This review is obviously not peer-reviewed and represents my opinion only.
Recently I deployed some Onset Hobo Light Loggers (https://www.onsetcomp.com/products/data-loggers/mx2202/) mounted to a frame next to a top-end multispectral light sensor (https://insitumarineoptics.com/ms9/). Onset Light loggers are relatively cheap, and log for long periods of time, making them a common tool in the Marine Scientist’s toolbox. However a short-coming of the loggers are that they measure in Lux (a measure of illuminance) rather than irradiance (a measure of energy, which is more relevant to the photobiology of organisms). A second issue is that they foul with algae thereby reducing the light measurement, whereas top-end sensors have wipers and antifoulants to reduce this.
To address the first issue with lux, the sensors are often converted using a standard conversion factor of 0.0185 (https://www.apogeeinstruments.com/conversion-ppfd-to-lux/) which is appropriate for the loggers above water in direct sunlight. But there is concern that changes in the light spectrum under water with depth and turbidity, may mean these conversion factors are not accurate.
To address the second issue, researchers often only use the first few days of data or clean the sensors intermittently, but how many days until the reading become unreliable?
Below is a 20-day deployment using only daylight hours, with sensor measurements converted to PAR.
Using the standard conversion factor you can see some divergence between the sensors after about day 4, but otherwise the Hobos were surprisingly reliable, given how many factors can affect light intensity.
Put as a proportion (below) and using the same standard conversion factor, you can see the error is pretty reliable for the first 4 days, with a little variation on the 4th day.
To correct for low but measurable turbidity (0.5 – 0.8 NTU), I found using a slightly higher conversion factor at 0.022736 (r^2 = 0.9847) for the first 4 days produced the best fit. The error during these days remained with +/- 11%.
I should also mention that the MS8 sensors are not perfect. They have a wiper that stops fouling but only have 8 spectral bands between 400 – 700 nm. This means that there is a bit of course interpolation between each of the bands to derive the PAR. The newer version contains 9 bands and covers the visible spectrum better.
Bottom-line. Calibrated Hobo loggers provide a pretty decent estimation of the downwelling irradiance for the first 3-4 days. For clear-water shallow reefs, a standard conversion factor of 0.0185 would provide a decent estimation of PAR. The amount of fouling could vary between reefs, but given this was deployment was done in one of the most inshore reefs of the GBR, it probably couldn’t be much worse.