Aerosol and Cloud Detection Using Machine Learning Algorithms and Space-Based Lidar Data

John E Yorks; Patrick A Selmer; Andrew Kupchock; Edward P Nowottnick; Kenneth E Christian; Daniel Rusinek; Natasa Dacic; Matthew J Mcgill

NASA NTRS · 20210013395

Aerosol and Cloud Detection Using Machine Learning Algorithms and Space-Based Lidar Data

John E Yorks·Patrick A Selmer·Andrew Kupchock·Edward P Nowottnick·Kenneth E Christian·Daniel Rusinek·Natasa Dacic·Matthew J Mcgill

View original source ↗Export metadata ↓

Abstract

Clouds and aerosols play a significant role in determining the overall atmospheric radiation budget, yet remain a key uncertainty in understanding and predicting the future climate system. In addition to their impact on the Earth’s climate system, aerosols from volcanic eruptions, wildfires, man-made pollution events, and dust storms are hazardous to aviation safety and human health. Space-based lidar systems provide critical information about the vertical distributions of clouds and aerosols that greatly improve our understanding of the climate system. However, daytime data from backscatter lidars, such as the Cloud-Aerosol Transport System (CATS) on the International Space Station (ISS), must be averaged during science processing at the expense of spatial resolution to obtain sufficient signal-to-noise ratio (SNR) for accurately detecting atmospheric features. For example, 50% of all atmospheric features reported in daytime operational CATS data products require averaging to 60 km for detection. Furthermore, the single-wavelength nature of the CATS primary operation mode makes accurately typing these features challenging in complex scenes. This paper presents machine learning (ML) techniques that, when applied to CATS data, enable detection of atmospheric features during daytime operations with a horizontal resolution of 5 km compared to the 60 km horizontal resolution often required for daytime CATS data. A Convolutional Neural Network (CNN) trained using CATS standard data products also demonstrated the potential for improved cloud-aerosol discrimination compared to the operational CATS algorithms for cloud edges and complex near-surface scenes during daytime.

Explore related subjects

lidar aerosol cloud machine learning

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

John E Yorks, Patrick A Selmer, Andrew Kupchock, Edward P Nowottnick, Kenneth E Christian, Daniel Rusinek, Natasa Dacic, Matthew J Mcgill. 2021-05-07. Aerosol and Cloud Detection Using Machine Learning Algorithms and Space-Based Lidar Data. https://ntrs.nasa.gov/citations/20210013395

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related reports

NASA NTRS2024

Differential Absorption Lidar for Searching Water Sources on Mars

Water vapor has been detected in the Martian atmosphere by multiple orbiting instruments. The Atmospheric Chemistry Suite (ACS) on the ExoMars Trace Gas Orbiter (TGO) observed H2O mixing ratios reaching up to 50 ppmv at altitudes of 100-120 km during global dust storms, while levels remained low (<2 ppmv) during other seasons. The Neutral Gas and Ion Mass Spectrometer (NGIMS) on the Mars Atmosphere and Volatile EvolutioN (MAVEN) spacecraft revealed that water transported to the upper atmosphere is dissociated by ions, producing atomic hydrogen that escapes into space, contributing to Mars’ water loss. This transport is seasonal, peaking in southern summer and intensifying during dust storms. Additionally, the Mars Reconnaissance Orbiter’s (MRO) imaging spectrometer detected hydrated minerals on slopes, suggesting that liquid water may intermittently flow on present-day Mars. However, an observational gap exists between high-altitude water vapor and surface water due to limitations in spatial resolution and a lack of measurements in the lower atmosphere. To address this gap, we propose using an airborne differential absorption lidar (DIAL) to search for water sources. DIAL provides high-resolution measurements both day and night, bridging the observational gap between high-altitude water vapor and surface water, thus enhancing our understanding of water transport and loss on Mars. Absorption lines of water vapor in the 2.7 μm and 1.8 μm bands have been selected in this study. Simulations show that both lines are capable of detecting water vapor sources with reasonable system parameters.

lidar↗

NASA NTRS2022

Vertical structure of biomass burning aerosol transported over the southeast Atlantic Ocean

Biomass burning in southwestern Africa produces smoke plumes that are transported over the Atlantic Ocean and overlie vast regions of stratocumulus clouds. This aerosol layer contributes to direct and indirect radiative forcing of the atmosphere in this region particularly during the months of August, September, and October. There was a multi-year international campaign to study this aerosol and its interactions with clouds. Here, we report on the evolution of aerosol distributions and properties as measured by the airborne high spectral resolution lidar (HSRL-2) during the ORACLES (Observations of Aerosols above Clouds and their intEractionS) campaign in September 2016. The NASA Langley HSRL-2 instrument was flown on the NASA ER-2 aircraft for several days in September 2016. Data were aggregated at two pairs of 2° × 2° grid boxes to examine the evolution of the vertical profile of aerosol properties during transport over the ocean. Results showed that the structure of the profile of aerosol extinction and microphysical properties is maintained over a 1 to 2 d timescale. In the 3–5 km altitude range, 95 % of the aerosol extinction was contributed by particles in the 0.05–0.50 µm radius size range with the aerosol in this size range having an average effective radius of 0.16 µm. This indicates that there is essentially no scavenging or dry deposition at these altitudes. Moreover, there is very little day-to-day variation in these properties, such that time sampling as happens in such campaigns may be representative of longer periods such as monthly means. Below 3 km, there is considerable mixing with larger aerosol, most likely continental source near land. Furthermore, these measurements indicated that there was often a distinct gap between the bottom of the aerosol layer and cloud tops at the selected locations as evidenced by a layer of several hundred meters that contained relatively low aerosol extinction values above the clouds.

lidar↗

NASA NTRS2021

Aerosol and Cloud Detection Using Machine Learning Algorithms and Space-Based Lidar Data

Clouds and aerosols play a significant role in determining the overall atmospheric radiation budget, yet remain a key uncertainty in understanding and predicting the future climate system. In addition to their impact on the Earth’s climate system, aerosols from volcanic eruptions, wildfires, man-made pollution events and dust storms are hazardous to aviation safety and human health. Space-based lidar systems provide critical information about the vertical distributions of clouds and aerosols that greatly improve our understanding of the climate system. However, daytime data from backscatter lidars, such as the Cloud-Aerosol Transport System (CATS) on the International Space Station (ISS), must be averaged during science processing at the expense of spatial resolution to obtain sufficient signal-to-noise ratio (SNR) for accurately detecting atmospheric features. For example, 50% of all atmospheric features reported in daytime operational CATS data products require averaging to 60 km for detection. Furthermore, the single-wavelength nature of the CATS primary operation mode makes accurately typing these features challenging in complex scenes. This paper presents machine learning (ML) techniques that, when applied to CATS data, (1) increased the 1064 nm SNR by 75%, (2) increased the number of layers detected (any resolution) by 30%, and (3) enabled detection of 40% more atmospheric features during daytime operations at a horizontal resolution of 5 km compared to the 60 km horizontal resolution often required for daytime CATS operational data products. A Convolutional Neural Network (CNN) trained using CATS standard data products also demonstrated the potential for improved cloud-aerosol discrimination compared to the operational CATS algorithms for cloud edges and complex near-surface scenes during daytime.

lidar↗