Presentation Information
[AAS08-P05]Assessing Measurement Consistency of the IMPROVE Elements in PM2.5 Over the Past Three Decades
*Nicole P Hyslop1, Jiayuan Wang1, Xiaolu Zhang1, Yongjing Zhao1 (1.University of California Davis)
Keywords:
atmospheric particulate matter,element measurements,x-ray fluorescence,measurement quality
The Interagency Monitoring of Protected Visual Environments (IMPROVE) network has monitored the chemical composition of fine particulate matter (PM2.5) in rural U.S. areas since 1988. Concentrations of 24 elements (Na, Mg, Al, Si, P, S, Cl, K, Ca, Ti, V, Cr, Mn, Fe, Ni, Cu, Zn, As, Se, Br, Rb, Sr, Zr, and Pb) have been measured since the network’s inception. Over the years, changes in analytical instruments, reporting conventions, and methods have introduced both obvious and subtle discontinuities in the data, which can confound the interpretation of long-term trends, particularly for elements with concentrations near detection limits.
To assess the long-term consistency of the elemental measurements, we analyzed data from 67 IMPROVE sites that have been operational since before 2000 and have greater than 75% data completeness. We first calculated the critical limit (Lc) values for 21 elements (Mg, Al, Si, P, K, Ca, Ti, V, Cr, Mn, Fe, Ni, Cu, Zn, As, Se, Br, Rb, Sr, Zr, and Pb) to harmonize the reported minimum detection limits (MDLs), which were calculated differently in different instrumental eras. We performed concurrent-measurement comparisons (elements ratio from X-ray analysis and elements–ions pairs from ion chromatography) for selected elements to evaluate measurement consistency. Based on detection rates and measurement consistency, elements were grouped into four categories reflecting their suitability for long-term trend analyses.
The empirically derived Lc values agreed closely with the theoretically estimated MDLs originally reported for most elements, supporting the general validity of the historical MDLs. Si, S, K, Ca, Fe, and Br were classified as “well detected”, with above 80% detection rates across sites and consistent time-series across the range of measured concentrations. Al, Ti, Mn, Zn, Sr, Pb, and Na were categorized as “robustly detected”, with detection rates of 50-80% and consistent trends at higher concentrations (above 50~75th percentile). Mg, P, Cl, V, Cr, Ni, Cu, Se, As, Rb were classified as “serviceable”, exhibiting inconsistencies across instrumental eras; for these elements, we recommend restricting analyses to specific instrumental eras and sites with detection rates exceeding 80%.
By harmonizing Lc values as an alternative to MDLs and systematically evaluating measurement consistency over three decades, this study provides practical guidelines for IMPROVE data users to minimize potential pitfalls and improve the reliability of long-term trend analyses.
To assess the long-term consistency of the elemental measurements, we analyzed data from 67 IMPROVE sites that have been operational since before 2000 and have greater than 75% data completeness. We first calculated the critical limit (Lc) values for 21 elements (Mg, Al, Si, P, K, Ca, Ti, V, Cr, Mn, Fe, Ni, Cu, Zn, As, Se, Br, Rb, Sr, Zr, and Pb) to harmonize the reported minimum detection limits (MDLs), which were calculated differently in different instrumental eras. We performed concurrent-measurement comparisons (elements ratio from X-ray analysis and elements–ions pairs from ion chromatography) for selected elements to evaluate measurement consistency. Based on detection rates and measurement consistency, elements were grouped into four categories reflecting their suitability for long-term trend analyses.
The empirically derived Lc values agreed closely with the theoretically estimated MDLs originally reported for most elements, supporting the general validity of the historical MDLs. Si, S, K, Ca, Fe, and Br were classified as “well detected”, with above 80% detection rates across sites and consistent time-series across the range of measured concentrations. Al, Ti, Mn, Zn, Sr, Pb, and Na were categorized as “robustly detected”, with detection rates of 50-80% and consistent trends at higher concentrations (above 50~75th percentile). Mg, P, Cl, V, Cr, Ni, Cu, Se, As, Rb were classified as “serviceable”, exhibiting inconsistencies across instrumental eras; for these elements, we recommend restricting analyses to specific instrumental eras and sites with detection rates exceeding 80%.
By harmonizing Lc values as an alternative to MDLs and systematically evaluating measurement consistency over three decades, this study provides practical guidelines for IMPROVE data users to minimize potential pitfalls and improve the reliability of long-term trend analyses.