HR: 1330h
AN: A32B-0146 [PDF]
TI: Addressing Global Questions With \emph{in situ} Measurements: Defining Accuracy Using Both Advances in
Data Reduction Algorithms and Developments in Laser Systems
AU: * Engel, G S
EM: gsengel@huarp.harvard.edu
AF: Harvard University, 12 Oxford Street, Cambridge, MA 02138
AU: Anderson, J G
EM: anderson@huarp.harvard.edu
AF: Harvard University, 12 Oxford Street, Cambridge, MA 02138
AB:
Data reduction and data analysis algorithms can introduce statistically significant systematic bias and loss of precision in
results of both satellite and airborne \emph{in situ} measurement results.
Because data from many instruments must be used to create a global mapping, reducing these hidden systematic errors in
\emph{in situ} instrumentation is crucial to validating satellite data and to integrating \emph{in situ} results into global
climate models.
Biases in the \emph{in situ} measurements must be eliminated before the result can be considered accurate.
Additionally, inter-comparison among \emph{in situ} instrumentation requires careful review of all collection, reduction and
analysis algorithms to eliminate differences in temporal and spatial offsets as well as extrapolation to the appropriate
timescales to compare instruments.
Typically, the \emph{in situ} community does not archive raw data nor publish retrieval and reduction algorithms in such a
way that they can be verified and reviewed; however the global nature of current atmospheric questions requires this change.
In flight inter-comparisons between results obtained from related instruments are necessary but not sufficient to resolve
differences in measurements and in uncertainties; details of analysis techniques must also be compared to ensure the
agreement or disagreement between instruments is well-understood.
Simply observing agreement or disagreement is not sufficient.
Having documented, traceable paths to compare laboratory calibrations and analysis to flight data will lead to improvements
in instrumentation and retrieval algorithms, thereby improving the credibility of atmospheric data.
We will show raw data from Cavity-Enhanced Absorption Spectrometers using Integrated Cavity Output Spectroscopy (ICOS) and
Cavity Ringdown Spectroscopy (CRDS) and demonstrate statistically significant improvement in second generation fitting and
retrieval algorithms.
Improved lineshape models and singular value decomposition of the baseline have improved internal consistency and precision
of the ICOS data, and empirical weighting of the least squares fit has improved retrieval of decay constants from CRDS while
removing documented biases from the data.
Within this framework, we will analyze the uncertainty in the results from both a theoretical and instrumentation perspective
to show how each aspect of the instrument design and algorithm can affect the final result and the uncertainty in that
result.
Accordingly, Data Analysis and Reduction algorithms should be subject to the same peer review process and documentation
process as physical instrumentation.
DE: 1610 Atmosphere (0315, 0325)
DE: 1694 Instruments and techniques
SC: Atmospheric Sciences [A]
MN: 2003 Fall Meeting