| Title: | An interface to the NOAA Operational Model Archive and Distribution System |
|---|---|
| Description: | An interface to the National Oceanic and Atmospheric Administration's Operational Model Archive and Distribution System (NOMADS) that allows R users to quickly and efficiently download global and regional weather model data for processing. rNOMADS currently supports a variety of models ranging from global weather data to an altitude of 40 km, to high resolution regional weather models, to wave and sea ice models. It can also retrieve archived NOMADS models. rNOMADS can retrieve binary data in GRIB format as well as import ascii data directly into R by interfacing with the GrADS-DODS system. |
| Authors: | Daniel C. Bowman [aut, cre] |
| Maintainer: | Daniel C. Bowman <[email protected]> |
| License: | GPL (>= 3) |
| Version: | 2.0.2 |
| Built: | 2024-12-02 03:12:07 UTC |
| Source: | https://github.com/r-forge/rnomads |
Automatically download forecast data from the National Oceanic and Atmospheric Administration's Operational Model Archive and Distribution System (NOMADS) and read it into R.
This can be done in two ways: reading ascii data directly from the server using the DODS-GrADS system (all operating systems, just needs an internet connection) or downloading binary files in GRIB format (linux only, or until a native R reader for GRIB is available).
The grib capability of rNOMADS uses an external series of routines called wgrib2 to read operational model data; get wgrib2 at http://www.cpc.ncep.noaa.gov/products/wesley/wgrib2/.
The package will also attempt to call another external routine called wgrib if the user wishes to read GRIB1 files; get wgrib at http://www.cpc.ncep.noaa.gov/products/wesley/wgrib.html.
| Package: | rNOMADS |
| Type: | Package |
| Version: | 2.0.0 |
| Date: | 2014-05-15 |
| License: | GPL v3 |
Daniel C. Bowman [email protected]
NOMADS website:
http://nomads.ncep.noaa.gov/
wgrib2 download page:
http://www.cpc.ncep.noaa.gov/products/wesley/wgrib2/
wgrib2 reference:
Ebisuzaki, W, Bokhorst, R., Hyvatti, J., Jovic, D., Nilssen, K,
Pfeiffer, K., Romero, P., Schwarb, M., da Silva, A., Sondell, N., and Varlamov, S. (2011).
wgrib2: read and write GRIB2 files. National Weather Service Climate Prediction Center,
http://www.cpc.ncep.noaa.gov/products/wesley/wgrib2/
wgrib download page:
http://www.cpc.ncep.noaa.gov/products/wesley/wgrib.html
#Getting temperature for North Carolina, USA, #6-12 hours ago depending on when the latest model run was. #Get values at the ground surface and at the 800 mb level #Then make a contour plot of the surface temperature. #We use GrADS-DODS here for compatibility. #Using the Global Forecast System 0.5x0.5 model ## Not run: urls.out <- GetDODSDates(abbrev = "gfs_hd") model.url <- tail(urls.out$url, 1) #Get most recent model date ## End(Not run) #Get most recent model run ## Not run: model.runs <- GetDODSModelRuns(model.url) model.run <- tail(model.runs$model.run, 1) ## End(Not run) #Get ground temperature for the 6 hour prediction variable <- "tmp2m" #temp at 2 m time <- c(2,2) #6 hour prediction lon.dom <- seq(0, 360, by = 0.5) #domain of longitudes in model lat.dom <- seq(-90, 90, by = 0.5) #domain of latitudes in model lon <- which((lon.dom >= 360 - 84) & (lon.dom <= 360 - 74)) - 1 #NOMADS indexes start at 0 lat <- which((lat.dom <= 37) & (lat.dom >= 32)) - 1 #NOMADS indexes start at 0 ## Not run: model.data.surface <- DODSGrab(model.url, model.run, variable, time, c(min(lon), max(lon)), c(min(lat), max(lat))) ## End(Not run) lev <- c(8, 8) #800 mb level variable <- "tmpprs" ## Not run: model.data.800mb <- DODSGrab(model.url, model.run, variable, time, c(min(lon), max(lon)), c(min(lat), max(lat)), level = lev) ## End(Not run) #Make results into arrays ## Not run: model.array.surface <- ModelGrid(model.data.surface, c(0.5, 0.5), "latlon") ## Not run: model.array.800mb <- ModelGrid(model.data.800mb, c(0.5, 0.5), "latlon") #Make a contour plot of the temperature around North Carolina, USA: ## Not run: contour(x = model.array.surface$x - 360, y = model.array.surface$y, model.array.surface$z[1,1,,] - 273.15, xlab = "Longitude", ylab = "Latitude", main = paste("North Carolina Surface Temperatures for", model.array.surface$fcst.date, "GMT in Celsius")) dev.new() contour(x = model.array.800mb$x - 360, y = model.array.800mb$y, model.array.800mb$z[1,1,,] - 273.15, xlab = "Longitude", ylab = "Latitude", main = paste("North Carolina Temperatures at 800 mb for", model.array.surface$fcst.date, "GMT in Celsius")) ## End(Not run)#Getting temperature for North Carolina, USA, #6-12 hours ago depending on when the latest model run was. #Get values at the ground surface and at the 800 mb level #Then make a contour plot of the surface temperature. #We use GrADS-DODS here for compatibility. #Using the Global Forecast System 0.5x0.5 model ## Not run: urls.out <- GetDODSDates(abbrev = "gfs_hd") model.url <- tail(urls.out$url, 1) #Get most recent model date ## End(Not run) #Get most recent model run ## Not run: model.runs <- GetDODSModelRuns(model.url) model.run <- tail(model.runs$model.run, 1) ## End(Not run) #Get ground temperature for the 6 hour prediction variable <- "tmp2m" #temp at 2 m time <- c(2,2) #6 hour prediction lon.dom <- seq(0, 360, by = 0.5) #domain of longitudes in model lat.dom <- seq(-90, 90, by = 0.5) #domain of latitudes in model lon <- which((lon.dom >= 360 - 84) & (lon.dom <= 360 - 74)) - 1 #NOMADS indexes start at 0 lat <- which((lat.dom <= 37) & (lat.dom >= 32)) - 1 #NOMADS indexes start at 0 ## Not run: model.data.surface <- DODSGrab(model.url, model.run, variable, time, c(min(lon), max(lon)), c(min(lat), max(lat))) ## End(Not run) lev <- c(8, 8) #800 mb level variable <- "tmpprs" ## Not run: model.data.800mb <- DODSGrab(model.url, model.run, variable, time, c(min(lon), max(lon)), c(min(lat), max(lat)), level = lev) ## End(Not run) #Make results into arrays ## Not run: model.array.surface <- ModelGrid(model.data.surface, c(0.5, 0.5), "latlon") ## Not run: model.array.800mb <- ModelGrid(model.data.800mb, c(0.5, 0.5), "latlon") #Make a contour plot of the temperature around North Carolina, USA: ## Not run: contour(x = model.array.surface$x - 360, y = model.array.surface$y, model.array.surface$z[1,1,,] - 273.15, xlab = "Longitude", ylab = "Latitude", main = paste("North Carolina Surface Temperatures for", model.array.surface$fcst.date, "GMT in Celsius")) dev.new() contour(x = model.array.800mb$x - 360, y = model.array.800mb$y, model.array.800mb$z[1,1,,] - 273.15, xlab = "Longitude", ylab = "Latitude", main = paste("North Carolina Temperatures at 800 mb for", model.array.surface$fcst.date, "GMT in Celsius")) ## End(Not run)
This function interfaces with the programming API at http://nomads.ncdc.noaa.gov/ to download archived NOMADS model data.
The available models can be viewed by calling NOMADSArchiveList without arguments.
The data arrives in grib (gridded binary) format that can be read with ReadGrib.
Some of these files are in GRIB format, others are in GRIB2 format; select the appropriate file type when calling ReadGrib.
ArchiveGribGrab(abbrev, model.date, model.run, pred, local.dir = ".", file.name = "fcst.grb", tidy = FALSE, verbose = TRUE, download.method = NULL, file.type = "grib2")ArchiveGribGrab(abbrev, model.date, model.run, pred, local.dir = ".", file.name = "fcst.grb", tidy = FALSE, verbose = TRUE, download.method = NULL, file.type = "grib2")
abbrev |
Model abbreviation per |
model.date |
The year, month, and day of the model run, in YYYYMMDD format |
model.run |
Which hour the model was run (i.e. 00, 06, 12, 18 for GFS) |
pred |
Which prediction to get (analysis is 00) |
local.dir |
Where to save the grib file, defaults to the current directory. |
file.name |
What to name the grib file, defaults to |
tidy |
If |
verbose |
If |
download.method |
Allows the user to set the download method used by |
file.type |
Determine whether to get GRIB1 ( |
grib.info$local.dir |
The absolute path to the grib file that was downloaded. |
grib.info$file.name |
The name of the grib file that was downloaded. |
grib.info$url |
The URL that the grib file was downloaded from |
Daniel C. Bowman [email protected]
CheckNOMADSArchive, NOMADSArchiveList, ReadGrib
#An example for the Global Forecast System #Get data for January 1 2014 #Temperature at 2 m above ground #3 hour prediction # using GRIB abbrev <- "gfsanl" model.date <- 20140101 model.run <- 06 pred <- 3 ## Not run: model.info <- ArchiveGribGrab(abbrev, model.date, model.run, pred, file.type = "grib2") ## End(Not run) ## Not run: model.data <- ReadGrib(model.info$file.name, c("2 m above ground"), c("TMP")) #Transform to grid ## Not run: gridded.data <- ModelGrid(model.data, c(0.5, 0.5)) #Get surface temperature in Chapel Hill, NC lat <- 35.907605 lon <- -79.052147 ## Not run: profile.data <- BuildProfile(gridded.data, lon, lat, TRUE) ## Not run: print(paste0("The temperature prediction in Chapel Hill was ", sprintf("%.0f", profile.data[1,1] - 272.15), " degrees Celsius.")) ## End(Not run)#An example for the Global Forecast System #Get data for January 1 2014 #Temperature at 2 m above ground #3 hour prediction # using GRIB abbrev <- "gfsanl" model.date <- 20140101 model.run <- 06 pred <- 3 ## Not run: model.info <- ArchiveGribGrab(abbrev, model.date, model.run, pred, file.type = "grib2") ## End(Not run) ## Not run: model.data <- ReadGrib(model.info$file.name, c("2 m above ground"), c("TMP")) #Transform to grid ## Not run: gridded.data <- ModelGrid(model.data, c(0.5, 0.5)) #Get surface temperature in Chapel Hill, NC lat <- 35.907605 lon <- -79.052147 ## Not run: profile.data <- BuildProfile(gridded.data, lon, lat, TRUE) ## Not run: print(paste0("The temperature prediction in Chapel Hill was ", sprintf("%.0f", profile.data[1,1] - 272.15), " degrees Celsius.")) ## End(Not run)
Takes the output of ModelGrid and extracts data at a specific point, performing interpolation if required.
