2.3
A Technique for Improving the Discrimination between Precipitation Types using the ASOS Present Weather Sensor
Charles G. Wade, NCAR, Boulder, CO
The ASOS Present Weather Sensor is the Light-Emitting Diode Weather Indicator (LEDWI). It detects the presence of precipitation by directing an infrared light beam over a 1-meter path towards a photo diode receiver. As a particle of rain or snow passes through the beam it creates a shadow which modulates the frequency of the light falling on the receiver. The shadow varies depending on the size and fall speed of the particle. When many particles fall through the beam, a jumble of frequencies is produced. A spectral analysis of the received signal reveals how much energy or power is contained in various frequency bands. By comparing the energy contained in the "high" frequency band (1000-4000 Hz), with the energy received in the "low" frequency band (75-250 Hz), the LEDWI algorithm is able to determine whether the precipitation is rain or snow, and estimate the intensity of the precipitation.
When the precipitation is entirely rain or snow, the LEDWI algorithm generally does an excellent job of determining precipitation type. Problems begin to occur when the precipitation is other than purely rain or snow, such as mixed rain and snow, drizzle, or snow or ice pellets. During these events the algorithm may falsely identify the precipitation type, or it may report "unknown" precipitation.
For a number of years, NCAR has been evaluating the performance of the LEDWI at its winter weather test facility, located just outside of Boulder, CO. During each precipitation event, a record is made (by either a human observer or a video camera) of the precipitation type, quantity (using a number of automated and manual gauges), and of the start and stop times for precipitation. The LEDWI data is then analyzed and compared to this record. A unique method for displaying the raw LEDWI data (high and low frequencies) has been developed that results in an improved method for discriminating between precipitation types. This method shows that it is possible to detect drizzle as distinct from rain, to identify mixed precipitation, and to identify snow and ice pellets. The purpose of this paper is to describe this improved method for identifying precipitation type, and to suggest that it can be used to improve the accuracy of the ASOS precipitation identification.
Session 2, Advances in use of observational data
Monday, 10 January 2000, 1:30 PM-5:15 PM
Previous paper Next paper