Looking up a maximum value in a column where the columns vary in the number of rows

The file I'm working on is available at the following link. It might be easier to look at the sheet itself. Cell E47 has the formula that isn't working as expected, it only returns a result of Z irrespective if what the rest of the table is doing.

http://files.me.com/jkuetemann/2ucihs

The sheet looks up the correct maximum values and the correct pressure group for a planned time (the only issue there is if a user keys in a value higher than the NDL it returns a PG higher than is allowed).

Thank you so much, DonkeyOte. I've been beating my head against this for ages.

Ah, yes. I forgot about those hidden values. They were from an earlier stab at my issue. Clearing those cells solved that problem.

The maximum values on the chart aren't subject to change which is why I substituted the 9.99E+307 for the actual maximum value (it helped me see better what was going on).

I wasn't going to store the column references which is why I combined the formulae In E43 and E44. For now they remind me of where the references are derived from.

=INDEX(I4:I29,1+COUNTIF(INDEX(J4:U29,0,E61),"<"&E46))

The 1+COUNTIF seems to cause a hiccup where there are a few cells at greater depths that skip a pressure group and are represented be an arrow. I tried clearing the cell but have had to use the preceding lower value in the cell to be skipped to get the expected response. For example, the 120 foot column, cell S12 contained and arrow to signify to a viewer to skip the cell and move to the next cell S13 if the required time was over 11 minutes. Is there any way to skip a cell that isn't a number? Or is it easiest to simply hide the value of 11 in cell S12 using formatting? The other issue that seems to be caused by the 1+COUNTIF is if a value for planned time exceeds the NDL it returns a pressure group one higher than allowed by the NDL. This is really dummy-proofing as you should never plan a dive time longer than the NDL unless you'd like to get bent .

I may have bitten off more than I can swallow and don't want to wear out my welcome with you. Just let me say you've been very helpful and I appreciate it. Let me outline how the whole thing I hope will eventually work.

The sequence for the table is this:
1) Dive to a target depth for a given time produces a given pressure group
2) Surface interval starts with the pressure group at the dive's end. After a specified time produces a new lower (closer to A) pressure group. This is the table to the right of the one you have assisted me with to this point.
2) The new pressure group following the surface interval and the depth of the subsequent dive are used to find the RNT and ANDL values in table three. This is the one immediately below table two.

For a subsequent dive the RNT is added to the planned time and the ANDL is simply a reference to you 'adjusted' NDL. The tables are cyclic. I've learned quite a bit in this exercise thus far about combining various functions. Thanks. Latest iteration is attached.