As usual the satellite data for the month of March was quickly available at the beginning of the April. This timing difference is one reason why I am doing more of my updates from the satellite data, but I am still watching the station data. Maybe this is why it really caught my attention the way it did when both the CRU and GHCN anomaly jumped in the month of March. This is in contrast to both the UAH and RSS measurements detecting a drop in global temperature anomaly.
So I decided to dig into this a little more to see why the different methods diverged so strongly. Since so many of the warmists focus solely on the station data it is important to understand what is happening to it. I am convinced that the problems inherent in the station method make it the least reliable source of monitoring the climate, but since so many do use it, it is useful to understand where it is detecting the warming.
At the beginning of the year all four sources were within 0.4 °C of each other. That spread has now increased to 0.6 °C. That is an enormous divergence for a 3 month period. So I decided to take a closer look at where the warming was being detected. To so this I chose to compare the UAH and the GHCN sets. Both of those sets have useful breakdowns of the regional anomaly by hemisphere and the land and ocean.
I started with the oceans and nothing really interesting showed up. This seemed like a good starting point because the stations have poor coverage over the oceans so a few readings can really skew the data, but nothing too substantial showed up.
The offset between the two doesn’t concern me very much as a different period baseline will generate such an offset. They do not show the exact same behavior as each other, but it is interesting that the GHCN shows the SH ocean warming while the UAH shows it cooling so far this year.
When I plotted the land data I went back to check the data because I thought there must have been an error, but there wasn’t. The GHCN shows radical warming in the NH, but only on land.
From February to March the UAH showed that the land anomaly increased by 0.1 °C, but the GHCN showed an increase of 0.45 °C. More than 4 times the change that the UAH showed.
Part of this difference could be that the different methods have radically different averages for the month of March, but when I looked back for the March NH Land anomalies since 2000, I found an even bigger surprise.
The GHCN shows more than double the temperature range for the month of March since the year 2000. The anomaly range of the UAH is 0.81 °C for the past 11 years. The GHCN shows a range of 1.68 °C for the same time frame. In 2008 the temperature difference between the two methods was an enormous 1.6 °C.
What is even more interesting is that the GHCN land temperature should resemble the CRU behavior as that is land only and they use so many of the same stations to generate their results, but the GHCN global land even makes the CRU data for the month of March look stable.
In this case the GHCN global land temperature is showing double the temperature range that the CRU does over the same period. Including that absurdly high March of 2008 temperature of 1.83 °C.
While the GHCN temperature data isn’t quite random, it is getting disturbingly close to that. It’s behavior is so different than the other temperature sets that it might as well be random for its usefulness. Despite the offsets that exist between the UAH and the CRU temperature data, they show much more comparable behavior than the GHCN. Both the CRU and UAH (global land only) show that this was the coldest March since before 2000. The GHCN is simply out to lunch. Based on the differences they are posting for version 3 of the GHCN, it will be even worse when they switch to that. At that point there will be no further need to refer to the GHCN as a measure of global temperature.