2D processing changes June 2009

 
A note on changes to 2D-processing for NCAR/NSF aircraft in ICE-L, START-08 and PACDEX projects
 

Background:

There is no standard for processing of 2D-data from PMS 2D-C (25 and 10-micron resolution) and PMS 2D-P probe (200-micron resolution). RAF is currently working on upgrades to 2D-processing as we strive for producing better and more easily usable data.

For the three projects (ICE-L, START-08 and PACDEX), RAF has produced spectra using the so-called 1-D probe emulation which essentially gives the particle size based on the maximum shaded pixels during any time of a particle passage during the laser beam.

For very small particles the Depth-of-Field is associated with considerable uncertainty. From theory, the Depth-of-Field is well defined; however, the correction for sample volume changes very dramatically for small particle sizes. Thus if there is a small error in sizing, the concentration of a particle may change dramatically. In the original ICE-L, START-08 and PACDEX data, RAF set the depth of field to the fixed distance between the probe arms, and it was left up to individual users to correct the size distributions using the Depth-of-Field of their choice.

RAF is currently working with MMM (Aaron Bansemer and Andy Heimsfield) on implementing much more sophisticated 2D-algorithms. This is a development project, both as far as algorithms are concerned and as far as implementation into RAF-s standard processing. Accordingly it cannot be implement in the short term.

Changes to RAF-s 2D-processing and contents of re-run of data:

RAF has made six main changes to the processing in the ICE-L, START-08 and PACDEX data that will be re-released shortly. They are:

(1)   A netCDF attribute has been added to the concentration spectra variables C1DC called DepthOfField, which contains a vector with the depth of field per channel (similar to CellSizes).

(2)   The 2D-particle concentration spectra have calculated using this Depth-of-Field. Accordingly the concentrations in the smaller bins are now higher than previously.

(3)   The 2D-particle concentration spectra have had the concentrations in the first two bins excluded. The reason is the above-mentioned uncertainty arising in bins 1 and 2 due to sizing errors and the very small Depth-of-Field in these bins. Users can still calculate particle concentrations in bins 1 and 2 by using the theoretical Depth-of-Field in conjunction with the 2D-particle count spectrum; RAF consider that this is too uncertain.

(4)   The 2D-C total particle concentration, CONC1DC, now only include the particles in bins 3 and larger. The concentrations will thus differ from previous values due to the changed Depth-of-Field and the omission of bins 1 and 2.

(5)   By excluding bins 1 and 2 of the 2D-C concentration spectrum, the nominal minimum size (lower bin limit) of bin 3 is 62.5 micron (for the 25-micron resolution 2D-probe). Thus the total particle concentration from the 2D probe is CONC1DC of all particles greater than 62.5 micron (nominally).

(6)   An error was found in the algorithm for accepting particles ending in a blank slice; this has been corrected with the result that more particles have been accepted.

Release of re-processed data:

The re-run of the data is currently underway, and this is a relatively short process. Merging in other special data, as was previously done, will take additional time. The same applies to uploading the data for web access.

Users with urgent needs, e.g. for submitting publications with a fixed short-term deadline, can contact Janine Goldstein (janine@ucar.edu) for rushed data access (minus merged data sets etc.). However, unless you have such an urgent need we request that you wait until the completed data sets are on the web.

On behalf of RAF - Jorgen Jensen
jbj@ucar.edu
303-497-1028 (office)
720-206-6478 (cell)