70 mph isn't 90 mph - the speed at which the building should have been designed.
But I presume a "micro burst" might actually have localized winds much faster.
That looks like a precast wall system perhaps. If a metal deck gets sucked up and off the corner the walls become destabilized very fast.
This happened some years ago in the Joplin, Missouri, storm where a Home Depot building had all its' perimeter tilt walls collapse.
We had a micro-micro burst happen in the parking lot at work about 10 years ago. Broke the rear window in my car, no damage to the front windshield, other vehicles had the same type of damage.
I would not call that an engineering failure of the car designer or manufacturer. Sometimes things happen.
@azcats, based on the photos I have seen so far, I would agree, this appears to be panelized wood roof with tilt walls, as is common for these types of buildings in AZ.
I ran the ASCE 7 Hazed tool set to report ASCE/SEI 7-10, Fig. 26.5-1C and Figs. CC-1–CC-4, and Section 26.5.2, incorporating errata of March 12, 2014, on Fri Jul 26 2024. Report is attached.
Based on this report wind design is based on 105 mph horizontal.
This link explains how to convert horizontal wind load to vertical force on a roof per ASCE 7. Short story is that vertical down loading from microbursts is not explicitly accounted for.
Building loads from microbursts is a current research subject that has not made it's way into ASCE7 yet, and so is not yet accounted for in our building codes.
The risk category can be increased from I to IV, that increases the horizontal wind from 105 to 120 mph, but still leaves you depending on the structures design margin in a direction which ASCE 7 does not require any particular strength (other that that required for the roof to be a work surface, equipment loads, and for maximum ponding).
I think there is a downward wind load that's emerging in the design standard of late, but back when, there really wasn't much consideration of downdraft loads, and in southern areas of the US, the larger structural elements can be designed for 12 psf roof live load, while the smaller area elements are designed for 20 psf roof live load. These are viewed as "reroofing" and service loads, that's their origin or the intent of the load magnitudes.
(Ignoring the potential for rain loads to control via either code or a clogged drain and a high placement of an overflow scupper.)
I would comment that's it's probably a rare roof that's been "proof tested" with a full 12 psf or 20 psf live load at any point in its lifetime, so the potential for design errors to rear their head is there, as well. (Burnaby, Station Square, etc).
Lawson has a good set of articles on drainage, if anyone is interested.
I'm betting uplift removed the diaphragm and any chance those walls had to stay standing.
Those panelized roofs were spongy - worked on a couple of them right out of school in the early 90s. I don't remember considering uplift. 2x4 joists spanning 8'.
I agree regarding uplift. The individual trusses can also gang up on the beam column connections or possibly the intermediate beam line connectors such that the beams are picked up off the columns. Once a section of roof lifts, it could pull in on the wall panels (which snapped just above slab level, about four feet above grade, or typical railcar bed height) simultaneously with the horizontal gust force directly on the panel. That, however, would require a robust diaphragm connection at the wall panel. A problem that I see with the sketches above is that they presume the bldg corner points to remain fixed.
In the eastern most bays, it appears that a series of columns were pushed over as though the diaphragm remained attached.
I also considered construction flaws in the panels which could leave them vulnerable. The most likely, aside from poor concrete, is the rebar mat being away from the surface. If the installer did not use high enough chairs, the upper mat (inside surface) could be closer to the center line. I don't know how much this would affect the panels ability to withstand the horizontal forces in this situation. Worse yet would be if the lower mat (outer surface), was raised up to high. This obviously would severely reduce the panels capacity to withstand direct wind pressure.
Rebar mats in damaged section appear to be along panel centreline but what was their cast location?
Columns bent over
Intermediate beam connections, possible uplift vulnerabilities.
