Ice Dams with New Construction
![](https://secure.gravatar.com/avatar/359632f053b6e0519f10844beb40f0b9?s=52&d=https%3A%2F%2Fimages.greenbuildingadvisor.com%2Fapp%2Favatars%2F200x200%2Fw.png&r=g)
So, first bit of snow and cold weather (St Paul, MN) and we’ve ice dams. Our house is nearly complete – rocked and primed, most cabinetry and trim installed. Photos: https://www.instagram.com/bamasotan/
and http://bamasotan.us
The 3rd level (attic) is finished and conditioned. The crawl spaces behind the knee walls are conditions (insulation is against the underside of the roof deck from eve to peak). All HVAC ducting in upper levels is within the insulation envelope. House is currently kept at about 60°f.
Blower door after insulation & prior to sheetrock is 1.8 @ ACH50. (Initial was 4.6 and then they went through finding leaks and said that 1.8 was the best they could do).
Insulation is:
Walls – 3″ Closed cell foam + blown fiberglass to fill out the cavity.
Roof – 7″ (R-49) closed cell foam. Some bays are vented (durovent), some are not (hot roof). Roof is mostly truss, some solid rafters then sheathing then cedar shingles directly on sheathing (numerous MN builders have stopped using cedar breather saying that they’ve seen no difference when it is or is not used which I’m guessing is because it doesn’t provide nearly enough air flow to be effective).
Lower edges of roof do have Grace I&W.
Ice dams are on both east and west sides of house. I need to get a drone up to see south and north facing roof segments. About 95+% of other homes in the area are free of ice dams so far this year so our having them is somewhat unusual.
Possible causes? Solutions?
Thanks,
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
![](https://www.greenbuildingadvisor.com/app/themes/greenbuildingadvisor/dist/img/modules/detail-library/thumb1.png)
![](https://www.greenbuildingadvisor.com/app/themes/greenbuildingadvisor/dist/img/modules/detail-library/thumb2.png)
Replies
Assuming your ventilation and insulation are good. In the beginning here don't over think it. You can see the depressions in the snow and missing snow in your picture. Investigate that.
Temp heat in garage or such bypassing into attic as fire wall not installed.
I have had subs open the attic scuttle and leaf it open for hours.
The first substantial snowfalls of the year often generate ice cycles that dry up or drop of and never reappear.
Is there solid blocking in the transition from wall to ceiling in your picture ?
You probably probably have thought of these things but I will through them out there.
> numerous MN builders have stopped using cedar breather saying that they’ve seen no difference
Over the short warranty periods they are concerned with, there probably isn't any noticeable difference.
Or perhaps, whatever the underlayment, cedar shingles aren't a great choice for roofing.
A shot of the east side this morning.
all those melty spots are places you have some kind of warm air leaking up and heating the roof. I’d be especially suspicious of that spot on the far right which looks to me like something is open under that spot.
The vertical groove-like depressions may be thermal bridging, but are more likely to be from channels that are carrying warm air from leaks out near the exterior walls. Sometimes these line up with interior walls that are perpendicular to the exterior walls, but in your house I’d be suspicious of the duravent — see if some warm air getting into some of those channels out near the eaves.
I like to take periodic picture of my home’s roof in the morning when there is frost on the roof. I always do this before and after any air sealing work if I can. You can spot leaky places, and you can see improvements your air sealing work has made. I can clearly see my two recessed cans (installed by the previous owners) in a part of a cathedral ceiling, for example, in the frost-free areas of the otherwise frost covered roof. The frost melt pattern from those looks like a spot over the can, and a slightly smaller vertical path up to the ridge.
Heavier snow tends to mask smaller leaks. Frost is best, light snow is next best, for identifying smaller leaks. If you can post an indoor and outdoor pic of the same roof it may be possible to make some guesses as to what might be leaking.
Bill
>"Roof – 7″ (R-49) closed cell foam."
That means you have a very low ~R8 - R8.5 framing fraction, low enough to cause melt-out along the dimensional lumber rafters. R30 rock wool in a 2x8 rafter bay would have about the same ice-damming potential as what you have, despite that fact that your house meets code minimums on an R-value basis.
It's probably just fine on trusses with 2x4 or 2x6 top chords, as long as there is at least a couple of inches of foam between the top chord and interior, but not where there is dimensional lumber rafters with 7" of penetrating rafter as a thermal bridge.
