Although the troposphere is often thought of as the layer of the atmosphere closest to the Earth's surface, the planetary boundary layer (PBL), the lowest layer of the troposphere, is actually the part that most significantly influences the climate near the surface. in 2018 decadal planetary science surveyABP was proposed as a important scientific question which has the potential to improve storm forecasting and climate projections.
“The PBL is where the surface interacts with the atmosphere, including moisture and heat exchanges that help drive severe and changing weather conditions,” says Adam Milstein, a technical staff member in Lincoln Laboratory's Applied Space Systems Group. “The PBL is also where humans live, and the turbulent movement of aerosols along the PBL is important for air quality that influences human health.”
Although vital for studying weather and climate, important features of the PBL, such as its height, are difficult to resolve with current technology. For the past four years, Lincoln Laboratory staff have been studying PBL, focusing on two different tasks: using machine learning to create 3D-scanned profiles of the atmosphere and resolving the vertical structure of the atmosphere more clearly to better predict the droughts. .
This PBL-focused research effort builds on more than a decade of related work on fast, operational neural network algorithms developed by Lincoln Laboratory for NASA missions. These missions include time-resolved observations of precipitation structure and storm intensity with a constellation of small satellites (TROPICAL ZONE) as well as Aqua, a satellite that collects data on the Earth's water cycle and observes variables such as ocean temperature, precipitation and water vapor in the atmosphere. These algorithms recover temperature and humidity data from satellite instruments and have been shown to significantly improve the accuracy and usable global coverage of observations over previous approaches. For TROPICS, the algorithms help recover data that is used to characterize a storm's rapidly evolving structures in near real time, and for Aqua, they have helped augment forecast models, drought monitoring, and fire prediction.
These operational algorithms for TROPICS and Aqua are based on classic “shallow” neural networks to maximize speed and simplicity, creating a one-dimensional vertical profile for each spectral measurement collected by the instrument at each location. While this approach has improved observations of the atmosphere to the surface in general, including PBL, laboratory staff determined that new “deep” learning techniques are needed that treat the atmosphere over a region of interest as a three-dimensional image to improve the details of the PBL. further.
“We hypothesize that deep learning and artificial intelligence (ai) techniques could improve current approaches by incorporating better statistical representation of 3D images of atmospheric temperature and humidity into solutions,” says Milstein. “But it took some time to figure out how to create the best data set: a combination of real and simulated data; “We needed to prepare to train these techniques.”
The team collaborated with Joseph Santanello of NASA's Goddard Space Flight Center and William Blackwell, also of the Applied Space Systems Group, on a recent study. NASA-funded effort demonstrating that these retrieval algorithms can improve PBL detail, including a more accurate determination of PBL height than the prior state of the art.
While better knowledge of PBL is widely useful for increasing understanding of climate and weather, a key application is drought prediction. According to a Global Drought Snapshot Report published last year, droughts are a pressing planetary issue that the global community must address. The lack of moisture near the surface, specifically at the PBL level, is the main indicator of drought. Although previous studies using remote sensing techniques have examined soil moisture To determine drought risk, studying the atmosphere can help predict when droughts will occur.
In an effort funded by Lincoln Laboratory technology-office/climate-change-technology-national-security”>Climate change initiative, Milstein, along with lab staff member Michael Pieper, are working with scientists at NASA's Jet Propulsion Laboratory (JPL) to use neural network techniques to improve drought prediction in the continental United States. While the work builds on existing operational work JPL has done incorporating (in part) the lab's operational “shallow” neural network approach for Aqua, the team believes this work and deep learning research work focused on PBL can be combined to improve even more. the accuracy of drought prediction.
“Lincoln Laboratory has been working with NASA for more than a decade on neural network algorithms to estimate temperature and humidity in the atmosphere from infrared and microwave space instruments, including those on the Aqua spacecraft,” he says Milstein. “During that time, we have learned a lot about this problem by working with the scientific community, including learning about the scientific challenges that remain. “Our long experience working on this type of remote sensing with NASA scientists, as well as our experience using neural network techniques, gave us a unique perspective.”
According to Milstein, the next step of this project is to compare the deep learning results with data sets from the National Oceanic and Atmospheric Administration, NASA and the Department of Energy collected directly at the PBL using radiosondes, a type of instrument flown on a weather satellite. balloon. “These direct measurements can be considered a kind of 'ground truth' for quantifying the precision of the techniques we have developed,” says Milstein.
This improved neural network approach promises to prove drought prediction that could surpass the capabilities of existing indicators, says Milstein, and be a tool that scientists can rely on for decades to come.
