Non-Supercellular large hail growth live and pre-forecasting.

[Eagle Eye]

Finished my report on non-Supercellular large hail forecasting

Non Supercellular large hail 

I have done 7 examples and a conclusion from my mini study. 

1)

28th May 2022 

5cm (about 1.97 in) hail report from developing cells south of a Supercell. You have the cell to the left of it and it’s actually the cluster in the right centre. Notice the feeder cell to the north acting as a flanking line cell for hail growth development in the main updraft. The large hail growth ‘embryo’ comes from that feeder cell which likely increases updraft strength and width because of increasing inflow strength into the storm. Strong inflow is correlated very well with updraft width. Being south of a Supercell it would’ve feeded into the Supercell, meaning that the inflow relative to the storm would’ve been through the updraft which may have limited maximum hail size. Most of the crosswise vorticity would have been close to the surface through the inflow and that would’ve artificially induced stronger rolling drag (maybe). Though the updraft width looks fairly good hence 5cm hail was reached but likely, larger could’ve been reached without the drag. There are multiple embryo stages in the storm connecting to the main updraft and those updraft embryos link up with the main updraft. Given the random processes within a storm, some mixing possibly occurred between the developing hail. This would’ve meant that forming hail could’ve been ‘passed’ through updrafts so time in updraft could’ve been increased by having multi-updraft development. Multiple updrafts can therefore contribute to larger hail as they will counteract the short growth time in a fast singular updraft due to updraft mixing. So with combining weak updrafts into the main updraft, there’s help for larger hail development. 

The initial modelled environment suggested that most activity would be advanced ahead of where this eventually formed. The Swiss model suggested moderate low-level-shearing (LLS) with slightly strong deep-layer-shearing (DLS). That’s not a fantastic environment for producing large hail and in fact you generally don’t get severe hail in an environment like this. That suggests that something else helped with large hail development along with the pre-mentioned multi-updrafts that formed. SBCAPE is modest at best so slower updrafts formed which meant that hail growth could’ve spent a while in the updrafts. Sometimes, you need faster updrafts and sometimes you need slower updrafts and it just depends on how much energy there is. I suspect that if there’s moderate amounts of energy and moderate updraft speed then you would want a slower updraft despite the weaker energy. Note the large boundary layer (just above it) buoyancy which would’ve counteracted for the weaker energy and contributed along with the multiple updrafts for that large hail growth. Modelling capping is always a difficulty in northern Italy especially near the coasts but with strong buoyancy, that created the lift needed to break the mixing of cap up from the Mediterranean and so models sometimes struggle in that region as to exacts in terms of breaking the cap. There appears to be a secondary vorticity lobe created which may have aided these storms on the south side of the one generated over the Alps and so, the orography likely aided forcing for storm development and strength beyond the models' best guess. 

Similarly, the AROME had difficulty modelling the development and breaking of the cap. With the hi-resolution CAM’s, it appears they struggle with handling of cap despite being more high-resolution. So if there’s a large cap in place, I may also suggest considering non hi-resolution models which may handle it better, likely because they have weaker cap in place because of their lower resolution. The AROME has fairly decent CAPE in place but seems to find it difficult to predict storms in its own CAPE maxima and is potentially over modelling the strength of capping advection with a warm nose up from the coast. Very interesting that this seems to be a running worry with the models in terms of their handling of cap. So, definitely keep that in mind when using CAM’s. 

The lower-resolution ECM meanwhile appears to have some of the best handling of the convection with the developments south of the Supercell. Like I mentioned before, that is likely due to its supposedly better handling of cap. 

 

Note that the hail growth zone here is quite close to the 700hPa layer, so is fairly low down. This would’ve put fairly good CAPE in the hail growth sector of the updraft. So, the zero line is not needed to have been worried about on this specific setup. This is one of the benefits of spring storms on large hail development compared to summer storms. 

The Swiss-HD model did rather well at working out cap breaking but was the only real CAM to forecast the breaking of the cap well. Cloud and capping are definitely 2 things that lots of CAM’s struggle with forecasting so I would not forecast based off reflectivity and CAPE alone. Using buoyancy to work out lift maxima areas, I tend to find is a good start. 

2)

23rd June 2022 

5cm hail from fairly shallow topped storm in Montenegro late afternoon. South of the main storm in Montenegro. Located slightly south of Podgorica in southwestern Montenegro. This is a theme so far, being located south of a stronger storm but producing larger hail than that storm. At a guess, it could be along a shortwave where the inflow is being forced into the updraft, more parallel to the storms advance due to the stronger storm meaning the southern mode will push more towards it than the airflow and it stays more parallel to the inflow into the updraft. That may increase the relative entrainment CAPE but keep mid-upper layer updraft speed the same. Meaning more energy but also the same amount of time in the hail growth zone. So, this is a similar outcome to the first example despite not having multiple-updrafts this time. There’s definitely some stuff to the south which may be acting as another momentum kicker for rapid increase in strength of the storm in general. The updraft width would’ve been widened because of the artificial inflow increase as well so hail growth area would’ve been widened. 