BuildProfile(gridded.data, lon, lat, spatial.average)BuildProfile(gridded.data, lon, lat, spatial.average)
gridded.data |
Data structure returned by |
lon |
Longitude of point of interest. |
lat |
Latitude of point of interest. |
spatial.average |
Whether to interpolate data using b-splines to obtain value at the requested point ( |
It is much more efficient to download a large chunk of data and extract profile points from that as opposed to downloading individual small model chunks in the vicinity of each point of interest.
profile.data |
A levels x variables matrix with data for a given point. |
Daniel C. Bowman [email protected]
#Get temperature profile in Chapel Hill, NC. #First, define each location lon <- -79.052094 lat <- 35.907562 #Get latest GFS 0.5 model, use analysis forecast ## Not run: model.url <- CrawlModels(abbrev = "gfs_hd", depth = 1)[1] pred <- ParseModelPage(model.url)$pred[1] ## End(Not run) #Get levels pressure <- c(1, 2, 3, 5, 7, 10, 20, 30, 50, 70, seq(100, 1000, by = 25)) levels <- paste(pressure, " mb", sep = "") #Variables - temperature and height only variables <- c("TMP", "HGT") ## Not run: grib.info <- GribGrab(model.url, pred, levels, variables, model.domain = c(-85, -75, 37, 32)) grib.data <- ReadGrib(grib.info$file.name, levels, variables) gridded.data <- ModelGrid(grib.data, c(0.5, 0.5)) profile <- BuildProfile(gridded.data, lon, lat, TRUE) plot(profile[,2] - 273.15, profile[,1], xlab = "Temperature (C)", ylab = "Height (m)", main = "Temperature Profile above Chapel Hill, NC") ## End(Not run)#Get temperature profile in Chapel Hill, NC. #First, define each location lon <- -79.052094 lat <- 35.907562 #Get latest GFS 0.5 model, use analysis forecast ## Not run: model.url <- CrawlModels(abbrev = "gfs_hd", depth = 1)[1] pred <- ParseModelPage(model.url)$pred[1] ## End(Not run) #Get levels pressure <- c(1, 2, 3, 5, 7, 10, 20, 30, 50, 70, seq(100, 1000, by = 25)) levels <- paste(pressure, " mb", sep = "") #Variables - temperature and height only variables <- c("TMP", "HGT") ## Not run: grib.info <- GribGrab(model.url, pred, levels, variables, model.domain = c(-85, -75, 37, 32)) grib.data <- ReadGrib(grib.info$file.name, levels, variables) gridded.data <- ModelGrid(grib.data, c(0.5, 0.5)) profile <- BuildProfile(gridded.data, lon, lat, TRUE) plot(profile[,2] - 273.15, profile[,1], xlab = "Temperature (C)", ylab = "Height (m)", main = "Temperature Profile above Chapel Hill, NC") ## End(Not run)
This function checks to see if data exists for a given date and model. It checks for both GRIB1 or GRIB2 files.
CheckNOMADSArchive(abbrev, model.date = NULL)CheckNOMADSArchive(abbrev, model.date = NULL)
abbrev |
Model abbreviation per |
model.date |
The year, month, and day to check for data, in YYYYMMDD format.
