Has anybody done a room by room Manual-J type heat loss/gain calc that included the site orientation & shading factors, etc? A 2 ton might be enough, too much or not quite enough, depending on the true summertime solar gain factors. (I suspect 2 tons is overkill for most decently insulated homes with all low-E glazing in NW Ontario, but certainly not all.)
At anything above ~ R5 or so wall-R is of almost no consequence from a cooling point of view, since even through low-E windows the solar gains through the windows will dominate gains through the walls. Roof gains & window gains are going to be 90% of the cooling load.
South facing windows with low solar gain factors have comparatively low gain due to solar geometry: The bulk of the heat & light are mostly reflected at high incident angles of mid-day sun, and roof overhangs may partially or fully shade those windows. There is still significant gain, but southwest & west side glazing has a bigger effect on peak loads, since the sun is lower in the sky when shining on them, and the air-temp is higher in the afternoons. The solar gain factors of even low-E windows varies quite a bit, and modeling the solar gain factors, particularly of the SW to W side windows correctly can change the peak load numbers by quite a bit.
Roof insulation types & finish roof materials can also affect the peak loads. Rock wool & cellulose are opaque to infra red, whereas fiberglass is somewhat translucent, and the peak temp in the insulation layer can be a few cm in from the top under a hot roof. Darker composition shingles have higher solar gain, but about the same emissivity as light colored shingles, and run significantly hotter. Metal roof finishes also vary by quite a bit from a gain/emissivity point of view.
Are the ducts in the crawlspace insulated? Are they air sealed at every seam & joint? If not you may run into some condensation issues, and duct losses cooling off the crawlspace may be of modest cooling benefit to the first floor, but not a lot, and it would add to the total tonnage requirement. Are the crawlspace walls/floor insulated & air sealed?
Don't expect anyone to do the more rigorous analysis for free, and I suspect only 1 in 10 are even up to the task. Being an energy nerd, I'd run a
Hot2000 simulation on the place to come up with both the heating & cooling load. It's time consuming to do the data entry, but may point you to ways of cutting the load that are cheaper (and more comfortable) than upsizing the compressor.
Is the floor plan open enough to consider going ductless (mini/multi-split)? Split systems are inherently more efficient (no duct gains or losses), and inverter drive ductless are more efficient & more comfortable due to fully modulating compressor speeds and automatically adjusting to the load. (And if your furnace is propane rather than natural gas, a heating/cooling heat pump would cut your heating costs, even if sized to only keep up with the heating load down to -5C or so.) A 2 ton high SEER high HSPF single-head heating/cooling mini-split would likely come in under $5500, maybe $4000 if cooling only. (Add ~$1000-1500 for a 2-head.) Many 1.5 story houses that have only moderate gains on the first floor can be cooled adequately with a single head mini-split on the upper floor if air is able freely convect up/down the stairwell.
At 2.5 tons or lower the choices of central-air compressors are relatively few, but 1-3 tons is the sweet spot for ductless, with many options to choose from. Oversizing an inverter-drive ductless by even 50% on cooling tends to yield higher average efficiency, since they run more efficiently at part load than when running flat-out. Whether there's "pay back" on the efficiency bump from upsizing just for cooling depends on the length of the cooling season, which isn't super-deep or long in your location. But if it's being used for both heating & cooling, in many locations it's cheaper to heat with a ductless at 0C than it is to run condensing natural gas. It's unlikely that even a 2.5 ton would be oversized for your design condition heat load (but it might handle your full load at -5C or even -10C.) Efficiency drops off pretty dramatically below -10C or so, but most are still more efficient than electric baseboards even at -20C. At -25C or lower, forget about it.
