1. There's more than 3 samples. And they picked the locations for taking these samples very carefully. 2. They're measuring different things, obviously. One is

1. There's more than 3 samples. And they picked the locations for taking these samples very carefully. 2. They're measuring different things, obviously. One is measuring the mixture of gasses, the other is pure temperature. The gas measurement is good for getting a picture of the entire system because it allows us to use basic chemistry, which is well understood, to gain a picture of the overall temperature change by looking at how much gas has escaped the ocean. The ocean temperature data, while very useful, can still be subject to both sampling errors and problems with the analysis as far as how to interpret it in aggregate. Especially since you have to put overall weather patterns, like El Nino / La Nina, etc. into your analysis. There's always going to be uncertainty into how you should do analysis of the data over so many disparate points.