If |
available.models$date |
What date the file is for, in YYYYMMDD format. |
available.models$model.run |
At what hour (GMT) the model was run. |
available.models$pred |
What predictions are available |
available.models$file.name |
List of file names for available model dates, runs, and predictions |
Daniel C. Bowman [email protected]
NOMADSArchiveList, ArchiveGribGrab
#See what is available for January 1 2014 abbrev <- "gfs4" model.date <- 20140101 ## Not run: gfs.available.models <- CheckNOMADSArchive(abbrev, model.date) #Determine all available North American Mesoscale models in the archive #This will take some time ## Not run: nam.available.models <- CheckNOMADSArchive("namanl")#See what is available for January 1 2014 abbrev <- "gfs4" model.date <- 20140101 ## Not run: gfs.available.models <- CheckNOMADSArchive(abbrev, model.date) #Determine all available North American Mesoscale models in the archive #This will take some time ## Not run: nam.available.models <- CheckNOMADSArchive("namanl")
This function determine which instances of a given model are available for download.
CrawlModels(abbrev = NULL, url = NULL, depth = NULL, verbose = TRUE)CrawlModels(abbrev = NULL, url = NULL, depth = NULL, verbose = TRUE)
abbrev |
The model abbreviation, see |
url |
A URL to use instead of using the abbreviations in |
depth |
How many model instances to return.
This avoids having to download the entire model list (sometimes several hundred) if only the first few instances are required.
Defaults to |
verbose |
Print out each link as it is discovered.
Defaults to |
This function calls WebCrawler, a recursive algorithm that discovers each link available in the URL provided.
It then searches each link in turn, and follows those links until it reaches a dead end.
At that point, it returns the URL.
For the model pages on the NOMADS web site, each dead end is a model instance that can be examined using ParseModelPage or have data retrieved from it using GribGrab.
urls.out |
A list of web page addresses, each of which corresponds to a model instance. |
It is a good idea to set depth to a small number rather than leave it at the default value.
Some models (such as the Global Forecast System) have a large number of instances, and crawling each one can take a lot of time.
I recommend depth = 2, since the first URL may not have an active model on it yet if the model is still being uploaded to the server.
In that case,the first URL will contain no data, and the second URL can be used instead.
Daniel C. Bowman [email protected]
WebCrawler, ParseModelPage, NOMADSRealTimeList, GribGrab
#Get the latest 5 instances #for the Global Forecast System 0.5 degree model ## Not run: urls.out <- CrawlModels(abbrev = "gfs_hd", depth = 5)#Get the latest 5 instances #for the Global Forecast System 0.5 degree model ## Not run: urls.out <- CrawlModels(abbrev = "gfs_hd", depth = 5)
This function interfaces with the NOMADS server to download weather, ocean, and sea ice data.
The available models can be viewed by calling NOMADSRealTimeList and NOMADSArchiveList.
The data arrives in ascii format, so this function can be used to retrieve data on any operating system.
DODSGrab(model.url, model.run, variable, time, lon, lat, levels = NULL, display.url = TRUE)DODSGrab(model.url, model.run, variable, time, lon, lat, levels = NULL, display.url = TRUE)
model.url |
A model URL for a specific date, probably from |
model.run |
A specific model run to get, probably from |
variable |
The data type to get. |
time |
An two component vector denoting which time indices to get. |
lon |
An two component vector denoting which longitude indices to get. |
lat |
An two component vector denoting which latitude indices to get. |
levels |
An two component vector denoting which levels to get, if applicable. |
display.url |
If |
model.data |
A structure with a series of elements containing data extracted from GrADS-DODS system. |
Daniel C. Bowman [email protected]
GetDODSDates, GetDODSModelRuns, GetDODSModelRunInfo
#An example for the Global Forecast System 0.5 degree model #Make a world temperature map for the latest model run ## Not run: #Figure out which model is most recent model.urls <- GetDODSDates("gfs_hd") latest.model <- tail(model.urls$url, 1) model.runs <- GetDODSModelRuns(latest.model) latest.model.run <- tail(model.runs$model.run, 1) #Download worldwide temperature data at 2 m variable <- "tmp2m" time <- c(0, 0) #Analysis run, index starts at 0 lon <- c(0, 719) #All 720 longitude points lat <- c(0, 360) #All 361 latitude points model.data <- DODSGrab(latest.model, latest.model.run, variable, time, lon, lat) #Make it into a nice array and plot it model.grid <- ModelGrid(model.data, c(0.5, 0.5), "latlon") image(model.grid$z[1,1,,]) ## End(Not run)#An example for the Global Forecast System 0.5 degree model #Make a world temperature map for the latest model run ## Not run: #Figure out which model is most recent model.urls <- GetDODSDates("gfs_hd") latest.model <- tail(model.urls$url, 1) model.runs <- GetDODSModelRuns(latest.model) latest.model.run <- tail(model.runs$model.run, 1) #Download worldwide temperature data at 2 m variable <- "tmp2m" time <- c(0, 0) #Analysis run, index starts at 0 lon <- c(0, 719) #All 720 longitude points lat <- c(0, 360) #All 361 latitude points model.data <- DODSGrab(latest.model, latest.model.run, variable, time, lon, lat) #Make it into a nice array and plot it model.grid <- ModelGrid(model.data, c(0.5, 0.5), "latlon") image(model.grid$z[1,1,,]) ## End(Not run)
This function determines which GFS forecast is closest to a given date. It returns which forecast precedes the date, and which forecast follows the date. Thus a user can average the two forecasts together to provide a precise forecast for a given date.
