An inwards characterisation large to small. Standard stuff.
If you want to supply literature refs to support or deny please use the numbers.
- A large, evaporatively cooled, mostly water faced object, with an apparent water/ice modulation controlling system, heat engine.
- A close companion, The Moon, that is slowly working its way up against gravity by tidal friction, away from its neighbour, Earth. Anti-gravity at work in nature. Gravity always attracts, right?
- Two ‘sets’ of temperature instrumentation available.
- One a set of single point sampling instrument inside the quasi-turbulent friction boundary with surface impacts layer at 2m off the ‘ground’.
- The other an overhead, volume integrating, slow (~100 minutes) global scanning, view with larger coverage.
- Two different viewpoints, two (at least) differing data figures series.
- Above the boundary layer the weather systems are much more laminar with long distance features that do not require as high sampling rate as the layer below.
- High quality, long term, methodology allows the lower instrument sets to estimate the higher laminar layers well thus increasing accuracy of local predictions.
- Unknown wide area boundary layer volatility characteristics means we cannot be sure of any decadal or longer changes in the past that effect calibration or data now. The global data is not there to verify wider usage of any previous long or short term calibration data.
- UHI is easy to see as a slow long term rise within a geographic volume as being outside the lower capture resolution of local area instruments. We cannot reliably connect offsets with those instruments from the outside. Ideally they should be in a different data set for ‘a’ to ‘a’ comparisons.
- Each individual point instrument is probably well calibrated (with some exceptions).
- Each site is mostly acceptable though quality values vary widely.
- Relative1 can only be directly compared to its absolute1. Then absolute1 to absolute2 and thence back up to relative2.
- Even merging two sites requires not only any ‘DC’ offsets but also any in reference period differences to be accounted for. Too low a length of reference period to be pure. Too few samples.
- The satellites algorithms are agreeing with each other quite well. Good validation for both.
- The point instrument sets are likewise quite close.
- Unfortunately 15 and 16 are still diverging. This has to be addressed sometime. Each are measuring the same thing, just one upwards, one downwards.
- A cross calibration exercise is likely to increase the accuracy of both.
- An examination of the above error sets should lead to more accurate back casting from current data.
- End logical executive summary. Details to follow. Quality control methodology is low to poor at present, especially for long period back-casting. Present data is high quality but more sub-sampling needed.