Reanalysis of Tuesday, October 23rd, 2018 Localized eastern Southern New England Tornado Outbreak

An extremely rare and well un-forecasted localized tornado outbreak occurred across eastern southern New England Tuesday afternoon into early Tuesday evening. The National Weather Service in Norton, MA confirmed two tornadoes; North Providence/Lincoln, RI and another in Norton, MA (both were spawned from the same supercell). Both tornadoes were rated as EF-1’s (the scale ranges from the weakest EF-0 to a violent EF-5). In addition to these two confirmed tornadoes, several waterspouts were also observed (two off the RI coast and two more on Cape Cod). Thunderstorms also produced small hail in Hubbardston, MA and Union, CT. Before we get into the science behind this event I want to take a little bit of time to discuss the waterspouts.

Waterspouts are tornadoes which form over the open water. However, if that body of water is landlocked it is defined as a tornado (an example would be from May 15^th, 2018 when a tornado which formed over a reservoir in Barkhamsted, CT was defined as a tornado as the reservoir was landlocked). There are also two types of waterspouts; fair-weather waterspouts and tornadic waterspouts. Each is defined based on the characteristics in which the phenomena develop. Fair-weather waterspouts are typically not associated with thunderstorm activity. These waterspouts form at the base of a cumulus or a stratocumulus cloud. The two sources involved here are; 1) strong wind shear and 2) the instability is generated due to the temperature gradient between the warmer waters and cooler airmass. With this said these phenomena or typically most common during the fall months. Tornadic waterspouts are formed via the same processes as a tornado would over land. They are generated from supercell thunderstorms and the processes involve strong wind shear and the instability source is airmass characteristics. Tornadic waterspouts may also be tornadoes which form over land and move offshore. In the case of yesterday, we were dealing with tornadic waterspouts which leads to this event being classified as a localized tornado outbreak.

Now we can get into the fun stuff…science and meteorology!

As stated during the open, this event was not forecasted. As we proceed with the post we will look at meteorological data from the day before which will show that the signals were there, however, one would have dug deep into the data to latch onto this potential. Then again, who would think/forecast the possibility of tornadoes when we’re dealing with temperatures mainly in the mid 50’s with dewpoints in the 40’s to around 50 (especially around these parts)? To prevent this post from becoming one of my typical war-and-piece novels, for data I am going to focus on the NAM (North American Mesoscale) forecast model with a focus on the afternoon run (18z or 2:00 PM EDT) as well as the Tuesday morning (12z or 8:00 AM EDT) run of the HRRR (High Resolution Rapid Refresh). The three forecast stations selected will be Chatham, MA (KCHHH), Boston, MA (KBOS) and Providence, RI (KPVD). Unfortunately, I only had HRRR data for Chatham, MA and was not able to get Providence so I went with Boston, MA. Finally, we'll look at a few parameters from the SPC mesoanalysis data that afternoon.

Analyzing the 18z NAM from Monday afternoon, the forecast was for a vigorous shortwave trough and associated shortwave energy to dig into New England during the day on Tuesday. Associated with the shortwave trough was an exceptionally cold (especially for October standards) mid-level airmass characterized by 500mb (about 18,000’ off the ground) temperatures into the lower -20’°C range! This lead to the forecast of very steep 700mb-500mb (~10,000’ to ~18,000’) lapse rates (lapse rate is defined as change of temperature with height) on order of 8 C/KM (indicating that for every altitude increase of 1 km the temperature would decrease nearly 8°C). That is a remarkable decrease of temperature with height and despite surface temperatures which would only warm into the mid 50’s with surface dewpoints into the 40’s to around 50 would eventually yield robust mid-level instability:

Sufficient wind shear was also present on the NAM for Tuesday, however, there was a bit of a disconnect between when the strongest directional shear (change of wind direction with height through the troposphere) would occur and when the highest instability would occur. In retrospect, the NAM ended up underdoing both how long the directional wind shear would be present and the degree of directional wind shear that would end up occurring. This will be shown a bit later. The forecast from the NAM showed 100+ knots of westerly wind over the region at 250mb (~38,000’ off the ground) with an impressive jet streak of 110+ knots just to the southwest of southern New England. This likely placed southern New England in the left front quadrant of the jet streak, enhancing upper-level divergence (this favors enhanced upward vertical motion…rising air). At 500mb the forecast was for 70+ knots of westerly wind across southern New England with a jet streak of 80+ knots west of the region with the region on the nose of this jet streak which would further enhance the degree of upward vertical motion. At 850mb the forecast was for 30+ knots of south-southwesterly wind to develop across RI and southeastern MA. Note how at 250mb and 500mb winds were more west. While I did not show this here, winds at 700mb (~10,000’ above the ground) were more west-southwest. Winds were transitioning from south-southwest at 850mb to west at 500mb and 250mb which yields the directional wind shear (wind changing direction with height):

As far as the forecast for directional shear is concerned, the NAM on Monday was not particularly impressive come late Tuesday afternoon. The below is the 18z NAM bufkit forecast sounding for Providence, RI valid 5:00 PM Tuesday afternoon. There are a few things of note:

1) The predominately unidirectional (same direction) winds with height through the troposphere, however, there is some degree of directional wind shear present mainly in the lowest 1000' of the atmosphere where you have surface winds forecasted to be south-southwest with winds above that becoming more southwesterly.