A 2x12 rafter bay with a full fill has an R13.5-ish framing fraction, which is a significant reduction in heat flow and melting potential compared to R8.5. If the raftered sections are 2 x 10s or 2 x 12s filling the gap between the foam and ceiling gypsum with blown cellulose or fiberglass will decrease the thermal bridging by quite a bit. If they're 2x8s an interior side layer of 1 or 1.5" of EPS behind the ceiling gypsum would increase the framing fraction's R-value to about that of a 2 x 10 or 2 x 12 respectively.
How would I-joists perform instead of dimensional lumber in a flash (and batt or fill) assembly? As long as the appropriate % of ccspf for the climate zone is used?
Thanks,
Alan
It's the same problem from an ice damming point of view, but with 1/2 -2/3 narrower melt-out ridges, so lower water-melt volume overall.
With a climate-appropriate foam + fiber solution it's always a lot better than a cc foam-only solution, whether I-joisted or milled lumber, since the lowest R-value path is higher-R when it's longer. In zone 6 Minneapolis 4" of HFO blown closed cell foam (R27-R28) and 7" of cellulose (R26-ish) would work from a dew point control point of view at the foam/fiber boundary, with the 11 inch path through the I-joist or 2 x 12 running a bit north of R13.
An unvented roof can meet code performance on an U-factor basis with 6" of continuous 1lbs foil faced polyiso on the exterior, which would be north of R30 over the rafters even if derated for temp- WAY better than ~R13 for ice dam mitigation.
Alternatively, 4" of exterior polyiso and 5.5" of cellulose or R21 fiberglass batts in 2x6 rafters (or sculpted to fit I-joists) would also meet code on a U-factor basis, with good dew point control at the roof deck for climate zone 6 or lower, and a framing fraction north of R25. That's about twice as good as R13-ish thermal bridging of a 2 x 12 from an ice dam perspective.
Better, but an unvented roof assembly in a cold climate is a roll of the dice in my opinion. Cathedral ceilings either need to be parallel chord trusses with deep blown insulation or need warm side rigid insulation in addition to batts or blown. Providing a ventilation channel is a must for assured performance.
Thanks. It took me a while to get the info on this. The majority of top cords are 2x4 and 2x6 with a few 2x8 and infrequently 2x10. The insulator said that all top cords have at least 1.5" of foam on all sides and from what we can see that is correct.
What is the intent of the cardboard like pieces extending up from the knee wall in the first photo? This doesn’t look the way I would imagine on a soffit to ridge vent. I think you said the space behind the knee walls is conditioned, it just doesn’t look like it is set up to be.
There definitely looks like some sort of air leak in the lower right of the 3rd picture.
As Dana discussed there are significant downsides to spf only roofs. You look to have some sections with a lot of framing, like tripled rafters.
That being said, I’ve damn development is somewhat complex. You said other house don’t have them. HVe the other house held snow the same as yours or seen significant melt out at times?
My roof is 3” HFO, r30 rockwool, 1” polyiso furring over 2x10’s. I can see striping in melt pattern because the r value difference is still large, r15/16 vs r51.
I was kind of annoyed to see that, but it makes sense.
I have two very similar houses next to mine with the same roof pitch and exposure, but insulated ceilings installed by the original 2009 builder. My roof ends up holding snow/ice for WAY longer than the other two.
Here's a drone shot of the east facing roof from yesterday afternoon. What I'm seeing:
- Several places with hot spots near the heal and then a line of lessor melt progressing towards the ridge. I'm guessing that these are likely hot and moist interior air leaking in to the durovents?
- A place just to the right of the rightmost dormer with a melt line nearer the ridge. Likely thermal bridging with a truss?
- A couple of hot spots near where the leftmost dormer meets the roof (the most prominent is on the south side and is visible in another photo that I don't have with me).
A higher resolution version is: http://bamasotan.us/wp-content/uploads/2019/12/B1200RoofInsul-100.jpg
??
Matt, I'm not sure about the cardboard. I'd asked that when they first put it in and never received a satisfactory reply so am asking again.
A thermal image of the roof when there is no snow on it would be better. But probably not so easy from a drone.
Do you have any photos of the installed insulation prior to drywall? At least check out the spaces behind the knee walls, that cardboard detail is very suspicious. How did they do blocking between the vertical wall and the vents down at the rafter tails?
There is absolutely one suspicious spot I circled in red.
Try to correlate the melted out bands with the known internal structures.
The ice damns don't seem terrible at this stage. I think you are always going to get some amount with a cathedral ceiling.
How does the other side of the roof look? The melting bands seem pretty wide. This could be tripled rafters. If this is the west facing roof, I wonder if there is warm air rising off the south wing and running up through the vents. It looks like there is less meltout as you move away from the wing. A logging thermometer at the soffit might be interesting. I am also sort of just reading tea leaves.
W,
Let's back up a second....
Ice Dams form due the following reasons in order:
1. Conditioned Warm Air hitting underside of sheathing (poor air sealing)
1a. Heat loss at ceiling (inadequate insulation)
1b. Leaky Ducts
.
.
.
2. poor ventilation under sheathing that would otherwise bring in cool air to offset snow's R-value
3. adequate ventilation but sun-warmed air traveling up a wall and into the soffit that melts snow.
Not likely Cause #2:
Your interior photo shows some skimpy ventilation baffles (small gap and doesn't cover entire cavity). This appears to be on the southeast part of roof. However, the exterior photo shows some snow cover in this area so it doesn't seem to be too big of an issue. Also, with spray foam a 'hot roof' assembly would suffice without ventilation baffles, in theory. So, I am going to guess it is not Cause #2.
Not #3:
The fact that there appears to be just as much snow melt above the front porch than the rest of the roof tells me that #3 likely isn't an issue either. Also, given that this is an east facing wall means its not seeing all-day sun like a south wall would. Furthermore, the Tyvek is bright white and would reflect a lot of the sun's heat.
I suspect #1 and #1a:
1: )
1.8 ACH on such a large house still means a lot of total air movement (probably over 1000 CFMs!) If the bulk of these air-holes are at the ceiling plane- and they probably are- then you will have a lot of air leakage at the roof- even with spray foam. I can't imagine trying to create a continuous air barrier using spray foam between rafters, knee walls, dorms, etc. This is a design issue not necessarily an insulation issue.
1a.) I am also going to challenge the depth of the spray foam. I seriously wonder if they really installed as much as they say they did. The reason I say this is because the other part of the house (the Dutch Colonial style wing that faces south) has considerably more snow on its roof than the main house. I suspect it has thicker insulation. It probably also has rafters rather than trusses. A 2x10 rafter has an r-value of around R-12 compared to the truss in the photo- which may only be around R-4.
I would be curious to see the pre-drywall photos too if you have them.
Otherwise, BEAUTIFUL home and the ice dams aren't too bad... Good Luck!
I forgot to add... given your assembly, Leaky Ducts are probably not the culprit but worth checking them.
Rick, I think you are right that this is a east facing roof, so the low likely hood event of warm rising air seems totally out.
This shows a few photos of the SPF going in.
http://bamasotan.us/2019/07/insulation/
http://bamasotan.us/2019/07/blower-door-test-1/
This is an very complicated home with a ton of mechanicals, so spray foam is particularly difficult to do well. The only other quality signal I see is that the spray foam operator is not wearing the correct supply air respirator.
Seems like the best thing you could do is a get a home energy specialist to come evaluate the place. They can use a couple blower doors if needed to really draw the place down and make leaks easier to find. They can also do a positive test with smoke.
You said "quality signal". I clicked through and skimmed faster than I could consciously register, looking for the spray foam.
My eye hung on once sentence, though, which is about all I need to read.
"There’s still some work to do to fix the trusses they had to cut for the chimney."
https://i0.wp.com/bamasotan.us/wp-content/uploads/2019/07/B1200XInsul-125.jpg?zoom=1.5&resize=600%2C362
*facepalm*
In Matt’s markup, I think the two rectangles are probably sistered rafters or similar. Maybe with no sealant between and too much of an air gap.
The circle is probably a hole in a top plate or some other kind of penetration. I’d check that first.
Bill
Some shots of the truss system.
From above. East is up.
From NE looking SW.
The rafters that are very closely spaced are probably either under insulated, or not correctly ventilated, or both. I’ve seen that before and even had that happen in my own house. They create channels to carry warm air and explain the vertical line melt pattern you were seeing. You can figure out which is the worst offender based on which one shows the most snowmelt above it.
Bill
Here's a shot from yesterday (Tue) noon. After 5" of new snow Mon & Mon night.
Snow is an okay poor mans IR camera. Get someone there with a good ir camera, blower door, and smoke machine if you really want to know what is going on.
