National Weather Service Duluth summary of the event is here.
Team Dominator live stream coverage is here. 2 hours 18 minutes into the video is the start of the tornado. Their summary video is also viewable.
Brandon Copic's timelapse of the tornado is here. It was helpful in allowing me to determine where the tornado was in my video (generally not visible).
I had been watching the prospects for severe weather in central and northern Minnesota for about a week ahead of this date. Medium range models showed a weather disturbance moving through (500mb shortwave trough and a surface trough) along with enough low-level moisture to generate some moderate CAPE values. I had made the decision to chase sometime before the day started. It was a work day, and that limited the time I could chase, but the storms were forecast to be nearby, so I could afford to wait until the end of the work day.
The first storms initiated far to the west, along the North Dakota-Minnesota border. That area had west-southwesterly winds, which made tornado prospects quite a bit less, so I wasn't interested in those storms (unless they moved into areas with more southerly near-surface winds). Within an hour, more storms initiated much closer to home, and I could see them by simply stepping out the front door. Heidi Pinkerton, a photographer, decided to join me today, and she was on her way. Once she arrived, we put her vehicle in the garage and took mine on the chase.
The first goal was to get to the southern storm of a pair of storms that had developed immediately west of Grand Rapids (the ones I could see from home). I drove southwest on Highway 6 toward Remer. Hoping to get ahead of the storm before its core reached Highway 6. We made it rather safely and encountered no hail. Maximum hail size was forecast to be around 3" or larger, so I did not want to be in the cores of any storms. While we were on our way to Remer, another storm developed to our south, and this would effectively remove our ability to stay in front of the Remer storm because it would soon move over Highway 200 between Remer and Hill City. To avoid this new storm, I drove northwest out of Remer on a county road and had a look at the updraft base of our first target storm.
First supercell with a wall cloud seen through a clearing on Highway 6 between Deer River and Remer.
The back edges of these bases exhibited a shearing motion I've seen on other tornado days.
Here I explain to Heidi what is going on with this wall cloud.
Wall cloud just north of Remer, MN.
Summary of the storm movement up to their tornado-producing phase.
Once we had stayed at our first shooting location long enough for the southern storm to pass, I evaluated the prospects for a new target. The first storm was now in the outflow of the southern storm. Should I wait for an opportunity to get in front of the southern storm? I looked to the east and southeast, and the view in that direction was not good. Rain was wrapped around the rear flank updraft base, and all I saw was flat gray everywhere. It didn't look like it would be easy to catch back up. These storms had what I thought was better low-level southerly winds and a more favorable hodograph for tornadoes, but there was a lot of cell interaction going on, so I decided not to continue chasing these initial storms. There were more discrete storms to the southwest, and getting in front of those would be much easier. I hoped that we would still be far enough east for the healthy southerly low-level winds. As it turned out, that southern storm became a tornado-producer (weak tornado near Cromwell), but it would have been really difficult or impossible to get out in front.
We drove southward on Highway 6 out of Remer toward Emily and Outing. There were other cells coming through the Brainerd Lakes region, and there was a nice tail-end-Charlie near Park Rapids. We stopped at a west-facing boat landing south of Emily, and I took a moment to survey the situation. There was a closer cell, but it certainly didn't look very discrete, and it wasn't well-separated from the others. The decision to continue southward was an easy one.
An intermediate cell just west of Outing, MN.
The tail end storm was looking a lot more interesting now, and radar showed it had a couplet developing. Now, I wanted to get to this storm rather quickly, and when the radar updated, I suddenly realized its core had crept up closer to Highway 6. Time to expedite southward and get in front! The velocity couplet persisted and intensified, and as we got a couple miles north of Crosby, a tornado warning was issued. I was happy that we were on this storm! Now, our main priority was to find a place were we could see it. Highway 6 runs mostly through a canyon of trees, and we couldn't see any part of the storm except for the anvil above us. We needed to get out of the precipitation region and find a view to the west-northwest. We found one just about a mile north of downtown Crosby at a mine pit lake overlook. The storm had already producing a tornado for about ten minutes when we pulled up to the lake, but we could not see it from our location. The cloud base was still mostly below the treeline, and the action area had a bit too much rain in it.