  

This time, with weaker cap and convergence or a trough, the CAM’s had a better view though mainly on the southern cell. They struggled on struggling the northern cell despite giving good energy there. There’s fairly good energy already in place so moderately fast updraft speeds but with larger energy it still would’ve made for some fairly good hail. Being on the southern side of stronger LLS but in an area of strong DLS meant that it would’ve been in that area of typical C shaped hodographs when you include the strengthened inflow as well. Potentially, the inflow layer could’ve been increased as well, another contributor to large hail. Being left of the low-level-lapse-rates (LLLR’s) maxima is typical for storm development along a buoyant airflow for larger hail than weaker buoyancy compared to the larger LLLR’s. So combining buoyancy with LLLR’s is better than the LLLR’s on their own with the better CAPE being left of the better LLLR’s which suggests better buoyancy is there as well. So forecasting purely based off either, for hail production, can lead to non development in the strongest potential hail production areas. 

3)

28th June 2022 

7.5 cm (about 2.95 in) hail in possible non-supercell. Looks to have been a developing Supercell, may not have been a Supercell yet. Multiple updrafts that were attempting to combine from initially semi-discrete thunderstorms providing for good hail growth mixing and total updraft width strengthening. Not all clusters like this will produce that large and damaging hail but the initial separation of updrafts meant that the combination of them worked to force some large and destructive hail to form. Hail would’ve stayed in updrafts for enough time and with enough energy to produce that significant hail. So, the clustering up of initially separate updraft embryos appears to be an efficient way of producing large hail by increasing the total theoretical updraft width significant whilst also increasing momentum and forcing into the storm. That would’ve increased EL’s and therefore the entrainment CAPE that it would’ve taken up which would’ve helped in hail production. The importance of slower updraft speeds with relatively large energy is less here given total amount of updraft and therefore faster updraft speeds will only help contribute to larger fall speeds along with larger hail. Drag will be assumed to be the same for now because it is easier to always assume that the drag is the averaged out shearing or just the LLS for the sake of this. 

he Swiss-HD probably had a very good handle on the forcing for clustering and the total energy as far as I can see. If you can use it, it appears to be one of the best CAM’s available for storm mode and initiation. It is interesting to see the LLS striations modelled by the model as well, showing that it handles the planetary boundary layer (PBL) very well with the exchange between viscosity and laminar flow. This is a more classical weak LLS day and I appear to have forgotten to get the DLS though I imagine that would’ve at least been about 20 knots stronger. Once again on the left of the LLLR maxima but this time better mixed in for even stronger LLLR’s. So that also suggests a fairly low down zero layer. Significant CAPE modelled suggests about 60 m/s hail fall speed if we assume 30 m/s drag which is a strong hail fall speed and also does suggest fairly big hail as did happen. When you include the buoyancy mixed with that LLLR modelled and it’s no wonder near 3 inch hail was reported from this cluster of storms though without that updraft width I imagine only about 1.5 inch hail would’ve happened. 

It appears to be based slightly ahead of/along a warm front. Something to note if that kind of setup happens in case it’s a frequent setup for large hail producing storms and good forcing for clusters. 

4)

12th June 2021 

Multiple 2cm (about 0.79 in) hail reports from late evening storm. The one near Cordoba with the inflow notch next to it which likely increased momentum. Generally, that seems to be the only real reason as to why lots of 2cm hail reports were able to occur for a time. South of a big cluster of messy storms though but none appear to be Supercellular. 

Located in area of weak LLS but also weak DLS. Kind of what you would expect considering it was only 2cm hail in the first place. Along with being on fairly good LLLR’s. Not much needed to analyse this specific setup. It was more just a coincidence of enough flanking cell power for low-end severe hail to occur. 

5)

18th August 2022 

Thunderstorm pushing into the back of a messy MCS. Likely going into the visible rear inflow jet (RIJ) slightly west of Pavullo nel Frignano. 3cm (about 1.18 in) hail. A cell forming into the RIJ will be able to take advantage of the inflow into the updraft to increase energy into the updraft which would increase potential hail size. Similarly to how we discussed before with inflow into the updraft. As long as updraft width remains stable and wide then large hail is possible when the inflow follows the buoyant zone trajectory. Given storm direction is steered by the stratiform inflow then inflow and storm direction will be relatively similar in this case.  

