If they are getting 0.02% more accurate temperature readings, couldn’t that account for a 0.02% “increase” in recorded temperatures?
First, the margin of error in the calculations using the data from ground based stations is far more than that even using their own figures. For decades it had been acknowledged that temperatures peaked during the 1930s and that even the strong El Nino in 1998 did not come close to reaching the temperatures from that time period. Over the past ten years however “for greater accuracy” the major databases have been adjusted downward during previous warm periods.
The raw data is still available, but the way that they calculate average temperatures has been “adjusted”. It is fairly easy to give the high and low temperatures for one location... calculating the average temperature is slightly more complicated. Temperatures are taken at equal intervals... say every 1/2 hour and then all of the readings are added together and then divided by the total number of intervals. This gives an average. Then the averages for the recording stations are added together and divided by the number of stations.
This is fairly straight forward and if one was uses data from the same recording stations and if the same instruments have been used one would assume that the data would show historical trends that could be graphed and would give a relatively accurate representation of what overall temperature trends have been. The problem for the alarmists is that if one takes the raw data available and performs these types of calculations and makes graphs from the results... the temperatures from the early and mid 1930s come out significantly warmer than more recent temperatures.
The biggest problem with this type of analysis is that it really only gives meaningful results if data from the same stations are being used and also if the instrumentation is accurate, has not been changed, and it has been set up correctly at appropriate sites. Unfortunately, most of these conditions are often not met. For example stations set up originally in rural locations which are now surrounded by asphalt, concrete, and structures show warming. Also thermometers must be changed over long periods of time and newer instruments often give different readings... hopefully, but not always more accurate readings. So it is reasonable to make some adjustments to the data for greater accuracy on a site by site basis. Most of the biases introduced tend to give false warming trends.
To get the results that they want they have come up with more complicated ways to calculate average temperatures world wide. First they claim that they are breaking the world into a grid and giving each section equal representation. Then they “cherry pick” the data by leaving out or adjusting the raw data from recording stations that do not fit the narrative. Then urban areas which tend have more recording stations and also tend to be warming because of the “urban heat island” effect are still over represented. It has been found on numerous occasions that “scientists” have estimated the temperatures of millions of square miles from remote areas such as Siberia and Northern Canada based on raw data taken from instruments hundreds of miles away located in urban areas.
To make this type of subterfuge more difficult to discover the “scientists” often refuse to release their methods or even disclose the stations that they are getting the data from. It becomes a garbage in garbage out situation. This is why the satellite data which is far more difficult to manipulate, uses much higher resolution, and was designed to be more accurate than ground based stations is much more appropriate to use for determining worldwide temperature trends.
The satellite data clearly shows that there has been no worldwide warming trend for nearly twenty years now.