Laserfiche WebLink
<br />FR = 1.04607 - 0.OO29590*R + 0.OOOO506*R2 <br /> <br />(3) <br /> <br />I <br />I <br />I <br />I <br />I <br />I <br />I <br />I <br />I <br />I <br />I <br />I <br />I <br />I <br />I <br />I <br />I <br />I <br />I <br /> <br />4.2 Early Modifications <br /> <br />A wind advection scheme was developed using the radar's velocity-azimuth display (V AD) <br />vertical wind profile, which compared radar estimates from upwind range bins to those fmm <br />range bins directly over the gauges. This scheme revealed no improvement. In fact, the <br />advection scheme increased variability beyond 100 km of the Cleveland radar. Moreover, the <br />use of a range of assumed snowflake fall speeds did not improve analysis results. This logic was <br />removed from the SAA after the first year of effort. <br /> <br />The SAA was modified to convert occultation and hybrid scan files from binary representation to <br />ASCII. This allows text editor adjustments of the values to account for surface features not <br />detected by the digital terrain data-based algorithms that generated the initial files. It also allows <br />use of the closest beam to the ground that does not generate ground clutter. <br /> <br />A simple range correction scheme was devised before testing at Albany (see section 5). Such a <br />range correction was specified by the memorandum of understanding with the OSF and was <br />described in Super (1998). This scheme applies a second-order polynomial multiplier beyond a <br />stated range (like 35 km). The snow precipitation rate was: <br /> <br />S = (Z/a)**(l/p), <br /> <br />(2) <br /> <br />where a = 150 and P = 2.0 for the north-central U.S. The range correction factor FR was, for <br />range R > 35 km, <br /> <br />Determining the appropriate coefficients for use with this correction scheme is a challenge that <br />must be met for each climate zone. Widespread synoptic storms need less correction than <br />shallow storms. An optional routine to remove speckle from the SAA accumulations was added <br />using a median filter on arrays of 3 km by 3 degrees of reflectivity data. <br /> <br />4.3 Real-Time Data Ingest <br /> <br />The early version of the SAA operated on archived Ze data, obtained on tapes from the National <br />Climatic Data Center (NCDC). These data had the following resolutions out to a 230-km <br />maximum range: 1 km in range, lOin azimuth, and 0.5. dB in reflectivity. Later, the algorithm <br />was converted to use real-time data from the National Severe Storms Laboratory's (NSSL) <br />Waming Decision Support System (WDSS), which was hosted at several of the testing Weather <br />Forecast Offices. <br /> <br />A later conversion allowed the use of real-time data from a NIDS vendor. These NIDS data were <br />available for all radars in the north-central U.S. Despite the reduced reflectivity resolution (see <br />section 3.1) and limited elevation scans (lowest four tilts) of these data, SAA accuracy did not <br />suffer appreciably. <br /> <br />12 <br />