In August, dozens of scientists from across the United States descended on the small island nation of São Tomé and Príncipe. Nestled on the equator off the coast of western central Africa, São Tomé was an ideal location to study the phenomenon we had all gathered to observe: a seasonal plume of smoke from agricultural and forest fires that gets lofted by the prevailing winds from the African continent to over the southeast Atlantic Ocean. As part of the NASA field campaign Observations of Aerosols above Clouds and their Interactions, or ORACLES, our aim was to better understand how all that smoke over the ocean affects the amount of sunlight that gets absorbed in the atmosphere and at Earth’s surface.

Aerosols—small airborne particles, like smoke, desert dust, and sulfates from power plants—affect the amount of energy the southeastern Atlantic Ocean gets from the sun, not only by absorbing and reflecting sunlight directly, but also through its effects on clouds. A large expanse of very bright low clouds covers much of the southeastern Atlantic, very similar to the clouds off the coast of California that create San Francisco’s characteristic fog. Smoke can change the properties of these clouds in various ways, including brightening the clouds by creating lots of small droplets, which, interestingly, make the clouds less likely to drizzle and thus stick around for a longer time. Both of those changes allow the clouds to reflect more sunlight, creating a cooling effect.

As anyone who’s been outside on an overcast day knows, clouds play a major role in regulating the amount of the sun’s energy that gets to Earth’s surface, so any changes in the clouds over the southeast Atlantic and those like them across the globe can have big implications for Earth’s energy balance. It is well-known that the heat-trapping effect of man-made greenhouse gas emissions have led to a net warming over the 20th and early 21st centuries. However, unresolved scientific questions about the potential cooling effects of aerosol-cloud interactions over the past century represent a large fraction of the uncertainty in estimates of how much humans have affected the present-day climate.

For ORACLES, NASA’s P-3 Orion aircraft was our primary transport for measuring the smoke-cloud system. On the P-3 we have a set of instruments that can be broadly separated into two categories: in-situ and remote sensing.

In-situ instruments, like those in the picture collage below, measure things in place through air inlets. For example, we have particle counters that can measure the number and size of smoke particles in a plume, and cloud probes that can measure how much liquid water is in a cloud.

In contrast, remote sensing instruments sense things remotely; that is, they tell us about the properties of clouds and smoke from far away, like how we use a telescope to observe stars. In our case, we use instruments like a radar to look at precipitation and a lidar (a laser that provides information about a what’s between the plane and the ground) to look at the smoke plume’s structure.

Clouds play a major role in regulating the amount of the sun’s energy that gets to Earth’s surface, so any changes in the clouds over the southeast Atlantic and those like them across the globe can have big implications for Earth’s energy balance.

Of course, the in-situ instruments that measure clouds aren’t much use when flying through smoke above the clouds, and when we fly high to get good lidar profiles, we can’t get in-situ smoke measurements. In addition, some of the remote sensing instruments don’t work well when high clouds are present, and the smoke and low clouds aren’t always in the same place from one day to the next. How do we balance all these competing objectives to produce a flight that collects high-quality, usable data? That’s where the forecasting and flight-planning team comes in.

As a graduate student at the University of Washington in Seattle, my role in ORACLES is to look at model forecasts from computer simulations and satellite imagery and then use flight-planning software to create flight plans that will meet our scientific objectives. On what we call routine flights, that mostly means picking altitudes and aircraft maneuvers rather than locations, because for these flights we always stick to the same north-south track to build up statistics that can be used to compare our observations with various computer models.

One example of the choices that have to be made here is whether to do stacked legs, in which we fly over the same location at different heights, or sequential legs, which let us cover more ground because we don’t need to backtrack and instead gives us observations at slightly different locations that might be harder to interpret. A similar choice has to be made when we switch between altitudes: we can ramp down and cover a lot of ground, or do a square spiral and get a vertical profile over the same location.

The other type of flight we call a flight of opportunity, in which we have more latitude in choosing our flight location to sample interesting features, or to avoid pitfalls like high clouds, that are identified by the models.

We were also able to combine flight plans so that the flights of opportunity could resample air that we observed a day or two earlier. Ideally, to study how the smoke evolves during the course of its journey over the Atlantic, we would be able to follow it as the winds push it westward and downward over a period of days. Unfortunately, this is not at all practical in an aircraft with nine hours’ worth of fuel. Instead, we can run a weather forecast model to predict where the air we sampled during a routine flight will end up in a few days. Then, like an advanced game of connect-the-dots, we can design our next target of opportunity flight to hit the right location and altitude to resample that air to see how it’s evolved.

Our August 17 flight of opportunity was a bit special because, rather than return to São Tomé, the P-3 landed on Ascension Island in the middle of the South Atlantic Ocean so we could do some joint flights with a British team studying similar science questions. On the way to Ascension, we planned our track to intersect the new (forecasted) locations of a few different smoke plume air parcels that we sampled on August 15.

Now that the 2017 ORACLES deployment is over, the task ahead of us will be to analyze the data we collected in flights like the August 15-17 resampling mission to produce new scientific insights into this unique smoke-cloud system. Within a year, all of our data will become public at https://espoarchive.nasa.gov/archive/browse/oracles so that other researchers across the country and around the world will be able to contribute their own research and generate new ideas and solutions. The data from last year’s deployment, which took place in September and was based out of Walvis Bay, Namibia, is already available. However, we’re not done with data collection just yet: We’ll be heading back into the southeast Atlantic next year for one last deployment, this time in October to characterize the end of the southern African fire season.

This piece was originally published on the NASA Earth Expeditions blog.