Sniffer dogs as an emerging approach for water leakage detection | Water Supply (2024)

Jiazhi Zhong;

Dongdong Chen

Article history

Advanced Search

Listen

Listen

Listen

Water network leakage control poses a critical challenge for the global water industry in its pursuit of Sustainable Development Goal 6.1, which aims to ensure universal and equitable access to safe and affordable drinking water for all.

The extent of water loss resulting from leakage varies across different countries. Across various municipalities within the United States, the leakage rate primarily inhabits the spectrum of 10–25% (Folkman 2018). Conversely, countries like Japan demonstrate commendable performance with leakage rates as low as 5% (Abd Rahman et al. 2018). Alarmingly, the global volume of NRW has been calculated to be 126 billion m³ per year. (Liemberger & Wyatt 2019).

Multiple factors contribute to network leaks, including aging infrastructure, insufficient maintenance, and human error. The consequences of network leakage are particularly pronounced in remote and less developed areas, where it can exacerbate local water shortages, pose public health risks, perpetuate water poverty, and contribute to inequity (Bhagat et al. 2019). Furthermore, the control of pipeline leakage presents a formidable challenge in shrinking cities, which often face limited funding and lack access to modern inspection technologies (Faust et al. 2016).

Currently, in response to the pressing need for pipeline leakage detection, engineers have developed numerous innovative technologies, such as smart water meters and hydrophones. However, these technologies often exhibit a high reliance on intelligence and information levels, making their application in complex and harsh environments challenging. Simultaneously, they struggle to provide adequate coverage for large areas, including rural and remote regions.

Sniffer dogs offer distinct advantages as they can swiftly and effectively cover large areas, including rural and remote locations where traditional leak detection methods face challenges. These areas are often inaccessible due to natural conditions or lack of device access. The deployment of sniffer dogs presents an emerging technology practice that holds immense potential for overcoming the limitations of conventional methods.

This paper review the progress of using sniffer dogs in water leak detection, summarizes and analyses the application and management models of sniffer dogs, and identifies several key issues for further research to promote the development and refinement of relevant studies.

Listen

Sniffer dogs are working dogs trained to utilize their exceptional sense of smell to detect specific substances, often surpassing other detection methods in terms of reliability and speed. For example, using an olfactometer, research showed that dogs were able to detect a target odour present in one part per billion in the presence of a distracting odour at a concentration of 20 parts per million (Waggoner et al. 1998), and it is reported that dogs can smell some odours at one part per trillion (Pearsall & Verbruggen 1982). Sniffer dogs have already found application in gas pipeline leak detection (Adenubi et al. 2023), COVID-19 virus detection (Eskandari et al. 2021), identification of specific substances in sewage (Reynolds & Reynolds 2018), and the separation of sewage from rainwater systems (Murray et al. 2011). More recently, their utilization in water leakage detection has gained prominence.

Listen

Listen

During the implementation of sniffer dogs for leakage detection in different regions, a range of diverse and effective operational modes have been investigated. These operational modes can be categorized into four based on two dimensions: whether sniffer dogs are orienteering used and whether the leakage event is clear.

Listen

One notable example is Western Australia, where local water companies employ a Springer spaniel named Kep for routine inspections. Kep utilizes its exceptional sense of smell to track chlorine in tap water, enabling the swift identification of leak points within a mere 10 min. During the 2020–2021 pipeline inspection missions, spanning approximately 3,800 km, Kep was responsible for covering approximately 250 km (6.58%) of the total distance. Moreover, Kep's efforts accounted for an estimated water saving of 197 million L (7.30%) out of the total 2.7 billion L saved during the year's mission (Orr 2021). Notably, sniffer dogs demonstrate a favourable cost-to-output ratio and have been found to be at least 10 times faster than traditional acoustic detection methods.

Listen

An exemplary case of this mode is observed in Queensland, Australia, where the local water company equips sniffer dogs named Halo and Danny with radio tracking collars during their working hours. The sniffer dogs begin their exploration from one end of the designated area, freely traversing and repeatedly searching the region. This process generates a map illustrating their behavioural trajectory. By analysing the behaviour patterns, the location of suspected leak hotspots can be identified. Subsequently, staff members employ methods such as triangulation to further confirm the precise leak point.