GetClosestGFSForecasts(forecast.date, model.date = "latest", depth = NULL, verbose = TRUE)GetClosestGFSForecasts(forecast.date, model.date = "latest", depth = NULL, verbose = TRUE)
forecast.date |
What date you want a forecast for, as a date/time object. It must be in the GMT time zone. |
model.date |
Which model run to use, in YYYYMMDDHH, where HH is 00, 06, 12, 18.
Defaults to |
depth |
How many model instances to return.
This avoids having to download the entire model list (sometimes several hundred) if only the first few instances are required.
Defaults to |
verbose |
Gives a detailed account of progress.
Defaults to |
forecasts$model.url |
URL to send to |
forecasts$model.run.date |
When the model was run. |
forecasts$back.forecast |
Nearest forecast behind requested date. |
forecasts$fore.forecast |
Nearest forecast after requested date. |
forecasts$back.hr |
How many hours the back forecast is behind the requested date. |
forecasts$fore.hr |
How many hours the fore forecast is in front of the requested date. |
Daniel C. Bowman [email protected]
RTModelProfile, BuildProfile, GribGrab
#Get the exact temperature profile of Chapel Hill, NC #by performing a weighted average of GFS model forecasts. #Figure out which forecasts to use forecast.date <- as.POSIXlt(Sys.time(), tz = "GMT") ## Not run: forecasts <- GetClosestGFSForecasts(forecast.date) #Get levels pressure <- c(1, 2, 3, 5, 7, 10, 20, 30, 50, 70, seq(100, 1000, by = 25)) levels <- paste(pressure, " mb", sep = "") #Variables - temperature and height only variables <- c("TMP", "HGT") #Location lon <- c(-79.052083) lat <- c(35.907492) #Get the data for each resolution <- c(0.5, 0.5) grid.type <- "latlon" ## Not run: back.profile <- RTModelProfile(forecasts$model.url, forecasts$back.forecast, levels, variables, lon, lat, resolution = resolution, grid.type = grid.type) fore.profile <- RTModelProfile(forecasts$model.url, forecasts$fore.forecast, levels, variables, lon, lat, resolution = resolution, grid.type = grid.type) temps <- cbind(back.profile$profile[[1]][,2], fore.profile$profile[[1]][,2]) - 273.15 heights <- cbind(back.profile$profile[[1]][,1], fore.profile$profile[[1]][,1]) time.gap <- forecasts$fore.hr - forecasts$back.hr exact.temp <- (temps[,1] * abs(forecasts$fore.hr) + temps[,2] * abs(forecasts$back.hr))/time.gap exact.hgt <- (heights[,1] * abs(forecasts$fore.hr) + heights[,2] * abs(forecasts$back.hr))/time.gap #Plot results plot(c(min(temps), max(temps)), c(min(heights), max(heights)), type = "n", xlab = "Temperature (C)", ylab = "Height (m)") points(temps[,1], heights[,1], pch = 1, col = 1) points(temps[,1], heights[,2], pch = 2, col = 2) lines(exact.temp, exact.hgt, col = 3, lty = 2) legend("topleft", pch = c(1, 2, NA), lty = c(NA, NA, 2), col = c(1, 2, 3), legend = c(forecasts$back.forecast, forecasts$fore.forecast, as.character(Sys.time()))) ## End(Not run)#Get the exact temperature profile of Chapel Hill, NC #by performing a weighted average of GFS model forecasts. #Figure out which forecasts to use forecast.date <- as.POSIXlt(Sys.time(), tz = "GMT") ## Not run: forecasts <- GetClosestGFSForecasts(forecast.date) #Get levels pressure <- c(1, 2, 3, 5, 7, 10, 20, 30, 50, 70, seq(100, 1000, by = 25)) levels <- paste(pressure, " mb", sep = "") #Variables - temperature and height only variables <- c("TMP", "HGT") #Location lon <- c(-79.052083) lat <- c(35.907492) #Get the data for each resolution <- c(0.5, 0.5) grid.type <- "latlon" ## Not run: back.profile <- RTModelProfile(forecasts$model.url, forecasts$back.forecast, levels, variables, lon, lat, resolution = resolution, grid.type = grid.type) fore.profile <- RTModelProfile(forecasts$model.url, forecasts$fore.forecast, levels, variables, lon, lat, resolution = resolution, grid.type = grid.type) temps <- cbind(back.profile$profile[[1]][,2], fore.profile$profile[[1]][,2]) - 273.15 heights <- cbind(back.profile$profile[[1]][,1], fore.profile$profile[[1]][,1]) time.gap <- forecasts$fore.hr - forecasts$back.hr exact.temp <- (temps[,1] * abs(forecasts$fore.hr) + temps[,2] * abs(forecasts$back.hr))/time.gap exact.hgt <- (heights[,1] * abs(forecasts$fore.hr) + heights[,2] * abs(forecasts$back.hr))/time.gap #Plot results plot(c(min(temps), max(temps)), c(min(heights), max(heights)), type = "n", xlab = "Temperature (C)", ylab = "Height (m)") points(temps[,1], heights[,1], pch = 1, col = 1) points(temps[,1], heights[,2], pch = 2, col = 2) lines(exact.temp, exact.hgt, col = 3, lty = 2) legend("topleft", pch = c(1, 2, NA), lty = c(NA, NA, 2), col = c(1, 2, 3), legend = c(forecasts$back.forecast, forecasts$fore.forecast, as.character(Sys.time()))) ## End(Not run)
