Laserfiche WebLink
<br />30 <br /> <br />.......... <br />Ul <br />" <br />E <br />-20 <br />"0 <br />Q) <br />Q) <br />0.. <br />Ul <br />-0 <br />C <br />~ <br /> 10 <br />0:: <br />> <br />U <br /> " <br /> 0 <br /> 0 <br /> <br />7' <br /> <br />7. 7. <br /> <br />n <br /> <br />7. <br />7. <br /> <br />7' <br />n " 71 <br />7' "",,nnn 7 n <br />7' 7. 7. 7. JI n 7. n <br /> <br />nn 7. <br />" n " 'h. <br />H.7. 7. 7' <br />Jl7H. <br />n7.77. <br /> <br />n <br />n <br /> <br />7' <br /> <br />7' <br /> <br />10 20 30 <br />TUS wind speed (m/s) <br /> <br />Figure 10. - OVR 750-mb wind speeds versus corresponding TUS <br />wind speeds. <br /> <br />Scatterplots of the HJ CLW versus CVR 700- and 750-mb relative humidities and wind speeds <br />are respectively given on figures 11 and 12. The figures show that generally, higher CLW <br />measurements were associated with higher relative humidities and wind speeds. Comparisons <br />at other pressure levels were substantially more noisy. <br /> <br />Figure 13 contains the scatterplot for HJ CLW values versus INW 700- and 750i-mb wind <br />speeds. Similarly, figure 14 portrays TUS relative humidities for those levels. The results are <br />similar to those for CVR in the case of wind speed but are noticeably more noisy in the case of <br />TUS relative humidities. <br /> <br />Regression results <br /> <br />Initially, sounding information from CVR was employed in MLR to estimate HJ CLW, the <br />dependent variable. In particular, relative humidities and wind speeds from several levels were <br />inserted as independent variables in stepwise MLR. AI; indicated in table l(a), a multiple <br />correlation coefficient of 0.67 was obtained with the two independent variables, (1) the mean of <br />700- and 750-mb relative humidities, and (2) the mean of 700- and 750-mb wind speeds. The <br />sample size was 87 and no more than 1 h of overlap was allowed between 7-h CLW periods. <br /> <br />Table 1 also contains regression results in which sounding data from INW and TUS were <br />included as possible independent variables in stepwise MLR. An improved multiple correlation <br />of 0.79 was obtained when INW 750-mb wind speed was selected as the second independent <br />variable, as indicated in table 1(b). Why a noticeably higher multiple correlation was obtained <br />with the INW versus the CVR wind is not obvious. The sample size was again, 87, so the results <br />were probably not an artifact of a small sample. <br /> <br />9 <br />