19th International CODATA Conference
Category: Knowledge Discovery
Key Aspects Of Data And Information For Climate
Change Science
Alexander Sterin, Ph.D. (STERIN@METEO.RU)
Russian Research Institute for Hydrometeorological
Information – World Data Center, Russia
The paper mentions several aspects related to methodology of the observed data
processing and analysis, when the final goal is to obtain new knowledge of climate
changes. This subject is of key interest, in connection with the problem of
global warming. It is not limited by purely research aspects, but is significant
for the decisions in social, economical spheres, as well as for politics.
The possible strategies of constructing
informational products based on observed data, will be considered.
The ways how to process and to analyze
data, will be discussed, in particular:
-
which statistics to use, traditional
or robust, for simple processing? The robust statistics is a good instrument
to eliminate the effects of outliers in climate-related statistical calculations.
On the other hand, one should be careful in using them, as soon as he studies
climatology of extremal events.
-
how to generalize data, to obtain
large-scale products for climate studies?
-
the data quality problems (flagging data values,
chance to make “step back” in data quality check)
-
how to calculate trends in climate series, (Least
Squares Method versus alternative techniques in trend calculations, trend
sensitivity problems and examples, the phenomenon of trend differences in
temperature for surface and in troposphere)
-
the inhomogeneities in climate data (are the “corrected”
series really correct, and “uncorrected” series really incorrect? – possible
approaches and some samples)
-
how to compare time series, obtained
by various groups, and is it possible to eliminate the uncertainty in large-scale
climate signals, using the ensembles of independent climate products? An
ensemble approach to large scale climate signals estimation.
The ten plus one principles of
climate monitoring, which must be taken as key strategy of empirical climate
analysis and requirements to climate data, will be listed and commented.