The tornado-producing storm just as we arrived at Portsmouth Mine Pit Lake at 6:23 PM. A tornado was in progress at this time.
The storm had a forward flank shelf cloud, and with its very low LCLs, it gave the storm the look like the air in the forward flank inflow region MIGHT be a little too cool for a tornado, but this was not certain. The shelf was not racing out ahead of the storm, and it really looked more moist than cold. The bases out in front of it were not that high. We stayed at our location as the mesocylcone approached, and once it got closer to our location, the rotation was pretty clear. I wish I had immediately gotten out my tripod and started a timelapse, but I was not decisive enough. I was looking at the velocity couple on my phone and monitoring the precipitation area approaching us.
A picture taken with my still camera.
A cell phone picture.
Just a few minutes later, a forward flank shelf cloud is more visible.
The shelf cloud is moving from right to left into the rear flank of the storm.
By about 6:40, the storm was close enough that we could see the action area under the base as well as a nice inflow band pointing in from the northeast. The original forward flank shelf cloud was still present, but it was somewhat less prominent now as it wrapped in behind (to the north of) the action area. The storm was starting to look ripe for tornadogenesis (I had enjoyed a couple Tornadogenesis brews at Klockow Brewing Company the night before, so maybe that's good luck).
Around 6:40, an inflow tail is visible on the right. A rear flank downdraft is emerging on the lower left.
A rear flank downdraft punched into the updraft base of the storm, and cloud tags could be seen rising more rapidly into the updraft. At the center of this, I could see a lowering developing behind the cascading motion of the RFD. There was enough motion that I suspected there could be a developing tornado in there, but I couldn't confirm it. As it turns out, the National Weather Service survey had the beginning of the tornado track right where I was looking. I was shooting video, but I put the video camera away right when I should have had it on tripod so that I could look back and see the motion. If I had it on tripod, I would be able to look back at video of this important stage of development. Instead, I shut it off and ran back to grab my still camera. I shot some stills, which show this same area, but stills don't show motion.
I wasn't paying attention to anything else in the area, but this is about the same time that storm, which was to the south of us at Remer, began to produce a tornado. Whatever the case, I think this storm was the better one to be watching.
Descending, cascading motion is developing in the RFD along with fast rising motion at the RFD gust front.
The center of this picture shows the center of rotation-- a rapidly rotating wall cloud.
I stopped video and took a still photo of the storm. A funnel develops on the lower left.
Cropped from the previous shot. This is the initial development of the tornado. Damage starts where I'm pointed.
A look at the radar scan of the storm at this time.
A tornado-producing storm also developed near Cromwell. Had I stayed with the initial cell, we may have ended up here.
I put the still camera away and brought out the video camera again. By this time, rain had filled in the action area, making it completely opaque as it began to cross the road. In order to maintain our view, we needed to get through Crosby before we got behind.
The tornado is not visible now because of rain wrapping into the rear flank downdraft.
We took Highway 6 through Crosby and then toward Deerwood. Just outside of town, we found a clearing over a lake with a view to the northeast, affording us a look into the heart of the RFD, right towards where the intense velocity couplet was on radar. I noticed an intense area of shear and rotation, leading me to speculate that there was likely a tornado in there, but I could not see a condensation funnel to the ground because the center of circulation had rain. As I continued to watch, the speed at which low clouds moved through that area led me to believe that there had to have been a tornado in there. By this time, it simply looked like a longer, straight, wet RFD. Indeed, other storm chasers were on the opposite side of the circulation, and the backlighting allowed them to see a dark condensation funnel to ground (something that our front lighting did not allow us to see). The tornado was located immediately underneath the more intense area of rotation that we saw. It's possible that the upper left part of the tornado is visible in the picture below.
On Highway 6 between Crosby and Deerwood. Rain is wrapped around the tornado cyclone, making the tornado itself mostly invisible to us.
Perhaps we could see a part of the tornado from time to time, but it generally was not visible.
Last video capture before we get moving through Deerwood.
