[daip] Re: Question about DELZN

[Leonia Kogan]

Richard Dodson wrote:

>Hi
>  
>
>>>If I understand what is plotted (with aparm(1)=0) it is the the
>>>VAL+ATM. Should it not be VAL _and_ ATM? Then I should see the line go
>>>through the points. Plotting VAL+ATM means that VAL is _always_
>>>scattered around the model. I don't understand!!
>>>      
>>>
>>If aparm(1)=0 DELZN plots VAL + ATM(RANT) which should follow polynomial presentation of the
>>atmosphere over IANT. ATM(RANT) is the found polynomial for atmosphere over RANT.
>>    
>>
> <>
> Of course, I agree on this. But so would plotting ATM only. So why
> plot VAL+ATM?



I repeat again: VAL as it derermined in SN table = ATM(IANT) - 
ATM(RANT); consider here that there is no clock error;

DELZN fits the polinomials to both ATM(IANT) and ATM(RANT)

So the polynomial ATM(IANT) fits VAL + polinomyal ATM(RANT)

THis means in particular that the identical error in the polynomial 
ATM(IANT) and ATM(RANT) will not change the fitting.

> <>Should not the ATM line go through the VAL points?
>
> Do you agree?


NO I DONOT AGREE    LOOK MY PREVIOUS COMMENT. VAL as read from the SN 
table is not
the ATM(IANT.) instead it is difference of atmosphere at IANT and RANT.


>>When the measurement are not completely independent (the elevation range is not big enouth,
>>the array is not big enough ...) the solution for ATM(IANT) and ATM(RANT) can be distributed differentely.
>>
>>    
>>
>
>Agreed, because you only measure differences. But, on that issue, does
>that not mean that it should be possible to get solutions even in this
>case? The solution will be degenerate, but
>the differences (all that we really need) should always be the same.
>
>Agreed?
>
>I wonder did you attach the plot I asked for? It did not come through.
>
> thanks,
>      Richard
>  
>


LK.