Personally I still want to see pictures of the final foam in the knee walls and understand what was happening with the cardboard.
Here are some shots from yesterday. The melt offs on the east side (front) now appear to be thermal bridging more than vent areas. I'm hoping to get some shots this morning head on with each to give a better idea of how each lines up with the windows to give a better idea of proximity to trusses.
Is it possible that it's both? That initially it melts off more along vents that have warm air in them and then melts off along trusses due to thermal bridging? To Dana's point, I believe the trusses are either 2x4 or 2x6 upper cords so s/b (big assumption with the folks who did our insulation) covered enough with foam to not be an issue?
I'm running in to numerous problems w/ GBA on image uploads. It won't allow some images to upload (and no pattern regarding file size... strange) and then sometimes does a bunch of compression on those I can upload (again, no pattern, some don't get compressed at all, others get totally grunged... and no rhyme or reason regarding file/image size).
So, I'll try to upload to Houzz or Bamasotan later this morning which are both more reliable.
Some of the melt out seems to line up with doubled up framing, but not all of it. The section highlighted in red in the photo seems unexepected based on the framing, unless something else was added later.
We still haven't seen any pictures of the installed foam. Those would be really helpful. If you don't have any, cut an access panel into the east kneewall and take some photos.
Thanks Matt. I'm working on some shots of the installed foam as well as the vent shoots just before foam.
Was there an insulation inspection on this house? MN code, at least in the 7 county metro area I believe calls for one.
Yes, there was an insulation inspection. HOWEVER, given that he inspected insulation in a chimney chase, apparently without a depth probe and without looking very closely, and signed off on it after which I found all cavities to have either barely code depth or less. One had only 1" in a portion of one cavity and another 3" wide cavity had foam along one side but you could see sheathing along the other side. My confidence in anything the inspector says isn't high.
This is the same inspector who insisted that we install a 1:1 MUA if we installed a 10,000 CFM ceiling exhaust fan (EG, to use w/ open windows).
Did they insulate inside a chimney chase, where there will be fire embers and hot exhaust on the inside, and hot bricks on the outside, with highly flammable spray foam?
Also: Why exactly would you want a 10k CFM ceiling exhaust fan?
It's a chimney chase for an Isokern fireplace so it is supposed to be insulated.
A whole house fan plus open windows has several benefits. It brings in a lot of fresh air and on many days can cool a house with a lot less energy than mechanical AC. Very common outside the U.S. (and at previously in the U.S. before people decided that we should close our houses up and use mechanical ventilation for everything).
I am going to agree with inspector on not allowing the fan unless it had make up air or interlocked to all combustion appliances. Interlocking to a to a fireplace is not easy.
I can’t imagine that a house fan pencils out in MN. There are pretty high latent loads and insufficient day/night temp swings. They make sense in dry hot climates. When you pull cool but humid air in you end up with a cool clammy house. Not to mention the number of windows you need to open to allow 10,000cfm in seems totally impractical.
There are millions (likely 10's (or 100's?) of millions) of these in use throughout the world. They are not an issue. There is only one known death caused by a whole house fan and that was a bunch of kids high on pot and decided to turn it on to clear the air before their parents got home and forgot to open windows. Even on this one the pot was a contributor and he likely would not have been harmed had he not been high.
There has never been a report in the U.S. of an inspector not allowing one and thousands of them are sold and installed every year. They are actually recommended by USDOE and the State of CA.
We've used natural cooling / fresh air intake like this for decades in MN with good results and no harm.
Don’t count on inspectors to find everything, especially when you get into unusual or complex stuff. As a consulting engineer, I frequently get into arguments with inspectors when they don’t understand complex systems.
If you are looking to build a really high performance home, you need someone with experience to get it right. That usually means a specialized builder, or a consulting firm that specializes in such things. You can expect the typical city building department to catch every detail, or a regular building unfamiliar with high performance structures to do everything optimally.
Bill
Yep. The really sad/irritating thing is that we were originally going to build a very energy efficient house. Our Architect and Builder were both on board and we had Rachel Wagner as our consulting architect for energy and building science. Unfortunately our Architects got cold feet after they saw her wall structures and refused to move forward. I chose the path of marital harmony (my wife really didn't want to change architects) rather than continue with energy efficiency.
