The response of the UNIPCC to the revelation that its 2007 prediction that the Himalayan Glaciers would disappear by 2035 has been
- To claim the allegations are voodoo science.
- To apologize, but say it is a one-off and insignificant.
Most reporters accept this response. For instance The Economist.
Let up put this into context.
Consider three (hypothetical) scenarios form the UK.
1. The police investigation into a (possibly) racially-motivated murder is flawed, leading to the acquittal of the accused. The Chief Superintendant blames it on lack of funds for staff training, having previously said race was not a motive.
2. A profitable, listed company goes bust as a result of long-term massaging of the figures. This occurs three months after a respected accountancy firm signs off the annual accountants with no adverse comments. A senior partner says that the auditors were denied access to certain data, but had a signed note from the CFO that another accountancy firm had reviewed that data as part of a management-consultancy exercise. The CFO claims that the company was sound, and has an independent audit to prove it.
3. A highly-rated hospital turns out to have significantly higher death-rates than the average. The hospital chief executive says that it is due to having to cut back on the cleaning, having previously stated that the figures were flawed and politically biased.
In every case, the press and opposition politicians, would be asking for independent enquires (to assess the extent of the problem and to make preventative recommendations for the future), the suspensions of those involved and the sacking of the top person in the organization. So why no such questions, when there is a serious procedural failing in (probably) the most important scientific report of all time? A report that could adversely impact the living standards of billions of people should be to the highest scientific standards ever achieved.
It is not a minor mistake to misquote and embellish a tract from a campaigning group – and then say the forecast is highly likely, without any statistical analysis. This report is written by top Phd’s in their field, not first-year undergraduates. They should know how to assess reports, and draw accurate conclusions based on the evidence.
Further, whilst it is right for the UN IPCC to recognize the error and apologize, it is not for that organization to say, without internal investigation, that this is an isolated incident.
There should be an independent audit of all the report, to make sure that it is uniformly based on clearly-defined scientific standards. The starting point of an audit should be an evaluation of the laid-down scientific standards, and the documented internal control procedures for evaluating the adherence to those standards.
Areas of a full audit might include:-
- That the report is a balanced assessment of the current state of the science, at least noting competing views where it comes down on one side.
- Any statistical probabilities to be verified by trained statisticians.
- Assumptions, where made, should be identified.
- Measurement errors compared to the changes measured.
- The robustness of conclusions over differing timescales. For instance the correlations between increase in CO2 and temperature changes should not be over a defined period, but should test for a decade
- Gaps in the knowledge identified and put into the context of known factors and measurement errors.
- To note the relative standpoints of lead authors of parts of the report in respect to the established science. That is to whether they have recent, novel or controversial standpoints. And to the extent to which this influenced their review comments.
- For recently published peer-reviewed articles central to the aspect, whether it firmly establishes new ground in the debate.