring 2022

Using Harvested Rainwater:
Should We, or Shouldn’t We?

Giles Green FWM Soc., Associate Technical Director at Zeta Compliance Services, former chairman of the WMSoc
Technical Committee and co-author of the WMSoc technical guide on safe management of harvested rainwater
shares some thoughts on the subject.

Hardly a day goes by without there being a programme on the television illustrating some wonder of the natural world and we have become so accustomed to the spectacular that it can seem a little disappointing if there is not something new to science or, at least, never filmed before. Sadly, almost all these programmes end on a serious note, warning that all this awe-inspiring magnificence is in serious and imminent danger of being degraded or lost completely owing to mankind’s activities causing habitat destruction, pollution, climate change or a pernicious combination of all three. It is not, however, the purpose of this paper to discuss our carelessness and its effect on the glories of nature, but one particular ramification which has been exercising many in the facilities management sector and that is how to assess the risks associated with harvested rainwater and to manage them safely.

A Good Reason

It is common for new building projects to include a rainwater harvesting facility, reportedly often as a condition of planning permission and this is logical when we consider that water demand is generally increasing, both per person and with more people to use their tonne-plus per week. This, combined with what seem to be more erratic patterns of rainfall leading to episodes of excess and shortage, means there are bound to be periods when the supply of water held in reservoirs is put under real strain; indeed, it is said that in much of the country consumption exceeds supply in the summer so reservoir levels typically fall from April to autumn. Whilst it is obvious that harvesting rainwater is also subject to weather variation and a prolonged period of low rainfall will empty rainwater collection tanks as well as reservoirs, harvesting and storing rainwater is equivalent to increasing effective reservoir capacity. There is, however, a problem, which is that many harvested rainwater systems are not used because the risk assessment indicates they are highly dangerous and because there is little guidance on how to operate them safely.

Legionellosis Risk Assessment

In the UK we have robust health and safety at work legislation and we have become generally, and quite rightly, conscious of waterborne microbiological risk from legionella. As a consequence, we look to our water systems to be hot or cold, never warm or cool; active not stagnant, though we sometimes conflate physical movement with freshness and; clean, sometimes even categorising a small amount of fairly inert sediment in a tank of stone-cold water with a high throughput as being as a serious and urgent danger. We often then pay rather less attention to aerosol generation and exposure and we frequently completely overlook the effectiveness of controls and how well they are implemented and managed.

If this rather critical summary of the shortcomings of assessments of one, clearly defined and well-understood, risk has any truth in it (and I suggest that it does), it has to be a concern that assessing the risks from all hazards in a system which has no temperature control, is inherently susceptible to contamination and stagnation might lead to greater confusion. It was with this in mind that the WMSoc produced its guide to the use of harvested rainwater systems, published in March 2022.

It is often said there is a legal obligation to carry out a legionellosis risk assessment of any water system, somewhat implying there is something unique about legionella. In fact, there is no law specifically requiring a legionellosis risk assessment and the laws which create the obligation apply equally to any and every other pathogen which might be present. The difference is that legionella thrive in many man-made water systems and have led to outbreaks of serious illness whilst most others do not and have not. Nonetheless, it is useful to consider legionellosis as a particular case and assess the risk by appraising the risk factors which are set out in BS 8580 1 and to consider other pathogens together in a more general way.

Legionellosis risk occurs when legionella infect the lungs and the means by which that risk can be assessed is described clearly and in detail in BS 8580 1. First, the likelihood of legionella entering the water is considered, then the likelihood of it increasing to significant concentrations in the system; next is the degree to which the water is dispersed in droplets fine enough to be inhaled, an aerosol, then; the likely extent of exposure to that aerosol. Finally, the susceptibility of whoever might inhale that aerosol is considered.

When this approach is applied to rainwater harvesting systems, it will often be the case that one or more of the risk factors is almost completely absent. For example, the occurrence of legionella in water which has condensed out of air seems unlikely, even though it is the case that rain does contain impurities. Once the water has been captured and is in an underground collection tank it is not likely to experience legionella growth conditions, having arrived at a temperature considerably below 20°C and having no source of heat above ambient temperature. Tanks above ground which are exposed to sunlight, however, are likely to gain heat and might even form a thermocline with a layer of warm water floating above a cooler mass below, making temperatures taken via drain valves meaningless in terms of legionellosis risk.

Considering these two risk factors together, could the risk assessor reasonably dismiss capture and storage in an underground tank as contributing to the risk and accept a collection tank with considerable levels of contamination as being satisfactory? And if conditions disfavour growth, how would stagnation contribute to the risk?

The protection against heat gain and increased temperature might cease once the water leaves the collection tank and growth might become significant in distribution systems. This will apply within buildings where water is used for toilet flushing for example, but it will be particularly pronounced where the water is used for irrigation because water in pipes exposed to even weak sunlight will quickly become much warmer than the air (just think about how it feels to step out of and into shadows on cold bright days) and that is a risk factor easily overlooked.

Aerosol formation at the point of capture and in the collection tank need to be recognised in the risk assessment but can be designated as a low or even negligible legionellosis risk factor if it is accepted that occurrence and growth are both insignificant (and taking into account that the tank is effectively enclosed). Aerosol formation at the points of discharge, however, is likely to be more important because the water will have passed through the distribution system where growth might be more favoured and because it is where the greatest degree of exposure takes place. Flushing a toilet or urinal both break the water surface and must be considered to create some aerosol, even though the main characteristic is the formation of a surface flow of water and coarse droplets; drip irrigation might also generate some fine water droplets but this is minuscule compared to the aerosol generation of spray irrigation or pressure jetting. There might also be other industrial processes which use harvested rainwater and each will have its own aerosolgenerating tendency and each will need to be considered accordingly.

