Historic Tornado Record in NJ

[hm2]

Hello, all. I'm working on investigating the "modern" tornado database for NJ. I haven't attempted to go back before 1950, yet. However, a few years ago I was in contact with Grazulis about it. He sent me some information and it seemed a bit contradictory to some of the stuff I had on hand. Anyway...that's for another time.

I went through Tornado Archive, SPC, NCDC Storm Data, Damage Assessment, and the Rutgers tornado datasets to try to piece together tornado outbreak days (really any day with 2 or more documented or have a long track) and determine tornadic miles. In some cases, it is quite straight forward; and in other cases, it's a complete mess. The first thing to understand is that every dataset, even the SPC, likely has an issue or two that needs your attention. Some of the things I have noticed:

1. For tornadoes in the record that are counted as 1 in the SPC database but with a long track, will be counted by NCDC with each county the long track covered. So, for example, March 10, 1964 is counted as 1 long track tornado in SPC (59.1 miles), but in the NCDC storm data it's 5 tornadoes (for each county it passed through from Gloucester to Monmouth. As we have seen recently, it is possible to have a rotational area span 50+ miles and drop tornadoes along its path. But it's most likely that both SPC/NCDC method is wrong.

2. Tornado archive will use lat/lon to count tornadoes by state. You have to be careful because in some events they picked up on a neighboring state and counted it as NJ. Also, the same method can make them miss something. Finally, they will use damage assessment to fill in potential tornadoes that didn't get an official rating. For example, on Halloween night 2019, they have more than 1 tornado in the record because there were multiple suspected areas of damage. However, Mount Holly only confirmed and rated 1 tornado for that event in NJ (there was another in PA).

3. Rutgers has a lot of issues with how they count tornadoes. They don't understand that in the SPC database, there is a column that tells you how many states are affected by the tornado and if it's counted more than once. I found several examples where they miscalculated something because of this issue.

So, having said all that, I was wondering if any of you had any additional information on tornadoes since 1950 that could be helpful in piecing together the actual history further than what's available. There are some events where the single long trackers need further investigating. Below is my list and anything highlighted in red is considered suspect for its track length.

 

 

 

 

[hm2]

Well, I just realized that this may be in the wrong forum. Lol 🤦 also, my current EF/F ratings are preliminary and do differ from the usually higher SPC ratings. Those higher ratings that exist are usually for the ones missing official ratings in storm data. So these are subject to change.

[Rainshadow]

We use to be able to get a print out of confirmed tornadoes from SPC. I don’t know if the NWS can electronically still get it. Maybe Sarah Johnson (new WCM) can help you.

[hm2]

I thought I would share some older graphics I've updated and some newer ones I'm working on. These are still a work-in-progress, esp. the two outbreak images.

First is a general count, as best I see it through the various sets:

 

Here is that information partitioned by month/decade:

 

Here is a new chart I'm working on that counts single day events with either a long track tornado or more than 1 tornado:

 

Finally, this is a busy chart that I'm still trying to piece together. Tornadic miles on the right axis vs tornado count on the left axis of each event since 1950:

[mb2231]

Hey! I'm glad I came across this post.

I used to work in the GIS field, this seems like a perfect application for it. ESRI has ArcGIS and there is a FOSS called QGIS.

The multi-county issue is something that you can definitely take care of in a GIS software, Here is the link to the SPC's shapefile. The state counting issue is also something you can handle with GIS. You'd basically take an outline of NJ and count the tornadoes inside the boundary (use paths rather than lat/lon points).

One thing in working in GIS was that we had to do a lot of manual intervention (as you're seeing) with datasets. They aren't always correct. I haven't looked at that SPC shapefile in awhile but if I remember correctly it was pretty well tabulated.

[hm2]

On 4/16/2023 at 4:40 PM, mb2231 said:

One thing in working in GIS was that we had to do a lot of manual intervention (as you're seeing) with datasets. They aren't always correct. I haven't looked at that SPC shapefile in awhile but if I remember correctly it was pretty well tabulated.

The SPC file is good overall. It's mostly the users of it not doing it properly, and the biggest offender of this that I have seen is the Rutgers http://climate.rutgers.edu/stateclim/climatologies/njtornado.html source. It's not terrible but there are several problems.

As for the SPC itself, I have two major questions:

1. How did they assign a rating to the tornadoes that are missing an official rating in storm data?

2. They have a storm from 1958 improperly marked as 1 state, but it's 2. It starts in PA and moves into NJ.

 

[Captain Lou]

Other factors to think about are technology now vs not too long ago (radar, cell phones, more spotters etc.) and population. South Jersey in particular especially south and east of current route 295 mostly entered its development stage in the 1960s and increased from there. Therefore less people around to witness them.  NJ overall has almost doubled in population since 1950. Point being the tornado numbers are probably much higher especially prior to the 1980s than charted. My 2 cents anyway..

[hm2]

On 4/22/2023 at 7:56 PM, Captain Lou said:

Other factors to think about are technology now vs not too long ago (radar, cell phones, more spotters etc.) and population. South Jersey in particular especially south and east of current route 295 mostly entered its development stage in the 1960s and increased from there. Therefore less people around to witness them.  NJ overall has almost doubled in population since 1950. Point being the tornado numbers are probably much higher especially prior to the 1980s than charted. My 2 cents anyway..

 I don't think you'll find a single person that disagrees with that.

[colonel_kurtz]

PA, NJ & DE among 11 states & DC that did not record a single severe weather report during May. 

During Apr 142 total reports for our tri-states. Talk about a complete 180.