This function checks the GrADS data server to see what dates and model subsets are available for model specified by ABBREV
GetDODSDates(abbrev, archive=FALSE, request.sleep=0)GetDODSDates(abbrev, archive=FALSE, request.sleep=0)
abbrev |
A model abbreviation as specified in |
archive |
Whether the model is on the NCEP real time server ( |
request.sleep |
Seconds to pause between HTTP requests when scanning model pages. |
This function determines which dates are available for download for a particular model through the GrADS - DODS system.
Once the user determines which dates are available, the output of this function can be passed to GetDODSModelRuns to determine which model runs can be downloaded.
model |
The model that was requested. |
date |
A list of model run dates available for download. |
url |
A list of URLs corresponding to the model run dates. |
Sometimes, sending lots of HTTP requests in rapid succession can cause errors.
If messages resembling "Error: failed to load HTTP resource" appear, try request.sleep = 1.
The code will take longer to execute but it will be more likely to finish successfully.
Daniel C. Bowman [email protected]
#An example for the Global Forecast System 0.5 degree model #Get the latest model url and date abbrev <- "gfs_hd" ## Not run: urls.out <- GetDODSDates(abbrev) ## Not run: print(paste("Most recent model run:",tail(urls.out$date, 1))) #Get model dates from the GFS archive abbrev <- "gfs-avn-hi" ## Not run: urls.out <- GetDODSDates(abbrev, archive = TRUE, request.sleep = 1)#An example for the Global Forecast System 0.5 degree model #Get the latest model url and date abbrev <- "gfs_hd" ## Not run: urls.out <- GetDODSDates(abbrev) ## Not run: print(paste("Most recent model run:",tail(urls.out$date, 1))) #Get model dates from the GFS archive abbrev <- "gfs-avn-hi" ## Not run: urls.out <- GetDODSDates(abbrev, archive = TRUE, request.sleep = 1)
Given a URL from GetDODSDates, find which model runs are available for download on the GrADS - DODS system.
GetDODSModelRunInfo(model.url, model.run)GetDODSModelRunInfo(model.url, model.run)
model.url |
A URL for a model on the GrADS - DODS system, probably returned by |
model.run |
A specific model run, probably returned by |
This routine grabs information about the latitude, longitude, and time coverage of a specific model instance.
It also finds data about levels (if present) and lists all the available variables (though they may not have data in them).
The user can refer to this information to construct calls to the DODS system via DODSGrab.
model.info |
Information provided by the GrADS - DODS system about the given model instance. |
Daniel C. Bowman [email protected]
GetDODSDates, GetDODSModelRuns, DODSGrab
#An example for the Global Forecast System 0.5 degree model #Get some information about the latest model url and date, real time server abbrev <- "gfs_hd" ## Not run: urls.out <- GetDODSDates(abbrev) model.url <- tail(urls.out$url, 1) model.runs <- GetDODSModelRuns(model.url) model.info <- GetDODSModelRunInfo(model.url, tail(model.runs$model.run, 1)) print(model.info) ## End(Not run)#An example for the Global Forecast System 0.5 degree model #Get some information about the latest model url and date, real time server abbrev <- "gfs_hd" ## Not run: urls.out <- GetDODSDates(abbrev) model.url <- tail(urls.out$url, 1) model.runs <- GetDODSModelRuns(model.url) model.info <- GetDODSModelRunInfo(model.url, tail(model.runs$model.run, 1)) print(model.info) ## End(Not run)
Given a URL from GetDODSDates, find which model runs are available for download on the GrADS - DODS system.
GetDODSModelRuns(model.url)GetDODSModelRuns(model.url)
model.url |
A URL for a model on the GrADS - DODS system, probably returned by |
This function determines which dates are available for download for a particular model through the GrADS - DODS system.
