The chart doesn't say _when_ to keep kids in vs. let them out; it instead quantifies risk. When I've read national coverage of regional variability in when to have indoor recess, the trend that has stuck out in my mind is that indoor recess basically happens whenever the day's low drops about 10 degrees below local _average_ low (for winter, that is); wind and precipitation is obviously a significant factor as well. Local average low is surprisingly variable -- it's not impossible to find two locations separated by a dozen or so miles with local average lows that vary by 10 or more degrees.
Areas with very low local average lows in winter will routinely send their kids out in temperatures which are rarely if ever experienced in more temperate regions and, when experienced in more temperate regions, may lead not just to indoor recess but even to school closure. If you sit around trying to figure out why they don't do it the same way in New Jersey that they did it in Michigan, well, yeah, you're gonna be confused. But if you factor in the local conditions, it makes a lot more sense.