I should say at the outset that the main problem here is that I do not have (and cannot get from Volvo Penta) an "official" spec for my engine's raw water pump (engine is a Volvo 2003, 28HP, max. 3200 RPMs). Here's what the bucket tests showed:

(1) raw water from exhaust at wide open throttle/neutral @ 3200 RPMs = 4.5 gallons per minute (gpm)
(2) raw water from impeller pump, same engine settings as above = a bit less than 4.5 gpm (I think we spilled some water as the hose we attached to the pump accidentally chafed on the alternator belt, cutting a hole in it! Didn't realize until a bit later)

Essentially, therefore, output is the same both at the end of the raw water system (from the exhaust), and the beginning of the raw water ssytem (from the pump, before it enters the heat exchanger and then exhaust elbow). Separately, djmarchand here noted to me that the manual for his Yanmar 3GM30F 27 HP engine (3400 max. RPMs) is specced at 7 gpm @ 3400 --similar sized engine to mine, with slightly higher max. RPMs. So, my raw water output should be somewhat under that --perhaps 5 gpm? However, I don't know if such raw water flow specs are typically a "minimum" flow or a "normal" flow. A friend on the dock with a new 28HP Beta Marina engine showed me his engine manual, which notes the raw water flow rate for that engine as a "minimum" (we did a bucket test on his engine at the specified RPM, and the output was a good bit above the minimum).

So, I'm wondering if my 4.5 gpm output is really low, or just kind of low. In any case, comparing to the Yanmar specs above, my bucket tests appear to suggest the pump isn't pushing enough water. In gear while underway, performance is very predictable --we run fine up to 1900 RPMs, but do run a bit warm --then, over 1900, predictably, it begins to overheat within 10 or so minutes. So, my guess --only a guess, since I don't really know how water flow and heat transfer works in an engine-- is that cooling is fine up until 1900; then the heat produced by the engine at this higher RPM increases at a faster rate, and the amount of cooling water isn't increasing fast enough to keep up with the greater increase in heat after 1900 RPMs?

Anyway, seems like the next logical step is to remove the pump. Anyone have any other thoughts on this saga? I've never removed a water pump, but it appears to be a matter of removing two bolts (my workshop and service manuals both say simply, "Unclamp the hoses and remove the pump").