Once the user determines which dates are available, the output of this function can be passed to GetDODSModelRuns to determine which model runs can be downloaded.
model.run |
A list of model runs available for the requested date. |
model.run.info |
Information provided by the GrADS - DODS system about each model run. |
To get model run information for archived analysis models, pass URLs directly from NOMADSArchiveList directly to GetDODSModelRuns.
Daniel C. Bowman [email protected]
GetDODSDates, DODSGrab, GetDODSModelRunInfo
#An example for the Global Forecast System 0.5 degree model #Get the latest model url and date, real time server abbrev <- "gfs_hd" ## Not run: urls.out <- GetDODSDates(abbrev) model.url <- tail(urls.out$url, 1) model.runs <- GetDODSModelRuns(model.url) print(paste("Latest model run", tail(model.runs$model.run.info, 1))) ## End(Not run) #Get model dates from the GFS analysis archive abbrev <- "gfsanl" model.url <- NOMADSArchiveList("dods", abbrev = abbrev)$url ## Not run: model.runs <- GetDODSModelRuns(model.url) print(model.runs$model.run.info) ## End(Not run)#An example for the Global Forecast System 0.5 degree model #Get the latest model url and date, real time server abbrev <- "gfs_hd" ## Not run: urls.out <- GetDODSDates(abbrev) model.url <- tail(urls.out$url, 1) model.runs <- GetDODSModelRuns(model.url) print(paste("Latest model run", tail(model.runs$model.run.info, 1))) ## End(Not run) #Get model dates from the GFS analysis archive abbrev <- "gfsanl" model.url <- NOMADSArchiveList("dods", abbrev = abbrev)$url ## Not run: model.runs <- GetDODSModelRuns(model.url) print(model.runs$model.run.info) ## End(Not run)
This function interfaces with the programming API at http://nomads.ncep.noaa.gov/ to download NOMADS model data.
The available models can be viewed by calling NOMADSRealTimeList.
The data arrives in grib (gridded binary) format that can be read with ReadGrib.
GribGrab(model.url, pred, levels, variables, local.dir = ".", file.name = "fcst.grb", model.domain = NULL, tidy = FALSE, verbose = TRUE, check.url = TRUE, download.method = NULL)GribGrab(model.url, pred, levels, variables, local.dir = ".", file.name = "fcst.grb", model.domain = NULL, tidy = FALSE, verbose = TRUE, check.url = TRUE, download.method = NULL)
model.url |
The address of a model download page, probably from |
pred |
The list of predictions (or model times) determined by the specific model from |
levels |
A list of model levels to download. |
variables |
A list of model variables to download. |
local.dir |
Where to save the grib file, defaults to the current directory. |
file.name |
What to name the grib file, defaults to |
model.domain |
A vector of latitudes and longitudes that specify the area to return a forecast for. This is a rectangle with elements: west longitude, east longitude, north latitude, south latitude. |
tidy |
If |
verbose |
If |
check.url |
If |
download.method |
Allows the user to set the download method used by |
grib.info$local.dir |
The absolute path to the grib file that was downloaded. |
grib.info$file.name |
The name of the grib file that was downloaded. |
grib.info$url |
The URL that the grib file was downloaded from |
This requires the external programs wgrib2 and\or wgrib to be installed (depending on whether the files are in GRIB2 or GRIB format).
Currently, these routines are only available on Unix/Linux systems.
When a native R reader for GRIB is available, it will be integrated with rNOMADS.
Daniel C. Bowman [email protected]
CrawlModels, ParseModelPage, ReadGrib
#An example for the Global Forecast System 0.5 degree model #Get the latest model url ## Not run: urls.out <- CrawlModels(abbrev = "gfs_hd", depth = 1) #Get a list of forecasts, variables and levels ## Not run: model.parameters <- ParseModelPage(urls.out[1]) #Figure out which one is the 6 hour forecast #provided by the latest model run #(will be the forecast from 6-12 hours from the current date) ## Not run: my.pred <- model.parameters$pred[grep("06$", model.parameters$pred)] #What region of the atmosphere to get data for levels <- c("2 m above ground", "800 mb") #What data to return variables <- c("TMP", "RH") #Temperature and relative humidity #Get the data ## Not run: grib.info <- GribGrab(urls.out[1], my.pred, levels, variables) #Extract the data ## Not run: model.data <- ReadGrib(grib.info$file.name, levels, variables) #Reformat it ## Not run: model.grid <- ModelGrid(model.data, c(0.5, 0.5)) #Show an image of world temperature at ground level ## Not run: image(model.grid$z[2, 1,,])#An example for the Global Forecast System 0.5 degree model #Get the latest model url ## Not run: urls.out <- CrawlModels(abbrev = "gfs_hd", depth = 1) #Get a list of forecasts, variables and levels ## Not run: model.parameters <- ParseModelPage(urls.out[1]) #Figure out which one is the 6 hour forecast #provided by the latest model run #(will be the forecast from 6-12 hours from the current date) ## Not run: my.pred <- model.parameters$pred[grep("06$", model.parameters$pred)] #What region of the atmosphere to get data for levels <- c("2 m above ground", "800 mb") #What data to return variables <- c("TMP", "RH") #Temperature and relative humidity #Get the data ## Not run: grib.info <- GribGrab(urls.out[1], my.pred, levels, variables) #Extract the data ## Not run: model.data <- ReadGrib(grib.info$file.name, levels, variables) #Reformat it ## Not run: model.grid <- ModelGrid(model.data, c(0.5, 0.5)) #Show an image of world temperature at ground level ## Not run: image(model.grid$z[2, 1,,])
Given zonal (East-West) and meridional (North-South) wind speeds, calculate magnitude of wind vector and azimuth from north, in degrees.
