...not really, after a closer look (which also explains the price difference, I guess). Fortunately I found a supplier and got a few LMC6484AIN op-amps.
So far I've assembled and tested the "first stage" (sampling only the averaged light intensity), seems to work fine (I'm using Arduino Micro):
Indeed, I had to put the cell phone's LED-flashlight very close to the photodiode to see some smallish increase in the sampled value. I should have some breaks in the clouds tomorrow for proper testing.
The diode is really tiny, here it is with a pendrive and a chunky 3.3 µF film capacitor:
Beside the current (1-second) seeing value and the 10- or 30-second average, I'm also going to display an "average of the best percentage", which I think is more meaningful for us imagers. For example, with my current setup I usually shoot a 15-second video and then stack 30% best frames. The expected image quality should correspond with the average of seeing values of the best 5 seconds out of 15 s.
Another change is using the timer interrupt to calculate the average (and update display/send data to PC) over exactly 1-second intervals ‒ and not by polling the ADC until 10 000 samples are collected. As you can see in the first photo, there were just 7754 read-outs in the last second.