88th Annual Meeting (20-24 January 2008)

Wednesday, 23 January 2008
Evaluation of an MPL Cloud Detection Algorithm as a Reference for Ceilometer Testing Within the ASOS PI Program
Exhibit Hall B (Ernest N. Morial Convention Center)
Aaron J. Poyer, NOAA/NWS, Sterling, VA
Poster PDF (713.1 kB)
The current standard cloud height indicator (CHI) for the Automated Surface Observing System (ASOS) is the National Weather Service (NWS) K220 laser ceilometer (K220). This ceilometer detects clouds to a height of 12,000 feet. In late 1998, the manufacturer discontinued production of the K220. The vendor agreed to maintain the existing ASOS ceilometers through 2007. New ceilometers will need to be deployed to the ASOS network accordingly.

The NWS awarded a contract for development of a K220 replacement ceilometer on January 05, 2007 to Vaisala Inc. of Woburn, Massachusetts.

In support of the acquisition and qualification of this replacement ceilometer, which is now specified to a range of 25,000 feet, it was determined that a research LiDAR (Light Detection And Ranging) sensor would be necessary to establish reference cloud heights for assessment of the candidate replacement ceilometer. The ASOS Product Improvement (PI) Program has acquired a Micro Pulse Lidar (MPL) from Sigma Space of Lanham, Maryland, which has an algorithm capable of cloud base detection beyond the specified range of the replacement ceilometer.

The test approach was to compare cloud bases identified by the MPL's cloud detection algorithm with those indicated by all available references (human, K220, radiosonde) as a means to validate the MPL as an independent reference for measuring cloud height and coverage. Comparisons utilized synchronized individual cloud scan data from the K220 and the MPL along with human observed cloud layers during conditions with uniform non-ragged cloud bases. The K220 cloud height reports and those of the MPL's cloud detection algorithm were compared for cloud bases from 2,000 feet to 12,000 feet. Human observed cloud heights as well as available radiosonde flight data were compared to those of the MPL's cloud detection algorithm, for heights throughout the range of the K220 replacement ceilometer, during all covered events. Human observations were taken at minimum intervals of one hour during business hours and at more frequent intervals when observers were on site for other weather events.

Testing was conducted at the NWS test and evaluation facility in Sterling, Virginia from January 2007 through September 2007.

Supplementary URL: