Since the lamp is shorter than the distance to the measuring points (90 cm), one may start from a point-like light source. Then, the UV-power (12 W, see data sheet) is distributed on a spherical surface 4 * pi * r² = 10 m². It follows the lighting density 12 W / 10,0000 cm² = 120 µW /cm².
PLEASE take care to differentiate between a theoretical calculation of the UV intensity and the effective or REAL intensity you'll get. UV intensity is determined by several factors, not only the technical description of an emitter!
That means:
1. Age of the emitter (output reduces over time)
2. Number of ignitions (each ignition causes output decrease in dependence of the ballast type, lamp type and way of connection)
3. Ballast type (is the ballast made for, or is optimised to work perfectly with this emitter?)
4. Ambient temperature (lamp output is extremely dependent to ambient air temperature respectively air cooling situation!)
5. actual running time (every emitter needs some time to get it's defined internal gas pressure... normally 2 minutes for conventional low pressure emitters and up to 30 minutes for amalgam lamps)
This are the most important facts to take care of when trying to get an information about acquired UV intensity. For a pure theoretical approach in a scientific work Herbert is absolutely right. To determine the effective intensity or e.g. calculate the dose a microbe was irradiated with this method will lead to false values/ data!!!
Best solution may be to use a radiometer and to have the experimental data. Depending on the surface size and shape the data may vary. In my opinion, light flow (J/m2) is more important that W/cm2.