Layer Analysis in ARCIONS John Merrill University of Rhode Island

1
Layer Analysis in ARC-IONSJohn
MerrillUniversity of Rhode Island

• Outline of new approaches to finding layers in
  sonde profiles and analyzing their occurrence.
  
• Filtering, to produce an unperturbed profile, is
  avoided.
  
• Instead a straightforward, if heuristic approach
  is used.
  
• This is applied here to data from the Arc-IONS
  ozonesonde network.
  

2

• Examples of unlayered and layered profiles
  shown.
  
• The tropopuase and lowermost 2 km of the
  stratospheric are indicated.
  
• Averaged MR in 100 m intervals, buth with
  appropriate choice of parameters the analysis can
  be applied to unsmoothed data.

3

• Begin with a vertical line intersecting the
  profile.
  
• Think of the line as the stile of a window, which
  is moved across the distribution in increments of
  the tolerance interval, 7 ppb, say, creating a
  set of parallel lines.
• Groups of three or more adjacent intersection
  points make up a putative layer edge, here marked
  with a crimson sign.
  
• A layer is defined to have edges with opposite
  slopes, with a limited separation between the
  edges at or near the peak.

4

• Begin with a vertical line intersecting the
  profile.
  
• Think of the line as the stile of a window, which
  is moved across the distribution in increments of
  the tolerance interval, 7 ppb, say, creating a
  set of parallel lines.
• Groups of three or more adjacent intersection
  points make up a putative layer edge, here marked
  with a crimson sign.
  
• A layer is defined to have edges with opposite
  gradient, with a limited separation between the
  edges at or near the peak.

5

• The sign and magnitude of the gradients (of the
  edges) are shown to the right.
  
• The extreme values for two tropospheric layers
  are shown. A lower stratospheric layer peak is
  off scale.
  
• The search does not extend above the lowermost
  stratosphere.
  
• The amplitude, thickness and mean vertical
  derivative are found simply.

6

• It is acknowledged that the bump in the
  unlayered profile could be counted as a layer.
  Careful specification of whats Not A Layer may
  suffice.
• Vertical distributions will be plotted relative
  to the thermal (WMO) tropopuase.
  

7
Tropopuase-relative height vs. Amplitude (ozone
mixing ratio increment) for westernmost sites.
8
Tropopuase-relative height vs. layer thickness
for four interior sites.
9
Scatter plot of gradient vs. layer amplitude for
easternmost sites.
10

• Diminished layering noted at some sites. Note
  also reduced contrast betweeen the upper
  troposphere and lower stratosphere.

11

• Weaker dominance of LS vs. UT layers is common at
  sites well east of the mountains. This begs
  compare/contrast with dynamical diganostics
  and/or LID characterization.

12
Modest response time of ECC cell may lead to
underestimate of amplitude (extreme values).
13
Tropopause relative height vs. temporal
variability of ozone mixing ratio. Layer
occurrence a secondary factor (albeit
underestimated?).
14

• Event-specific geographical variation of layering
  is next target. Some relationships dont require
  this (left), while others do (right).

15
Finis
16
Acknowledgements

• Support provided by NASA as part of the AURA
  Validation effort.
  
• Kathryn George, an undergraduate at Slippery
  Rock U. supported over the summer by an
  NSF-sponsored site REU program.
  
• Data provided by valued colleagues and carefully
  archived by J. Witte.