Reanalysis of a 15-Year Archive of IMPROVE Samples

摘要

In summary, advances in measurement method may create discontinuities in data quality (e.g., less noise or smaller bias) that impact temporal trends analyses. Measurement performance can also vary over time as a result of uncontrollable factors or slight adjustments to the instrumentation or measurement technique: changes in the commercial availability of filters and other consumables, retirement or reassignment of operators, degradation and subsequent replacement of instrument components, changes in mathematical algorithms, and replacement of calibration standards. Shifts are not always obvious in the measured values themselves, particularly when the measurements are of inherently fluctuating quantities such as ambient PM concentrations. Our reanalysis using a consistent protocol revealed method-linked shifts in concentrations over time for several elements measured in IMPROVE, many of them outside the reported uncertainties in the concentrations.