September snowstorms are rare in southwest Kansas. The biggest September snow of the 20th century for the Dodge City county warning area occurred on September 20-21, 1995. Parts of northwest  Kansas and northeast Colorado received 6 to 12 inches of snow. Much of the northern, central and western part of the Dodge City warning area received measurable snow. The highest amounts were in Ness and Lane counties where some places reported 6 to 7 inches. Widespread snow amounts of 1 to 3 inches were reported in other areas. Only trace amounts fell east and south of Dodge City. Most of the snow fell between 06 and 14 UTC  September 20. In areas north of Dodge City and Jetmore the heaviest snow fell before 10-11 UTC.

On September 20, 1995 at 12 UTC, a jet streak and associated surface front were  located across the southern Plains. The effective boundary pushed a little to the south because of convection, but essentially remained almost stationary through September 21st in Texas. Meanwhile,  a very cold upper trough dropped south/southeast out of Canada into the central and northern plains. A shortwave trough was located on the back side of this trough over the northern Rockies as shown by the 500mb chart.  There was a very tight baroclinic zone associated with the mid level front as shown on the 700mb chart. The 18 UTC surface chart shows dewpoints along the Texas coast in the upper 70sF. The 21 UTC surface chart  showed a surface front associated with the northern trough from central Colorado into Kansas and Iowa.

The 03, 06, 09 and 12 UTC surface charts show the rain-snow line moving south through Dodge City. Rain changed to snow in Garden City around 0730 UTC. Although Garden City missed the heaviest snow,  surface observations from there indicated thunder. The max snow depth was 3". Here is the VAD wind profile at Dodge City from 4 to 6 UTC. At 03 UTC, surface dewpoints were still in the 40sF across much of southwest KS. Even Hays, KS had a 39F dewpoint.

The 12 UTC sounding from Dodge City showed a nearly isothermal profile from the surface to 600mb. The elevated CAPE was about 170 j/kg. This yields very high 600mb mixing ratios from 5 to 6 g/kg with T/TD values around -2 to -3 C !! The 00 sounding can be found here for comparison. Western Kansas was in the right rear quad of the upper jet by 12 UTC.  The 12 UTC 500mb chart is here. The 12 UTC sounding was compared to a typical sounding in May.  Note that 29F/29F at 620mb yields the same theta-e was 75F/51F at 938 mb despite the mixing ratio being 55% less. This is because the potential temperature is 100F at 620mb and 84F at 938mb.                      

A comparison of the 00 UTC and 12 UTC soundings at DDC shows dramatic cooling below 650mb while temperatures around 620mb remained unchanged. There was modest cooling at upper levels but the upper levels remained very warm. In fact it was -13C at 500mb when the snow was falling. 1000-500mb thicknesses were very high for this event and varied from 563 dm at Dodge City at 00 UTC to 550dm at 12 UTC. In our northern counties the thickness varied from 559dm at 00 UTC to 545dm at 12 UTC. This is because the temperature
profile was nearly isothermal from the surface to 600mb, with very balmy 600-500mb temps.

Some folks talk about how warm ground temperatures preclude snow accumulation. This case serves as an excellent  example of how heavy snow will accumulate even when temperatures have recently been warm. Here are the highs, lows and average temperatures for Dodge City for September 1995.  The coolest minimum temperature leading up to the event was 52F on Sept. 8 and 50F on Sept 19. In fact the 4" soil temperatures at the Garden City Experiment Station were very warm. The minimum soil temperature from Sept 20 at 8 am to
Sept 21 at 8 am was 57F. The maximum was 72F. So 18 hrs before thundersnow was reported at GCK the soil temperature was around 72F. After the event at Garden City the 4" soil temperature was down to 57F.