Modelling pre-storm suggested a more consolidated MCS with less of a RIJ. Modelling struggles with inflow cells because the ability to create the stratiform structure of a strong MCS is very weak with models and I think it's something that needs to be worked on because that also affects any post MCS storms afterwards. Generally, you can assume that a MCS is strong enough, if you want to, for inflow cells to form and they should be more capable of sustaining hail growth in the updraft longer than general thunderstorms so if the parameters are there then they may produce some large hail. Located in a buoyant trail behind the MCS, it is clear that the models know that they can happen but for some reason don’t form them. The LLS and DLS in residence of an MCS are often very high and theoretically that LLS weakens max hail size but enough energy should counteract that somewhat. 

6)

5th June 2022 

4cm (about 1.57 in) hail from cells pushing up into stratiform region of a squall line. Pretty much the same as last time actually but with a stronger stratiform region of the MCS likely due to stronger outflow winds and the inflow slightly more perpendicular to the flow which may have allowed one or two of these to eventually grow into a Supercell actually. 

The Swiss-HD did a rather good job of handling this and even had a Supercell further along the line on the Italy-Austria border as you can see there. This can be expected with most of the vorticity being streamwise based off the cell movement against the storm movement, the cell moves across relative to the storm movement. Cell movement is the direction of the inflow and storm movement is the direction of the vorticity.  

7)

23rd June 2023 

6cm (about 2.36 in) hail induced from momentum kicker cell to the north. Inflow looks to be fairly strong and a wide updraft with a cell to the left combining to help attempt to increase updraft width. Cell to the north moving perpendicular to the storm before getting ‘trapped’ increases the strength and updraft speed and size. Resulting in the growth of large hail because of strong energy and also a large hail growth zone available on the south side of the storm. Might be some fairly strong inflow to the south attempting to form some weak showers from the south which is feeding into the updraft so there’s a lot of energy and speed and messiness within the updraft.

I could not find a single model which had a good handle on this. Perhaps the Swiss 4X4 did fairly well in the evolution but with it being expected to be such a messy mode it’s really hard to model well. In fact, by all accounts it doesn’t really make sense as to why that size hail formed there at all from a modelling perspective, proves how much updraft strength and momentum can help. Though there was a local LLLR maxima to use up.  

Conclusion

In conclusion. It is very important to consider updraft width when live forecasting hail whereas updraft speed is theoretically less important than previously assumed as storms can affect their own relative energy. When considering LLLR’s, combine it with the buoyancy to get a picture of whether or not that hail environment is associated with a storm environment because if it isn’t, the LLLR’s won’t be used up anyway. Flanking cells and momentum kickers can increase updraft strength and width to introduce a hail producing localised environment even when the background signals are modest at best. 

It is important to note that updraft width does not have to be a singular cell but if an organised cluster of previous semi-discrete or discrete updrafts can form then the total updraft width can theoretically be used in a chaotic environment. So, if the models are showing a forced zone of buoyant air with good enough LLLR’s then, large hail is quite possible.  

On the south side of a Supercell where the storm is dictated by the flow of a Supercell then it may move parallel to the inflow and increase the updraft strength assuming buoyant airflow is taken from the inflow. A Supercell would take more advantage of the energy but often follows it’s own track and with splits makes it difficult to produce large hail. Generally, a Supercell will produce large hail because of the stronger entrainment CAPE and directional shearing with a stable updraft. 

When considering an MCS, the RIJ is a good place to look for remnant buoyant air behind or in the stratiform region with strong DLS and enough energy to allow for feeder cells to produce large hail. The stronger the MCS then theoretically, the bigger the feeder cells and the bigger their updraft width is so theoretically, the larger their hail is. 

From a pre-modelling perspective and the handling of stratiform regions of an MCS is often very badly done so as long as it’s a strong MCS, I would tend to assume a stratiform region with those feeder cells capable of following parallel to the inflow. Sometimes, they may not be parallel but more perpendicular and it depends on the pressure signals from an MCS and all I’m saying is good luck forecasting those from just models. 

Models sometimes struggle with capping problems and so look out for some very buoyant air in a capped area. That may be enough to break through capping even if it yields rather weak CAPE on the models and so that’s another modelling quirk. 

If available, the Swiss-HD has the best handle on the PBL which is very important in understand hail size below initial formation so using that is important if available. This is due to it’s handling of the transition between viscous and laminar flow which is very important in fluid models and convective modes with the transition of temperature between the two and lots more of important dynamical processes. The boundary layer is mostly beyond my current understanding though. 

The size of inflow layer will always be very important though, as will the zero line. The closer to the initial inflow layer, the better for stronger hail growth I imagine. That’s how I would use soundings for hail along with the inflow and shearing creating a C hodograph but there’s no specificied numbers for where the zero line should be and it’s dependent on the amount of CAPE it has above it as well as the inflow available to the beyond zero area.  

P.S.

There might be some accidentally double posted images here at the end.

Edited June 27, 2023 by Eagle Eye