The uncertainty of extreme weather is spurring both altruists and entrepreneurs to finesse our forecasts.
Good morning. Some 2.1 billion of us woke up today and checked the weather on our phones before getting out of bed. The average person will look at the forecast again three to five more times; nine if you wear an Apple watch. What you may not realise is that your glancing interest in the weather is worth billions.
Weather is one of the biggest business opportunities in the world and grows with each heatwave and flood. A combination of catastrophic weather and seismic advances in technology has catapulted the industry’s ability to analyse and collect data in new ways and from every corner of the world.
The latest weather entrepreneurs are coming from academia and legacy weather outlets – even gaming, Google and the America’s Cup. They promise to revolutionise weather reports and deliver more actionable forecasts through AI, hi-tech balloons, ocean-going robots and mobile-phone masts. The race is on to monetise climate change and help businesses and consumers navigate The New Weather.
Billions of dollars are pouring into the space. The market is projected to be worth $4.6bn (€4.2bn) by 2025. A few of these emerging companies are bold enough to claim they can predict the future; reassuring claims of clairvoyance in volatile climatic times are a tantalising sales pitch.
No wonder graduating meteorologists no longer aspire to be TV weather broadcasters. At the Penn State Department of Meteorology and Atmospheric Science – one of the top-ranked schools for meteorology in the US – only three graduates in the class of 2019 went into television. Forty per cent went to private-sector jobs working in data mining, weather-risk management, forecasting and geospatial intelligence, or to work for energy companies. It must pay better than doing the weather in Dubuque, Iowa.
Jon Nese, associate head for undergraduate programmes, says, “This is not the meteorology of 20 years ago. I see opportunities expanding with the change in weather. In the past three years, five undergraduate students have founded and sold start-ups.”
It’s the ultimate white space: no one owns the weather. Historically governments collected weather data on a large scale. Private weather companies would then employ in-house teams of meteorologists and data scientists to re-skin and repackage the data and sell it to clients whose survival depends on precision forecasts. Farmers need to know the optimal time to plant corn before a frost, airlines need to optimise their flight scheduling in advance of a blizzard and Walmart needs to plan pre-hurricane inventory (stocking up on beer and Strawberry Pop-Tarts, which see sales rise seven-fold before a hurricane). What’s happening now is fresh, young and hi-tech; it’s the difference between a chubby, bumbling weatherman and an armada of sleek, orange, unmanned surface vessels.
Cut to: those Day-Glo sailing drones. Saildrone occupies a former Naval Air station hangar (a few hangars over from Hangar One Vodka) on Alameda Island in San Francisco Bay. Founder and ceo Richard Jenkins has managed to create a sense of wonder and emotional connection around the brand’s collection of ocean data. He has raised more than $88m (€79m) in funding.
His ocean drones are fitted with sea, sky and horizon cameras and sensors that take atmospheric and ocean measurements. Saildrone coo Sebastien de Halleux says, “It’s important because this data has not been collected before and most of the weather we experience on land originates in the ocean.” The Saildrone weather forecast supplements publicly available weather data with the proprietary data calibrated by its global fleet of ocean robots sailing from Pole to Pole.
Jenkins started building boats as a teenager and completed his first Atlantic crossing at 16. He built a wind-powered vehicle that shattered the land-speed record in 2009 at 202.9km/h in the Mojave Desert. He founded Saildrone in 2012 after applying the underlying design principle (specifically wing technology) of his record-breaking vehicle and using it as the basis for the Saildrone prototype.
Saildrone deploys more than 30 solar and wind-powered unmanned surface vehicles (usvs) that operate in the ocean autonomously for up to a year without fuel and without disturbing eco-systems. In August, a lone Saildrone headed for Cowes Week regatta in the UK and completed the fastest unmanned crossing of the Atlantic to date. A fleet of six is currently measuring fish populations in the Bering and Chukchi Seas in Alaska – and water surface temperatures in the Arctic – that are critical to understand the effects of climate change.
The drones measure co2 in the atmosphere, track oil spills, observe ocean acidification and reach glaciers and ice edges inaccessible to ships. This autumn, in partnership with noaa and Bering Sea Fisheries Research Foundation, a fleet will monitor migration of Alaskan crab to check if they are, in fact, migrating to the designated protection areas.
Initially Saildrone revenue came from government contracts but it now seems to be tacking towards consumers. De Halleux says, “Saildrone makes money by selling premium weather packages to consumers who want high-definition forecast [temperature, wind, currents]. We also offer weather services to private companies and high-performance sports teams.”
Another weather venture with a daring macho pedigree is Climacell. Founded by Israeli Air Force pilots Shimon Elkabetz and Rei Goffer, and Israeli army special operations veteran Itai Zlotnik, it raised $80m (€72m) from Softbank, Jet Blue Technology Ventures (Jet Blue’s early-stage investment arm that funds tech and travel ventures) and others for its street-level micro-forecasts. It collects data from connected devices: mobile phones, cars, Internet of Things devices. It then sells this forecast to customers such as Jet Blue and American football team the New England Patriots (which wants to know the weather in case, for example, there’s a risk that lightning will touch down on the field and result in a game’s cancellation).
