Think about it this way, to be able to say that it takes place annually instead of i.e. biannually or monthly, you need a lot more than one sample per year. You need enough samples to know when it is or isn't occurring.

https://www.earth.com/news/unprecedented-collapse-panamas-oc... mentions a lot of date oriented measurements which suggest they probably have at least 52 samples per year, if not daily samples:

> The 40-year record makes the 2025 failure stand out. Average historical onset around January 20 contrasts with a March 4 threshold crossing in 2025.

> The cool season shrank from roughly nine weeks to less than two weeks. Minimum sea surface temperature (SST) rose from historical lows near 66.2°F to about 73.9°F.