When it comes to water safety, not all irrigation is created equal. Our recent study delves into the complex world of water reuse regulations, focusing on a critical aspect: how do we validate that reclamation plants are able to achieve performance targets necessary to provide water that is safe enough for agricultural use? The answer lies in quantifying so called log-reduction values, or LRVs, which quantify how effectively water treatment removes certain groups of microorganisms. One LRV equals a reduction of 90%, while four LRV reduce microorganisms by 99.99%.
The European Union's Regulation 2020/741 requires that 90% of water samples meet stringent safety standards. But here's the catch – the regulation doesn't specify exactly how to prove compliance and validate these 90%. No information is provided about how many samples have to be analyzed or which method should be used to evaluate the data. This ambiguity has left scientists and water treatment plant operators scratching their heads, trying to determine the best way to demonstrate their treatment plant meets performance targets.
Hence, we compared different statistical approaches to validate this 90% threshold. Our findings reveal that the method chosen can significantly impact whether a water treatment plant passes or fails validation tests. It's not just an academic exercise – the consequences are real. A false positive, which indicates the treatment performance is sufficient while in reality it is not, could lead to unsafe water being used, while a false negative, which assesses the treatment performance to be insufficient while in reality it is not might unnecessarily shut down a perfectly good treatment facility.
The study examined several statistical techniques, from straightforward binomial evaluations (coin flip) to more complex tolerance intervals. Each method has its strengths and weaknesses, but one approach stood out: Bayesian tolerance intervals. This method proved particularly suitable, and flexible for handling real-world complications, like unequal sample sizes, non-normally distributed data or measurements below detectable limits.
To put these statistical tools to the test, we collected data from a large wastewater treatment plant in Germany over more than a year. We analyzed samples for three key indicators of water quality: Escherichia coli bacteria, spores of Clostridium perfringens and somatic coliphages, a type of virus that infects bacteria. By applying different statistical methods to this real-world data, we could see how each approach performed under real world conditions.
One of the most intriguing findings was the benefits of using tolerance intervals. Unlike percentile calculations based on point estimates, tolerance intervals account for sample size uncertainty. This means they can provide reliable results even with smaller sample sizes – an effect that could lead to economic advantages in cases where the treatment performance highly exceeds existing regulatory values.
So, what does this mean for the future of water reuse? Our study suggests that tolerance might be worth considering for setting standards for analytic process validation. By incorporating tolerance intervals into legal guidelines, a more robust and flexible framework for ensuring water safety could be created. This could lead to increasing trust in water reuse practices, a critical step for its further adoption, and thus for addressing water scarcity issues around the globe.
But the implications go beyond just process validation. The statistical methods explored in this study could potentially be applied to other areas of environmental monitoring and public health, where high percentiles are used for setting quality standards, like bathing water or irrigation water quality assessment.
As we face increasing pressure on our water resources, the ability to safely reuse water becomes ever more crucial. Our research not only provides practical tools for validating water safety but also highlights the importance of rigorous statistical analysis in environmental regulation. By refining our methods for ensuring water quality, we're not just contributing to the discussion of the best available science for ensuring microbial safety– we're paving the way for a more sustainable and water-secure future.
KWB authors
