[Prev][Next][Index][Thread]

Re: Human vs. natural influences on the environment



In article <4vn2e4$9h3@igc.apc.org>, tomgray 
<tomgray@igc.apc.org> wrote:
>brshears@whale.st.usm.edu (Harold Brashears) wrote:
>>kowens@teleport.com (Jeff Owens) wrote for all to see:
>>
>>[edited]
>>
>>>There are thousands of subjects we do not have full 
understanding
>>>of, but that should not stop people from trying.  
Knowledge is
>>>gained by suggesting a model and then trying to disprove 
it.
>>>The real issue here is: do we have enough information to 
suggest
>>>models.  The answer from the scientific community is a 
clear YES.
>>
>>While I think it is always worth it to build models and to 
test them
>>(otherwise, I would have been in some other business), I do 
not
>>believe that means we must act on the output of these 
models at this
>>time.  If you have some reason to believe that the models 
under
>>discussion are accurate, tell everyone about those reasons.
>
>Well, far be it from me to quarrel with anyone's religious 
beliefs.
>
>I don't know that a strong belief in models is required, 
though, to
>simply note that we are clearly changing atmospheric CO2 
concentrations
>and therefore the earth's energy balance.  


Tom,

This is where I continue to disagree with you.  We know that 
we are changing the amount of CO2 in the atmosphere.  
However, we don't know if it will keep increasing in 
concentration, and we don't know if there is a threshold 
concentration above which "it is bad".  You are making a BIG 
assumption that we are changing the energy balance of the 
earth at this point, because there is no clear documented set 
of measurements that indicate that this is so.

Why am I worried about making the measurement of increases in 
global temperature?  Because anyone can formulate a 
mathematical or computer model.  Often, these models are very 
large and complex.  And ALWAYS, the computer code that makes 
up these models has errors in it.  I have been programming 
computers for many years now, and I can guarantee that any 
code with more than a very few hundred lines in it cannot be 
absolutely debugged and verified.  Thus, without making 
physical measurements to validate the model, you are on very 
shaky ground.  This necessarily means that we must be able to 
measure a definite change in global temperatures before we 
can select the most appropriate (or least error free) model 
upon which we can base future work.  Since the changes we are 
trying to make take decades to happen, I suggest that we use 
a good model when we get to the point of trying to move 
atmospheric temperatures in a given direction.  This would 
give us the advantage of choosing the most effective method, 
being able to measure our progress, and having confidence 
that we are on the chosen path long before we can see the 
results.

By the way, if you are thinking about using the argument that 
we know CO2 is bad, so we should reduce it to reduce its 
impact on global temperatures, let's go into this a bit.  If 
the atmosphere is that simple to change, why are we working 
on all of these models?  One possible reason occurs in the 
world of process control (I have experience in this area).  
There are some process variables in refineries that move in a 
direction opposite to what is intended, but only for a short 
amount of time.  These variables then move as expected. This 
effect is termed "inverse response".  We don't know enough 
about the atmosphere yet to know if there are any variables 
of this type, but if there are, it would be very good to know 
about it before we try to change them.  Obviously, if you are 
trying to move a variable that has inverse response, and it 
takes decades to see the effect, you can't get the desired 
effect without a good model, because the short term inverse 
response will cause people to invariably move in the wrong 
direction.  Think about it.




Follow-Ups: References: