args

Governments should require that funded climate data be posted

The full resolution is: "In all countries, governments should impose a condition on climate research grants and aid related to climate research that source data collected or analyzed under the grant, and all software developed under the government support shall be posted on the Internet within one month of publication or announcement of the results by any means." The resolution was abbreviated to meet the character limits, and the full resolution is the one to debate. The purpose of this resolution is address one of the issues raised by Climategate, the scandal in which e-mail and software at the Climate Research Unit (CRU) in East Anglia. http://www.climate-gate.org... It's not known whether the CRU data was exposed by a hacker or by a whistleblower, but however revealed, issues persist. The scientists were revealed to be trash-talking about climate crisis skeptics, and apparently conspiring to subvert the peer review process. Those issues are put aside here to discuss another problem, the concealment of software and data from the scientific community. The revealed documents includes a README file of a scientist, "Harry," trying to reproduce the climate data published by CRU, documenting enormous difficulty doing so. the file is posted at http://www.anenglishmanscastle.com.... CRU's mission is to obtain temperature data from various sources around the world, validate and correct the data, and convert it into a gridded format useful for scientific and practical purposes. The validation and correction steps are important because the raw data includes clerical errors, instrument errors, and errors due to the heat effects of new construction near the individual collection stations. "Gridding" converts the temperature data from the randomly located collection stations to regular increments of latitude and longitude using interpolation techniques. CRU performs all of the processing functions. For research on global warming, small errors are important because the total amount of global warming examined is on the order of only a degree per century. Moreover, scientists look for "natural experiments" in which local conditions may have local climate effects. For example, rapid growth of a city many increase local pollutants or local CO2 levels, and scientists like to examine the possible local effects on temperature. Britain has a Freedom of Information Act (FIA) similar to that in the United States. FIA requests can be filed to obtain certain documents and other data developed at government expense. In Britain, someone filed a request for the data used to support claims of CO2 global warming. CRU had great difficulty complying, Climategate revealed, because the software and data files were such a mess that they could not reconstruct the results they had published. he tale of woe begins with a guy copying 11,000 files and trying, unsuccessfully, to make something of them. He discovers, for example, that there are alternate files with the same name and no identification of which file is the one that should be used, or why. NASA has similar responsibilities for climate data in the United States, and a similar FIA request was filed for supporting climate data. After nearly three years, NASA has still not complied with the request, and a lawsuit is now threatened to attempt to force compliance. http://www.thenewamerican.com... I suspect that the problems of data compliance at CRU and NASA are due to professional incompetence, not a conspiracy to cover up errors they know to have been made. What has been revealed at CRU clearly shows incompetence. Moreover, there is nothing novel about incompetently written software. A product of human nature and schedule pressures is the method of hacking at software until it appears to work, then calling it done. In the commercial world, demands from users limit incompetence through calls for bug fixes, and ultimately user abandonment of one vendor in favor of another. Those mechanisms do not apply to climate data. In the case of climate research, the tendency will be to hack at the software until it meets the expectations of developer, in this case the global warming believers at CRU. They could be innocently making a dozen small errors that tend to inflate temperatures in recent times, and no one would question the results, because expectations are met. The remedy lies in immediate public disclosure. If the software must be posted regularly, which it will have to be because new results are released regularly, then peer pressure will greatly encourage sound software engineering practices like the use of software configuration control systems. Moreover, the details of the methodologies employed for processing and analysis will be subject to peer review. CRU deals mainly with data rather than climate models, however the resolution applies to climate modeling software as well. The basic physics of carbon dioxide only accounts for about a third of the global warming it is claimed to cause, and that's not enough to cause a climate crisis. The models contain multiplying factors that are not verified by experimental measurement. All of the mechanisms should be subject to peer review and public scrutiny. A few institutions have made their model code public, but only a very few. Aside from the concerns for good science and good professional practice, the public has a right to access what it paid for, for no reason beyond the fact that they paid for it. There are exemptions allowed in FIA legislation. The exemptions are for national security, independent proprietary data, and information sealed in lawsuits. None of the exemption apply to climate research. The requests to CRU and NASA were not denied under exemptions, they just not fulfilled. Requiring disclosure before publication or within a month after publication will guarantee that the public gets what it has a right to. Climate research strongly affects public policy, so while good professional practices are important in all areas, the situation addressed by the resolution is exceptionally important. The resolution is affirmed.