A Star Wars style satellite, antique airplane and a flock of ground truthers seek to improve weather forecasts
CARMAN, Man. — Hitch a high tech NASA satellite to a 1930s DC3 aircraft, add 10 teams of students and scientists prowling southern Manitoba fields and you’ll have a highly accurate climate prediction model.
The project encompasses organizations such as NASA, the U.S. Department of Agriculture and the University of Manitoba. It’s called Soil Moisture Active Passive (SMAP), and while the title may not sound too exciting, the mission itself is, especially for prairie farmers.
The NASA-SMAP-USDA satellite takes images that document soil moisture, and SMAP scientists use the data to build better, more accurate weather forecast models.
The intent is simple, but the technology is not.
This NASA satellite measures sunlight reflecting off the surface of the Earth as a means of determining soil moisture. The satellite has focused its attention on southern Manitoba this summer. | NASA illustration
Building SMAP climate models involves images from the NASA satellite launched in 2015, images from a low-flying DC3, more images from a lower-flying drone and a large volume of data gathered on the ground by an array of people and their instruments.
Soil moisture is a result of weather, but it also works the other way around. Weather is a result of moisture already in the soil, says Michael Cosh, USDA co-ordinator on the SMAP project.
Tyson Ochsner hammers a core sampler into the ground to obtain a volumetric calibration for soil moisture. Ground-truthing is critical for soil moisture research. | Michael Cosh photo Cosh is stationed at the remote sensing lab in Maryland, but he has spent time in Manitoba fields working with the projects, ground-truthing component.
“Measuring soil moisture is now one of our (USDA) top priorities because it affects our climate so significantly,” Cosh said.
“The benefit will be better weather forecasting and better climate modelling in terms of flood forecasting, reservoir management, irrigation, assessing drought conditions and understanding the surface of the Earth.
“When sunlight hits the soil, it can do only one of two things. If there’s moisture in the soil, sunlight causes it to evaporate into the atmosphere. If there’s no moisture in the soil, then the sunlight warms the soil.”
The obvious result is that heat build-up in drier soil raises the temperature of the air above the soil and causes air movement.
As well, the evaporating moisture will increase the amount of moisture in the air, which determines the immediate weather and longer-term climate.
“The science of it is that we have a water balance and an energy balance,” Cosh said.
“The energy balance is sunlight coming through the atmosphere, hitting the Earth and doing something. Some energy is reflected, some is absorbed.
“Of the energy that’s absorbed, the first thing it will do is cause water to evaporate from the soil and cause transpiration by plants. The second thing the energy will do is heat the soil. We are gaining a better understanding of these things by using satellite imagery.”
The SMAP antenna is too delicate to support self in Earth’s gravity, so it must hang suspended while researchers make the final connections before folding it up again to fit into a NASA rocket for its journey into space in 2015. | Thomas A Dutch Slager/NASA photo
The SMAP satellite, which was launched last year, originally used an active sensor that sent and received radar signals and a passive sensor that simply worked like a camera, collecting light signals from the Earth’s surface.
However, the active sensor stopped working shortly after it went into service, so the project is continuing with only the passive system.
“We did get some data from the active sensor, and the airplane uses an active sensor also, so it’s not as bad as it may seem,” he said.
“Plus, the NASA people are starting to collect active data from another satellite with a C-band sensor. C-band is another of the frequencies that can be used for high resolution soil moisture mapping. The Aqua satellite gave us soil moisture information with C-band as far back as 2002.”
Cosh said the satellites look for brightness temperature. People see red, green and blue, while near infrared is just past the red light waves, which some animals can see, but people can’t. The cameras see near infrared.
There are also ultra high frequencies that can be seen only with special equipment.
The L-band is located at 1.2 gigahertz and is the band SMAP uses to detect soil moisture.
Researchers can monitor these bands, correlate them to real data collected on the ground and then use a satellite to measure soil moisture. L-band has the least interference from vegetation.
Cosh said agencies around the world are running dozens of similar projects, but a major drawback is that they use many different sensors and formats for handling data. This means scientists cannot easily share results.
That’s bad news because a climate modelling researcher wants as many inputs as possible, but there must be a common base line.
To deal with this issue, Oklahoma State University bought samples of all the common sensors used around the globe and tested them against each other to establish a common language.
USDA researcher Steve Elliot installs a time domain reflectometer in a trench. Moisture sensors were located at depths of 5, 10, 20, 50 and 100 cm. | Michael Cosh photo
The benefit of the OSU initiative is that a climatologist developing a forecast model for a specific region can access a vast network of data from other regions of the globe that have similar characteristics.
“For example, a farmer in Manitoba scheduling his irrigation doesn’t care about South Africa,” Cosh said.
“He may be working with a good model that’s been made better because it includes relevant data from South Africa, but all he cares about is monitoring, modelling and forecasting on his own farm. A common global base line lets the researchers pull in more data when they construct these models. Oklahoma provides researchers with a standard location to study these different sensors and systems from around the world.
“We’ve seen that some sensors are very good at getting the mean numbers right, but when they get out to the extremes of very dry or very wet, they don’t perform as well.”
Cosh said SMAP has a small network of core validation sites in North America, such as the one in southern Manitoba that has been running the past two summers.
These sites are used to confirm and calibrate information from the larger number of data collection sites scattered around the continent.
There are parallel core validation sites in other parts of the world, including Argentina and Europe.
Once scientists know that a core site has reliable soil moisture information, they can run trials to see if the SMAP satellite estimates the soil moisture correctly.
This feedback lets them tweak the algorithms, which allows them to account for soil type, surface roughness, vegetation and a complex maze of other factors.
“Once we’ve accounted for all other factors, the only thing left for the satellite to deal with is raw soil moisture,” Cosh said.
“Tweaking the algorithms is an ongoing process. We have a little more than one year of data so far. Every time we get more data, we fine tune the algorithms again.
“To get the best data, we run something called an intensive observation period. We had two of these operations in the Carman area of southern Manitoba this summer, one in June and one in July. We’ll have the NASA DC3 flying the same area the satellite is flying and the drone is flying, all at the same time.
“Also at the same time, we have the ground crews walking a pre-determined pattern through the fields collecting ground data to verify the data we get from above. This lets us quantify truth at the ground and further refine the algorithm.”