The following is from an email sharing how the National Weather Service (NWS) measures a tornado's direction, path, width, etc.
For the most part tornado path width is determined by the measurable damage observed during the storm survey. Our WFOs will integrated into that assessment any additional evidence they can get (e.g., video, photos, radar data, survivor accounts) to make their best determination. That goes for all the characteristics of the tornado - path length, path width, EF-Scale rating, etc - that they report. Here is our Norman WFO's El Reno event web page - http://www.srh.noaa.gov/oun/?n=events-20130531
Below is the NWS policy guidance for our storm survey teams to utilize with regard to determining tornado path length and width. The full NWS Storm Data policy can be accessed here: http://www.nws.noaa.gov/directives/sym/pd01016005curr.pdf
7.40.7 Determining Path Length and Width. Path length (in miles and tenths of miles) and maximum path width (in yards) will be indicated for all tornadoes, including each member of families of tornadoes, or for all segments of multi-segmented tornadoes. The length in the header-strip is the length of that particular segment in that particular county/parish. The Storm Prediction Center, NCDC or a Storm Data user can determine the entire length of a multisegmented tornado by adding the lengths from each segment as well as using the latitude and longitude of that segment. Note that latitude and longitude are not available in the Storm Data publication, but are available on the internet in National Climatic Data Center and the Storm Prediction Center databases.
The tornado path length generally excludes sections without surface damage/disturbance, unless other evidence of the touchdown (e.g., a trained spotter report, videotape of the tornado over a plowed field, etc.) is available. The excluded section will generally not exceed 2 continuous miles or 4 consecutive minutes of travel time; otherwise, the path will be categorized as separate Tornado events. The beginning and ending locations of the excluded sections should be described as accurately as possible in the event narrative. In some cases, careful analysis and eyewitness descriptions will determine if separate tornadoes actually occurred within 2 miles or 4 minutes. Refer to Section 7.40.9 for related information on multiple-vortex tornadoes. Use the event narrative to describe whether a tornado “skipped” or was continuous in these types of cases.
The width in the header-strip is the maximum observed through the entire length of a tornado, or of each segment in a multi-segment tornado. Generally, in the absence of structural damage, broken small tree branches of at least 3 inches in diameter can be considered as a marker for tornado width (assuming this damage isn’t related to the rear flank downdraft). In arid regions where there is a lack of trees, other vegetation or landscape material will have to be used as a marker. To determine the tornado's maximum width, the Storm Prediction Center or Storm Data user must check each segment which is entered as a separate event.
The preparer is encouraged to include in the event narrative the average path width (in yards) of all tornadoes, especially for strong or violent tornadoes (EF2 damage or worse). Availability of average path width information in Storm Data benefits the scientific research community and other users.
The above information was shared by Chris Maier, NWS.