Mann’s Screw Up #7 – His “Trick”

There has been a lot of focus on “Mike’s Nature trick.” Unfortunately, the focus has mostly been on the word “trick.” It doesn’t matter what you call what Michael Mann did. It was dishonest.

To be clear, this “trick” was used in both Mann’s 1998 and 1999 paper (MBH98 and MBH99, respectively). I’m discussing it in relation to MBH99 because MBH98 had messy images. The trick is easier to see MBH99. Here is the temperature reconstruction as presented in it:


To make things easier to see, I dropped out all the colors except that of the reconstruction itself. I also zoomed in. That gave this image:


Focus on the sideways ‘S’ at the far right side. That shows while temperatures in the 1900s rose, there was a temporary dip in them. That’s basically what people expect if they’re familiar with the modern temperature record. It’s also not what Mann’s data actually showed. To demonstrate, here is a graph taken from a Real Climate post:


It’s a bit messy so here’s a somewhat crude attempt of mine to extract the temperature reconstruction:


You’ll note this graph doesn’t have that same ‘S’ shape. It shows temperatures at the end of the reconstruction ending on a relatively low note (for the 1900s) rather than a high one. The reason is this graph doesn’t use “Mike’s Nature trick.” His trick was used entirely to change that end of his graph so recent temperatures would look warmer (according to his reconstruction).

Now that we’ve established what the trick did, let’s look at what it was. To do so, we can turn to Michael Mann himself. In an inline response at Real Climate, he said:

In some earlier work though (Mann et al, 1999), the boundary condition for the smoothed curve (at 1980) was determined by padding with the mean of the subsequent data (taken from the instrumental record).

This terminology may not be familiar so I’ll explain. A “smoothed curve” is a graph whose data has been made smoother by combining a large number of data points into a smaller number. The way you combine those data points depends upon the type of smoothing you use. However, there is always more data in the middle of a graph than at the ends. How you handle that is called the “boundary condition.”

Once you understand that, Mann’s comment is fairly easy to understand. When smoothing his graph, Mann padded his series with data “taken from the instrumental record.”* That is, he took data from the instrumental record and added it to his reconstruction. He then smoothed the resulting series. Finally, he truncated his smoothed series at 1980 so nobody would be able to tell what he did.

That’s it. That’s the entirety of Mann’s trick. He added the instrumental record to his reconstruction so when he smoothed his data his reconstruction would end on a warm note instead of a relatively cooler one.

*Mann claims he used “the mean of the subsequent data.” Others say that’s not true, and he used the data itself. Who is correct doesn’t make a difference to this post.



  1. Hey Brandon –

    Nothing to contribute to the topic, I just wanted you to know people are following your work

  2. Thanks David Jay!

    So you know, I was planning on uploading the next post today, but it’s probably going to be pushed back a little. I’ve decided to write a post about something related to the Stephan Lewandowsky papers, and I expect that’ll go up first.

  3. In his book, he says a trick is a clever way to deal with a vexing problem. So what is the vexing problem?

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s