Files in this item

FilesDescriptionFormat

application/pdf

application/pdfVILLEGASBRAVO-THESIS-2021.pdf (7MB)
(no description provided)PDF

Description

Title:Development and verification of the NASA Multi-Angle Imager for Aerosols operational cloud mask
Author(s):Villegas Bravo, Javier Alfredo
Advisor(s):Di Girolamo, Larry
Department / Program:Atmospheric Sciences
Discipline:Atmospheric Sciences
Degree Granting Institution:University of Illinois at Urbana-Champaign
Degree:M.S.
Genre:Thesis
Subject(s):cloud mask
MAIA
MISR
multi angle
pollution
aerosols
particulate matter
cloud detection
thresholds
NASA
cloud
epidemiologist
MODIS
MAIAC
Abstract:The National Aeronautics and Space Administration (NASA) Multi-Angle Imager for Aerosols (MAIA) instrument is set to launch in 2022 with the mission of quantifying the epidemiological relationships between aerosols and human health. The MAIA instrument's primary product is a level 2 aerosol particulate matter concentration measurement collected over cloud-free pixels. The quality of this product heavily depends on the validity of the cloud mask. In this project, we present a cloud masking algorithm for MAIA constrained to its hardware. It consists of 7 observables that are tested against predetermined static thresholds. Both observables and thresholds are a function of scene type, which is a unique combination of sun-view geometry, day of year and surface type, including a novel surface classification scheme derived from the Multi-Angle Implementation of Atmospheric Correction Bi-Directional Reflectance Distribution Function (MAIAC BRDF) data set. The cloud mask algorithm works by checking if an observation exceeds or falls short of a threshold for any of the 7 observables, resulting in a cloudy or clear classification. The thresholds are derived to match the performance of the Terra Moderate Resolution Imaging Spectro-Radiometer (MODIS) high-confidence-cloud cloud mask to achieve cloud conservative behavior. The algorithm allows tuning of the conservativeness by introducing the quantities of Distance-to-Threshold, Activation Value and number of tests to activate. These user-specified parameters determine how much confidence is needed for a cloudy or clear classification. The results are presented for the Los Angeles primary target area. The overall agreement between the MODIS cloud mask and the MAIA cloud mask (MCM) is 92.9%. Of the 7.1% disagreement, 60% of it was due to false positives by the MCM, considering MODIS as the truth. The MCM is more than 90% in agreement with MODIS for deep non-sun-glint water and the first 11 of the 16 snow-free land surface types. It differs from the MODIS cloud mask the most over bright desert, mountains and coastlines due to false cloudy flags. It agrees well with the MODIS cloud mask for cumulus, stratus and high cirrus, with greater disagreements over cloud edges, smoke plumes from wildfires, and very thin cirrus. The MCM agrees well with the MODIS cloud mask (>85%) for most solar zenith angles between 25 and 53 degrees, viewing zenith angles less than 60 degrees, and relative azimuth angles between 105 and 135 degrees. Several recommendations for improving the MCM are discussed, and its advantages over the MODIS cloud mask.
Issue Date:2021-04-26
Type:Thesis
URI:http://hdl.handle.net/2142/110518
Rights Information:Copyright 2021 Javier Villegas Bravo
Date Available in IDEALS:2021-09-17
Date Deposited:2021-05


This item appears in the following Collection(s)

Item Statistics