There are dozens, hundreds more studies, graphs, and other data to answer that question. At this point, due to the overwhelming amount of research data available, this kind of question should be considered as made in bad faith.
This graph does not answer the question at all. Why is not the ideal level 100? 200? 300? 500? If you have never been at 500 how would you know if the outcome is actually detrimental overall? You can't make models when you don't have historical data outside of your training set, that's a very simple rule that everyone can understand.
Al Gore made ridiculous claims back in 2008 about how bad it would get in just 10 years (making it seem that the trend was linear, which it isn't) and he was completely wrong on most of them - so the question I raise is very much valid.
Who says this is the ideal level to be at?