Snow load intervention

Large drift on the back roof Following up on a short article about local ice dams and related fun during Boston's endless winter of 2015, I've decided to proactively attack the huge load of snow on the back roof and take some of the weight off before, say, it gets rained into and becomes even heavier.  No ice-dam or water problems because of the superinsulation, but none of this pack is about to slide off by itself due to the sturdy snow-rail underneath holding it back.

The pink box highlights where the siding had shrunk enough in the cold to gap open at an insufficiently overlapped joint.

Rafter deflection under snow load The triggering discovery is this: slight but visible deflection of the roof rafters, which by now are supporting something near the 40 pounds per square foot that typical residential roofs are engineered for.  And not just vertical load -- also close to 250 pounds of longitudinal stress along each panel, trying to slide it off the edge.

Leaving the entire pack up there would probably be okay, but why chance it?  There would likely be additional issues when it all finally started to melt, and draining all of this into the gutters would take a long time.  I'm also a little concerned about stress against the PVC toilet vent stack.

Ready for rooftop adventures So it's time to head for that topside adventure!  Preparation is a little involved: clearing the basement bulkhead for access to get the ladder out, and carving a notch into the pack thrown up from shoveling out the heat pump to foot the ladder into, but finally the ladder's in place and stable.

The task before us I'm greeted by an impassively ponderous wall of snow at the top, and the task at hand seems fairly immense.  Being careful to remain balanced on the [wiggly cheapass "homeowner grade"] extension ladder, the first few bites are taken out of it and flung far out into the yard behind.

No visible load on the gutter There doesn't seem to be any weight on the gutter, which is a relief.  Still nice and level.  It helps that it's not holding up 200 pounds of icicles, as many other gutters all over Massachusetts are.

Panels still in place There's also no sign of panel attachment compromise or sliding.  If anything, the ends are snugged up against the drip edge tighter than usual due to shrinking in the cold.

Odd ripply ice formation above rail Not exactly an ice dam, but some ice
Further excavation reveals more of the snow rail, and an odd ripply pattern of ice just above it.  The rail has a fair amount of ice packed in underneath as well, which extends down toward the edge.  Not exactly an ice dam, but it's clear that some small amount of melt and refreeze has happened under here.

The bit that hangs over the gutter is actually quite light and also ends in a thin ice-pack, which is easy to bang on to break up and remove over the small area right in front of me.  No discernible pressure on the gutter-guards.

The presence of the snow rail is why I can do any of this in the first place: it's my safety catch to prevent parts of the pack sliding off and taking me with it.  It's the foothold that I make sure to generously expose before venturing near any part of the edge.

Evidence of slight warming underneath A little farther up, the connection between the bottom of the pack and the roof panels feels fairly tenuous and there's evidence that it's just above freezing right here.  So over the main part of the house that isn't the open eave, it *is* a little warmer but not in any profound way like in a house with a poorly insulated attic.

  This makes sense, though, and it's not because there's something wrong with the insulation. Snow has some R-value, nominally about 1 per inch, although probably a little less as density goes up.  So two feet of variably-compressed pack adds R-20 or more to the roof assembly, which itself is about R-26 through the foam and sheathing to begin with.  So the roof panels themselves are now maybe 60% of the way through the thermal gradient between the 60F-ish of the attic and the 15F daily averages we've been having outdoors, putting it around 33F and perhaps more on a sunny day.

Which is exactly why more houses get ice dam problems under heavier snow accumulations, especially if things like ridge vents get blocked.  Even with good attic-floor insulation in a typical vented setup, enough heat can leak up to accumulate more under that big blanket and start melting what's on the roofing layer itself. 

Now that I think about it, that's probably the main reason this place never really had any ice dam problems pre-retrofit.  It had gable vents instead of a ridge vent, which would remain open under snow cover, and I usually let the whole upstairs run cold over winter anyway.  It also had very small eaves and no gutters back then, which would give water far less opportunity to refreeze and accumulate on its way off the roof.  It sported a few small icicles now and again, nothing major.

Some element of dumb luck, perhaps, but nice to be able to explain.

