The explore page is pretty inefficient and runs a lot of queries that can be slow, especially when the query has a lot of results. It’s been a while since we designed the explore page, and it could use an update. Regardless - unfortunately the explore page is also one of the more visited pages of iNaturalist and our partner sites, so these queries get run a lot.
If you are only using the place_id and quality_grade parameters, we will cache the first batch of 24 observations shown on the explore page for up to 1 hour. If still only using those two parameters, the total observation count as well as the identifiers and observers will be cached for up to two hours. If you add any other search parameters, or scroll past the first 24 results, then you should see more up-to-date results.
Identify is only available to logged-in users and is designed for more active workflows, so it does not have the same caching implemented. As iNaturalist grows, performing real-time queries becomes harder to do, and when at the same time we have more users running queries, the performance issues grow geometrically. Caching commonly-run queries is one way to help, and so this is something we’ll likely do more of, where appropriate.