Facts must be available in the form of irrefutable data and observations. The Earth is no laboratory where experiments can be repeated at will under well controlled conditions; on the contrary, any observations is made only once and in association with a multitude of disturbances that may or not be correlated with the target parameters. Therefore, the ability to develop and verify scientific climate theories is limited.
Physical parameters describing the climate can be measured and monitored. They are similar to those used to characterize the current weather but their evolution must be observed over a much longer period of time, centuries and millennia. But the significance of changes is evaluated by statistical analysis of long time series.
Before satellites and remote sensing technologies, weather parameters were recorded locally with various exactness and precision. And there are only few locations on Earth where time series have been systematically recorded, most of them were located in industrialized countries of the Northern Hemisphere. This is why all data gathered prior to the 1960's must be taken with caution when compared, grouped, smoothed and averaged. Furthermore, no direct measurements are available before the 18th century. All knowledge about climate history (paleoclimatology) is derived from subsidiary observations (proxies) such as tree rings, trapped gas bubbles in glaciers samples, geological observations, etc.
Scientists have developed a huge volume of data and interpretations thereof. With such wealth of information one can ask if it is at all possible to achieve a complete overview without getting bogged into impossible details. Therefore, simplification is necessary; however, this should not lead to exaggerations.
Raw data is collected locally, by remote sensing, or in the form of proxies that need to be aggregated to render an honest physical representation over the whole globe. As a simple citizen with access to the Internet it would be beyond my area of competence to put in question the unavoidable data massaging. It is the role of scientists to do so in the most honest way as possible. But we need to remember that such data massaging takes place all the time, would it be just by performing the simplest calculation of averages among measurements made at different points of time, or across different locations. Different data quality will be obtained by using more or less adequate statistical distribution models. Also, we need to remember that scientists are normal human beings, with their biases, tribal dependencies, and – possibly – political agenda.
Independent parameters need to be identified that, together, characterize the climate; but it may take centuries or even millennia to ascertain their mutual dependencies.