Made it up to the ridgeline I opt for a mostly "top-down" approach to clearing, which means I have to make my way upward at first and finally reach the ridgeline.  I'm mostly leaving a few inches of pack intact on the panels instead of trying to clear them completely, but things up here are getting a bit slippery and trying to slide down anyway.  This is the main reason for going top-down -- to leave the big pack undisturbed below where I'm working, as a safety buffer that I can plant my feet against and not have them slip out.

I only go down on my ass once or twice during this entire operation, and quickly learn how to identify the most "sticky" areas of what I've shoveled and how to keep my feet wide apart and up/down for better stability.  This type of shoveling requires more body twisting and conservation of energy while still flinging loads far enough out to land in the yard away from the house, as I generally don't move my feet around while working on a given area.  So it's rather slower than, say, doing the driveway. 

Usual half-slide situation on front roof Peeking over the ridge reveals the usual half-slid situation on the front part of the roof, where the bulk of its pack has sort of broken away at the top but then failed to slide completely off and is stuck.  I take an experimental prod at it with the shovel, and manage to trigger a couple of panels' worth into sliding all the way down and off!  This observation gets earmarked for further investigation later.

Little melted cave around vent stack The toilet vent is usually warmer than the outdoors, and has melted a cute little cave around itself.  This is good, because otherwise the part of the pack above it would be pressing downward against the pipe.  This is a frequent concern on standing-seam roofs as snow can slide down to put pressure against penetrations, and one that I haven't done anything about yet.  But it's here that an idea begins forming for a solution to this; see below.

26 inches at the stack The pack is about 26 inches near the stack, almost completely burying it.

The thing that looks like a cap isn't closing off the pipe, as the whole point is that it stays open to the outside air.  It's the plastic cover from a 25-pack of CD-Rs, which happens to be just a little larger than the PVC pipe and is screwed on to leave a gap above and around the rest of the pipe.  Just to keep big stuff from falling in, and possibly helps hold in the warmth coming up from the house and septic tank a little better when it's windy.  PVC or not, the stink pipe is still a bit of a thermal bridge.

About half finished About half finished at this point.  Looks ugly, but most of that oppressive weight is going to be safely on the ground soon enough.

Finally done! That was a workout Done!  Or done enough, at least.  I've left anywhere from one to five or six inches in place, partially to facilitate walking around on and partially to keep a little of that free extra R-value and mitigate the night-sky radiation issue.  It's also not worth trying to dig down completely flush with the panels; even though I've got a plastic shovel it's not worth banging up the finish on the metal any more than necessary.

It does seem prudent, however, to completely clear the panel above the stink-pipe to minimize what comes down against it during melt.  Which we're still convinced is never going to happen.  In the meantime, development of a better solution can proceed.

Shed looks less scary now Dismal weather prospects or no, the shed-dormer roof looks considerably less scary now.  At this point in mid-February, despite most Bostonians believing that this winter will never end, it probably won't get a whole lot more snow.

Pushing front-side pack to slide down Okay, now it's time to have some fun.  This picture links to a short video of pushing part of the front-side pack off and having it go most of the way down.  This only works near the right and left ends of the roof, and not so much in the middle -- could be related to that mild heating effect from inside, hard to say.  But having seen part of this avalanche off earlier, I decide that I have to try and snag some video of what it looks like.  I suppose I'm hoping for something more spectacular, but I guess I get what I get.  What I can't push off with an extended-handle broom will simply continue to sit there until it warms up enough to slide off on its own, but this nonetheless helps take some of the load off this part of the roof too.

The mound in the front yard is taller than I am by now, and it won't be a surprise if some of that water winds up in the basement once the ground underneath gets totally saturated.  At this rate I'll probably still have snow on the north side in July.

Neighbor's stoop awning torn off I shouldn't complain about any of this.  Here's what can happen when high roof loads are ignored -- my neighbor didn't do any clearing on his place, and as the season wore on the ice dam built up enough on his front awning that it eventually got torn right off the house.  Trashed the stair railing on its way down, and now he's got a mess.  His take on it?  "That's why we have insurance."

_H*   150217