Luther, TX F2 Tornado – June 5, 2005

Stats

Path length: 7.8 miles

Width: 1300 yards

Fatalities: 0

Injuries: 0

Rating: F2

Counties: Borden, Howard

Tornado Path

SPC coordinates:

Start 1: 32.57 / -101.3 End 1: 32.5 / -101.32

Start 2: 32.52 / -101.4 End 2: 32.48 / -101.45

Corrected Coordinates Based on NWS Track Map and Analysis of Aerial and Satellite Imagery:

Start: 32.5756 / -101.3383 End: 32.5110 / -101.4315

Introduction

During the evening of June 5, 2005, a massive, discrete supercell meandered along a boundary across the arid plains of west-central Texas. For more than five hours, it crawled generally south and west. With a forward speed of just 5 mph, it spent this time over almost total wilderness in Borden and Howard Counties. The storm spat out six tornadoes, one of which is the focus of this summary. A special thanks to the National Weather Service (NWS) Midland/Odessa for sending us some of the photos in this article.

A reflectivity and velocity radar loop of the storm.

Summary

At 7:08 pm CDT, the fourth and strongest member of the tornado family formed 11.4 miles northeast of Luther. According to a National Weather Service (NWS) narrative, locals described seeing a multi-vortex structure. It drifted south-southwest over County Road 1785, with, per the National Climatic Data Center (NCDC) entry, “a wide swath of damage to mesquite trees, road signs, fences, and utility poles.” A comparison of 2004-2005 aerials from the National Agriculture Imagery Program (NAIP) showed that a very isolated property set back three-quarters of a mile from the road was heavily damaged. At least one unidentifiable structure was demolished.

Initially, NWS surveyors theorized that the tornado dissipated south of County Road 1785 and that the rest of the damage path was a new circulation. However, they soon received photographic evidence that showed a continuous, if briefly weakened, funnel moving across this land. Additionally, a close look at aerial and satellite imagery months later that would have been unavailable to surveyors seems to show a continuous damage path. It is worth noting that this is still listed as two separate F2 tornadoes in the Storm Prediction Center (SPC) database between Borden and Howard Counties. Still, all other official narratives specify it as a single twister.

The tornado soon reintensified and curved further southwest across County Road 328. NWS meteorologists found a 1,300-yard (0.74 mile) wide swath of damage to mesquite trees and fences along the rural thoroughfare. Per the event narrative, “Large drifts of wind blown mud accumulated several feet high along barbed-wire fences on the north side of the tornado’s circulation. A tin shed also was severely damaged.”

The tornado ground across plowed fields for another two and a half miles, entering far northern Howard County. The mesocyclone aloft began to occlude, dragging the maelstrom toward the west-southwest. In a majority of cases, this cycle toward dissipation weakens a tornado until it fades. Here, the opposite happened. The windfield constricted to less than a third of a mile across (roughly 550 yards wide) and became exceedingly violent. Three hundred hay bales, each weighing about a ton, disintegrated into a hailstorm of straw projectiles.

The shrapnel-laden vortex dragged over County Road 41 and into mesquite-populated rangeland. The debris tore limbs and bark from the extremely tough trees. Per Storm Data, “Large mesquite trees were reduced to stumps that only stood a few feet tall, and were partially debarked.” The swath of debarking was noted as 250 yards in width. Small chunks of pavement “up to six inches in diameter” were stripped from the roadway and deposited in the field.

The twister edged deeper into the thicket, leaving a muddy trail visible months later in NAIP images. Here, it encountered an old steel railroad boxcar being used for storage that sat along the side of a trail. The large, multi-ton object initially bounced along the ground, with large craters noted by surveyors. It then broke into three pieces deposited up to 980 yards (0.56 miles) downwind and up a 40-foot incline. As of the latest available imagery in Google Earth (November 2022), these rusting heavy objects remain to this day in the wilderness where the tornado left them more than 18 years ago.

Dissipation occurred a half mile later, prior to 7:50 pm CDT, 4.9 miles NNE of Luther. No injuries or fatalities occurred, and according to official information, no residences were affected. While there was no traditional F2 level damage to structures, meteorologists at the NWS Midland/Odessa determined that the magnitude of the context, namely vegetation and the boxcar, warranted a significant rating. The total path length measured 7.8 miles.

Discrepancies:

We gathered information for this event from the SPC & NCDC Databases, the June 2005 Storm Data Publication, the NWS Midland/Odessa event page, and analysis of aerial and satellite imagery and found the following differences:

One or Two Tornadoes?

As discussed in the summary, first impressions from surveyors were that the damage was caused by two separate tornadoes. However, photographic evidence proved that the entire track was caused by a single twister, and we found support for this when analyzing aerial and satellite imagery. The final text narrative from the NWS Midland/Odessa to the NCDC and SDP explains that they determined this to be a single tornado. However, the SPC still logs two separate F2 tornadoes, divided between Borden and Howard Counties. It is worth noting that the coordinates listed in the SPC and NCDC for those tracks are also displaced up to a few miles from the actual swath of destruction.

Path Length:

The SPC/NCDC/SDP list a total path length of nine miles.
Analysis of aerial and satellite imagery indicates a path length, accounting for all twists and turns, of 7.8 miles.

Duration:

The SPC and SDP list a time of occurrence from 7:08 to 7:40 pm CDT, making for a duration of 28 minutes.
When comparing radar to the damage track, the start time of 7:08 pm CDT appears feasible. However, it would be more likely that the tornado ended between 7:45 and 7:50 pm CDT, given the location and rotational strength of the velocity couplet on radar. This makes for a duration of 33 to 38 minutes.