Dr Jennifer Marohasy writes about the Technical Advisory Forum on the Australian Climate Observations Reference Network (ACORN-SAT); whose purpose seemed to be to cover up the Bureau of Meteorology’s incompetence, not to shed light upon its practices and malpractices that include the heinous fudging of meteorological data to fit in with the lucrative support of the Climate Change meme.
Dr Marohasy explicitly identified Rutherglen in Victoria as one weather station whose data had been homogenized out of reality; producing a warming trend when the raw data from an unchanged weather station indicate a slight cooling.
The BOM’s persistence with statistical smudging of meteorological data in order to make it “nice” for further analysis is fundamentally flawed. The process fabricates data; based on subjective judgement (bias) of the data processor. Meteorological data’s inherent discontinuity cannot be bridged by statistical methods without introducing artifacts of the statistician’s own making.
BOM has not even tried to preserve what would be continuity in meteorological records because every time the equipment to ostensibly measure certain weather parameters is changed; the physical quantity actually being measured also changes. Moreover, temperature cannot be measured directly. Temperature is not a physical quantity; it’s an “intensity” whose strength can only be quantified (approximately) by its effect on other things. i.e. by proxy. Different types of instruments use different proxies, and while they all give the same reading in the stable environment of an idealised test cell; their transient response to perturbances can differ substantially.
The World Meteorological Organization (WMO) has long been aware that shift in instrumentation technologies would plausibly produce “observed” changes when there are indeed none. But it washed it hands of the problem and left it to national bodies to decide. It seems that those didn’t care about the real consequences that that would have on the continuity of weather data for climate analysis. Independent researchers have bothered; running various instrumentation technologies side-by-side for long periods. A recent example was identified by Pierre Gosselin in his NoTricksZone blog. Electronic instruments resulted in a bias to warmer temperatures compared to using older instruments in the same place and time.
The discontinuity of data series cannot be ignored. One cannot “adjust” short data series with others to splice them into one long data series. That’s not even valid if instrumentation does not change, but if there was a break in data collection. Any interpolated data is a fabrication. Using any fabricated data for statistical analysis will bias the statistical results, depending on the method used to fill the gaps.
But BOM does more than just fill gaps. BOM adjusts temperatures up and down, depending on how much they judge it is necessary to make those historical measurements fit in with the new ones. And they do it even when there has been no change in weather station instruments or surroundings. To anybody with a moderate respect for the integrity of data, the BOM’s homogenization process reeks of incompetence and the Forum’s attempts to gloss over the unscientific handling of data within BOM doesn’t cover the stench.
Unlike Dr Marohasy, I find the “need” for post hoc homogenization to be specious. And to fundametally produce nought but statistics biased by fabricated data. Historical, short-term time series lacking measured alignment to later technologies should be analysed as independent, complete sets, taking into account discontinuities as discontinuities; not to be in-filled by averages of an arbitrary flavour. This makes the statistics very much harder to do. And producing a rational, easy to understand result will entail a deep understanding of statistical methods, not just a mechanistic application of whatever method seems to fit.
If there is to be any hope for producing a continuous set of data, it is the newer technologies that need to have their readings adjusted so that they show what historical instrumentation would show.