Answer
Stripes / Not meant to be condescending. In the first couple of answers there were remarks about substitution and then there were strange ways that substitution seemed to be implemented by some posters.
I think #11 (Goldenrivett) already dealt with that mis-interpretation of substitution.
Reading and understanding ought not be difficult; what is difficult is FINDING things in EASA rules. An example is the reference in #3 – it leads properly to the AMC and GM book of EASA, but that is only part of the whole complex of rules. The issue date seems to be 25 oct 2012 (but the first page literally states “Initial issue” ) – so, is this a current reference or not?
The initial check I made was with my company regulations for the situation as described by the OP.
Those rules state literally – The touchdown zone RVR (may be replaced by midpoint RVR if touchdown not available) is always controlling. If reported and relevant, the midpoint and stop-end RVR are also controlling. … etcetera …
Now, as you correctly pointed out, EASA rules nowhere mention the substitution of one RVR with another one.
However, EASA does state that all Low Visibility Operations are subject to Special Approval. The company seeking a Special Approval must present their Operations Manual to the national authority for authorization – with that approval, that OM becomes the law for the company pilots.
What we, as pilots, cannot check is whether the inclusion of the substitution rule is a sort of unintentional leftover from previous rules, and not weeded out by the authority during their inspection, or whether it is an intentional amplification of the meagre wording in the EU rules.
Using Air Ops Easy Access lumps all the books together, but what a load of electronic paper to rummage through. Of course, search function, but what if you are looking for “low temperature correction” , they call it “temperature compensation”, “substitute” or “replace”? Tedious!
With all these caveats it is hard to give a definitive answer to the OP, he really should get his company to check it with their national authority.
Your reasoning in #22, second paragraph, about the similarity with the situation where an RVR drops below limits after you have descended below 1.000 HAT seems correct, but actually it is not from a statistical risk standpoint, such as regulators regard.
The time from 1.000 ft HAT until touchdown is around a minute and a half. Only one aircraft is in that position at any one time. The crew is bound by required visual reference rules. The risk for just that one aircraft to do look and see in case one RVR drops below limits is acceptable.
Suppose you are number 1 of the 200 aircraft bound to land at that airport. All the other aircraft behind you get the message about the RVR below limits before they reach 1.000 ft HAT.
With your reasoning about similarity of situation, all the other 199 aircraft would also go for a look-see and that exactly is the risk that the authorities do not find acceptable.
Wording can be misleading – EASA OPS states that when 2 or more RVR’s are available on a runway, then one may be U/S. With this wording you do not get the idea that Cat 3 runways, in principle always have three RVR’s and only in exceptional cases would one be classified Cat 3 with only two RVR transmissometers (perhaps very short runways?). If the original runway only has two RVR positions, is it a TDZ and MID, or a TDZ and Stop-end? If TDZ and MID, then the opposite RWY would not have a TDZ RVR. If it is a TDZ and a Stop-end, and the TDZ is U/S, Stop-end 75, wow, what a problem if your required TDZ RVR is 200 m, or even 300 m? Would an XAA dare to say, hey, Stop-end is above limits, let all those 200 aircraft inbound just have a look see?
In my post I did not confuse T/O and landing minima, I only mentioned RVR 200 as limit for landing, as per OP post, the mid and end RVR were just example of values.
The problem of the OP is exactly for aircraft with a rather high TDZ RVR requirement, my aircraft with 75/75/75 for both take-off and landing is not so complicated (if all systems are go).