Cold climate residential buildings take a lot of effort in the construction to perform thermally. There is a proven method for ceilings that include ventilation, sufficient R-value and minimized thermal bridge. These 3 elements work together to assure performance in all conditions, you have redundancy to handle unusual periods of extremes. If for instance the roof ventilation is the only thing keeping ice dams from forming after a snowfall you have failed miserably in the construction process. That the ventilation channels are carrying away heated air from the building envelope is a sign of failure.
Going to have to agree with this.
A hot roof approach produces ice dams. It's implicit. You can reduce the *rate* that they form by packing ten times as much insulation in there as you otherwise would, and breaking every thermal bridge, but unless the ratio of R-value provided by the snow to the R-value provided by the insulation is so low that the bottom of the snow can stay above freezing, it will melt.' The amount of insulation necessary to completely prevent ice dams rises to infinity as you approach 32F ambient. Venting saves you from this math and saves a ton of money on insulation.
Snow is about R-1 per inch.
For 12 inches of snow (R-12), to avoid ice dams entirely at 0F outdoor temperature and 70F indoor temperature, you will have to have
((70-0)/(32-0)) = 2.19x as much R-value from insulation as from snow, or R-26
Seems easy enough!
Now how about at 10F?
((70-10)/(32-10)) = 2.72x as much R-value from insulation as from snow, or R-33.
And 20F?
((70-20)/(32-20))= 4.17x as much R-value from insulation as from snow, or R-50.
How about at 30F?
((70-30)/(32-30)) = 20x as much R-value from insulation as from snow, or R-240.
The amount of insulation will go up linearly with the thickness of snow, and will scale inversely with the number of degrees between outdoor ambient and 32F.
The *rate* of ice-dam formation is different. If your roof is packing R-240 of insulation and R-12 of snow at 31F, ice dams will form eventually, but extremely slowly, with very low heat flux per second.
Venting adds a control: the heat flux through the sheathing is both drawn away into vented air from wind, and if significant, generates its own convective 'wind' from stack effect. More often than not, even a small amount of soffit-to-ridge venting completely eliminates the prospect of ice damming from that part of the roof.
On a roof like this where much of it cannot be vented due to valleys, dormers or other stuff - best to not vent at all or to still vent where possible?
Still vent where possible is ideal. Ice will melt to water in proportion to the area of roof that is above freezing.
Just take care. Reverse-lap situations, if any are inadvertently created, will usually lead to some degree of leaking. Any pitched roof of typical construction where water driven by gravity has the freedom to make a small puddle instantly becomes a terrible roof. Complex roofs usually require a lot of metal flashing applied carefully in the correct spots in the correct order, to keep water reliably flowing down into the gutter and out of the attic.
Ice & water shield and other peel-and-stick membranes attempt to deal with the leaks generated by ice damming by establishing a water-tight surface, rather than address the cause. They can be very effective with a bit of roof pitch, but they're not quite as effective as flat-roofing membranes, and they're expensive. They also don't do anything about icicles and icefall when the dam eventually breaks off.
The nuclear option here is spending a good amount of money every winter on heat tape for the whole area of unheated eaves where water otherwise cools back down to freezing. For obvious reasons, this site is generally opposed.
Burninate
Do you have a good link that discusses this principle/calculation further?
At first glance, its slightly counter intuitive.
But that it takes only a little escaping heat to melt snow when its closer to 0C° than when its really cold does make sense.
What I'm wondering is how closely it tracks reality. Say its -0.1 °C. It won't take much heat to melt snow at that temp, but water also won't freeze as fast at that temp.
I'm wondering if evidence confirms that ice-dams are more common when its just below the freezing mark?
In the last discussion of your build here - https://www.greenbuildingadvisor.com/question/ice-dams-totally-preventible-or-not
You got a lot of advice on this, but it looks like other considerations prevailed. Now you have ice dams.
As a fix, here's what I gather would help:
Decide that your roof is not finished yet. Put in two or even four layers of 1" EPS, with the offset seams taped, to break thermal bridges, and another layer of synthetic underlayment (taking care to pick something cedar-compatible per manufacturer). Then put in some 1x3 battens (vertical, then horizontal if necessary for attachment) to establish a 0.75" vent channel, and the cedar cladding on top of that.
Special care and improvisation with flashing required to avoid reverse lapping.
Update and photos...
In looking at each of the 13 visible melt lines on the east facing roof it appears that some line up with conjoined truss (double & triple). I'll do a separate post on this problem.
Others appear to line up with the space between.