Jupiter is the start-up to keep an eye on, especially if you believe in Arthur C Clarke’s maxim that “any sufficiently advanced technology is indistinguishable from magic”. Jupiter claims its model predicts catastrophic risk for up to 50 years in the future. Judging by the team they have in place, they feel like the grown-ups in the new weather universe. Founder Rich Sorkin is a Bill Gates-style nerd who has worked on the earliest form of big data with Nobel winning Yale Professor Bill Nordhaus, as well as with Google and Paris aerospace. His team includes Nobel prize winner Betsy Weatherhead (really), who served on the intergovernmental panel for climate change, which won the Nobel Peace prize in 2007. While it’s not like having Malala on the team, it’s an impressive bench that includes atmospheric scientists and experts on hurricanes, cyclones, flood prediction, fire and risk; there are also former leaders from noaa, Nasa, even the CIA.
Sorkin says, “If you want to solve the tough problems, you need smart people.” In 2017 he raised $10m (€8.9m) to build Jupiter before he had built anything and has been selling services to insurance companies and infrastructure clients ever since. He has subsequently raised another $23m (€21m). Three insurance companies participated in this latest round.
“We’re focusing on impacts to industrial infrastructure, real estate, financial services or a business asking: ‘What’s the probability that my hotel is going to be underwater in South Beach in seven years at some point during that year?’” he says.
“Risk prediction is the biggest opportunity. The whole planet is exposed to impacts of climate change. Very few people understand what that means in practice. Entire financial systems will have to adjust to these risks. Anyone with an exposed asset will have to figure out what to do about it. We provide the information for making those decisions.”
Given all the risk data he has to hand, why does Sorkin still live in San Francisco, one of the cities most vulnerable to severe future earthquakes? “On a regular basis I have conversations with my family about where the best place for everyone to live is. It’s a family decision that we’re still here. It’s not a one-dimensional problem. You have to balance career and quality of life. People like to live near their friends too,” says Sorkin.
For now Jupiter’s promises seem the most fantastic and, if correct, they have the greatest capacity to prepare the world for the effects of unpredictable weather.
But another disaster prediction start-up has proven to be a Theranos of weather prediction. One Concern, recently the subject of a New York Times investigative report, exaggerated its technology and earthquake modelling capabilities, resulting in a prediction product that many believe could jeopardise lives in a disaster instead of saving them.
It raises the question: do investors understand the complicated weather and climate modelling behind the curtain? Do they know what they are investing in? Just because Silicon Valley takes an interest in something, it doesn’t mean it’s actually useful.
Data scientists now hold the keys to the kingdom because they actually know and control what’s behind the curtain. That’s why entrepreneur and data scientist Donald Giuliano is ebullient about the possibilities ahead: “We’re about to enter the golden age of meteorological application,” he says. Ten years ago he and his partners started and sold Weather Fusion, the first in the industry to collect automated US wind and hail data to help insurance and roofing companies understand specific storm damage.
The company was bought out by Core Logic and Giuliano is now back after a five-year hiatus. He says, “I’m excited to be back in the weather business. AI and computing capabilities, graphical techniques: everything is coming together to enable us to do things with data that were unthinkable five years ago.” Giuliano and his partner Matt Van Every just launched Canopy, a venture focused on determining the impact of severe precipitation, floods and other extreme perils.
In the US the costs of flooding and storms in the Midwest alone this spring were estimated to exceed $3bn (€3bn). Worldwide, floods have become the most common and deadly natural disaster, affecting 250 million people every year. So predicting them accurately is a lifesaver and a money-spinner.
That’s also why Fathom, Google and other data visionaries are focusing on flood risk. Google’s Crisis Response team have been successful using machine learning to predict floods far enough in advance to deliver early warnings days ahead. According to software engineer Sella Nevo, who leads Google’s Flood Forecast team, “We estimated that the most promising place to create life-saving impact was flooding, over other natural disasters.”
Turns out response time is the most crucial element. “Early warning systems can reduce flood deaths by a third,” says Nevo. The pilot project in India, where 20 per cent of global flood-related fatalities occur every year, had 90 per cent accuracy in predicting where and when a flood would come. “We are able to provide between 12 and 24-hour notice and are always right about floods arriving.” Google plans to scale this AI to other countries and to other natural disasters, such as fires and earthquake aftershocks.
“It’s an exciting time to use technology to make a difference,” says Nevo. “It’s equally critical we provide that value for free to any government in the world, regardless of their financial status.” Google’s efforts in India are a good example, where more and improved data is practical and helpful to people’s lives. “But data is only useful if it can be used,” says Mark Govett from the noaa Earth System Research Laboratory. “We have 100 times more data than before. Only 1 per cent is being used.”
