A Detailed Isobaric Surface Analysis

Though the detailed surface analysis of the previous article (not yet here) makes many weather features evident, hand-plotting and analyzing is a laborious task - so much so that doing this for NWS North American surface charts (with generally less detail and much less data) once required teams including an analyst for each region. Now their surface charts are computer-plotted (example), with isobars often objectively-analyzed. This method, though not presently with quite the flexibility of hand-plotting, is now possible using a PC. For this, I recommend Digital Atmosphere. Using it, you can choose dimensions and color of your base map, add many geographical parameters of your choice, and select among many station model options; though plots can overlap if too densely-packed (a problem sometimes unavoidable hand-plotting even).

Below is a 00 UTC 16 NOV 1998 surface analysis using Digital Atmosphere:

85% of stations are plotted so cluttering is minimized, though all are used for the isobaric analysis. Isobars are analyzed using their nearest neighbor method using a smoothing coefficient of .20 and no additional smoothing passes. This provided what I feel is the best objective analysis - isobars are not too jagged, nor so smooth such that some features such as the ridge/trof couplet over the Dakotas to northern Wisconsin are not visible. A local Low is analyzed in the trof extending northwestward from the main Low over southeast South Dakota, and a local High northwest of Lake Superior. Altimeter settings rather than sea level pressures are plotted, which adds detail for this map (because more were reported and plotted) and quite possibly is more representative for this situation. (Altimeter settings use difference among station readings and the standard atmosphere, and a standard atmosphere from station elevation to sea level. Sea level pressure estimates use temperature at a station for adjusting station readings using a specific lapse rate. Thus, altimeter settings would be perfectly representative for a map of flat terrain, and sea level pressures would be perfectly representative if temperatures varied with the specified lapse rate on variable terrain.)

More so than suggesting the idea of letting a PC do the laborious tasks, I illustrate in this article a detailed isobaric analysis and advantages of using as much data as possible. Below I present hand-analyzed isobars on the 00 UTC 16 NOV 1998 chart with 15%, 40%, and 80% of data plotted and a preliminary analysis using Digital Atmosphere for station plots. (You can compare these charts with those at some WWW sites for station densities typically used.) These analyses were admittedly done after the previously-shown chart. Thus I ignored what I knew about the complete analysis the best I could. Place your mouse over letters on each for comments and reasoning for my analysis.

Below is the chart with 15% of data plotted. With relatively few observations, the best which can usually be done is a smoothed analysis. This is generally linearly-interpolated - the best reasons for doing anything significantly different is bad data and drawing isobars closest together where winds are strongest. Place mouse over words or the script L for more commentary.

Below is the chart with 40% of data plotted. More detail is visible.

Below is a preliminary analysis for the chart with 80% of data plotted. This is my initial sketch of a reasonable final analysis (which the 80% chart is - all data could be plotted, but that clutters the map, mainly with observations near those already plotted). The region of the cyclone is focused on. The map letter comments mention many considerations for final analysis.

Below is the chart with 80% of data plotted. I think this isobaric analysis is better than that on the hand-plotted map (shown in an article not yet here) because it contains more data - altimeter settings rather than sea level pressures. Main features are similar though, with a few new ones. Clicking on the D with an asterisk offers a further discussion.

It may be questioned why the cold front is drawn well ESE of IML with its temperature reported as 73°. That could be a bogus report, though perhaps there was a local flow in the front range of the Rockies causing subsidence and adiabatic warming - or because of local topography, the cold air did not displace the warm air so quickly there. That is unusual after a cold front, but often occurs after warm fronts (lighter air not easily replacing denser air). Note in the METAR reports below that MCK briefly became warmer (after the supposed frontal passage) with WNW winds from the direction of IML between 02 & 04 Z, and IML became much cooler (as reported).

 KIML 152058Z 31006KT 15SM SCT050 BKN150 23/11 A2967 RMK T02300111
 KIML 152159Z 00000KT 15SM SCT050 BKN150 22/14 A2965 RMK T02200136
 KIML 152159Z 00000KT 15SM SCT050 BKN150 22/14 A2965 RMK T02200136
 KIML 152353Z 00000KT 15SM SCT050 BKN150 23/08 A2967 RMK T02300083 DAYTIME HIGH 23.3 CELSIUS
      24HR LOW MISSING
 KIML 160058Z 30006KT 15SM BKN040 BKN120 10/02 A2970 RMK T01000020
 KIML 160058Z 30006KT 15SM BKN040 BKN120 10/02 A2970 RMK T01000020
 KIML 160259Z 28010KT 15SM BKN040 BKN120 04/M06 A2973 RMK T00401057

 KMCK 152353Z AUTO 07003KT 10SM CLR 16/06 A2966 RMK AO2 SLP044 T01610056 10211 20161 55001 TSNO
 KMCK 160053Z AUTO 19003KT 10SM CLR 14/06 A2966 RMK AO2 SLP047 T01390061 TSNO
 KMCK 160153Z AUTO 25006KT 10SM CLR 08/04 A2967 RMK AO2 SLP057 T00830044 TSNO
 KMCK 160253Z AUTO 29010KT 10SM CLR 10/06 A2969 RMK AO2 SLP057 T01000056 53009 TSNO
 KMCK 160353Z AUTO 29008KT 10SM CLR 12/04 A2972 RMK AO2 SLP061 T01220039 TSNO
 KMCK 160453Z AUTO 27006KT 10SM CLR 11/03 A2975 RMK AO2 SLP067 T01060033 TSNO
 KMCK 160553Z AUTO 30010KT 10SM CLR 09/02 A2977 RMK AO2 SLP075 T00940022 10156 20083 402111017 
       51027 TSNO
 KMCK 160653Z AUTO 32009KT 10SM CLR 07/02 A2980 RMK AO2 SLP083 T00720017 TSNO


 KLBF 152156Z 19009KT 10SM CLR 19/M01 A2963 RMK AO2 SLP033 T01891011 $
 KLBF 152256Z 18007KT 10SM CLR 16/M01 A2965 RMK AO2 SLP038 T01611011 $
 KLBF 152356Z 00000KT 10SM CLR 11/M01 A2964 RMK AO2 SLP040 T01111011 10211 20111 56003 $
 KLBF 160056Z 00000KT 10SM CLR 09/M01 A2967 RMK AO2 SLP053 T00941011 $
 KLBF 160156Z 29003KT 10SM CLR 07/M01 A2968 RMK AO2 SLP059 T00721011 $
 KLBF 160256Z 32009KT 10SM CLR 08/M01 A2970 RMK AO2 PRESRR SLP065 T00781011 51019 $

Surface flow was weak at all locations - which supports a more gradual temperature decrease as at MCK & LBF rather than the more abrupt one for IML. Evening cooling (sunset was about 2325 Z) may differ because of local effects, though the surface winds and clouds do not seem to be much of a factor. The front seems to be clearly past LBF at 00 Z or in the process of passing, judging from temperatures & winds - and seems not passed at MCK with a light southerly breeze at 0053 Z and not much cooler than at 2353 Z. So maybe the front was a little further NW in the locale, though if anything this illustrates how fronts are simply small transition zones separating large air masses of basically different characteristics. Any more commentary regarding this aspect would only be appropriate with a separate mesoscale analysis

Hopefully seeing how analysis changes as more stations are included is helpful. An advantage of objective analysis such as Digtial Atmosphere provides is that all data can be used in the analysis, even if plotting it would make the map unreadable.

Text and embedded images are copyright of Joseph Bartlo, though may be used with proper crediting.