That is for the 1st, the one for the fourth is even warmer. Don't know why this site posts the graph from three days ago, after it shows the present one in the reply box.
There are records of warm ocean from the Aztecs and from ancient writings of polynesian tribes. Those are the first results.
None of that matters compared to the dishonesty of the graph the OP posted. Has anyone noticed the subset of years that this information is portraying? 1971 to 2000. The question now becomes, "Why just those years?" Why not a comparison of all historical data? The answer is pretty simple.
That was the period used to calculate the baseline - the zero level on the graph. The data displayed was for Saturday, Dec 01, 2018. It was compared to the average temperatures of the baseline
A date chosen to highlight the message they want to convey. How about they chose the baseline from ALL historical data, not just a subset that makes their graph confirm what their argument is. It is an old trick with graphs. They count on people not taking note of the parameters and subsets used to create them.