Conclusions

The center evaluated several real-time optical particulate monitors during the August 2000 wildfires in Montana. The real-time instruments were collocated with an EPA federal reference method gravimetric sampler in both Missoula and Hamilton, MT, to determine the accuracy of the instruments. The instruments were also compared to an EPA equivalent method PM10 instrument (a Rupprecht & Patashnick TEOM). This work was part of an ongoing evaluation the center has been conducting since 1998 to survey commercial off-the-shelf instruments to monitor smoke particulate. Two other publications, Laboratory Evaluation of Two Optical Instruments for Real-Time Particulate Monitoring of Smoke (9925-2806-MTDC) and Evaluation of Optical Instruments for Real-Time Continuous Monitoring of Smoke Particulates (0025-2860-MTDC) provide details on real-time instrument evaluations in both laboratory and field situations.

The particulate concentrations measured in Missoula and Hamilton ranged from low levels (less than 10 µg/m³) to very high levels (higher than 500 µg/m³). All the real-time instruments were capable of estimating particulate levels in this range. Two types of real-time instruments (light scattering and light absorbing) were used.

The performance of the light-scattering instruments when compared to an EPA FRM sampler were:

Met One GT-640 (serial No. 5278)–Underestimated concentrations by 0.94 times (y-intercept = -6.15 µg/m³)
Met One GT-640 (serial No. 1466)–Overestimated concentrations by 1.67 times (y-intercept = -30.26 µg/m³)
MIE DataRam (serial No. 2327)–Overestimated concentrations by 2.34 times (y-intercept = -32.43 µg/m³)
MIE DataRam (serial No. 2484)–Overestimated concentrations by 2.61 times (y-intercept = -45.92 µg/m³)
Optec NGN-3 nephelometer–Overestimated concentrations by 1.43 times (y-intercept = -11.04 µg/m³)
Radiance Research nephelometer (serial No. 0102)–Overestimated concentrations by 1.52 times (y-intercept = -6.88 µg/m³)
Radiance Research nephelometer (serial No. 0101)–Overestimated concentrations by 1.43 times (y-intercept = -6.59 µg/m³)

The performance of the light-absorbing aethalometer was:

Andersen aethalometer (370-nm channel)–Underestimated concentrations by 0.051 times (y-intercept = 0.349 µg/m³)
Andersen aethalometer (880-nm channel)–Underestimated concentrations by 0.027 times (y-intercept = 0.111 µg/m³)

The correlation coefficients for all the regression results were excellent for all the instruments.

Several of the instruments were used both in Missoula and Hamilton. This allowed us to determine whether the instruments performed differently based on possibly different optical properties of the smoke particulates at the different locations. Several of the instruments had considerably more samples taken when deployed in Hamilton. However, the Optec NGN-3 nephelometer took the same number of samples in Missoula and Hamilton. Results from the Optec NGN-3 nephelometer when compared to the FRM results were:

Optec (Missoula)–Overestimated concentrations by 1.44 times (y-intercept = -11.90 µg/m³)
Optec (Hamilton)–Overestimated concentrations by 1.42 times (y-intercept = -9.67 µg/m³)

The Missoula real-time instruments (corrected based on the above results) were compared to the MT DEQ TEOM for 1-, 8-, and 24-h running averages. The TEOM (configured to estimate PM10) typically overestimated the corrected results from the real-time instruments (configured for PM2.5). Results of the TEOM compared to the corrected values of the MIE DataRam were:

1-h average–Overestimated concentrations by 1.06 times (y-intercept = 8.35)
8-h average–Overestimated concentrations by 1.13 times (y-intercept = 4.15)
24-h average–Overestimated concentrations by 1.23 times (y-intercept = -1.23)

Top

Top

Back | Next

Table of Contents

Cover Page

Visitor

since June 26, 2003