Laserfiche WebLink
<br />I <br /> <br />I <br /> <br />In the longer term, we would like to collect data and develop sile-specific VPR corrections <br />(Huggins, personal communication). Such corrections hold promise for more accurate OPEs in <br />mountainous terrain (see section 5). For further improvement. it would be desirable to sample <br />actual VPRs near the radar (if precipitation exists there) and use them to extrapolate from the <br />lowest beam 10 the surface allong ranges in real time. Such a correction would demand <br />considerably more computer resources. <br /> <br />Just distinguishing rain from snow, <br />operationally. can be a daunting task. because <br />the meteorological pnx'esses controlling the <br />phase al the surface arc often subtle. Even <br />aflcr an event, researchers may differ as to <br />which processes were most influential (e.g., <br />Sleigerwaldt, 1998; Kain et aI., 2(XX)). We <br />adopted a simple scheme that classifies radar- <br />sensed precipitation into three types: rain, melting ice or slush (bright hand). and snow. This <br />modification is the most ambitious one undertaken in the effort to convert lhe SAA into it new. <br />more generic PAA. The scheme is based on vertical profiles of tcm~rature and moisture and is <br />described below. <br /> <br />I <br />I <br />I <br /> <br />6.3 Prt'dpitation T,)'pe Classification <br /> <br />I <br />I <br />I <br /> <br />f\1ix.cd-phase precipilation has heen another <br />major hindrance to SAA accuracy. As related <br />in section 4.1. Ihe SAA was designed for <br />"dry" snow only. If the melting level is aloft, <br />tbe algorithm will convert data from rain areas <br />10 snowfall S because (here were no crileria to <br />ascertain if snow is melting, causing errors in <br />SD estimates. Furthermore, in such instances. <br />Ihe radar bright band will ex.ist. compounding <br />the problem by producing significant radar <br />QPE overeslimalion (figure 6). The unusually <br />warm I 999-2()(x) season had an abundance of <br />such cases. Therefore, some method was <br />needed to identify such situations and <br />segregate these different precipitation <br />regimes. <br /> <br />I <br />I <br />I <br />I <br />I <br />I <br />I <br />I <br />I <br />I <br />I <br /> <br /> <br />Figure 6.-$ accumulation over 6 hours for the <br />KABR radar, Aberdeen, South Dakota (at center). <br />Ring of $ > 0.25 inch is largely overestimation <br />caused by bright band elleet just below the melting <br />(0 oc) level. which was about 2.2 km Above <br />Ground level (AGL) (reached by 0.50 beam center <br />at approximately midrange. or 120 km). <br /> <br />19 <br />