The radar loop showed a strong velocity couplet from the Crosslake area down to Glen, where a final, weak tornado was observed. As far as the other chasers go, they saw the tornado until it crossed Highway 210 between Deerwood and Aitkin. By the time it did that, we had left our viewing location. We turned south and drove to Bay Lake because 210 would take us into the core of the storm and also get us behind it.
Radar loop of reflectivity and velocity during the tornado-producing phase.
Had to get a photo of Auntie Em's Coffee Shop in Bay Lake.
We continued to drive southeast parallel to the storm, but we were falling a bit behind, due to the need to circle around the various lakes in the area. We drove to Farm Island Lake and looked northeast again toward where we thought the action area might be, but we saw nothing particularly interesting. We zig-zagged over to the north side of Mille Lacs Lake and drove toward Glen. While we were rounding the north side of Mille Lacs Lake, a tornado report appeared on RadarScope. The storm produced its last tornado while we were well behind it.
Farm Island Lake just west of Bennettville. Any tornado would be in the distance, behind rain to the right.
We arrived at Malmo and continued to pursue the storm eastward, hoping to get back ahead of the rear flank outflow behind the storm. Once or twice we could see an area that looked a little promising, but we never got out of northerly winds, and as we got to Dad's corner (on Highway 65), the sky took on a much more stable appearance with north winds seemingly everywhere. We found ourselves among a larger group of chasers at this point. Most continued eastward, but I was convinced the storm was done producing tornadoes. I decided we could drop south and wait for a new storm that was to the west of ours. The outflow boundary from our previous storm looped a bit back northward in the direction of our next storm, so there appeared to be a small chance that the new storm could get interesting. It never took on as interesting a look as the previous storm did, but it did have a nice wall cloud for a bit (I did not get pictures), and it also had a pretty intense signature for hail. I did NOT want to play with that and respectively kept my distance. We let it cross 65 to our east, then drove south in front of it a little bit to take a cell phone shot. I turned the car back north with the intention of finding what the extent of hail was with this storm.
Radar view of our last supercell of the day.
Our final supercell of the day as it pulls east of us on Highway 65 near McGrath.
We drove back northward through McGrath, and several miles later, large hail was visible in the grass along the side of the road. We stopped to pick some up and examine it. A few of the stones had a maximum dimension approaching three inches. Most of these were somewhat flat and had long lobes of ice extending from a more round center part of the hailstone. These stones apparently spun on their trips through the updraft, creating spikes radiating outward from the center. Once we had pictures of hail, and nearly had the blood sucked out of us by the dense swarm of mosquitoes, we continued northward, stopping again to shoot the sunset.
Hail "flake" north of McGrath.
A collection of hailstones-- up to 2.5" or so.
A nice sunset.
Our last stops of the evening were at the Holiday station in McGregor to grab some snacks in case Zorbaz at Big Sandy Lake was no longer serving food. They were serving food, so we ate a late dinner at Zorbaz. We got back to my house about 11:30.
To try and figure out where tornadoes were, I went back and examined the video that Team Dominator had gotten and posted on YouTube. Based on what I had heard from bystanders at the mine pit overlook, I got the impression that they had seen their tornado a bit farther west as the storm was approaching us, and before we could see under the rear updraft base. However, I studied the video closely and geolocated some of the shots. To my surprise, the tornado they filmed was right where we were looking while we were stopped between Crosby and Deerwood. It was IMMEDIATELY behind the rain, and maybe the upper part of the condensation funnel was visible to us at times.
We would have had a better view of the tornado, and we would have been able to confirm ground contact if we had driven into the core of the storm. This would have taken us into large hail. Reed Timmer's truck suffered a cracked windshield due to hail, so I'm glad I avoided this area, even though it meant I was not sure about seeing a tornado. I also figured out the most likely location of the path. It turns out my path was a little bit too far north on its western side. One of my geolocated tornado spots may have been incorrect, and the NWS damage survey indicates the tornado was well to the west of their position on County Road 30.
Based on my assessment of Reed Timmer's video and Brandon Copic's video, this is my estimate of the tornado path.
Approximate route for the chase. Google says it was 191 miles of driving. The drive northwest from Remer was overdone by Google.
Full analysis showing sight lines I and other chasers had to triangulate locations.