Summary

The public’s exposure to disinfection byproducts (DBP) in swimming pools is of growing concern globally. DBPs are the inevitable result of chemical reactions between the disinfection agents (e.g., chlorine) added to swimming pool water and organic or nitrogeneous matter that is naturally present or introduced by bathers. DBP formation and concentrations in pool water and the ambient air (for those that are volatile), vary according to the type of treatment used. However, no data are currently available to systematically compare these treatment approaches under controlled conditions in a single pool. This study aimed to provide such data. The objective was to explore the impact of various water treatment processes on DBP levels in swimming pool water and ambient air in order to determine the effectiveness of each and to help identify best practices to reduce exposure to them.

The swimming pool used for this research enabled the various procedures of its treatment system to be put into operation for variable durations, both separately and simultaneously. Two campaigns to measure DBP concentrations (water, air) were carried out (campaigns A and B) to understand the impact of four different changes made successively to the standard water treatment process, at different times of the year and under changing attendance conditions, i.e.:

T1: additional activation of ultraviolet rays (UV);
T2: halt of air stripping with continuation of air extraction from the buffer tank;
T3: halt of air stripping and suppression of air extraction from the buffer tank;
T4: suppression of the polyaluminium silicate sulphate (PASS) flocculant.

The first stage in the analysis of results took into account the impact of every change in treatment for each class of DBP examined successively over each of the two campaigns. UV rays have a marked impact, causing a potentially high increase of emerging DBPs, e.g., halonitromethanes (HNM), haloketones (HK), and trihalomethanes (THM) in the water and subsequently in the air. However, they also contribute to reducing the level of chloramines in the air and of N-nitrosodimethylamine (NDMA) in the water. The results highlight the positive impact of air stripping in reducing volatile contaminants. The use of PASS does not appear change the presence (quantity and speciation) of DBPs, except for the THMs, which appear to increase slightly with the use of this flocculant.

The second stage in the analysis of results details specific aspects related to variations in contamination levels between the various sampling points. In particular, it deals with: (i) exacerbation of contamination levels in pool water in relation to the potable water network that supplies the establishment; (ii) the existence of seasonal variations in contamination levels (which basically doubled from one campaign to the other); (iii) the relationship between DBP levels measured at different times at a water intake situated between the end of the treatment process and the outlet into the pool with the levels measured directly in the pool; (iv) the existence of spatial variations in air contamination in the swimming pool.

To the best of our knowledge, no other data have been obtained using this original and systematic experimental approach. However, this type of data is clearly relevant to the always desirable and strongly supported goal of reducing exposure to DBPs in swimming pools and developing a better understanding of the dynamic of these contaminants. Managers are sure to find the information very useful in implementing more effective treatment processes and strategies (developing better practices), in particular, to better protect the health and well-being of swimming pool staff, given the concerns surrounding DBP exposure. In that sense, the investigation results will also raise awareness of the need to make choices in treatment procedures that take into account the specific context of each swimming pool and the related issues, in this case, DBP contamination.

Note that this is an exploratory study, and therefore, the results cannot be generalized to validate (or invalidate) the use of different processes that have been considered in other contexts. For that, further study is required.