Laserfiche WebLink
<br />a function of wavelength and the coefficients Gl, G2 and C. Therefore, <br />in order to evaluate the Amax, a series of weighting functions and <br />response functions for wavelengths from 10 to 4000 km were plotted, <br />Fig. 3.2. From these families of curves, initial band-pass coefficients <br />Gl, G2, Cl and C2 may be selected which provide the appropriate macro- <br />scale and mesoscale weighting functions. These coefficients are then <br />analyzed by a routine which tabulates a series of weights and responses <br />at different separation distances and computes the band-pass and total <br />response functions and the value of Amax. The composite of response <br />functions (Fig. 3.1) shows the characteristics of a selected set of <br />band-pass coefficients. When the most appropriate band-pass filter is <br />designed, it may be used in the objective analysis. <br /> <br />In this research, two sets of filters were used in all objective <br />analyses. A filter with Amax at 500 km was used for all larger scale <br />analyses that incorporated the 0000 GMT NWS rawiosonde data using a grid <br />of ~x = 10. All detailed meso net analyses (~x = 0.20 longitude) which <br />focused on the Texas HIPLEX soundings alone used a filter having Amax at <br />about 200 km. Rawinsonde stations were separated by 80 to 110 km. <br />Sensitivity analyses of the objective analysis to band-pass filter on <br />the scale from 150 to 400 km showed that this filter provided the most <br />meteorologically acceptable contour analyses and provided an optimum mix <br />of macroscale smoothing with mesoscale detailed structure. <br /> <br />3.2 Automated Graphical Analysis <br /> <br />In order to assimilate the large volume of information generated <br /> <br />in the multilevel, multifield objective analysis, graphical display <br /> <br />routines were developed which: <br /> <br />43 <br />