Water is life. And, unlike resources such as oil, there is no alternative once it runs out. If we are going to safeguard the world’s water supply, then we have to understand and manage every aspect of it. The preciousness of water should mean that wasting it is a crime, as is not fully exploiting information communications technologies (ICT) to intelligently ensure water security for the whole world.
Water is a limited resource, but recent advances in science and technology, and, in particular big data and the Internet of Things (IoT), provide opportunities to prevent its waste. By gathering detailed measurements and leveraging analytics, it is possible to develop an end-to-end picture of our water supply to safeguard its quality and security.
First comes understanding
Big data and IoT will allow us to do it, but the understanding will have to come first. That’s why IoT must be part of the big data solution to the world’s water woes. A measure/analyze approach is the best strategy to improve water management.
Water is life. And, unlike resources such as oil, there is no alternative once it runs out. If we are going to safeguard the world’s water supply, then we have to understand and manage every aspect of it.
First, we must decide what we are going to measure. Different situations require different strategies — cities versus farms, for example. One obvious measurement might be the flow rates in water mains. Some cities have started to deploy automatic meter reading to collect data from customer water meters. In irrigation systems, it might be the water levels in canals. Often, those systems are distributed across many hectares of land, complicating the measurement infrastructure needed to track water movement from the source, such as a reservoir or river, to its ultimate destination. Just as importantly, we need to be able to measure how much water is lost along the way.
IoT devices and their associated ICT play an important role in performing the actual measurements and sending the information back to the analytics tools on a steady basis. Sometimes even a simple IoT device that records when a pump is running can provide crucial data for the broader picture. We can’t dismiss the communications requirements; in some cases, the systems to transmit the information may be costlier than the actual measuring devices. But many real world applications of these technologies are already emerging: For example, the city of Long Beach, California has reduced its water consumption six percent since the drought started. These new meters can detect illegal watering in real time and have helped to cut some homeowners’ use by 80 percent.
Knowing where and why water is being wasted is key to managing it in an agricultural environment, whether it’s to nurture seeds or sustain livestock. Within a city, it might be finding dripping faucets or leaking fire hydrants so they can be repaired. Additionally, the amount of water that never reaches customers and is lost due to crumbling infrastructure cannot be underestimated.
Next up, analytics
Recent advances in science and technology, and, in particular big data and the Internet of Things (IoT), provide opportunities to prevent water waste. By gathering detailed measurements and leveraging analytics, it is possible to develop an end-to-end picture of our water supply to safeguard its quality and security.
The second part is the analytics. Whether on farms or in cities, it is important to know what is normal for a particular water system. Historical measurement data can be compiled to derive a baseline model. Individual measurements don’t work in isolation; analysis based on multiple data sources provides a better picture of the overall water situation for a given area. Other observations such as current temperature and weather history can supplement the direct measurements. Sophisticated change point and anomaly detection algorithms are essential and critical. Once the model is built, the system can generate alerts and notifications when there’s some sort of deviation from what’s expected. Anything abnormal can be relayed to the appropriate people who will now have actionable information to take corrective steps. These systems are most effective with a continuous stream of measurements so they can build a comprehensive model of the system state. Infrequent or irregular measurements complicate the analysis and render it less effective.
The impact of finding one leak can be enormous; if it goes unchecked for days or weeks a tremendous amount of water may be wasted. If a leak can be detected quickly a great deal of water loss can be prevented. Similarly, identifying particular water customers or zones that have anomalous consumption patterns can make a large difference in overall water usage. It may be that a water main is leaking and wasting water to an entire subdivision. And then imagine if residents had up-to-date access on how their municipal water system was working: they may be more inclined to conserve, for example, because they know the reservoirs are running low, and otherwise the city would have to buy water from another municipality at higher rates. That’s where data can help control water usage. On an individual level, if a resident can see that their usage is substantially above the average, they may feel social pressure to change their behaviour.
The nature of big data
Bringing together everything that is needed will be a multi-disciplinary endeavour — that’s the nature of big data.
Bringing together everything that is needed will be a multi-disciplinary endeavour — that’s the nature of big data. A lot of parts already exist in other industries, including sensors, IoT-level embedded computing, wide-area networking over Wi-Fi or cellular, and the cloud computing-based analytics to crunch the data that is collected.
Ultimately, water management, quality and security can only be achieved through the ability to make decisions based on end-to-end analysis. Continuous reliable data collection, communication and analysis leveraging big data generated by IoT will enable intelligent decisions at the right time, in real time.
About the Authors
Ron Hiller is the founder of BLX.io, which was established to advance agricultural automation. Previously, he founded Quantiva to bring statistical machine learning techniques to web performance management and has led a variety of innovative software development, networking, and performance analysis projects at Bell Labs and other companies. An IEEE veteran of 35 years, he is a member of the Big Data Community as well as a long time member of IEEE Computer, Cloud Computing and Communications Societies.
Dr. Mahmoud Daneshmand is professor of business intelligence analytics at Stevens Institute of Technology. He is an expert in big data analytics, Internet of Things/sensor and RFID data streams analytics, data mining, machine learning, probability and stochastic processes, and statistics. He has more than 35 years of teaching, research and publications, consultation, and management experience in academia and industry, including Bell Laboratories, University of Texas, University of Tehran and New York University. Daneshmand holds key leadership roles with IEEE Journals Publications as well as IEEE Major Conferences. He is a member of the steering committee for IEEE BDI; leader of IEEE Big Data Standardization; and, chair of the steering committee of the IEEE IoT Journal, among others.