Emerging capabilities of small satellite constellations, commercially built and constantly refreshed, will make weather data affordable, more impactful, and ultimately create millions of new data points.
Images are based on NASA's Blue Marble and City Lights satellite pictures. Courtesy: Living Earth Desktop
With thousands of satellites, and airborne and surface-based sensors constantly taking the pulse of our planet, weather forecasting has been fueled by Big Data since long before ‘Big Data’ was even a thing. And yet, today’s weather forecast still lacks the reliability required by the disaster response team positioning its precious resources, the electric utility predicting power demand, the commercial airliner optimizing flight paths, and the average citizen planning a weekend getaway.
A closer examination reveals why this is the case: despite the seemingly ample amounts of weather data collected, much of it never makes it into the forecast models, and most of it misses the inner workings of our atmosphere that have the largest impact on the weather we experience. In large part, this failure to adequately sample the atmosphere can be traced back to a decades-long lack of innovation in weather satellites that has us launching yesterday’s technology today — a troubling trend that commercial industry is set to turn around in a big way.
On the face of it, the amount of environmental data collected each day is mind-boggling massive — over 20 terabytes by the National Oceanic and Atmospheric Administration alone. That equates to 1,200 piles of typed paper, each stacked as high as the Eiffel Tower, or more than double the total printed collection of the Library of Congress. The world of weather forecasting has surely benefitted from the ever-increasing environmental intelligence gathered about our planet — the vast majority of it now from satellites — with today’s five-day forecast about as accurate as a three-day forecast was 20 years ago.
However, we still lack the data necessary to routinely predict high-impact weather events with adequate accuracy and lead times. We see this year in and year out with snowstorms along the East Coast of the United States, with widely varying forecasts even less than 24 hours before the first snowflakes arrive. We see this with hurricanes as well, with last fall’s Hurricane Joaquin the most recent example. Model projections showed Joaquin striking anywhere from South Carolina to New Jersey, when ultimately the storm did not make landfall at all, tracking well off the coast despite contributing to historic flooding in South Carolina.
The missing links
With so many observations from a multitude of space-based, ground and airborne instruments being funneled into increasingly advanced weather models powered by more and more computing capacity, why do we still struggle to accurately predict the most impactful weather events with the lead times necessary to properly prepare and protect property, lives and businesses?
One major reason is that, despite the deceivingly large amount of data collected on a daily basis, the vertical structure of our atmosphere is still severely under-sampled. Large, billon-dollar weather satellites, flown by a handful of major government space programs around the world, provide crucial data for weather forecasting. Yet their vertical resolution — that is, the distance in altitude between data points — is typically no better than 1 to 3 km, which is an order of magnitude too coarse to see small-scale changes that can have huge impacts on our weather. Not to mention, many of the existing sensors can’t even see through clouds. That is a lot of missing data given that, at any moment, about two-thirds of the Earth is covered with clouds.
A true 3-D cat scan of our atmosphere would reveal sharp changes with altitude in temperature, pressure, humidity and other variables. These vertical gradients determine the stability of the atmosphere, which has profound impacts on the development of clouds and storms. However, our ability to observe the detailed structure of the vertical atmosphere is currently limited, primarily to about 900 weather balloons sparsely scattered across the globe and launched only once or twice per day. Furthermore, the vast majority of these balloons are launched over land, whereas most weather systems originate over the ocean, and very few balloons cover the developing world, remote areas or the poles.
How smallsats can change things
Enter the emerging capabilities of small satellite constellations, commercially built and constantly refreshed. From low-earth orbit, our much lighter and less expensive satellites will drastically increase the number of vertical atmospheric profiles collected each day, with observations approximately every 100 meters in altitude from the Earth’s surface up into the ionosphere. That equates to over 8 million observations per day of temperature, pressure and moisture, which will dramatically improve forecasts for all kinds of weather, and especially for storms that do the most damage to people, property, and the bottom line of numerous businesses in numerous sectors.
As an industry, it is critically important that we deliver not just on the quantity of data promised, but also on the high quality of data required by numerical weather prediction centers worldwide. That is why despite their small size and low power requirements, our satellites will carry the only sensor of its kind that is powerful enough to probe all the way down into the lowest three kilometers of the atmosphere and to the surface, where most of our planet’s weather occurs, and are designed to last seven years or more. After all, poor-quality and unreliable data, no matter how much of it, will only serve to degrade the forecast, not improve it.
There is a precedent for the kind of commercialization we are now witnessing in the area of space-based atmospheric weather data. In fact, there are several examples in which solely governmental satellite and space activities transitioned to a team effort, with the commercial sector augmenting government programs that had often fallen behind, lacked innovation and seen costs skyrocket. Satellite telecommunications was the first to undergo such a transition, starting as far back as the 1960s. The most parallel example, though, is that of satellite imaging.
An artist’s rendering of PlanetiQ's satellites, designed to capture information from GPS signals?
Private sector to the rescue
When an aging and overworked fleet of government satellites led the demand for satellite imagery to outpace supply about 15 years ago, the US government partnered with the private sector to promote a commercial industry for earth observation. Soon it was the likes of DigitalGlobe and GeoEye that were not only fulfilling the mission-critical needs of the defense and intelligence communities, and doing so in a more cost-efficient way, but also laying the groundwork for the online and mobile mapping tools we can’t imagine living without today.
And now it is a new generation of maturing start-ups, including Planet Labs and Terra Bella (formerly Skybox Imaging), that are starting to image the world with swarms of small satellites that can provide the rapid refresh rates needed to support a host of new applications. The launch frequency required for such constellations will be aided by the emergence of SpaceX, Virgin Galactic and other space launchers — yet another industry that, only a decade ago, was solely the purview of governments.
In a strikingly similar situation to that of satellite imagery a decade and a half ago, today’s government weather satellite programs are delayed and over budget, with data gaps already occurring, and innovation hamstrung by competing budget priorities and development cycles that last 10 to 15 years from requirements setting to launch. There is an alternative to this status quo, however, and it’s in the form of a commercial weather satellite industry now poised for rapid growth. Much as it has done for satellite imaging, the commercial sector’s miniaturization of space-based sensors and components will enable the launch of many more weather satellites for a much lower price.
The result will be a dramatic increase in atmospheric data worldwide, which when coupled with steady improvements in weather and climate modeling, increasing computing capacity, and the emergence of new tools for custom-tailored analytics, promises an unprecedented increase in the quantity and quality of actionable information supporting the government, energy, agriculture, insurance, transportation and retail sectors, to name a few. Only now are we beginning to fathom the magnitude of social and economic benefits possible from better forecasts of hurricane tracks, optimized ship routing, reduced spending on aviation fuel, more efficient agriculture practices, and many other applications.
In some ways, the market opportunity for satellite weather data may even exceed that of imagery. Not just because our weather seems to be getting more extreme, or that businesses and entire industries are increasingly recognizing the value of weather and climate information in reducing risk and making better decisions. But also because, by its very nature, weather data is perishable. Almost as soon as an observation is collected and delivered into a weather forecast model, the next forecast cycle requires a fresh look at the latest conditions.
It is this growing demand for precise and frequent data, with dense global sampling from the top of the atmosphere down to the Earth’s surface, that has set the stage for a transformation of the satellite weather data marketplace. As other industries have proven before, commercial augmentation of government systems will make weather data more affordable, more impactful, and ultimately create millions of new data points to support numerous sectors with the information they need to better prepare for whatever Mother Nature may offer next.
– Timothy Puckorius, Executive Vice President, Business Development
Dan Stillman, Senior Manager, Corporate Marketing & Communications, Meteorologist