Splash artifacts in FSSP measurements—observations and flow modeling studies

The focus of this study is splash artifacts that contaminate FSSP measurements of cloud droplets. These artifacts can be generated when large cloud particles hit the forward edge of the sample tube on FSSP instruments. Some of the fragments from these collisions can be swept through the laser beam and lead to erroneous measurements of cloud droplets. This paper describes observational data that demonstrate the occurrence of this effect in liquid and mixed-phase clouds. This paper also describes air flow modeling studies that show the velocity field is affected by the presence of the sample tube in the FSSP and suggest that the trajectories of some splash fragments can intercept the laser beam.

The observational data were obtained from two projects: Instrumentation Development and Education for Atmospheric Science (IDEAS-3, August and September 2003 in Colorado) and the Rain in Cumulus over the Ocean (RICO, December 2004 and January 2005 in Antigua). The NCAR C-130 aircraft was used for both of these projects. During IDEAS 3, two FSSP-100 probes were mounted side-by-side on an under-wing instrumentation pod. One was a standard probe operated by the NCAR Research Aviation Facility (RAF) while the other was from GKSS (Germany). External modifications to the GKSS probe involved removing the sample tube and replacing the standard hemispheric caps on the ends of the arms with bullet-shaped tips. Both probes used the Signal Processing Package electronic interface upgrades (from Droplet Measurement Technologies, Boulder).

An example is shown in Figure 1, from IDEAS-3 in mixed-phase clouds at -13C. Concentrations from the two FSSPs showed close correspondence except when large snow particles were present in high concentrations, 21:48:00 to 21:48:20. During that time, concentrations in the RAF standard FSSP were about five times greater than the GKSS probe. Figure 2 shows particle size distributions as ten second averages from the FSSPs. Note the large discrepancy in the fourth panel, 21:48:05 to 21:48:15. We suspect that snow particles hitting the front of the standard FSSP sample tube generated a large number of small fragments, many of which passed through the laser beam and were counted. During rain penetrations from IDEAS-3 flights (plot not shown), the standard FSSP concentration exceeded GKSS when larger rain drops were present. Again, we think this is evidence of splash artifacts created by large drops hitting the sample tube in the standard FSSP.

FLUENT was used to simulate airflow and particle trajectories at the front end of both the standard and GKSS probes. Figure 3 shows an example of trajectory calculations for 50 um water drops in a flow of 100 m/s. Some of the drops hit the leading edge of the sample tube and are deflected into the laser beam. Although the physics of drop rebound and breakup mechanisms are not adequately represented in this simulation, the process suggests an explanation for the differences between data from the standard and GKSS probes.