Is the block and page buffer completely seperate, or can a block scan use page buffer if the record is already there? (i.e. will the block pool need to be the table size)

I don't really understand your question...

DB2 has its own buffer pool(s) to cache pages that are accessed and potentially modified by DB2. Additionally, the underlying operating system may have a file system cache. Both are not related. For some platforms, you can tell DB2 to use direct I/O and, thus, bypass the file system cache.

Or Are you asking about block based bufferpools...Block based bufferpoools are the small part of the bufferpool (depends on u how much % of that u can give to block ) used to read in blocks and it has the same page size of the tablespace

Yes I'm talking about a table space that I have designated NUMBLOCKPAGES 75000. This is 50% of the total tablespace size. Are the 2 areas treated as separate entities ?

you cannot have NUMBLOCKPAGES for a tablespace. You can assign a part of bufferpool to be NUMBLOCKPAGES which can make that many pages of bufferpool as Block based . And the tablespaces associated with that bufferpool can have access the NUMBLOCKPAGES. But until your database is executing queries with large result set I dont see any reason to have BLOCK based bufferpool

Block area of a block-based bufferpool is only used for sequential prefetch operations. Page area is used for other types of IO operations. In that sense you can say they are treated as separate entities. If too many pages that are to be read by a sequential prefetch are already in the page area, the block area will not be used.

By the way, it has little to do with large (or small) result sets, contrary to what azs0309 is saying.