meteoForecast is a package to access outputs from Numerical Weather Prediction models both in raster format and as a time series for a location. The current version works with these services:

• GFS
• MeteoGalicia
• NAM
• RAP

New services will be added gradually, using this collection of meteorological data sources as reference.

This package has been developed in the framework of the European Project PVCROPS.

The development version is available at GitHub:

## install.packages("remotes")
remotes::install_github("oscarperpinan/meteoForecast")

The stable version is available at CRAN:

install.packages('meteoForecast')

library(meteoForecast)

grepVar will help you to find the variable you need. If no day is defined, it will return the variables available in the current version of the service.

grepVar('cloud', service = 'gfs')

[1] "Cloud_water"        "Cloud_mixing_ratio"

However, you may be interested in the variables available in previous versions. It that is the case, choose a certain day to get this information:

grepVar('cloud', service = 'gfs', day = as.Date('2014-09-15'))

[1] "Total_cloud_cover"  "Cloud_water"        "Cloud_mixing_ratio"

You can use the empty string to retrieve the full list of variables provided by the service.

grepVar("", service = 'meteogalicia', complete = TRUE)

Each function admits an argument named service to choose the service. After loading meteoForecast the default service is Meteogalicia as reported with:

mfService()

[1] "meteogalicia"

You can choose a different service (for the running session) using mfService with the name of the service.

mfService('gfs')

Option service changed to gfs

mfService('meteogalicia')

Option service changed to meteogalicia

• mfProj4 provides the projection of a service:

mfProj4('nam')

[1] "+proj=lcc +lat_1=25 +lat_0=25 +lon_0=-95 +k_0=1 +x_0=0 +y_0=0 +a=6367470.21484375 +b=6367470.21484375 +units=km +no_defs "

• mfExtent provides the extent of a service (using the class Extent defined by the package raster):

mfExtent('meteogalicia', resolution = 36)

class       : Extent 
xmin        : -49.18259 
xmax        : 18.789 
ymin        : 24.03791 
ymax        : 56.06608

• getRaster gets a forecast output inside a bounding box and provides a multilayer raster data using the RasterBrick class defined in the package raster.

  wrf <- getRaster(var = 'temp', day = '2014-01-25', run = '00')
      

  library(rasterVis)
  
  levelplot(wrf, layers = 10:21)
      

• getRasterDays uses getRaster to download the results cast each day comprised between start and end using the 00UTC run.

  wrfDays <- getRasterDays('cft',
                        start = '2014-01-01',
                        end = '2014-01-05',
                        box = c(-2, 35, 2, 40))
  
  hovmoller(wrfDays, par.settings = BTCTheme,
            interpolate = TRUE, panel = panel.levelplot.raster)
      

getPoint, getPointDays, and getPointRuns get data for a certain location and produce a time series using the zoo class.

• getPoint

  ## temperature (Kelvin) forecast from meteogalicia
  tempK <- getPoint(c(0, 40), vars = 'temp')
  ## Cell does not coincide exactly with request
  attr(tempK, 'lat')
  attr(tempK, 'lon')
  ## Units conversion
  tempC <- tempK - 273
      

  xyplot(tempC)
      

• getPointDays

  ## Time sequence
  radDays <- getPointDays(c(0, 40), start = '2013-01-01',
                          end = '2013-01-15')
  
  xyplot(radDays)
      

• getPointRuns

  ## Variability between runs
  radRuns <- getPointRuns(c(0, 40), start = '2013-01-01',
                          end = '2013-01-15')
  xyplot(radRuns, superpose = TRUE)
      

## variability around the average
radAv <- rowMeans(radRuns)
radVar <- sweep(radRuns, 1, radAv)
xyplot(radVar, superpose = TRUE)