First paper:
1. Biased sampling. Satellites have this problem because they do not simultaneously sample the oceans. The samples are sequential.
2. From the pdf:
Changes in instrumentation and station location introduce biases, especially in systems not specifically designed for long term climate modeling.
Well I agree with that. That would make the whole temp record (land and sea) before the advent of satellite monitoring suspect. Given that there is only 30 years of satellite data no firm conclusions can be obtained about climate changes since only 1/2 of the PDO cycle has been monitored and it may very well be that the PDO/ENSO has been aliased into the CO2 "effect".
I am unaware of any effort to make such a correction. Since d/dt effects may affect the numbers, rising edges should be compared to rising edges etc. Or else peak to peak. The problem of noise is exacerbated in such circumstances however. Making two or three cycles a better gauge. In another 100 or 200 years we may know something.
====
Another point is the claimed accuracy of 1.3 mm in 1,330 km seems beyond reason. That is 1 part in 1E9. Including clock variation, atmospheric variation, orbital variation, etc. Not to mention variation in reflectivity which will affect lock. Since an FM system will lock on the stronger of two signals with as little as 1 db difference between them the results could be rather twitchy unless the lock was looser (adding to the error band). Or it could be very sea state sensitive (adding to the error band).
http://en.wikipedia.org/wiki/TOPEX/Poseidon
The claim from the Wiki is that the satellite position can be measured to within 2 cm. That puts a lower bound on the error. To claim sea surface height measurements over a track that varies in time and space to 1.3 mm is frankly not credible.
===Now they go on further to state that in "average sea state" (could they provide a number? Sea states have numbers.) The accuracy is 4.2 cm to 1 sigma. Statistically a significant variation would need to be 2 sigma. That would be 8.4 cm. They claim to be able to reduce the error to 2.5 cm with ground processing (5 cm at 2 sigma) - note wiki says 3.3 cm. This does not account for wind. Variations in ocean currents. Track variations. Decadal variations (such as PDO/ENSO) And the fact that the measurement although area averaged over a 300 mi diameter circle do not measure a number of spots on the ocean simultaneously. etc. Also note that the reported rise is 3.0 mm a year which is just at the edge of the 2 sigma (significant) claimed error band. Which means the rate could be double what is reported or zero.
Which has been claimed in a sceptic paper I consider credible. I believe I left a link to it some where above. That would be the Swedish Prof. whose name begins with an M.
Also note that they go to the tidal gauge network for drift calibration of the atmospheric compensator. The drift is claimed to be 1.2 mm a year for the 5 years before 1997. That would be .6 cm. That is statistically undetectable. They also do not give the direction of the drift so it is possible to evaluate the direction of possible bias if their drift rate is in error. Then they state that after 1997 the drift rate jumps to 5.2 mm a year. That is detectable over a 10 year period. But again they do not give the direction of the correction in order to evaluate what it is doing to sea levels if the correction is in error. Shoddy.
Now Tom. You could probably do as well or better if you put your QC inspector hat on. Give it the design review treatment for medical, nuclear, or aircraft eqpt. Rip it apart.
I can tell you this. That paper in so far as I have read it would Never, ever, pass a FAA/Medical eqpt. review.
I don't have time right now to go further, but so far I am not impressed.
I'll get to the second paper later.
Glaring piece of missing data in the first page of the second paper. The tidal gauges show 1.8 mm a year over a century. And the satellite shows 3.3 mm during its time in orbit - what were the tidal gauges showing during the satellite period in orbit and if there is a difference what is the cause? The paper needs two charts. They have only one. The critique the paper I posted above made.
Can you read anything about this stuff with a sceptic eye Tom?
If the tidal gauges show the same rate of change as before (accounting for seasonal and decadal variations) and the satellite record is more accurate then we can reasonably assume that the tidal gauges for the past century need a correction by 2X (roughly). The splicing of two different measurement methods is in fact an error. As the paper I posted suggests.
I'm not interested in carrying this further without better data.
BTW the satellite data shows no correction for PDO/ENSO. i.e. decadal variations.
Another bit of crockery to throw at the masses. I am not saying that the numbers in the two papers are wrong. I'm saying: case not proved.
Color me unimpressed.