Alternating decision trees for cloud masking in MODIS and VIIRS NASA sea surface temperature products

Katherine A. Kilpatrick, Guillermo P Podesta, Elizabeth Williams, Susan Walsh, Peter J Minnett

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Identification and exclusion of clouds from satellite-based infrared fields is critical to achieve accurate retrievals of sea surface temperature (SST). Historically, identification of clouds has been driven primarily by a few uniformity tests involving a small number of pixels, brightness temperature range tests, and comparisons to low-resolution gap-free reference fields. Collectively these tests are adequate at identifying large, upper-level, very cold cumulus clouds, and uniformity tests identify moderately sized patchy cumulus clouds. But the efficacy of cloud identification often decreases at cloud edges, for small or thin cirrus clouds, and for the lower, more uniform stratus clouds, for which cloud-top temperature can be comparable to that of the sea surface, particularly at high latitudes. The heavy reliance on stringent uniformity thresholds often also has the unintended consequence of eliminating strong SST frontal regions from the pool of best-quality retrievals. This paper presents results for an ensemble cloud classifier based on a machine-learning approach, boosted alternating decision trees (ADtrees), applied to NASA MODIS and VIIRS SST imagery. The ADtree algorithm relies on the use of a majority vote from a collection of both "weak" and "strong" classifiers. This approach offers the potential to identify more cloud types and improve the retention of SST gradients in best-quality SST retrievals and also provides a per pixel confidence estimate in the classification.

Original languageEnglish (US)
Pages (from-to)387-407
Number of pages21
JournalJournal of Atmospheric and Oceanic Technology
Volume36
Issue number3
DOIs
StatePublished - Mar 1 2019

Fingerprint

Decision trees
MODIS
NASA
sea surface temperature
Temperature
Classifiers
Pixels
cumulus
Thermal gradients
pixel
Learning systems
product
decision
VIIRS
Luminance
Satellites
stratus
Infrared radiation
cirrus
brightness temperature

Keywords

  • Algorithms
  • Classification
  • Climate records
  • Remote sensing
  • Satellite observations

ASJC Scopus subject areas

  • Ocean Engineering
  • Atmospheric Science

Cite this

Alternating decision trees for cloud masking in MODIS and VIIRS NASA sea surface temperature products. / Kilpatrick, Katherine A.; Podesta, Guillermo P; Williams, Elizabeth; Walsh, Susan; Minnett, Peter J.

In: Journal of Atmospheric and Oceanic Technology, Vol. 36, No. 3, 01.03.2019, p. 387-407.

Research output: Contribution to journalArticle

@article{2fd9cc385ba44d50996d755fa898b9c9,
title = "Alternating decision trees for cloud masking in MODIS and VIIRS NASA sea surface temperature products",
abstract = "Identification and exclusion of clouds from satellite-based infrared fields is critical to achieve accurate retrievals of sea surface temperature (SST). Historically, identification of clouds has been driven primarily by a few uniformity tests involving a small number of pixels, brightness temperature range tests, and comparisons to low-resolution gap-free reference fields. Collectively these tests are adequate at identifying large, upper-level, very cold cumulus clouds, and uniformity tests identify moderately sized patchy cumulus clouds. But the efficacy of cloud identification often decreases at cloud edges, for small or thin cirrus clouds, and for the lower, more uniform stratus clouds, for which cloud-top temperature can be comparable to that of the sea surface, particularly at high latitudes. The heavy reliance on stringent uniformity thresholds often also has the unintended consequence of eliminating strong SST frontal regions from the pool of best-quality retrievals. This paper presents results for an ensemble cloud classifier based on a machine-learning approach, boosted alternating decision trees (ADtrees), applied to NASA MODIS and VIIRS SST imagery. The ADtree algorithm relies on the use of a majority vote from a collection of both {"}weak{"} and {"}strong{"} classifiers. This approach offers the potential to identify more cloud types and improve the retention of SST gradients in best-quality SST retrievals and also provides a per pixel confidence estimate in the classification.",
keywords = "Algorithms, Classification, Climate records, Remote sensing, Satellite observations",
author = "Kilpatrick, {Katherine A.} and Podesta, {Guillermo P} and Elizabeth Williams and Susan Walsh and Minnett, {Peter J}",
year = "2019",
month = "3",
day = "1",
doi = "10.1175/JTECH-D-18-0103.1",
language = "English (US)",
volume = "36",
pages = "387--407",
journal = "Journal of Atmospheric and Oceanic Technology",
issn = "0739-0572",
publisher = "American Meteorological Society",
number = "3",

