Download presentation
Presentation is loading. Please wait.
1
R Package for Smoke Monitoring Data
Jonathan Callahan - Mazama Science Pacific Wildland Fire Sciences Lab (PWFSL)
2
Motivation Air Resource Advisors need up-to-date, QC’ed data and analyses. Real-time PM2.5 monitoring data from: AirNow -- permanent AIRSIS -- temporary (Iridium) WRCC -- temporary (GOES) Uniform processing and display: Process and QC raw data from AIRSIS, WRCC Handle UTC/local timezones Provide a selection of plots to better understand the data
3
PWFSLSmoke R package R is a free, open source statistical programming language. PWFSLSmoke is a free, open source package for R. PWFSLSmoke makes it much easier to: Process raw data Work with aggregated datasets Apply algorithms Generate maps and plots Write scripts to automate data analysis
4
Demonstration of PWFSLSmoke Package
2017 Wildfire Smoke Demonstration of PWFSLSmoke Package
5
Smokiest Spots
6
Daily Mean PM2.5
7
Nez Perce Reservation
8
Nez Perce Reservation
9
Unhealthy All Day (average of 10 monitors)
10
Monitors on the Reservation
11
Nez Perce Reservation
12
Monitors near the Reservation
13
Hourly PM2.5 and Nowcast
14
“Hourly AQI” Levels
15
PWFSLSmoke R package -- Features
Ingest, parse, QC and reformat data from: AirNow, AIRSIS, WRCC Apply algorithms to PM2.5 timeseries: rolling means, daily statistics, Nowcast, AQI Handle time: UTC, local time, daylight savings, sunrise-sunset Combine/Subset monitors by: monitorID, state, time, values, anything in the metadata
16
PWFSLSmoke R package -- Plotting
Static plots: Timeseries, maps, daily/hourly barplots, diurnal cycle plots Interactive plots: Timeseries, maps Work with “engineering data” from AIRSIS/WRCC: QC plots, pollution rose plots
17
PWFSL Monitoring Database (for 2018)
Pregenerated files for easy access within the package. Realtime “latest data” files for: Airnow, AIRSIS, WRCC Archival files for: EPA (monthly, ) AirNow (monthly, 2016-present) AIRSIS (annual, 2010-present) WRCC (annual, 2010-present)
18
Potential Extensions -- New Sources of Data
New sources of PM2.5 data can be ingested by creating custom functions for the data processing steps: Ingest Parse Apply Quailty Control Additional hourly parameters (Ozone, NOX, etc.) could be supported in the same manner.
19
Potential Extensions -- Model-Data Comparison
PWFSL has an in-house R package to work with BlueSky output. Features include: Data ingest and conversion Aggregation of model runs into longer timeseries Timeseries algorithms Subsetting by area mask (polygon, distance from point, etc.) Static plots and maps Conversion of grid cell timeseries into “fake monitor” timeseries Model-data comparison with ROC analysis and metrics. Additional models could be supported with custom ingest/conversion functions.
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.