gam smoothing is
carried out using mgcv.polarPlot(mydata, pollutant = "nox", x = "ws", wd = "wd",
type = "default", statistic = "mean", resolution = "normal",
limits = NA, exclude.missing = TRUE, uncertainty = FALSE,
percentile = NA, cols = "default", weights = c(0.25, 0.5, 0.75),
min.bin = 1, mis.col = "grey", upper = NA, angle.scale = 315,
units = x, force.positive = TRUE, k = 100, normalise = FALSE,
key.header = "", key.footer = pollutant, key.position = "right",
key = TRUE, auto.text = TRUE, ...)wd, another
variable to plot in polar coordinates (the default is a column
date
if plots by time period are required.pollutant =
"nox". There can also be more than one pollutant specified
e.g. pollutant = c("nox", "no2"). The main use of using two
ortype determines how the data are split
i.e. conditioned, and then plotted. The default is will produce a
single plot using the entire data. Type can be one of the built-in
types as detailed in cutData e.g. TRUE (the
default) removes points from the plot that are too far from the
original data. The smoothing routines will produce predictions at
points where no data exist i.e. they predict. By removing the
points too far frTRUE three plots are produced on the
same scale showing the predicted surface together with the
estimated lower and upper uncertainties at the 95interval. Calculating the uncertainstatistic = "percentile" then
percentile is used, expressed from 0 to 100. Note that the
percentile value is calculated in the wind speed, wind direction
RColorBrewer colours --- see the openair
openColours function forweights applies a weighting to reduce their
influence. For examin.bin is > 1 it can be useful to show
where data are removed on the plots. This is done by shading the
missing data in mis.col. To not highlight missing data when
min.bin > 1 choose mis.col = "transparenangle.scale to another value (between 0 and 360 degrees) to
mitigate suTRUE. Sometimes if
smoothing data with steep gradients it is possible for predicted
values to be negative. force.positive = TRUE ensures that
predictions remain positive. This is useful for several
reasons. First, gam
function in package mgcv. Typically, value of around 100
(the default) seems to be suitable and will resolve important
features in the plot. The most appropriate choice of kTRUE concentrations are normalised by
dividing by their mean value. This is done after fitting
the smooth surface. This option is particularly useful if one is
interested in the patterns of concentrations for several
pollutants onkey.header = "header",
key.footer = "footer1" adds addition text above and below the
scale key. These arguments are passed to drawOpenKey via
qukey.footer."top", "right",
"bottom" and "left".drawOpenKey. See drawOpenKey for further details.TRUE (default) or FALSE. If
TRUE titles and axis labels will automatically try and
format pollutant names and units properly e.g. by subscripting
the `2' in NO2.lattice:levelplot and cutData. For example,
polarPlot passes the option hemisphere = "southern"
on to cutData to provide southern (rather than default
polarPlot
also returns an object of class ``openair''. The object includes
three main components: call, the command used to generate
the plot; data, the data frame of summarised information
used to make the plot; and plot, the plot itself. If
retained, e.g. using output <- polarPlot(mydata, "nox"),
this output can be used to recover the data, reproduce or rework
the original plot or undertake further analysis.An openair output can be manipulated using a number of generic
operations, including print, plot and
summary.
polarPlot surface data can also be extracted directly using
the results, e.g. results(object) for output
<- polarPlot(mydata, "nox"). This returns a data frame with four
set columns: cond, conditioning based on type;
u and v, the translational vectors based on
ws and wd; and the local pollutant estimate.
The function has been developed to allow variables other than wind speed to be plotted with wind direction in polar coordinates. The key issue is that the other variable plotted against wind direction should be discriminating in some way. For example, temperature can help reveal high-level sources brought down to ground level in unstable atmospheric conditions, or show the effect a source emission dependent on temperature e.g. biogenic isoprene.
The plots can vary considerably depending on how much smoothing is
done. The approach adopted here is based on the very flexible and
capable mgcv package that uses Generalized Additive
Models. While methods do exist to find an optimum level of
smoothness, they are not necessarily useful. The principal aim of
polarPlot is as a graphical analysis rather than for
quantitative purposes. In this respect the smoothing aims to
strike a balance between revealing interesting (real) features and
overly noisy data. The defaults used in polarPlot are based
on the analysis of data from many different sources. More advanced
users may wish to modify the code and adopt other smoothing
approaches.
Various statistics are possible to consider e.g. mean, maximum,
median. statistic = "max" is often useful for revealing
sources.
Wind direction is split up into 10 degree intervals and the other variable (e.g. wind speed) 30 intervals. These 2D bins are then used to calculate the statistics.
These plots often show interesting features at higher wind speeds
(see references below). For these conditions there can be very few
measurements and therefore greater uncertainty in the calculation
of the surface. There are several ways in which this issue can be
tackled. First, it is possible to avoid smoothing altogether and
use polarFreq. Second, the effect of setting a
minimum number of measurements in each wind speed-direction bin
can be examined through min.bin. It is possible that a
single point at high wind speed conditions can strongly affect the
surface prediction. Therefore, setting min.bin = 3, for
example, will remove all wind speed-direction bins with fewer than
3 measurements before fitting the surface. Third, consider
setting uncertainty = TRUE. This option will show the
predicted surface together with upper and lower 95intervals, which take account of the frequency of measurements.
Variants on polarPlot include polarAnnulus and
polarFreq.
Carslaw, D.C., Beevers, S.D, Ropkins, K and M.C. Bell (2006). Detecting and quantifying aircraft and other on-airport contributions to ambient nitrogen oxides in the vicinity of a large international airport. Atmospheric Environment. 40/28 pp 5424-5434.
Carslaw, D.C., & Beevers, S.D. (2013). Characterising and understanding emission sources using bivariate polar plots and k-means clustering. Environmental Modelling & Software, 40, 325-329. doi:10.1016/j.envsoft.2012.09.005
Henry, R.C., Chang, Y.S., Spiegelman, C.H., 2002. Locating nearby sources of air pollution by nonparametric regression of atmospheric concentrations on wind direction. Atmospheric Environment 36 (13), 2237-2244.
Westmoreland, E.J., N. Carslaw, D.C. Carslaw, A. Gillah and E. Bates (2007). Analysis of air quality within a street canyon using statistical and dispersion modelling techniques. Atmospheric Environment. Vol. 41(39), pp. 9195-9205.
Yu, K.N., Cheung, Y.P., Cheung, T., Henry, R.C., 2004. Identifying the impact of large urban airports on local air quality by nonparametric regression. Atmospheric Environment 38 (27), 4501-4507.
polarCluster for identifying features in
bivairate polar plots and for post processing and
polarAnnulus, polarFreq,
percentileRose for other ways of plotting directional data.# load example data from package
data(mydata)
# basic plot
polarPlot(mydata, pollutant = "nox")
# polarPlots by year on same scale
polarPlot(mydata, pollutant = "so2", type = "year", main = "polarPlot of so2")
# set minimum number of bins to be used to see if pattern remains similar
polarPlot(mydata, pollutant = "nox", min.bin = 3)
# plot by day of the week
polarPlot(mydata, pollutant = "pm10", type = "weekday")
# show the 95\% confidence intervals in the surface fitting
polarPlot(mydata, pollutant = "so2", uncertainty = TRUE)Run the code above in your browser using DataLab