Characterizing Variability of Synoptic and Cloud Regimes over the ARM Azores Site and the Northeast Atlantic C. S. Fish1, D. B. Mechem1, M. A. Miller2, S. E. Yuter2, and S. P. de Szoeke3

1Atmospheric Science Program, Department of Geography, University of Kansas; 2Department of Marine, Earth, and Atmospheric Sciences, North Carolina State University; 3College of Oceanic and Atmospheric Sciences, Oregon State University

The Azores lie in a region of substantial variability in cloud and synoptic properties, as demonstrated during a 20-month deployment of the ARM Mobile Facility (AMF) on Graciosa Island. We explore this variability using the clustering technique of self-organizing maps (SOMs). The SOM algo-rithm �nds twenty-�ve characteristic synoptic states based on monthly 500-mb geopotential height anomalies from ERA–Interim (reanalysis). Once these characteristic states are determined, other �elds from reanal-ysis, satellite data, and surface observations are projected onto these characteristic states. For every month, the clustering algorithm consis-tently identi�es four archetypical synoptic patterns based on the rela-tionship of the 500-mb geopotential pattern relative to the Azores: pre-trough, trough, post-trough, and ridge. The location and persistence of low clouds are strongly related to the location and intensity of the Ber-muda High. Post-trough and ridge patterns are the most conducive to liquid-phase clouds in the vicinity of the Azores and are accompanied by surface-level high pressure and subsidence, and the region, even in summer, is punctuated by frequent synoptic intrusions.

Introduction and motivation

Summary and discussion • The location and persistence of clouds over the Northeast Atlantic are strongly relat-ed to the location and intensity of the Bermuda High but also synoptic systems associ-ated with the Icelandic low. • We find evidence of synoptic “intrusions” over the Azores region in the cloud fraction and near-surface height �eld, but these occurrences are not associated with strong 500-mb forcing over the Azores themselves.

Synoptic Pattern Annual Cycle

• The SOM algorithm successfully identifies synoptic regimes and transitions on multiple timescales over the Northeast Atlantic.• More often than not, the Azores lie in a gradient of synoptic and cloud properties.

We focus on the following objectives:•Identifying preferential synoptic states over the NE Atlantic •Projecting cloud system properties onto these synoptic states to explore joint variability between cloud and synoptic states.

This research is supported by the DOE O�ce of Science grant DE-SC0006736.

RidgeZonalPre−troughTroughPost−troughUnclassified

1 2 3 4 5 6 7 8 9 10 11 120

10

20

30

40

50

60

70

80

90

100

Relative Frequency of Synoptic Patterns (annual anomalies)

Month

Freq

uenc

y [%

]

500-mb Geopotential Heights

1000-mb Geopotential Heights

Vertical Velocity at 500-mb

Pre-trough(Node 25)

Post-trough(Node 1)

Ridge(Node 5)

−50 2505100 6000[m] [m]

0.−0.1 0

1(1 cm s-1 up) (1 cm s-1 down)

Trough(Node 21)

1 20000[g m ]-3

ERA-Interim Liquid+Ice

MODIS TotalCloud Fraction

30

40

50

0 0.5 1

ERA-I CF

January Synoptic and Cloud Properties (monthly anomalies)

To examine the annual cycle, 500mb geopotential heights over the entire data set are used to ini-tialize the SOM algorithm.

Even though ridge conditions associated with the Bermuda high dominate the summer months, the Azores nevertheless experiences periodic intrusions of synoptic systems. This necessitates a closer examination of individual months, namely June and January as representatives of summer and winter, respectively.

June Synoptic and Cloud Properties (monthly anomalies)

1. Choose period of interest and �eld on which to run SOM analysis: June and January 500-mb geopotential heights from ERA–Interim, 1979–2012 (1020 and 1054 samples respectively, one for each day), or the entire January-to-December period

Analysis procedure

2. Calculate normalized anomalies3. Choose number of nodes: 254. Run SOM algorithm5. Project MODIS total cloud fraction and other ERA–Interim �elds onto SOM nodes

• All months contain some level of pre-trough patterns.

clearovercast

[%]

500-mb Geopotential Heights

1000-mb Geopotential Heights

Vertical Velocity at 500-mb

Pre-trough(Node 5)

Post-trough(Node 21)

Ridge(Node 25)

Trough(Node 1)

−50 2505100 6000[m] [m]

0.−0.1 0

1(1 cm s-1 up) (1 cm s-1 down)

10 20000[g m ]-3

ERA-Interim Liquid+Ice

MODIS TotalCloud Fraction

0 0.5 1

ERA-I CF

[%]

Janu

ary

June

All Pre-trough Post-trough RidgeTrough

Icelandic low is present in January, and the Bermuda high is present in both January and June

Post-trough and ridge patterns are the most conducive to clouds in the vicinity of the Azores and are accompanied by

surface-level high pressure and subsidence.

Joint density plot demonstrates no cloud fraction sensitivity to vertical motion, except for a weak dependence in the

trough and post-trough regimes.

Joint PDFs of ERA-Interim Cloud Fraction and Vertical motion at the Azores

• Winter months are primarily dominated by trough and post-trough patterns.• Summer months are dominated by ridge and zonal pattens.

Holton, 2004

RidgePost-troughTroughPre-trough

Nodes are catego-rized by which weather pattern is the most dominant feature over the domain.