The next five years will see a massive proliferation of new weather-related apps, and analytics. Weather historically has been taken for granted by consumers and businesses. However, this is changing. The annoying, even punishing, cost of being unprepared for weather events has come to be increasingly appreciated throughout society.
For most consumers, this appreciation revolves in large part around extreme weather events – Hurricane Katrina in New Orleans, Hurricane Sandy on the U.S. east coast. For businesses, the issues are far more pervasive and rooted in profit-and loss economics.
Weather Technology on the Cusp of Disruptive Change
The surge in new apps and information capabilities will arise from huge technological improvements in the key elements of weather forecasting and analysis. These changes can be divided into three principal areas:
1. Atmospheric Sensing
Almost all of the weather data currently disseminated by the U.S. government is gathered from satellites (estimated at over 90%). New information will, in part, derive from new satellites (primarily from the private sector) that enable high resolution vertical measurements of the atmosphere.
An example of this type of satellite venture is start-up Hypercubes (http://www.hypercubes.global/). Their satellites will incorporate hyperspectral sounders, sensing devices that can make infrared observations of the atmosphere, providing thousands of times more data observations than are available from older atmospheric sensing devices. (Hypercubes will actually focus on pollution tracking from space, but information from hyperspectral sounders will be widely used in weather forecasting and analysis.) In addition, there will be billions of other added sensing devices, such as those on UAVs and other airborne vehicles, as well as on IoT devices.
2. Prediction Model Science
Weather forecasting, along with other big data applications will benefit from new capabilities introduced by developments, such as Deep Learning, Machine Learning, Machine Vision, Cognitive Computing, and the like.
3. Processing Speed Improvements
Software based on high performance computing (HPC) components can provide order of magnitude improvement in the speed and precision of weather forecasting.
There are already signs in the marketplace of the latent demand for such improvements, both in commercial and consumer applications. In this article we consider some of the innovative, newer entrants into the market for consumer mobile weather apps – a later article will evaluate the commercial market.
Dark Sky a Weather Innovator
While weather forecasting was long an activity relegated to government agencies, this has changed significantly. On the consumer side there are several mobile apps available from major weather information providers, such as the Weather Channel’s weather.com (now part of IBM), but also including interesting recent entrants into the market.
These include companies such as Dark Sky, Sunshine, Arcus Weather, Carrot and many others. Dark Sky bills itself as offering “state-of-the-art technology to predict when it will rain or snow — down to the minute — at your exact location, and presents it to you alongside the most beautiful weather visualizations you’ve ever seen.”
Dark Sky is available for $3.99 for iPhone and iPad, and also works with the Apple Watch. The company claims it will warn you about impending weather changes up to an hour in advance. There are also longer term weather forecasts displayed.
In June 2015 Dark Sky released an update that incorporated crowd-sourced information capability, so that users could input info about actual weather conditions. The update included added notification and other features.
Dark Sky only works in the United States (including Hawaii, Alaska, and Puerto Rico) and the British Isles, at this time.
Underlying the Dark Sky app is the company’s web-based app, Forecast.io. The company lists 19 data sources that Forecast.io ingests information from, mostly affiliated with NOAA (National Oceanic and Atmospheric Administration) in the U.S. and the UK Met Office. They claim that they have algorithms to correct for data issues in these different sources and they have discussed their process for correcting radar images using neural nets and computer vision technology to achieve detailed information about the local precipitation outlook.
Dark Sky’s business approach includes revenue streams both from sales of the mobile app on the iTunes store as well as revenue from developers that use the Forecast.io API.
Sunshine – How You Feel About The Weather
Another mobile app that features crowd-sourced information is Sunshine. Sunshine’s differentiator is that it seeks out user personal reactions to the weather, e.g., does the user feel “hot” or “cold.” Knowing that one user might feel cold at 60 while another might feel warm enables Sunshine to personalize its weather messages.
Sunshine claims to have a substantial network of smartphones using its apps. It states, that it has “a network of millions of smartphones on the ground, producing results that are on average 3x more accurate.”
Arcus, Carrot – Forecast.io API Derivatives
Forecast.io states that thousands of developers have signed up to use their API for further iterations of weather-related information apps. Several of these apps are listed on the Forecast.io website. One of these is Arcus that developed an Android version of the hyper-local service, which was a gap in Dark Sky’s strategy.
Arcus claims that, “Arcus Weather is the most accurate weather app on Android!” It also heavily markets the ability to offer precise weather data for the individual user’s exact location.
Carrot is another weather app that is drawing data from Forecast.io. Carrot offers an entertaining marketing approach by having a whimsical, even sarcastic, bot that comments on the weather data. This is an approach that Carrot has used in other apps, such as for Fitness and ToDo.
Earth Networks/Weather Bug – Weather & The IoT
Earth Networks is not a new player, having been in the weather field for over 20 years, and offering the well-known WeatherBug mobile app. However, it is interesting because it illustrates another development, which is the proliferation of sensors and the potential impact of the IoT on the weather information industry.
The company maintains its Total Lightning Network. This consists of some 1200 sensors in 40 countries that can detect in-cloud lightning as well as cloud-to-ground strikes.
Earth Networks also provides a network that it describes as “10,000 exclusive neighborhood-level sensors” for detecting local weather. These are used to supplement information provided by government agencies.
Dark Sky is one company that is indicative of the most powerful trend in weather forecasting and analysis – both consumer and commercial – which is HyperLocal Weather. As for individuals, the weather you care about is the weather at the places you are going to be during the day. Similarly for businesses, in agriculture, for example, the weather that matters is the weather at your field, not the weather for the rest of the county.
Sunshine, Arcus and Carrot are three examples of companies that extend the features of apps and explore new marketing approaches to the classic, generally routine, presentation of the weather. Earth Networks illustrates one of the big opportunities, which is the integration of information relevant to weather analysis from billions of new sensor locations.
The weather app area is a robust one. As a highly respected expert, and friend of ours, has written: “There are virtually billions of applications of weather data for improving billions of decisions, with the potential to save companies hundreds of billions of dollars each year” (TruWeather Solutions, LLC blog, 10/29/15.) We fully expect that this explosion of weather app creativity will be mirrored in both the commercial and consumer spheres.