Common use of PRECISION Clause in Contracts

PRECISION. The movable measurement platform (satellite site shelter) was periodically deployed at the core site for collocated measurements with the results summarized in Table 3-2. These metrics are based on all data. Including only those concentration values exceeding ten times the MDL reported in Table 3-1 yields collocated precision of 1.0 μg/m3 (4.6%) for PM1.0 (N = 3) and 2.0 μg/m3 (9.5%) for PM2.5 (N = 10). All concentration values were greater than ten times the MDL for the collocated PM10 data set. In addition to the completely independent measurements, two PM2.5 samples were collected in parallel on most days. The last row of Table 3-2 shows the collated precision for those measurements which shared the same pump and timer but had independent flow control elements. This measurement captures a subset of the overall collocated variability. Bias and Comparability. One quality check for the PM gravimetric mass data is to test whether PM1.0 < PM2.5 < PM10 mass within the measurement uncertainty. Figure 3-1 shows scatter plots for PM1.0 and PM2.5 (Fig. 3-1a), and PM2.5 and PM10 mass (Fig. 3-1b). There were four cases where PM1.0 exceeded PM2.5 (by 0.5, 0.7, 0.8, and 1.0 μg/m3), and one case where PM2.5 exceeded PM10 (by 0.2 μg/m3). Assuming the 0.5 μg/m3 absolute precision estimated from PM2.5 samplers from the same sampling system (last row of Figure 3-2), these cases can be explained by measurement error. In addition to the above Supersite platform measurements, the Illinois EPA (IEPA) operated a full suite of NAAQS compliance monitors, including a PM2.5 FRM, at the 13th & Tudor (East St. Louis) at the monitoring site which shared the same physical footprint as the Supersite. The IEPA and Supersite PM2.5 gravimetric mass measurements were independently conducted at all levels (different field staff, audit devices, handling and storage facilities, and gravimetric mass analytical laboratories). IEPA samplers and continuous analyzers were included in the systems and performance audits conducted by DRI. Comparability between the IEPA FRM and Supersite Harvard Impactor (HI) gravimetric mass measurements are shown in Figure 3-2a for the period 4/14/01 through 3/31/2003 and excluding one value at (95.7, 88.8). A reduced major axis regression yielded: FRM = (0.92 ± 0.05)× HI + (−0.8 ± 0.7 μg / m3 ) The FRM and HI measurements are comparable with the HI biased high. This bias likely arises from differences in samplers, including but not limited to the impactor cutpoint curves (collection efficiency as a function of aerodynamic diameter). There are several outliers in Figure 3-2a which in most cases have been identified as anomalous values for the FRM measurement (based on comparisons to other onsite measurements including semicontinuous PM2.5 mass, and comparisons to St. Louis area network-wide FRM data). IEPA PM2.5 FRM data quality is demonstrated by comparison to additional FRM measurements conducted at the East St. Louis site by the ▇▇▇▇▇▇ group (WUSTL) in support of a field evaluation of the Thermo PM2.5 SHARP mass monitor (▇▇▇▇ et al. 2006). Figure 3-2b shows the collocated PM2.5 FRM gravimetric mass data for calendar year 2005. The collocated precision was 1.3 μg/m3. There are three outliers – 6/15/05 (IEPA=7, WUSTL =11), 10/1/05 (IEPA=23, WUSTL=14), 12/6/05 (IEPA=13, WUSTL=18). Based on comparisons to the onsite PM2.5 continuous mass monitors and other FRM measurements in the STL area, it appears the 6/15/05 outlier is WUSTL biased high, the 10/1/05 outlier is IEPA biased high, and the 12/6/05 outlier is IEPA biased low. After removing these three outliers the collocated precision for the CY2005 data is 0.8 μg/m3. This comparison is a robust determination of the measurement precision given that IEPA and WUSTL conducted completely independent measurements. Parameter Method N Mean ± Std Dev (μg) Nominal Blank Concentration(a) (μg/m3) MDL (μg/m3) PM1.0 mass cyclone 29 -5.3 ± 16.2 -0.2 2.0 PM2.5 mass HI 54 2.5 ± 7.0 +0.2 1.5 PM10 mass HI 30 3.4 ± 4.5 +0.2 0.9