NavList:

   

Reply

Herbert,

The question is whether there is any significant variation left after
accounting for the linear trend.  I'm not convinced there isn't.

Additionally, the real question is whether there's been an increase
in the last five years.  With 150 datapoints on one side of any
inflection and only five or ten on the other, it would be difficult
to find the signal.

Fred

On Jul 8, 2006, at 2:04 PM, Herbert Prinz wrote:

>
> Fred,
>
> Just look at some of these graphs, whether from tidal data or from
> satellites. They can be fitted much better with straight lines than
> with
> curved ones.
> While it is true that you can fit any curve with a straight line, you
> can fit it *better* with a curve only if it's really a curve.
>
> Herbert
>
>
> Fred Hebard wrote:
>
>> On Jul 8, 2006, at 11:31 AM, Herbert Prinz wrote:
>>
>>
>>
>>> Yes, and that's also confirmed world wide by satellite data, even
>>> for
>>> places in the middle of the ocean. But the global trend has been a
>>> linear increase of 2-3 mm / year.
>>>
>>>
>>
>>
>> Just to clarify the statistics of linear trends.  One can find a
>> linear trend whenever a response variable is increasing or
>> decreasing, even if there is statistically significant curvature in
>> the line.  For instance, a parabola starting from a minimum or
>> maximum will show a significant linear trend.  It additionally will
>> show a significant quadratic trend, indicating that the line is
>> curved as well as rising or falling.  However, if you look at points
>> symmetrically distributed around the inflection point of a parabola
>> (the minimum or maximum), then there will be no significant linear
>> trend, but the significant quadratic trend will still be there.
>>
>> What this boils down to is that the linear trend of increasing ocean
>> levels over the past century and a half does not indicate that the
>> trend is ONLY linear.  Those data may not be at variance with the
>> evidence for global warming.
>>
>>
>
> >


--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to NavList@fer3.com
To , send email to NavList-@fer3.com
-~----------~----~----~----~------~----~------~--~---

   

Reply