If you have a chance, try to catch a show at Ronnie Scott's Jazz Club. They do blues, jazz, funk, etc. and it's a really cozy venue - not a bad seat in the house!

The U.S. NWS is a truly world-class meteorological predictive service, perhaps singularly so. Its cost of operation is only ~$3-4/yr per taxpayer—equivalent to a single cup of coffee—and yields a truly remarkable return on investment (at least 10 to 1, and perhaps 100 to 1).
3/11

That being said, I think there's a lot of interesting work to be done on how a coral colony thermoregulates...I know some research has looked at how different morphologies can disapate heat more effectively. We still have lots to learn!

To your point about measuring "what the coral feels", I think the physical properties of the coral are going to heat up in a different way than the plastic housing of a logger. The easiest way to measure across different habitats is to measure the water column itself, so long as we do it correctly

Thanks for pointing this out! This is something we've mulled over as well. The problem with the models that overestimate temperature is that they are measuring temperature inside the housing, which acts like a greenhouse. Models that have a thermistor outside the housing are much more accurate.

There's much more in the paper, including modelling to estimate the amount of bias one can expect under different irradiance levels and with different shading methods. We hope this serves as a reminder to the coral reef research community to think carefully about how you deploy loggers!

Fortunately, there is a cheap and easy way to improve data accuracy: shade your loggers! We recommend putting them inside a PVC pipe, which shades them enough from the sun but allows water to flow freely around them. With proper calibration, even cheap logger models can give reliable data.

Moreover, if you want to directly compare temperature stress from your study site to previously-published data, you want to be sure you're comparing apples to apples. What if one logger was shaded and calibrated and the other wasn't? Clearly stating deployment details can improve data comparibility.

This has implications for research and management...if your logger is reading 2-3ºC above reality because it's in direct sunlight, you are overestimating heat stress at your study site. Remember, 1ºC above normal summer temps is enough to cause coral bleaching and mortality!

To summarize: more and more people are deploying loggers on coral reefs, but not many of us are explicitly stating whether they are properly shaded or calibrated. What's more, different models have different degrees of bias...some suffer bias even under low light, while others were unaffected!

Another way to visualize this is to plot the offset from the "true" temperature against irradiance on the x-axis. Here you can see the relationship is linear - as irradiance increases, the offset does too. But luckily, when you shade all models, this relationship disappears!

Below you can see that most loggers suffer solar bias - but some are much worse than others! In general, the brighter it is, the higher the error (that is, the logger overestimates the true temperature more when it's sunny).

We then performed field trials to see how much solar bias affects loggers. First, we determined that one model (SBE-56, pictured below) is unaffected by sunlight. We compared unshaded loggers to a SBE-56 on a shallow reef flat. We also measured irradiance to relate light levels with solar bias.