|
Post by eamonnshute on Jun 23, 2011 22:07:25 GMT 1
The Earth oceans have no fixed level to be characterised by a +/- 0.3 mm tolerance..... You can take the centre of the Earth as your datum - that is a fixed level! Or you can use the local ground as your reference - if the land rises or falls it can be measured very accurately with GPS.
|
|
|
Post by StuartG on Jun 24, 2011 2:03:32 GMT 1
Doesn't appear so, Eamonn... "We now know that the centre of the Earth lies about 6370 km from the surface. In fact, because the Earth is shaped like a flat ellipsoid – not a perfect sphere – the radius at the Equator (6378 km) is slightly longer than the radius at the poles (6357 km). However, the Earth is constantly changing shape as tectonic plates move, ice sheets melt and volcanoes erupt. Scientists would therefore like to determine accurate values for the position and velocity of the centre of the Earth, which provide the frame of reference for measuring how fast locations on the Earth’s surface are moving" environmentalresearchweb.org/cws/article/opinion/32445The Earth is not constructed that accurately. Its intrinsic accuracy of manufacture is not that good. "land rises or falls it can be measured very accurately with GPS" perhaps for an instant in time but it's own errors creep in to the calculation. "Error analysis for the Global Positioning System is an important aspect for determining what errors and their magnitude are to be expected. GPS errors are affected by geometric dilution of precision and depend on signal arrival time errors, numerical errors, atmospherics effects, ephemeris errors, multipath errors and other effects." en.wikipedia.org/wiki/Global_Positioning_System#Error_sources_and_analysis Do You remember my post about PI? radio4scienceboards.proboards.com/index.cgi?action=gotopost&board=debate&thread=915&post=11365I was looking for an accuracy that intrinsically isn't there. The accuracy is knowing that the accuracy isn't and can never be, that precise. I now know that precisely. StuartG
|
|
|
Post by marchesarosa on Jun 24, 2011 9:39:22 GMT 1
Not as easy as you may think, Eamonn.
The only reasonable to way to measure sea-level rise is by tide gauges - these automatically include the element of isostatic rebound, coastline subsidence etc.
Don't place your faith in satellites but in something a bit more tangible and with a relatively long record is my advice.
|
|
|
Post by StuartG on Jun 24, 2011 10:22:11 GMT 1
Good point MM... "If you were to stand on the ocean shore and try to measure sea level with a ruler, you would find it to be impossible -- the level changes by the second (waves), by the hour (tides) and by the week (planetary and solar orbit changes). To get around this, scientists try using tide gauges. A tide gauge is a large (1 foot [30 cm] or more in diameter), long pipe with a small hole below the water line. This pipe is often called a stilling well. Even though waves are changing the water level outside the gauge constantly, they have little effect inside the gauge. The sea level can be read relatively accurately inside this pipe. If read on a regular basis over a time span of years and then averaged, you can get a measurement of sea level. You can see that getting an accurate reading (for example, down to the millimeter level) is extremely difficult. " science.howstuffworks.com/environmental/earth/oceanography/question356.htmSeems a nice idea with a simple approach, effective. USA could do it around Her coastline, GB and Ireland, and EU theirs. Falklands, Australia & NZ perhaps Japan would join in. Agree on a standard design of gauge. Shouldn't cause political problems, if they are not sited in areas of military sites [say] StuartG
|
|
|
Post by eamonnshute on Jun 24, 2011 10:24:17 GMT 1
The only reasonable to way to measure sea-level rise is by tide gauges - these automatically include the element of isostatic rebound, coastline subsidence etc. Don't place your faith in satellites but in something a bit more tangible and with a relatively long record is my advice. Nonsense. according to Wiki Amsterdam has tide records dating back to 1700. How did these correct for isostatic rebound, which was unknown at the time? What is wrong with satellite measurements? If a satellite says that the position of a point at time T is X +/- Y then why should I not trust it?
|
|
|
Post by eamonnshute on Jun 24, 2011 10:38:17 GMT 1
the Earth is constantly changing shape as tectonic plates move, ice sheets melt and volcanoes erupt. Everyone knows that, especially the people who study sea-level changes. They know about the various uncertainties and can calculate their magnitude. Are you trying to say that you have spotted a problem that they have overlooked?
