Laserfiche WebLink
<br />ithin 25 km of the radar, and the fourth is 49 km distant, the only site farther than the nearest Cleveland <br />s te. So the Denver sites generally had less calculated advection because of less vertical distance between <br />e gages and the lowest radar beam. Horizontal winds may have been less for the Denver cases, although <br />t is supposition has not been tested. <br /> <br />he least Denver R values are for site 7 (about 1.5 km from protected gage 2), where NWS personnel <br />ade hourly snow board observations of SWE at an airport location protected only by their building. <br />ind effects are believed to have degraded those observations and hence their degree of association with <br />r dar measurements. Observations made by trained professionals cannot compensate for an inherently <br />P100r observing location. Unfortunately, most official observations in the United States are made at windy <br />I cations, which adds variance to the snow measurements. <br /> <br />5. SAA REFINEMENT <br /> <br />1[ask 4.b.(1) of the MOU was originally intended to refer to Denver and Albany area SWE and SD <br />observations. However, because the final edited Albany data set was not available until the summer of <br />1 96, Cleveland data had to be substituted in research report R-04-96 (see J. Klazura memo of September <br />5 1995). Refinement of the SAA will be reported herein for Cleveland and Denver, and initial work with <br />e Albany data will also be discussed. <br /> <br /> <br />5 1 SAA Refinement for Cleveland and Denver <br /> <br />~lper and Holroyd (1996) used the partia1 data sets theo available from the Cleveland and Deaver areas <br />i developing optimized Ze -S relationships. Specifically, their fIrst annual report on SAA development <br />u ed only the fIrst 3 months of 1995-96 winter data from Denver and the fIrst 2 months from Cleveland. <br />Subsequently, the same approaches discussed by Super and Holroyd (1996) have been applied, with one <br />e~~eption, to the entire winter's data sets from those two areas. The exception is that nearest neighbor <br />rFge bins directly over each gage were used in the analysis reported here rather than approximately 3- <br />b~ 3-km arrays for the reasons discussed in section 4.1. <br /> <br />111 storms with significant snowfall were used to make the latest estimation of the optimum Ze-S <br />r!lationShips for both the Cleveland and Denver areas. That is, the values for a and {3 were empirically <br />itermined for the usual power law relationship of equation (1). <br /> <br />te recommended relationship for Cleveland by Super and Holroyd (1996) was based on only 143 hours <br />1ith detectable (0.005 inch or more) snowfall at the gage closest to the radar at 36 km range. The values <br />'jere 318 for aand 1.5 for {3. The then recommended Denver values were 155 and 1.6, based on the 3 <br />~ages within 49 km of the radar, which provided 196 hourly pairs of radar and gage data. Figure 1 of <br />Super and Holroyd (1996) shows that radar snowfall estimates are insensitive to the {3 exponent for <br />;fflectivities less than about 30 dBZ. <br /> <br />Jow that all Cleveland and Denver data are available, new calculations of optimized a and {3 values have <br />bEn made based on the nearest neighbor range bin directly over each gage. Data sets are for all hours <br />ith at least 0.005 inch SWE detected by Alter-shielded Universal gages in protected locations. The <br />s ngle exception is Denver's site 7 (NWS WFO), which had little protection from the wind except for the <br />o building just upwind. A snow board was sampled each hour at site 7, and 0.01 inch was the <br />inimum SWE resolution (trace observations have been assumed to equal 0.003 inch). Gage charts were <br />I' duced with magnification. No minimum radar estimated value was specified. <br /> <br /> <br />12 <br />