Data Notification
Protocol and Data Processing Updates for Dust and Particulate Size Distribution (DP1.00017.001)
December 19, 2024
The Dust and Particulate Size Distribution data product (DP1.00017.001) has suffered from low validity since installation. Multiple changes were made to field protocols and data processing to address these issues. Staff guidance was improved for troubleshooting and maintaining the sensor, and software designs were addressed to improve connectivity and limit flow rate problems with external pumps. These changes were implemented throughout 2024 and should improve data quality going forward.
In processing, three quality tests were adjusted to reduce over-flagging.
- Range test: The maximum and minimum allowable range values were adjusted downward. They were originally set to the specifications of the sensor and far above typical observations. The maximum dust concentration for all size classes was adjusted down to 5 mg/m3 to more reasonably exclude erroneous values while still allowing for extreme events. Future updates will incorporate more flexibility to “soft flag” extreme events without removing the data. Concentrations of 0 mg/m3 and slightly below are no longer flagged. Although zero dust concentration is not plausible in atmospheric air, literature suggests allowing negative values to reflect uncertainty (Jiang 2010). The DustTrak sensor is known to sometimes get stuck at 0 mg/m3 (Rivas 2017), which tends to occur when the pumps cannot create adequate flow. These occurrences will no longer be flagged by the range test but should still be flagged by the persistence quality test.
- Spike test: The spike test was removed from processing. At NEON’s sampling frequency of 1Hz and typically very stable particulate concentrations, the spike test often flagged valid momentary increases in concentration.
- Measurement Validity test: Part of the Measurement Validity test was removed from processing. When any of the particulate concentrations read 0 mg/m3 (PM1, PM2.5, PM4, PM10, PM15), all concentration values were flagged. Given that the persistence test detects a stuck sensor, this test was deemed unnecessary and overly restrictive. The second part of the Measurement Validity that requires large dust size concentrations to be higher than small size concentrations was retained.
All provisional data have been reprocessed to incorporate the changes to quality tests and are available now. The full data record will be updated in RELEASE-2025. The removal of the spike test and associated quality metrics from the expanded download package will cause a brief discontinuity in data file columns between provisional data and released data (RELEASE-2024 and earlier). When RELEASE-2025 is issued, data file columns will be consistent for the full data record.
Please use the Contact Us form for further inquiries; please select DP1.00017.001 in the Data Product dropdown list
References
Jiang, N., Akter, R., Ross, G., White, S., Kirkwood, J., Gunashanhar, G., Thompson, S., Riley, M., & Azzi, M. (2010). On thresholds for controlling negative particle (PM2.5) readings in air quality reporting. Environmental Monitoring and Assessment, 166(1-4), 803-812.
Rivas, I., Mazaheri, M., Viana, M., Moreno, T., Clifford, S., He, C., ... & Querol, X. (2017). Identification of technical problems affecting performance of DustTrak DRX aerosol monitors. Science of the Total Environment, 584, 849-855.