|
|
|
Post by StuartG on Jun 24, 2011 10:41:35 GMT 1
|
|
|
Post by StuartG on Jun 24, 2011 10:50:36 GMT 1
# 6 1950–2000 * 6.1 1952: Severo-Kurilsk, Kuril Islands, USSR * 6.2 1958: Lituya Bay, Alaska, USA * 6.3 1960: Valdivia, Chile * 6.4 1963: Vajont Dam, Monte Toc, Italy * 6.5 1964: Niigata, Japan * 6.6 1964: Alaska, USA * 6.7 1976: Moro Gulf, Mindanao, Philippines * 6.8 1979: Tumaco, Colombia * 6.9 1980: Spirit Lake, Washington, USA * 6.10 1983: Sea of Japan * 6.11 1993: Okushiri, Hokkaido, Japan * 6.12 1998: Papua New Guinea # 7 2000s * 7.1 2004: Indian Ocean * 7.2 2006: South of Java Island * 7.3 2006: Kuril Islands * 7.4 2007: Solomon Islands * 7.5 2007: Niigata, Japan * 7.6 2010: Chile * 7.7 2011: New Zealand * 7.8 2011: Pacific coast of Japan en.wikipedia.org/wiki/Historic_tsunamisand that's just some of the big ones, how many others have there been? How accurate is an average tide height going to be? StuartG ps. Don't forget Japan and NZ are still experiencing after shocks, now, from their recent large quakes.
|
|
|
Post by StuartG on Jun 24, 2011 11:58:42 GMT 1
Dear Eamonn, please stop blustering, You are obviously an intelligent man, but please, please do read at least some of the posts. [including the annotations of web sites] YOU quoted me as saying.. "Today at 1:03, StuartG wrote: the Earth is constantly changing shape as tectonic plates move, ice sheets melt and volcanoes erupt." TOTALLY INCORRECT, I didn't say that, this chap did... "About the author Donald Argus is a research scientist at NASA’s Jet Propulsion Laboratory in Pasadena, California, US. This article first appeared in Physics World."
They may take ALL known variables into account for all I know, [they don't but still, they try] but they cannot measure their effect accurately [certainly not down to the millimeter level] StuartG ps. To give a reading accurate to the millimeter level [roughly 40 thou] it is necessary to be able to measure accurately to at least a tenth of that, 0.1 mm [roughly 4 thou]. At the present time it appears that measurements at the millimeter level are Assessments.
|
|
|
Post by speakertoanimals on Jun 24, 2011 12:04:11 GMT 1
We seem to have the SAME misunderstanding repeated, that we have to be able to measure the sea-level to an accuracy of 0.3mm to get an estimate of the yearly change to within 0.3mm.
We don't. It's basic stats. SO, we have very noisy individual measurements, even when a tide-gauge has eliminated wave action. We have the bloody obvious fact that sea-level changes with the tides, whcih vary during the month. We then have additional changes due to weather systems etc.
Basic stats: we have measured values distributed independantly and identically about the mean we are trying to find. Obvious that we should take the mean over the sample as an estimate of the true mean. But there is a statistical contribution to our error on the estimate of the mean, the famous 1/sqrt{N} factor, which means that the more samples we have, the better our estimate. But we also have the fact that our original measurements have a limited precision in the first place. But we're not doomed, because although this gives an error on each individual measurement, these errors are uncorrelated between different measurements.
Basic rule -- the bigger the sample, the better your estimate of the population mean using the sample mean.
The average sea-level change is different again, in that we have a set of measurements over time, with associated errors, and we are trying to fit a straight line to this to find the slope. Hence, depending on the size of the error bars on the individual measurements, and the length of time over which we have measurements (and whether or not the real plot IS actually linear in the first place), all combines to give an error on the slope of the best linear fit.
Don't be misled by a quoted slope with + or - 0.3mm per year, the 0.3 referers to an error in slope, NOT error on a single measurement!
