Inaccuracies in U.S. temperature records

A comment by Fred Svetz

Over the past 4 years, the National Weather Service has
replaced human weather observers with an automated weather
observing system called ASOS (Automated Surface
Observation System).  This system regularly underreports
daily maximum temperatures by 1-4 degrees because of
poorly designed software.  How will your research results
be affected by this artificial cooling of U.S.
temperatures?

only time will tell . . .

A comment by Robert S. Thompson

I have heard anecdotes in the other direction as well,
that is:  the gradual nationwide replacement of
thermometers by thermistors may be artificially inflating
temperature records.  On-going analyses of these data
will (I hope) identify the true temperature signal
(separate from instrumental and software considerations).

Warning about reliability of Global models

A comment by Rob Bracken

In understanding Global and even local "climate change", we are
confronted with serious difficulties of spatial aliasing and lack
of baseline data.  Systems which operate on large spatial and
temporal wavelengths are difficult if not impossible to quantify.
The difficulty is compounded when the system is soft, having many
poorly constrained variables and an incomplete understanding of
the system responses; as is the case with global climate change
models.  A model claims to contain enough knowledge of a system
that when one or more of the input parameters are changed, the
model responds in parallel to the true system.  A predictive
model must also know how the input parameters are changing.

Therefore, an accurate global change model must claim to have
accounted for all dynamics of the world climatic systems either
explicitly or implicitly.  An explicit accounting requires
current and future knowledge of all pertinent inputs and
responses of the system at all locations, i.e. the spatial
wavelengths extant in the Earth's system cannot be characterized
without a densely spaced global network of sensors comprehensive
of all appropriate variables and a complete understanding of the
associated effects.  An implicit accounting requires historical
knowledge of all pertinent parameters over periods of time long
enough to observe accurately many complete cycles of all
processes within the system; i.e. geologically dependent
variables having temporal wavelengths of thousands or millions of
years cannot be measured.

Because the prerequisites to a reliable predictive model cannot
be met, we must begin interpolating across gaps in data and
understanding.  This process involves introduction of various
tweaking factors which preclude the model from being
dispassionately tied to the data and physical law.  A model such
as this sometimes can be useful to bootstrap our understanding of
a system or to prove (after many decades of confirming evidence)
that we do, in fact, understand the system; but, to produce a
prediction upon which the re-organization of governments may
depend (vis-a-vis UNEP & IPCC) is altogether beyond the scope of
any global model that can be produced.  The USGS has, in the name
of true science, a moral obligation to state explicitly that we
do not yet know what the global climate trends are nor the
overall effect of mankind.  It would be absurd to begin
mitigating something that may not be a problem.

However, as the scale of the system decreases, it becomes easier
to observe trends and understand their causes.  Therefore, it
is valuable to discuss the climatic variations in a small area
such as the southwestern U.S.

U.S. Department of the Interior | U.S. Geological Survey
URL: http://geochange.er.usgs.gov/sw/changes/anthropogenic/ipcc-b/discuss.html
Page Contact Information: ESD Web Team
Page Last Modified: Tue 25-Nov-2003 15:27:02 MST
Accessibility, Privacy, and other policies and notices