According to the tracking collars, three-year-old Halo can run at speeds of up to 45 km per hour, while 14-month-old Danny can reach speeds of 25 km per hour. They can complete detection over a hectare of land in approximately 20 min. In general, they can patrol approximately 20 km of pipelines per day (Woodward 2019). Compared to conventional methods, this approach significantly improves efficiency.

Listen

A notable application of this model is observed in France metropolitan. Local water companies have established data collection systems using smart water meters. During the period between 3 and 5 a.m., when domestic water consumption is typically zero or nearly zero, any abnormal increase in minimal water consumption indicates a leak. When a community is suspected of experiencing leakage based on this anomaly, Veolia, a prominent company, deploys sniffer dogs. Veolia's canine section currently consists of five sniffer dogs. In just a year and a half, these dogs have successfully detected 450 leaks across various provinces in France, resulting in a total water savings of 25,000 m³ (Lefèvre 2022).

Listen

An illustrative example of this approach is evident in Little Rock, Arkansas. The local water utility employs ASTERRA's satellite-based synthetic aperture radar, which gauges the soil's dielectric constant. Given the nuanced differences in dielectric constants among various water types, such as groundwater and tap water, it becomes feasible to construct a satellite-generated hotspot map of network leaks, as depicted on the right side of Figure4. Following this, sniffer dogs are deployed to validate the identified leakage points based on the map. The seamless match between the results of the work with sniffer dogs and satellites has yielded impressive results. In the local area, the overall number of leaks detected per kilometre has increased by a remarkable nine times compared to traditional methods (Boozer 2021).

Listen

There are approximately 20 sniffer dogs deployed worldwide for water leakage detection. To facilitate the application of sniffer dogs, different regions have adopted specific business management models, which can be categorized as follows: (1) government exclusive, (2) water company sharing, and (3) private lease.

Listen

This model is primarily observed in Australia, where the state water department collaborates with specialized training institutions to train dedicated sniffer dogs. These dogs are considered government employees and operate exclusively within the state.

Listen

This model is predominantly found in France. The dogs are typically assigned to various locations within the company and are regarded as shared internal resources. Veolia serves as an example, as their leak-proof dog team was established in 2019 by former French army dog trainers. Since its inception, the team has operated in multiple provinces across France as well as in international locations like Morocco.

Listen

This model is primarily seen in the UK and USA, where local water companies engage specialized firms to provide sniffer dog rental services when required. Examples include Thames Water and Scottish Water.

As a rule, sniffer dogs for government exclusive typically exhibit higher efficiency, given that these dogs receive tailored training for specific regions. Conversely, the other two modes can yield substantial cost reductions, but the transitions of dogs often necessitate an adjustment period. Consequently, overall efficiency may experience a decline. Striking a balance between economic considerations and efficiency is a decision that each region should make based on its unique requirements and regulations.

Listen

Listen

Although sniffer dogs have made some progress in detecting leaks in the water supply network, the detection accuracy is as high as 90–96%, and have been proven can find pipeline leaks up to 4 m deep (Hamilton & Jones 2012). However, current use scenarios rely on chlorine in tap water as a characteristic identification compound. There are two main problems with this method. First, the detection is affected by ground materials and backfill, and chlorine must be able to permeate to the surface, although only a very low level of volatilization is needed. Second, some remote and rural areas have not adopted disinfection processes, and some urban areas have not used chlorine as a disinfectant. Therefore, further research into characteristic odour fingerprints is warranted.

Taste and odour (T&O) compounds in water are a major concern for the water sector, often leading to consumer complaints. A review (Zahraei et al. 2021) analysed over 60 T&O compounds detected in water between 2005 and 2020, encompassing lakes, reservoirs, water treatment plants, and water distribution systems. The most frequently observed compounds were classified into four groups: (i) Geosmin and 2-methylisoborneol, (ii) Methoxypyrazine derivatives, (iii) Haloanisoles, and (iv) others. Based on a literature review (Zhu et al. 2022), studies have explored the impact of different treatment processes on T&O compounds in water.

