Maybe use the Gravity Ordered Velocity Stress Score. That's still only an estimate of power since some of it is statistically based.
In the meantime, just use the same machine and the same incline every time and the initial values can be used as benchmarks whether they're accurate or not in absolute terms.
Watts aren't that necessary a metric for running anyway. Power is used a lot more often in cycling and in rowing, where factors such as wind conditions and temperature (which affects air density/aerodynamic drag) make day-to-day comparisons of performance and training status very difficult to measure by only relying on velocity (as in a time trial effort or in interval training), even if the same course is used each time. Comparing performances on different courses is even more difficult (specifically in cycling, where hills, clothing, and the quality of the road surface come into play) without looking at power (watts) as the standard of effort. For indoor cycling, measuring watts can be extremely useful. Even setting up the same bike on the same trainer with the same number of turns to the resistance unit and using the same gears can produce small workout-to-workout variations in resistance based on air pressure in the rear tube, how worn the rear tire is, how centered the rear wheel's contact patch is on the roller bar and gradual loss of resistance to the wheel from potential mechanical factors during a workout. So using a power meter is the most reliable way of determining true effort in that case.
But don't worry about trying to use (or even measure) watts in running. Pace, duration and RPE are where it's at.