You can easily imagine different scenarios -- very small measurement error on individual points, but not a very good fit to a straight line, so we'd expect a large error in the slope. Or we could have a truly linear theoretical line, but large measurement errors on individual points. Actually, its a problem that foxes most students when they first come across it -- they know how to do error propogation with simple algebraic relations between measured quantities and the answer they're trying to find, but estimating the error on a slope is a harder problem.
|
|
|
Post by eamonnshute on Jun 24, 2011 12:17:11 GMT 1
Dear Eamonn, please stop blustering, You are obviously an intelligent man, but please, please do read at least some of the posts. [including the annotations of web sites] ps. To give a reading accurate to the millimeter level [roughly 40 thou] it is necessary to be able to measure accurately to at least a tenth of that, 0.1 mm [roughly 4 thou]. I suggest you practice what you preach! The article about 55 Cancri and its planet shows variations in individual measurements of several parts per thousand, but by averaging we can measure the size of the dips, which are much smaller, at about 2 parts in ten thousand. There is some uncertainty remaining, but it is much less than the errors in individual measurements! (This is similar to what STA says at 12:04).
|
|
|
Post by speakertoanimals on Jun 24, 2011 12:20:55 GMT 1
No they DON'T, because what isn used in tidal gauges is measurement relative to some nearby land-based benchmark. Hence if whole area rebounding, this won't detect it. So what do you need? GPS and satellites perhaps.................... sealevel.colorado.edu/content/tide-gauge-sea-level No you don't because you seem to have no idea about stats and standard error propogation..............Or linear fits for that matter! That is why I made the remark about assuming scientists were stupid, because all this sort of analysis is BASIC to whatever science course you did where you first started to measure anything. Assuming that oceanographers and climate scientists have forgotten about all this when to comes to going out to read tidal gauges IS assumning they are incredibly stupid!
|
|
|
Post by speakertoanimals on Jun 24, 2011 12:46:46 GMT 1
you are the only one now who keeps insisting that to get a slope accurate to plus or minus 0.3mm/year, you need to be able to make an INDIVIDUAL measurement to a similar accuracy. WRONG.
Plus IF you could measure accurately to a tenth of a millimeter, then THAT is the error on an individual measurement, NOT at millimeter level!
Measurements at the miklimeter level are assessments -- nope, because the measure a CHANGE in sea level, you need at least two measurements. It's measuring the SLOPE, which you can't measure directly, only by combining individual level measurements. Hence it as much an assessment as any other quantity which isn't measured totally directly. but calling it an assessment is DAFT and misleading, since the colloquial usage of assessment implies an inherently uncertain process, whereas the point about errors and statistical analysis is that we have 'rigorously defined areas of doubt and uncertainty' (to quote Douglas Adams).
Frankly Stuart, you seem to have no idea what you're talking about on this one, as your original water-barrel post indicates.................
|
|
|
Post by StuartG on Jun 24, 2011 12:49:33 GMT 1
Yes StA, You've made the point before, I accept that the error is on the mathematical calculations of the data that is included. However, the practical measurement of tide height is a difficult, not as pointed out by people like me, but by real experts. My point to You is read the contents of the posts, especially the site annotations, here the experts state what sorts of accuracy they obtain, they acknowledge its difficulty, and use the word 'Assessment' against the final figure. So to recap we were talking about the IPCC figure and the way it is used to promote their ideas. The error on a single measurement is somewhat larger. In this case the Sea is part of the noise, we are trying to find the level of that noise over all the Oceans and Seas. See the tsunami pdf. That's just one variable. The Physworld article that I quoted to Eamonn environmentalresearchweb.org/cws/article/opinion/32445 by Donald Argus, here He goes through the various ways of obtaining an accurate position of the Earth's centre, in order to measure land surface to within 0.5 mm... He states... "This is a pretty big discrepancy given that geophysicists want to know how fast places like New Orleans are sinking to within 0.5 mm per year" ----- The error refers to 0.3 units of error for those mathematical calculations, so fine the maths is OK then, the problem is the practical measurement error of the individual measurements that go to make that plot. StuartG StuartG
|
|
|
Post by speakertoanimals on Jun 24, 2011 13:02:17 GMT 1
The point being? DO you now believe they can actually measure the rate as accurately as they claim, since that was your initial doubt?
What's the problem? As in any measurement, those doing it would LIKE to have better measurements, but given the inherent noise of sea-level, the ONLY way to get estimates of the rate of change are more tide-gauges, sea-bound buoys, satellite data etc. Improve the stats if you can't get rid of the inherent noise.
Is this relevant? What you need for sea-level is NOT the centre, but the surfaces that are at equal heights from a gravitational point of view (some fancy name for that!), because that is where the sea level would lie, on one of those IF there was nothing to disturb the water.
And since gravity varies slightly over the earth, as well as the land banging up and down...............................
The land movement 'problem' is because that is what old-style tidal gauges use.
|
|