Agreed, when the main reason to add a second sub is to even out the frequency response. You'll still benefit from the additional headroom with colocated subs, though not much when mixing subs like the OP is doing. Found this cool site that helps determine the additional dB gain from multiple subs, both identical and non-identical. They need to be in phase and of course playing the same material to use this chart (otherwise you need to use the "incoherent" version - in a nutshell the combined dB is mostly cut in half from these numbers). It also points out decreasing returns as you add more subs:
So 2 identical subs would increase headroom by 6 dB, adding a 3rd adds an additional 3.6 dB, and a 4th only adds 2.4 dB. Of course, those extra subs also further smooth the frequency response, so everyone in the room feels it! But if you add an SB2000 to an SB13u, because they are not equal power you won't get a 6 dB gain. Assuming an average 7 dB difference between the two, adding an SB2000 to an SB13u only gives you an extra 3.2 dB. Not bad, but probably not worth the price of the SB2000 if you just colocate it and only buy it to add some headroom.
One other thing to point out, these charts assume the sources are colocated. Once they get about a 1/4 wavelength apart, they won't be in phase anymore, and that additive power will decrease. I noticed this on my own setup running REW, with my SB2000's about 10 feet apart. That's the 1/4 wavelength for 30Hz, and sure enough, above about 35 Hz I saw a gradual decrease in the effect of adding the second sub, from 6 dB down to 4 dB @ 50 Hz and 3 dB (incoherent) pretty much after that. They weren't co-located, so different responses to room nulls between them started dominating at that point in my test, but I would imagine they would see coherent effects again as they got close to full wavelength frequencies in the 90-150 Hz range. Subs further apart than 10 feet would see coherent effects die off lower than 30 Hz.