Exposure to aerosol is a risk factor which often seems to be overlooked, both in risk assessments and in schemes of control and when it is considered it frequently leads straight to the use of personal protective equipment as a control measure, which the law explicitly states should be the last (actually the fourth) resort and in addition to, rather than instead of, the others.

Finally, BS 8580 1 advocates considering the susceptibility to infection of those who might be exposed, and on that subject, there is a clear message from the Department of Health, which says in Healthcare Technical Memorandum 04-01 that rainwater should not be collected for use on, or in, healthcare premises, though that is not explicitly in the context of legionellosis risk and that introduces the question of other pathogens.

Assessing the Risk from Other Pathogens

When writing the WMSoc guide to safe operation of harvested rainwater systems, non-legionella pathogens were discussed by the drafting group and this led to some alarm, not least when prospective candidate organisms were supplemented by some which had been identified by analysis, until it was realised that it is a schoolboy error to fail to distinguish between hazard and risk.

It is true that capturing rainwater incorporates a mechanism for collecting just about anything which might be lying around, but it is also true that it provides a very good way to wash the capture surfaces so most of the contamination can be diverted. Another important consideration is that the organisms likely to be washed into a rainwater harvesting system are typically not types which thrive in cold water containing few nutrients; this is true of most faecal organisms, which probably make up most of the likely harmful flora, and these die out quickly.

Whilst BS 8580 1 is specific to legionellosis risk assessment, the principles can easily be adapted to other organisms and the infections they can cause and this approach is used in BS 8580 2, which considers risk from waterborne pathogens other than legionella.

A good example of this is faecal organisms which can cause gastroenteritis which might well be present on rain capture surfaces and could enter the collection tank. Just like for legionella, this does not provide a favourable environment for growth but, unlike for legionella, it is positively hostile, so the first two risk factors of occurrence and growth are respectively higher and very much lower. That is not to say all organisms will die off, far from it, but the flora will quickly develop a cold water environmental profile rather than one with large concentrations of gastroenteritisinducing species.

The next risk factor in BS 8580 1 is aerosol generation and, whilst that might be significant, it needs to be broadened to include the routes by which other pathogens infect and this will vary from pathogen to pathogen, but it is likely to include inhalation, ingestion (including via the nose and eyes) and entry via wounds.

Exposure is clearly linked to the infection routes and is likely to be minimal in many instances. For example, it would be exceptional for infection via a wound to be a significant risk factor where harvested rainwater is used for flushing toilets but a risk assessment should consider unusual but foreseeable circumstances and that would include maintenance and repair, where wound infection is very much more likely and would require control measures. Jet washing with harvested rainwater, whilst advocated by one manufacturer in television and magazine advertisements a few years ago, probably ticks every box on the exposure scoresheet with sharing it with the neighbours as a post script.

Susceptibility to infection is rather difficult to appraise usefully in any detail, which returns us to the Department of Health’s statement which opened this section of the discussion.

Other Hazards

There are hazardous contaminants other than pathogens and, just as health and safety law applies to any such organism as much as to legionella and requires risk assessments, so it does to other hazards. These are categorised in the WMSoc guide as chemical, radiological and exceptional. Chemical hazards include lead from flashing around roofs, zinc from galvanised steel roofs and petroleum residues in car park run off. Other chemical contamination is typically unlikely and would usually be considered as exceptional.

The risk from chemical hazards can be assessed using the strategy from BS 8580 1, adapted for each hazardous contaminant and the means by which it can be hazardous, for example considering how lead can accumulate in living tissue if it is taken up by plants or ingested by animals drinking contaminated water.

There were two surprises for me when the guide was being discussed, the first being radon accumulating in underground tanks and the second that it had not occurred to me. Its potential as a hazardous contaminant in harvested rainwater was identified from experience in another but related situation where it has been discovered that venting is not effective at removing it (in part, no doubt because it is eight times denser than air). Being an alpha emitter with a half-life of 3.8 days, radon is particularly pernicious and is considered to be a significant cause of lung cancer in areas where it occurs and risk assessment, though making use of an adaptation of the now well-established rationale, will typically require specialised support in the form of geological information and radioactivity measurement.

Safe Operation

The starting point of safe operation has to be a sound risk assessment, and it is hoped the comments above have provided some useful guidance on carrying one out. If the outcome of that assessment is that the risk is insignificant, no more need be done other than ensuring the situation does not change and increase the risk. What is more likely is that the risk will be significant, in which case UK law requires exposure to hazardous substances to be eliminated, failing which, reduced to a safe level.

When devising ways in which exposure can be eliminated, the detail of the risk assessment and clear and separate consideration of each risk factor is particularly important. For example, if the risk assessment identifies a nearly zero risk of legionella occurring in rain and an effective low-flow diverting system to wash accumulated contamination off the capture surfaces rather than allowing it into the collection tank and that the tank is not susceptible to heat gain so it is too cold for most gastroenteritis-inducing organisms to survive, it might conclude that the water stored in the tank constitutes a negligible microbiological risk under normal operating conditions. In such a situation, there is nothing to eliminate to which to apply control measures. If it then identifies potential for heat gain and increased temperature in the distribution system and concludes that might create growth conditions for legionella (which, whilst unlikely, cannot sensibly be deemed to be completely absent at all times), that is where controls would be required and, for example, a UV disinfection unit treating the water drawn from the collection tank would not reduce the risk where it actually occurs.

In practice, of course, the scheme of control will tend towards caution because there is an underlying legal duty which requires all means to be applied to reduce risk to its lowest reasonably practicable level and an established interpretation that a risk can be latent rather than currently manifest as the presence of a specific hazard at a particular time. Whilst that is to be supported, it must always be coordinated with the risk assessment to check the controls are effective and do not distract from any risk factors which are more significant by creating a false sense of security.