Some of the more daring weather ventures out there entirely reject the idea that every new weather company needs its own proprietary data. These outliers – not the Nobel, noaa or Nasa crowd – piggyback on the numbers that already exist and are creating new rivalries in the one arena that truly matters: the App Store. Like a pop-chart battle, the place to be is number one on Apple.
According to app-market intelligence provider Sensor Tower, first-time weather app downloads exceeded 210 million in the first half of 2019 and consumers spent more than $42m (€38m) on them worldwide.
Dark Sky founders Adam Grossman and Jason LaPorte were first to recognise that most of us don’t care about things such as humidity percentage or wind direction and that beauty makes an interface easier and more pleasurable to use. They launched Dark Sky iOS app after a successful Kickstarter campaign in 2011, foregoing proprietary data in favour of using their proprietary aggregation of existing data along with pretty animations and short-term hyperlocal forecasts (will it rain or snow in the next hour?).
Dark Sky was ranked number 16 in US weather app downloads last quarter and sits at number four for total earnings in the category. Evangelists pay $3.99 (€3.60) to show off the Dark Sky down-to-the-minute precipitation alerts over dinner. It’s weather app as conversation piece. And people love it.
In 2015, Grossman and LaPorte partnered with technical innovation company Applied Invention, based in Cambridge, Massachusetts. Applied Invention’s founder is Danny Hillis, a scientist, inventor and entrepreneur of Thinking Machines and The Long Now foundation fame. Sensor Tower data estimates that Dark Sky has generated more than $8.6m (€7.7m) since 2014.
And while Dark Sky has been criticised for its actual accuracy, the truth is that even the leading forecasts – Weather Underground, The Weather Channel and AccuWeather – are only accurate 75 per cent of the time. Dark Sky gets it right 64 to 67 per cent of the time. But according to 2016 research from Forecast Watch, it turns out that users don’t care as much about accuracy as the developers would have you believe. If it predicts 14c or 16c, you’ll still know to wear a jacket. Dark Sky’s 2.8 million devoted users prove this point.
Data scientist Zachary Flamig shares his insider view on the opportunity in weather. Flamig has three meteorology degrees and builds weather apps for fun. He has just completed his post-doc at the University of Chicago’s Center for Data Intensive Science and recently joined Amazon as tech lead for its Sustainability Data Initiative. He believes data is overrated. Flamig says, “Make it fun and make it pretty. Same as Dark Sky.” He points to Carrot as “another great example of a consumer company that figured out that proprietary data and forecasting doesn’t matter”.
Carrot founder Brian Mueller is not in the weather business, he’s in the app business. A former aspiring screenwriter, he built the interface for his first Carrot app after watching online tutorials. He uses the industry’s own data to disrupt it. And he made the Carrot weather app provider-agnostic the forecasts can be powered with data from Willy Weather in Australia, Dark Sky, Climacell or AccuWeather. He was using The Weather Channel’s data until it eyed his ascent up the charts and quickly shut down his access to its information.
But the Carrot app is not for everyone. The user interface is nice and simple, the weather information is accurate and it is one of the few to offer games, alongside extreme forecasts for lightning strikes and storm-cell prediction. But whether or not you will like it depends on how much shtick you can stomach with your weather report. There’s an option for an AI bot that delivers weather along with cringe-inducing insults and you can choose for it to be more, or less, insulting. The app will curse at you, call you “meatbag” and talk about your fat sausage fingers.
It’s a point of difference that has helped drive approximately 360,000 downloads worldwide since 2016, generating $1.8m (€1.6m) in revenue for Mueller’s one-man operation. It’s number nine in the top weather app revenue-generator charts, ahead of global giant AccuWeather.
Now everyone is gunning for The Weather Channel’s number one spot and its 131 million users. Saildrone and Climacell both recently launched subscription consumer apps: with 20,000 and 26,000 downloads respectively. Both charge a small monthly fee for boutique features. For example, Saildrone can give sailors ultra-high-resolution data that includes wind and current information essential to winning regattas.
Still, there seems to be space for a disruptive weather brand with a slightly loftier goal. If you don’t want shtick, you care about the world and wouldn’t mind nicer fonts, where do you go?
The New Weather is about to change everything about our lives. Where is the global weather news hub that offers real-time premium content, speaks to people’s interests and concerns, seeks out the answers and lauds innovators who are solving problems on the edge of our shifting climate?
Nevo said that they know from Google search that people are constantly looking for weather information. Machine learning might be able to pinpoint all the flooding in the world but what’s the point if you can’t get that critical information to people who need it or want it? And why can’t everyone have the edge, not just those in peril, or businesses who need to protect their assets?
It feels like the most glaring and obvious opportunity in weather is to connect the dots of what exists already. In VC-pitch jargon: it’s Jupiter meets Dark Sky meets the feels and technical innovation of Saildrone.
Where is that beautifully designed weather app with a built-in news and storytelling platform that is well-edited and digestible, using eye-popping graphics and providing actionable and personalised intelligence to navigate your life in weather? This is what we need. That’s what needs to be built. The weather doesn’t need more data. Weather needs an editor. Watch this space.