Calculate Rolling Means

Source: R/monitor_rollingMean.R

Calculates rolling means for each monitor in ws_monitor using the openair::rollingMean() function

monitor_rollingMean(ws_monitor, width = 8, data.thresh = 75, align = "center")

Arguments

ws_monitor

ws_monitor object
width

number of periods to average (e.g. for hourly data, width = 24 calculates 24-hour rolling means)
data.thresh

minimum number of valid observations required as a percent of width; NA is returned if insufficient valid data to calculate mean
align

alignment of averaging window relative to point being calculated; one of "left|center|right"

Value

A ws_monitor object with data thaty have been processed by a rolling mean algorithm.

Details

  • align = 'left': Forward roll, using hour of interest and the (width-1) subsequent hours (e.g. 3-hr left-aligned roll for Hr 5 will consist of average of Hrs 5, 6 and 7)

  • align = 'right': Backwards roll, using hour of interest and the (width-1) prior hours (e.g. 3-hr right-aligned roll for Hr 5 will consist of average of Hrs 3, 4 and 5)

  • align = 'center' for odd width: Average of hour of interest and (width-1)/2 on either side (e.g. 3-hr center-aligned roll for Hr 5 will consist of average of Hrs 4, 5 and 6)

  • align = 'center' for even width: Average of hour of interest and (width/2)-1 hours prior and width/2 hours after (e.g. 4-hr center-aligned roll for Hr 5 will consist of average of Hrs 4, 5, 6 and 7)

Examples

library(PWFSLSmoke)

N_M <- Northwest_Megafires
wa_smoky <- monitor_subset(N_M, stateCodes='WA', tlim=c(20150801, 20150808), vlim=c(100,Inf))
wa_smoky_3hr <- monitor_rollingMean(wa_smoky, width=3, align="center")
wa_smoky_24hr <- monitor_rollingMean(wa_smoky, width=24, align="right")

monitor_timeseriesPlot(wa_smoky, type='l', shadedNight=TRUE)

monitor_timeseriesPlot(wa_smoky_3hr, type='l', col='red', add=TRUE)

monitor_timeseriesPlot(wa_smoky_24hr, type='l', col='blue', lwd=2, add=TRUE)

legend('topright', c("hourly","3-hourly","24-hourly"),
       col=c('black','red','blue'), lwd=c(1,1,2))

title('Smoky Monitors in Washington -- August, 2015')