I would guess that this is the last point in the frequency response with a passband gain that is still inside the ripple spec. Meaning if you use the 3-dB bandwidth, you will have an attenuation of 3 dB (and therefore a "ripple" of at least 3 dB assuming the ripple spec for the digital filter is less than 3 dB). On the other hand, if you use the ripple limit bandwidth, your attenuation at that point will be within whatever the specified ripple is.
I couldn't find the spec for the ripple in the data sheet, but consider the following example. If the ripple were 0.1 dB and the passband gain were 0 dB then the attenuation at the ripple limit point would be 0.05 dB and from there the gain would decrease until reaching the 3 dB point where the attenuation would be 3 dB.
There is a chance I am wrong about this, since it isn't a standard term I'm familiar with, but I'm posting the answer because it seems very likely to be correct. Maybe someone else with more experience can provide a better response or point out my error if there is one.
The reason they probably include both is that a digital filter designer would only be interested in the ripple limit bandwidth (which we would generally refer to as the passband edge when doing an equiripple design) while many analog filters are specified with their 3-dB (or 6-dB) bandwidth. Probably there was some kind of internal discussion about which one to use, and they wound up with both. But that's all speculation on my part.