}

TY - JOUR

T1 - Alternating decision trees for cloud masking in MODIS and VIIRS NASA sea surface temperature products

AU - Kilpatrick, Katherine A.

AU - Podesta, Guillermo P

AU - Williams, Elizabeth

AU - Walsh, Susan

AU - Minnett, Peter J

PY - 2019/3/1

Y1 - 2019/3/1

N2 - Identification and exclusion of clouds from satellite-based infrared fields is critical to achieve accurate retrievals of sea surface temperature (SST). Historically, identification of clouds has been driven primarily by a few uniformity tests involving a small number of pixels, brightness temperature range tests, and comparisons to low-resolution gap-free reference fields. Collectively these tests are adequate at identifying large, upper-level, very cold cumulus clouds, and uniformity tests identify moderately sized patchy cumulus clouds. But the efficacy of cloud identification often decreases at cloud edges, for small or thin cirrus clouds, and for the lower, more uniform stratus clouds, for which cloud-top temperature can be comparable to that of the sea surface, particularly at high latitudes. The heavy reliance on stringent uniformity thresholds often also has the unintended consequence of eliminating strong SST frontal regions from the pool of best-quality retrievals. This paper presents results for an ensemble cloud classifier based on a machine-learning approach, boosted alternating decision trees (ADtrees), applied to NASA MODIS and VIIRS SST imagery. The ADtree algorithm relies on the use of a majority vote from a collection of both "weak" and "strong" classifiers. This approach offers the potential to identify more cloud types and improve the retention of SST gradients in best-quality SST retrievals and also provides a per pixel confidence estimate in the classification.

AB - Identification and exclusion of clouds from satellite-based infrared fields is critical to achieve accurate retrievals of sea surface temperature (SST). Historically, identification of clouds has been driven primarily by a few uniformity tests involving a small number of pixels, brightness temperature range tests, and comparisons to low-resolution gap-free reference fields. Collectively these tests are adequate at identifying large, upper-level, very cold cumulus clouds, and uniformity tests identify moderately sized patchy cumulus clouds. But the efficacy of cloud identification often decreases at cloud edges, for small or thin cirrus clouds, and for the lower, more uniform stratus clouds, for which cloud-top temperature can be comparable to that of the sea surface, particularly at high latitudes. The heavy reliance on stringent uniformity thresholds often also has the unintended consequence of eliminating strong SST frontal regions from the pool of best-quality retrievals. This paper presents results for an ensemble cloud classifier based on a machine-learning approach, boosted alternating decision trees (ADtrees), applied to NASA MODIS and VIIRS SST imagery. The ADtree algorithm relies on the use of a majority vote from a collection of both "weak" and "strong" classifiers. This approach offers the potential to identify more cloud types and improve the retention of SST gradients in best-quality SST retrievals and also provides a per pixel confidence estimate in the classification.

KW - Algorithms

KW - Classification

KW - Climate records

KW - Remote sensing

KW - Satellite observations

UR - http://www.scopus.com/inward/record.url?scp=85063689115&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85063689115&partnerID=8YFLogxK

U2 - 10.1175/JTECH-D-18-0103.1

DO - 10.1175/JTECH-D-18-0103.1

M3 - Article

AN - SCOPUS:85063689115

VL - 36

SP - 387

EP - 407

JO - Journal of Atmospheric and Oceanic Technology

JF - Journal of Atmospheric and Oceanic Technology

SN - 0739-0572

IS - 3

ER -