We had a BPI energy consultant in on Friday. We had previously found that there was improper insulation below the dormers (knee wall below window was insulated but underside of the roof deck was not nor was the wall section between this area and the conditioned space behind the adjacent knee walls) and knew this needed to be corrected. He also found a couple of other similar spots that were uninsulated. We foamed all of these to correct them.
Our builder believes that this should have corrected all of the problems of warm interior air getting in to the rafter chutes as he believes the warm air was flowing from the conditioned space behind the knee walls to the space below the dormer windows, down in to the soffits, along the soffits and then rising up in to the rafter chutes. This is possible given areas that were open to air flow but could that have caused several inches of snow to melt within a day? Interior temps are about 60°f btw.
Photos coming in a couple of hours.
It’s possible that air leakage was thawing 3” of snow. Here’s why:
Thawing 3” of snow sounds like it would take a lot, but that’s not what’s happening here. Your roof is a little warm from the air leakage, so it essentially melts each snowflake one at a time as they fall — the snow never had a chance to accumulate to the full 3” in the warm areas.
If you had a very heavy snow, with the snow coming down rapidly, you might see some accumulation. In a more typical light snow fall that gradually builds up over hours, the snow melts as it falls and never has a chance to build up.
It doesn’t really take much of leak to keep the snow from accumulating on a roof. Check your roof in the mornings, especially after a frost, and you’ll see if the foam you added solved the problem.
Bill
Thanks. That's a good point. Here are a couple of photos from 2019.12.09 when snow was falling. First is between the middle and northernmost dormer, second is just north of that.
First photos - General insulation. First is of the west facing shed roof. Second is the area about 5' to the right of the rightmost dormer (5' north of the northernmost dormer). Third is to the left of the leftmost dormer (south of the southernmost dormer) and forth is just to the south of that (the triple truss is the same in each photo.
Probing the foam it appears to all meet spec of at least 7" in the roof sections and 2.5" in the wall sections (the walls were filled out with blown fiberglass). Top cords of trusses vary from 2x4 to 2x10. IIRC the triple truss top cord is 2x6 in the section of the photo.
Second photos. First is in the space to the right of the center dormer looking towards the north dormer. Second is the same space looking towards the center dormer and light coming from a flashlight in the space below the center dormer window. Supposedly warm air was being pushed from this space through the hole where light can be seen, in to the space below the dormer window, down in to the soffit space, across the soffit and then up in to vent chutes.
Third is the center dormer window. The greenish foam is the old foam that was against the wall below the window. The bluish foam is the new foam installed Friday against the underside of the roof sheathing as should have been done prior. The leak to the space behind the knee wall is to the left.
Also, how concerned should I be about this? I don't want to overreact but also want to make sure that this is corrected if it should be. That we had significant melting and ice dams when no other houses did is very unsettling. My concerns, somewhat in order, are:
- Warm moist air causing the sheathing (and shingles?) to rot.
- Ice dams causing damage to shingles or other stuff.
- Waste of energy
- Falling icicles could be dangerous, will damage plants below and could cause other damage.
- Aesthetically the ice dams and icicles look bad.
Thanks,
This project is perfect example of the value of airtight sheathing. You do all your air sealing at a single well thought out layer when everything is easy to access. Thick layers of spray foam are a 3d mess. I say this from experience with SPF in my own house. The mistakes are frustrating, but some of the ones I dealt with were somewhat forgivable. Only after the fact do realize that laying down a perfect layer of what is a sort of exploding paint in weird spaces is pretty damn challenging. Fingers crossed everything we see is HFO blown.
I mentioned it before, but if you want to solve this, get a really cold day, someone with a good ir camera, couple blower doors, and fog machine. I was out your way last Wed, and it would have been a good day for this work (remember to check the weather before jumping on a plane to MSP). Don’t worry about 50pa, have them run the fans all out when pulling suction. Apparently you can often hear the leaks when you do this all out approach. Fog will be useful for understanding the overall problem. You should be able narrow down the source of leaks somewhat by fogging a specific room. One of the big questions for me is there significant leakage from the 2nd floor into the soffits or it all from the the 3rd floor.
In the last photos, if I understand, they had sprayed the vertical wall under the window and not the sloped roof? Had someone baffled the vertical wall or was there sheathing there?
Energy waste doesn’t seem like an actionable concern at this stage. The rotting roof tends to just be sections, so do the best you can and deal with it if it unfortunately happens. The the guaranteed cures beyond leak testing/air sealing at this stage are worse than the potential future problems.