National Weather Service United States Department of Commerce

Major Grass Fire in Harvey County

February 10, 2006

- by Mary-Beth Schreck and Ken Cook

With an already dry and warm winter, the conditions on February 10th were prime for fires to get out of control quickly. Burn bans had been in effect for weeks for many counties, including in Harvey county for 4 weeks prior to the fire. This fire started at about 2:15 pm on Thursday February 9th, and was not completely extinguished until Saturday February 11th during the afternoon or evening. At least 8500 acres were burned, with only 1 building being damaged and no injuries. Smoke was observed in the west Part of Wichita on the 10th.

The fire began February 9th, and got out of control during the afternoon. That evening fire crews got the fire under control, but it was not completely out. Friday February 10th began with temperatures just above freezing and minimal winds. Fires should not get out of control very easily in this environment at 12z (see sounding). However, with northwest winds gusting to 30 knots and very low relative humidites expected in the afternoon, they would burn and spread quickly.

LAPS 12z sounding for NW Harvey County

Dew points were generally in the upper teens and lower 20s Friday morning. Relative humidities were already fairly low for the early morning hours.

LAPS 12z surface dew point image


By 2:00 pm, the inversion was gone, and temperatures climbed into the lower to middle 40s across South Central Kansas. Since December 17th there had been only 0.11" of precipitation (at ICT) and most of that fell on January 28th. So fuels were quite dry and ready to burn quickly. The LAPS sounding shows that the temperature profile is dry adiabatic through about 750mb, and winds were approximately 30 knots through the dry adiabatic layer (can be seen well in Bufkit using the Momentum Xprt winds not shown).

Notice too the dewpoint sounding. This can be appropriately forecast using the equal area method.

LAPS 20Z sounding for the same point as Figure 1 (at the site of the fire)

There are 3 steps to do this. Look at your morning sounding. In this case, we are using a LAPS sounding for KICT.

  1. Find the Mixing Height
  2. Draw a line parallel to the mixing ratio, creating an equal area between the areas above and below the intersection point
  3. Where this line intersects the "surface" is the forecast minimum dewpoint given that moisture advection is not expected to take place.
Dewpoints had fallen into the 10 to 15F range by 2:00 pml. Relative humidities by this time were around 15% across the region. This matched the forecast dewpoint using the equal area method very well.


As the fire got out of control again on the afternoon of the 10th, it could be seen on radar imagery. The smoke plume advanced southeastward with the northwesterly winds, and it could be seen from the NWS office as well. This radar loop shows the smoke plume quite clearly.


This visible satellite image shows the burn scar left by the fire.

A closer view of the burn scar.


This fire burned for 3 days under dry and windy conditions. Along with the burn bans that were in effect across the region, extreme fire danger was noted in the rangeland fire danger index product along with the HWOs and AFDs around this time. On Friday, the smoke from this fire moved into the western areas of Wichita. As a result of several received phone calls, an interest in how the plume of smoke would behave resulted in a run of the Hysplit model. This was run on Friday afternoon by calling the Senior Duty Meteorologist (or SDM) at HPC. This shows where the smoke would be dispursed with the current weather conditions. To the right is the result of this model brought into ArcGIS9.

Figure 8: Hysplit model run of smoke plume. Click on image for larger view of this animation.