You've asked a bunch of questions about driving the LEDs. It's important to understand how LEDs operate in order to answer your questions.
Basically, high power LEDs are (typically) driven by constant current drivers. You choose an LED, and then you determine what current you want to run it at, based on it's specifications. Then, you go choose a constant current power supply (LED driver) that can supply the current you want, and can push enough voltage to drive the number of LEDs you want in a single string. Wattage is usually a secondary concern.
The thing that causes most people to struggle with this is that for almost everything else we deal with in terms of electronics, the assumption is that voltage is the important variable. With LEDs, current is the important variable. Because so many consumer electronics are designed to operate on a constant voltage, most of our power supplies (like the one you probably experimented with) are constant voltage.
Typically, you would not drive HP LEDs on a constant voltage power supply, because it makes it hard to know for sure what's happening to the LEDs.
Books could be written about this, you might be best off doing some reading on the web about HP LED basics and then asking any questions you have.
You also asked some questions about lumens. Lumens are a measure of the total light produced by a luminary. There's no spatial component - i.e. distance or coverage area. Every photon that comes off the LED is included in that 7-10lm number. You mention "at 0 feet" or penetrating deeper - those two statements both imply a spatial relationship. In this hobby, we often measure light in terms of PAR. PAR assumes a given spatial relationship (i.e. "PAR at the bottom of the aquarium directly under the light" or something like that). You can't convert between PAR and lm without knowing the spatial distribution of the light source. This is why, for LEDs, it's important to know things like how the light is distributed (most raw LEDs are spec'd with a "viewing angle" that tells the entire width of the light). This is where optics come in to play - a tighter optic means a narrower distribution, and hence for a given lm, higher PAR over a narrower area.