Data, Data, Everywhere and Not a Drop to Drink

September 18, 2015 • Blog Post • By Atlantic Re:think


  • To help California residents conserve resources, companies are using Big Data to track water consumption habits
  • Companies are looking to connect Big Data solutions with the Internet of Things to create machines that can automatically conserve water and prevent drought conditions from spreading to other states

Updating Californias water infrastructure for the digital age may be the key to surviving the states historic drought


Dan Sonke thought his family was doing okay. He and his wife live in Roseville, a suburb of Sacramento, where they are raising four children. Sure, theyd heard about Californias drought crisis and knew it was important to conserve water, but their two-story, 1930s home had been renovated before they moved in. It had modern features like low-flow toilets and faucet aerators. They figured they were in good shape.


One day, however, Sonke received a notice that he could go online for an in-depth report on his water usage. He could see, at a glance, how his home compared with average households as well as with the most efficient homes his size. While his family wasnt quite in the top 20 percent, it was close.


"My first thought was, Okay, not too bad, at least were better than average, he says. Then I sent it to my wife, and her competitive streak burst out. Her first response was, What are we doing wrong?!"


The Sonkes are among the millions of people affected by Californias historic drought, which has heightened awareness of water as a limited resource and created a surge in efforts to evolve the technology of water-resource management. To be sure, the environmental factors are daunting-researchers say theres 40 percent less snowpack in the Sierra Nevadas this year than last, and they predict even less rain for California in 2016. But part of the problem is also that Californias water-management infrastructure is aging-it's often 20 to 40 years old or more.


Widespread use of smart meters and major infrastructure changes are still years, if not decades, away, analysts say. Which is why leading companies are focusing on Big Data analytics solutions that offer more precise data on how current water management methods work and shed light on ways to make them more efficient. The approach is being explored not only at the residential level, with people like the Sonkes, but also on large-scale, water-intensive farms that grow everything from grapes to almonds and inside water utility plants, which could someday help other states that are predicted to experience similar water crises within the next decade.


Leading companies are concentrated in the water-energy nexus, a term the California Energy Commission uses to describe the links that tie nearly 20 percent of the states electricity consumption to water-related uses. The idea is that if youre going to treat, transport, heat, use or dispose of water, then youre going to need some form of pump or other electricity-driven device to do it. All those devices have at least basic sensors and meters generating data that could be harnessed to identify and stop water leaks or stem overuse. Combine those insights with other existing data, and you can present it all in a way that inspires people to take action-including by tapping into their competitive spirit.


One company at the forefront of the residential sector is San Francisco-based WaterSmart, whose data is behind what Sonke saw on his bill. WaterSmart combs through about 760 million data points per hour and boils it down to the information residents receive. In Sonkes case, seeing the comparisons to similar homes made him look for problems-such as a leak he discovered in his backyard irrigation system. It also changed his familys behaviors. He began having his kids put buckets in the shower to collect the first sprays of cold water while waiting for the temperature to heat up. The buckets are now used to water outdoor plants, driving down use even more.



When Sonke clicks on his report today, he sees that while the city wants his water use to be 28 percent lower than in 2013, his family has already achieved 62 percent less usage.


"We do occasionally post things on Facebook: Weve done it, you can too," he says. "You can see how you compare to that top 20 percent. Honestly, were not there yet, but Im sitting here thinking, Theres got to be a way we can do better."


That's exactly the attitude the Big Data presentation is designed to inspire, says WaterSmart CEO Robin Gilthorpe, even for someone whose home infrastructure is nearly a century old.


"While we would love it if every meter in the entire nation and the world were a smart meter-one that's giving us data every five minutes and beaming it to us automatically-only about 18 percent of utilities have smart meters", Gilthorpe says. "We can work with smart meters and conventional meters. It doesnt matter. Because were doing the patterning on such a large scale, we can do analysis that gives us a pretty accurate idea, even without fine-grain data."


The gallons of water that families like the Sonkes are saving add up, especially when similar applications are used for large-scale agricultural properties. The water dribbling out of leaky pipes alone reduces total available drinking water by about 15 percent, according to reports in the San Jose Mercury News.


But the most dramatic potential savings can be found in Californias massive agricultural regions, which account for 80 percent of the states water use.


