Well, since the controversy about surface temperature data boiled over and not only climategate happened, but also the collection of data on surface stations by surfacestation.org, there has been a remarkable change in argumentation.
While this is nothing new to informed laymen or even scientists, but certainly for media types of all couleur, we should investigate it a little.
Instead of using land surface temperatures, climate scientists argue that sea temperatures are a lot more indicative of global temperature movement in long terms. I agreed with that and follow their train of thought on this from an engineers perspective. Sea water can be seen as a huge capacitor which can hold a lot of energy as thermal energy. Since there is so much volume of water on the planet, the discharging or charging of the capacitor water is slow (C very high, Change in Temperature low). It’s like this, when you put force on a toy car, it will move with the small force you give. If you put the same force on a real car (with no brakes or gears locked) it won’t move at all, or you at least have to wait (depending on the force) a long time till its visibly starts to move. The sea water is this real car.
Now, this means that it is a filter, it filters the small changes due to natural phenomena and only gives us long term trends. If you look at sea temperatures (you can play a bit here), you will recognize that they have a trend (slightly upwards). Now, natural scientists stop here and say, we have a temperature trend upwards.
The engineer in me asks, well, how did they get this data and did they change data gathering during its time. And, yes, they did. Before 2003, water temperature was measured by devices that had to be used by crew members on commercial ships. These ships usually cruise the Northern Atlantic and along the shores down around Africa. They don’t traverse the Southern Atlantic much, same for the pacific. So, data was sparse and restricted to small areas. Now, we have self-sufficient floating islands that measure temperature and thus we have a grid that actually gives useful data on a wide and divers sea surface area. Another advantage is that all kinds of errors due to handling are reduced, because once they are floating, they are going. Measurements done by crewmember on a ship are prone to a multitude of errors and thus should be neglected.
However, that leaves us with a pitiful small amount of reliant data, less than a decade and thus unsuitable for a long term analysis. We still have to wait a bit, I think.