2) The hodograph is like a "graphically" way of visualizing how winds are behaving with respect to height. When the hodograph is more straight in nature that is suggestive of a unidirectional wind profile. When there is curvature to the hodograph that suggests the presence of directional wind shear. On this forecast hodograph we have a predominately straight look, however, the slight curve suggests that slight change of wind direction with height in the lowest 1000'.

3) The forecast CAPE (Convective Available Potential Energy...a measure buoyancy of air parcels) in the 0-3km layer and 0-6km layer is quite noteworthy. Before proceeding further I want to discuss the significance of 0-3km CAPE and 0-6km CAPE and show exactly why it holds such high importance. There are several different computations of CAPE; surfaced-based CAPE, mixed-layer CAPE, most-unstable CAPE, normalized CAPE, and downdraft CAPE. In the case of this setup, analyzing the forecast for mixed-layer CAPE and surface-based CAPE the forecast from the NAM was for values between 200 J/KG and 300 J/KG. Typically these values aren't really eye opening when it comes to the potential for severe weather, however, when you have such strong wind shear in place (such as we did here) these values can be sufficient but it is very borderline. Without trying to go into crazy detail, each of the CAPE measurements are calculated via different processes. Surface-based CAPE indicates the total amount of energy available for an air parcel at the surface to accelerate upwards to the lifted condensation level (level at which water vapor condense and clouds develop...its the base of the cloud). Mixed-layer CAPE yields the average conditions of the lowest 100mb of the atmosphere and indicates the energy available for air parcels to lift to the level of free convection. 0-3km CAPE and 0-6km CAPE indicate the potential energy available to air parcels between the surface and 10,000' off the ground and surface to 18,000' off the ground respectively. First, we will look at the forecast for surface-based CAPE and mixed-layer CAPE values:

Now we will look at the forecast values of 0-3km and 0-6km CAPE along with what was just discussed regarding the shear:

Forecast CAPE values between 0-3km and 0-6km were to forecast to exceed 100 J/KG which is quite impressive, especially with the shear that was being forecasted.

Now we will move onto Tuesday's morning of the HRRR. We will look at the 12z run for both Chatham, MA and Boston, MA for later that afternoon. What we will see are soundings which are not only very opening in terms of indicating the potential for severe thunderstorms but for tornadoes:

Finally, let's take a look at a few parameters from the SPC mesoanalysis at 4:00 PM that afternoon. Mesoanalysis data showed a narrow area where the supercell composite parameter was between 0.5 and 1 with values of 2 just offshore. These values aren't very eye opening but this is a case where you have to look at more than just a raw number. You also have to sort of know/understand the parameters used within the calculation of the parameter. When taking into account the environment these values actually are a bit more eye opening than you would realize. Significant tornado parameter again wasn't eye opening but when taking into account the environment it was a bit more impressive. Also, note how these parameters were positioned over the area the tornadoes/waterspouts occurred. The most impressive out of all of this...0-3km CAPE/surface vorticity overlay. 0-3km CAPE values exceeding 125 J/KG is highly impressive as it indicated the potential for significant vertical stretching (an ingredient favorable for tornadogenesis...development of a tornado). The surface vorticity value of 4 indicated sufficient vertical rotation was present. Combine this with enhanced likelihood for vertical stretching and the result...tornadoes.

What can we take from all this? While the forecast models (in this case the NAM) hinted at the potential for strong-to-severe thunderstorms, the signal a day out wasn't overly strong but the ingredients were certainly present. Also, in the case of this event, as Tuesday morning progressed the signals became quite alarming as all the ingredients were going to juxtapose with a perfect timing of everything.

This is an event in which forecasters can learn heavily and thankfully there were no injuries. As the event was ongoing forecasters at the National Weather Service and local TV meteorologists did a fantastic job at getting word out and issuing warning.

2018 has been quite the year across New England in terms of tornadoes. It seems just about every setup capable of spawning a tornado has done so. While tornadoes in New England aren't common, having such an active year such as this is.