Preliminary conclusions can be drawn from statistical analysis, suggesting that Geosmin, 2-methylisoborneol, halomethane, and bromoorganics may serve as potential odour fingerprints of tap water. Studies have already confirmed the effective recognition ability of sniffer dogs for Geosmin and 2-methylisoborneol (Browne et al. 2006).

Recently, per- and polyfluoroalkyl substances (PFAS) have been detected in drinking water systems in various parts of the world, presenting potential health risks (Teymoorian et al. 2023). With the considerable attention on the volatility of PFAS (De Silva et al. 2021) and the prior research that has utilized domestic dogs as exposure sentinels (Rock et al. 2023), it is plausible that PFAS could also serve as a distinctive odour fingerprint. This is indeed a captivating perspective. The advantage of this approach lies in its ability to simultaneously detect PFAS and similar emerging contaminants while pinpointing leaks in the distribution network.

It is important to emphasize that there is no universal standard for the selection of identification compounds, and careful consideration must be given to the local context when choosing single or mixed identifiers to improve accuracy.

Listen

Presently, the training of a sniffer dog typically spans from several months to as long as a year and a half. The accompanying expenses can range anywhere from a few thousand to several tens of thousands of dollars. This encompasses the acquisition of the dog, trainer fees, training facility costs, and ongoing expenditures for sustenance and healthcare during their service period. In most cases, the lion's share of these expenses is attributed to the procurement of the dog, hiring a trainer, and securing suitable training. A noteworthy reference case is Aqua Water Company in Texas, where the deployment of a sniffer dog incurred an investment of approximately $20,000. Over the course of less than a year, this sniffer dog successfully identified 150 leak points, making it a resoundingly successful investment (Stromberger 2022).

Today, training sniffer dogs typically follows two approaches. First, training involves using a pure reference odour, and secondly, training includes various odour mixtures where the target odour acts as a shared element (Fischer-Tenhagen et al. 2017). The training methods relied on positive reinforcement, employing dog toys and small food treats as rewards. Each dog received its preferred food as a reward. If a wrong indication occurred, no reward was provided. Sniffer dogs are only authorized to carry out practical tasks when their training success rate consistently surpasses a specific threshold, typically exceeding 80%. This serves as a vital step in guaranteeing the success rate.

While sniffer dogs already offer a distinct cost advantage compared to many data-driven methods, there is still room for further cost reduction through the continuous optimization of animal training methods and duration. This avenue holds significant promise as an interdisciplinary research direction, calling for the involvement of animal behaviourists.

Listen

Through the summarization and analysis of relevant application cases, management models, and an evaluation of current issues, we believe that, despite certain challenges associated with the use of sniffer dogs for this purpose – such as the time-consuming training process and the identification of odour fingerprints – sniffer dogs have undeniably emerged as a promising technology for detecting leaks in water supply networks.

The water sector is currently facing a dilemma, with traditional inspection methods proving inefficient and heavily reliant on human resources. Advanced technologies, while promising, often come with high equipment and technical requirements and still require manual inspection for final positioning. The use of sniffer dogs for pipeline leakage detection does not aim to replace the existing system but rather serves as an effective supplement of existing technology. Sniffer dogs can effectively bridge the gap between needs and technical capabilities, thereby supporting the implementation of other technologies.

Integrating sniffer dogs into the existing technology package can be accomplished as a cost-effective technology module. By conducting technical and economic comparisons, the role of sniffer dogs in local contexts can be comprehensively evaluated. In remote and rural areas, sniffer dogs can serve as a sustainable and economically advantageous solution for achieving integrated water management. In cities and developed areas with strong monitoring equipment and technical capabilities, sniffer dogs can be utilized as a supplementary tool to further enhance advanced technology applications. The convergence with technologies such as machine learning, artificial intelligence, and real-time pipeline modelling is anticipated to further unlock the potential of this approach.

The adoption of sniffer dogs in water supply network leak detection also underscores the need for disruptive innovation, emphasizing the importance of exploring low-cost and sustainable solutions that incorporate local and community knowledge alongside the development of cutting-edge technologies.

Listen

All relevant data are included in the paper or its Supplementary Information.

Listen

The authors declare there is no conflict.

© 2023 The Authors

Cited by