Friday, February 17, 2012

Steve Goddard steals Pat Michaels' trick of deleting inconvenient data


Pat Michaels got in trouble for repeatedly deleting data that is inconvenient to his denialist viewpoint.

Now Steve Goddard is trying his hand at the same "hide the incline" trick.  He puts up a post with this graph to claim sea levels aren't rising:


and adds a link to the dataset that inconveniently omits the much more extensive datasource that comes with its own graphs.  Shall we help out Mr. Goddard?  Yes!  Just click the link, click "Time serie" and away we go!

First graph:

What do you know - same data (Envisat, no inverted barometer, no seasonality removed, no isostatic adjustment), but it comes complete with nifty smoothing and an inclined trendline.  Instead of just pasting in the url for this graph, Goddard downloaded the data and recreated a graph without inconveniently trending trendlines.

Still, it's not going up quickly, so that's something.  Or is it?

Same data, except with inverted barometer:

I've poked around about whether applying the inverted barometer adjustment is a good idea.  My best guess is it's unnecessary in the long term, but this short data set makes it helpful (see Realclimate).  Goddard denies it, but in general I find him unpersuasive.

How about other data?

Maybe Jason:

Maybe Multi-mission:



Plenty more variations out there for you to try yourself, but the pattern sticks.  Goddard cherrypicked a short dataset that most favored his position and then recreated the graph to hide the incline.  It then got picked up by Watts (who tempered it slightly) and spewed everywhere (I found it first at TigerHawk and traced it back).

The one useful conclusion to this is if you do feel it necessary read Goddard, do so by reading ReallySciency who's watchdogging him, including on this issue.