PowWow Energy, based in Redwood City, just won a $2.3 million grant from the California Energy Commission for on-farm testing of its Irrigation Advisor software. Like the WaterSmart concept, Irrigation Advisor targets growers by combining fragmented data-from traditional and smart electric meters, U.S. Department of Agriculture soil mapping, aerial imagery, local soil sensors and more-and then sending it to the cellphones of growers in almond fields, at wineries and on horse farms in a way that makes sense.


"We don't approach them directly and say, You have an energy problem. That makes them think they have to call the electrician", says Olivier Jerphagnon, PowWow Energys co-founder and CEO. "For example, if youre out hoeing and you use a megawatt of energy, you have a problem, but what does that mean? Instead, you should tell them, Hey, you have a leak, and this is where it is. Go fix it. If you give them that information, it becomes very valuable and they will go and fix it. They love looking at how much they are saving. The farmers are the original environmentalists." While companies such as WaterSmart and PowWow Energy are working with the data that exists-and positioning themselves for whatever comes next-some water-management utilities are learning that infrastructure installed as recently as the late 1980s can be dramatically improved with relatively little investment. At the Monterey Regional Water Pollution Control Agency, for instance, engineers like Tom Kouretas have been working with Oakland-based Candi Controls and Orangevale-based MC Engineering to find ways that Big Data analysis can solve everything from wasted lighting to broken water pipes.

Kouretas became aware of the possibilities when the water plant started taking a more in-depth data dive into its monthly power use. Seeing how data analysis could pinpoint wasted electricity, he started thinking about how it might also improve the water-delivery infrastructure. The current system is so low-tech that it starts with a grower faxing a water order to the plant, promising to take a certain number of gallons at a certain time of day from a "turnout", kind of like a giant spigot. The plants drivers then take daily, manual readouts to see if the growers did what theyd promised. If too many growers are running late or plug into the turnouts at once, the infrastructure strains and water lines break.


"What I expect to develop in the next year or two is a cloud-based, ag-water order system that has instantaneous monitoring," Kouretas says. "One more possibility - and this is really exciting - is to install some kind of predictive system to get hold of pressure spikes in the distribution system. Once that goes into the cloud, you can have a real-time model in the screen of the distribution system and see where its experiencing pressures."


"That's really where we want to get in using data instantaneously to keep the system healthy and minimize interruptions in service."


Even if utilities take similar steps one by one across California, and then nationwide, the results could be enormous. Gilthorpe says that across America, about 4,000 large water utilities and another 40,000 smaller ones are producing an incredibly fragmented data set, because each built its systems in its own way. With todays technology bringing all those data points together, he says, solutions are possible today that were unimaginable 10 years ago.



"The cloud gives us the ability to generate large economies of scale even when the implementation is small", Gilthorpe says. "If we get a request from Los Angeles to do a particular type of report or analysis, and its not really something thats strange or unusual, chances are that it will be useful for the other 4,000 or 40,000 utilities."


"That's important", he says, "because what is being called the California drought today will spread to other states. The problem is partly about water, but its primarily about population growth. When California experienced a severe drought during the 1970s, it had about 25 million residents. Today, that number is 40 million, compounding demand."


"That pattern is sort of repeated across the country", Gilthorpe says, citing a May 2014 report from the U.S. Government Accountability Office stating that freshwater shortages are expected in 40 out of 50 states within the next decadeeven if conditions remain average.


The next step in finding Big Data solutions will be to tie the newly understood analytics into the Internet of Things so that machines can act on Big Data automatically. Its one thing to receive a text message on your smartphone saying that a water pump is leaking on your farm. Its quite another for the app to talk to the pump and turn it off automatically. Jerphagnon and Gilthorpe both see the basics of that technology being deployed in the next five years or so.


"We're not there yet because the ecosystem of devices isnt there yet, but over the next couple of years, youll be hearing more about that from us", Gilthorpe says. "Once we can knit all that stuff together, it starts to get really interesting."


HPE CEO Meg Whitman: Heres to a More Connected Future

Blog Post

HPEs Spaceborne Computer Successfully Powers Up in Space and Achieves One TeraFLOP

Blog Post