plotProbcast: Surface plots for forecast information.

Description

Produces contour, image, or perspective plot of a forecast using loess prediction on a grid.

Usage

plotProbcast( forecast, longitude, latitude, nGrid = 65,
                 type = c("image", "contour", "persp"), ..., 
                 interpolate = FALSE, span = 0.75, maps = NULL)

Value

An image, contour, or perspective plot of the forecast.

Arguments

forecast: Numeric vector of forecasts.
longitude: Numeric vector giving the longitude of each forecast location.
latitude: Numeric vector giving the latitude of each forecast location.
nGrid: Number of grid points for loess interpolation. (Binning and interpolation are done on an nGrid by nGrid grid).
type: A character string indicating the desired plot type. Should be one of either "contour", "image", or "persp".
...: Additional arguments to be passed to the plotting method.
interpolate: A logical variable indicating whether or not a loess fit should be used to interpolate the data to points on a grid. The default is to determine grid values by binning, rather than interpolation.
span: Smoothing parameter for loess (used only when interpolate = TRUE). The default value is 0.75, which is the default for loess.
maps: A logical value indicating whether or not to include a map outline. The default is to include an outline if type = "image" and the fields library is loaded.

Details

If the fields library is loaded, a legend (and optionally a map outline) will be included in image plots.

References

C. Fraley, A. E. Raftery, T. Gneiting and J. M. Sloughter, ensembleBMA: An R Package for Probabilistic Forecasting using Ensembles and Bayesian Model Averaging, Technical Report No. 516R, Department of Statistics, University of Washington, 2007 (revised 2010).

Examples

Run this code

  data(srft)

  labels <- c("CMCG","ETA","GASP","GFS","JMA","NGPS","TCWB","UKMO")

  srftData <- ensembleData( forecasts = srft[,labels],
                            dates = srft$date, observations = srft$obs,
                            latitude = srft$lat, longitude = srft$lon,
                            forecastHour = 48, initializationTime = "00")

if (FALSE)  # R check

  bmaFit <- ensembleBMA( srftData, date = "2004012900", trainingDays = 25,
                         model = "normal")

  bmaForc <- quantileForecast( bmaFit, srftData, date = "2004012900",  
                                  quantiles = c(.1, .5, .9))

  obs <- srftData$date == "2004012900"
  lat <- srftData$latitude[obs]
  lon <- srftData$longitude[obs]

  plotProbcast( bmaForc[,"0.5"], lat, lon, 
                   type = "contour", interpolate = TRUE)
  title("Median Forecast")

  plotProbcast( srftData$obs[obs], lat, lon, 
                  type = "contour", interpolate = TRUE)
  title("Observed Surface Temperature")

  data(srftGrid)

  memberLabels <- c("CMCG","ETA","GASP","GFS","JMA","NGPS","TCWB","UKMO")
 
  srftGridData <- ensembleData(forecasts = srftGrid[,memberLabels],
      latitude = srftGrid[,"latitude"], longitude = srftGrid[,"longitude"],
                            forecastHour = 48, initializationTime = "00")

  gridForc <- quantileForecast( bmaFit, srftGridData, 
                    date = "2004021400", quantiles = c( .1, .5, .9))

  library(fields)

  plotProbcast(gridForc[,"0.5"],lon=srftGridData$lon,
     lat=srftGridData$lat,type="image",col=rev(rainbow(100,start=0,end=0.85)))
  title("Median Grid Forecast for Surface Temperature", cex = 0.5)

  probFreeze <- cdf( bmaFit, srftGridData,  date = "2004021400", 
                             value = 273.15)

  plotProbcast(probFreeze, lon=srftGridData$lon, lat=srftGridData$lat,
                   type="image",col=gray((32:0)/32))
  title("Probability of Freezing", cex = 0.5)

Run the code above in your browser using DataLab

Last chance! 50% off unlimited learning