Climate Change Humpty Dumptyism

“When I use a word, it means just what I choose it to mean—neither more nor less.”, Humpty Dumpty

Humpty Dumptyism: The practice of insisting that a word means whatever one wishes it to.

Fact: Something that has actual existence; an actual occurrence

Data: Factual information (such as measurements or statistics) used as a basis for reasoning, discussion, or calculation

I have previously written about facts in relation to global temperature measurements. Climate scientists recognize that not all of the data they collect are necessarily “facts”, because they have been collected at the wrong time, or are of questionable accuracy for other reasons. However, they are still data, in that they are measurements, even if they are known or suspected to be inaccurate, subject to the conditions surrounding their acquisition.

All of the data collected from measuring stations on a given day, or during a given month, constitute a dataset for that period. However, those datasets might not be complete, if data is not collected from all measurement sites for some reason. When a producer of temperature anomaly products selects data for analysis from a dataset, eliminating missing data, obvious data “outliers”, etc. they create a subset of the data which becomes their dataset for that period.

It is common practice among the producers of temperature anomaly products to “adjust” the data in their data sets to compensate for errors and suspected biases. In making these ‘adjustments”, their datasets become “estimate sets”, since the temperatures in the sets are no longer actual measurements. Rather, they are now estimates of what the data might have been if it had been collected timely from properly sited, calibrated, installed and maintained instruments. However, it is also common practice to continue to refer to these sets of temperatures as datasets.

NASA GISS typically “infills” their estimate sets with estimates of the likely temperatures in areas in which there are no measuring stations, or from which no data were collected for other reasons during the current period. It is typical to refer to the “infilled” estimate sets as datasets, despite the fact that most of the original data has been “adjusted” and missing data has been “infilled” with manufactured estimates.

The producers of the temperature anomaly products then compute a global average temperature value from their estimate sets for the period under analysis. This temperature value is then compared with the average estimated temperature value for the corresponding period (month or year) during a thirty-year reference period. The calculated temperature difference between the current period and the reference period is then recorded as the temperature anomaly estimate for the current period.

The “adjustments” made to the data prior to calculation of the anomaly for the period are the only “adjustments” made to the data. However, it is not uncommon for further adjustments to be made to the estimate sets over time, as discussed here and here. Again, it is certainly possible the estimate sets are accurate; and, that the subsequent anomaly calculations are also accurate at some point in the ongoing “adjustment” process. However, if that is the case, it would be very difficult to determine which of the “adjusted” values is accurate.