MagnitudeAzimuth(zonal.wind, meridional.wind)MagnitudeAzimuth(zonal.wind, meridional.wind)
zonal.wind |
A vector of zonal (East-West) winds, west negative. |
meridional.wind |
A vector of meridional (North-South) winds, south negative. |
winds$magnitude |
Magnitude of wind vector. |
winds$azimuth |
Azimuth of wind vector in degrees from North |
Daniel C. Bowman [email protected]
zonal.wind <- c(35.5, -2) meridional.wind <- c(-5, 15) winds <- MagnitudeAzimuth(zonal.wind, meridional.wind) print(winds$magnitude) print(winds$azimuth)zonal.wind <- c(35.5, -2) meridional.wind <- c(-5, 15) winds <- MagnitudeAzimuth(zonal.wind, meridional.wind) print(winds$magnitude) print(winds$azimuth)
This function takes output from ReadGrib and produces an array with dimensions: levels x variables x longitudes x latitudes.
This greatly reduces the size of the data set as well as makes it easier to manipulate.
ModelGrid(model.data, resolution, grid.type = "latlon", levels = NULL, variables = NULL, model.domain = NULL)ModelGrid(model.data, resolution, grid.type = "latlon", levels = NULL, variables = NULL, model.domain = NULL)
model.data |
Output from |
resolution |
Resolution of grid, in degrees if |
grid.type |
Whether the grid is in lat/lon or cartesian. Options |
levels |
The model levels to include in the grid, if NULL, include all of them. |
variables |
The model variables to include in grid, if NULL, include all of them. |
model.domain |
A vector c(LEFT LON, RIGHT LON, TOP LAT, BOTTOM LAT) of the region to include in output. If NULL, include everything. |
If you set the spacing of lon.grid and/or lat.grid coarser than the downloaded model grid, you can reduce the resolution of your model, possibly making it easier to handle.
z |
An array of dimensions levels x variables x lon x lat; each level x variable contains the model grid of data from that variable and level |
x |
Vector of longitudes |
y |
Vector of latitudes |
variables |
The variables contained in the grid |
levels |
The levels contained in the grid |
model.run.date |
When the forecast model was run |
fcst.date |
The date of the forecast |
Daniel C. Bowman [email protected]
ReadGrib, BuildProfile, RTModelProfile
## Not run: #Get some example data urls.out <- CrawlModels(abbrev = "gfs_hd", depth = 1) model.parameters <- ParseModelPage(urls.out[1]) levels <- c("2 m above ground", "100 mb") variables <- c("TMP", "RH") #Temperature and relative humidity grib.info <- GribGrab(urls.out[1], model.parameters$pred[1], levels, variables) #Extract the data model.data <- ReadGrib(grib.info$file.name, levels, variables) #Make it into an array gfs.array <- ModelGrid(model.data, c(0.5, 0.5)) #What variables and levels we have print(gfs.array$levels) print(gfs.array$variables) #Find minimum temperature at the ground surface, and where it is min.temp <- min(gfs.array$z[2, 1,,] - 273.15) sprintf("%.1f", min.temp) #in Celsius ti <- which(gfs.array$z[2, 1,,] == min.temp + 273.15, arr.ind = TRUE) lat <- gfs.array$y[ti[1,2]] #Lat of minimum temp lon <- gfs.array$x[ti[1,1]] #Lon of minimum temp #Find maximum temperature at 100 mb atmospheric pressure max.temp <- max(gfs.array$z[1, 1,,]) - 273.15 sprintf("%.1f", max.temp) #Brrr! ## End(Not run)## Not run: #Get some example data urls.out <- CrawlModels(abbrev = "gfs_hd", depth = 1) model.parameters <- ParseModelPage(urls.out[1]) levels <- c("2 m above ground", "100 mb") variables <- c("TMP", "RH") #Temperature and relative humidity grib.info <- GribGrab(urls.out[1], model.parameters$pred[1], levels, variables) #Extract the data model.data <- ReadGrib(grib.info$file.name, levels, variables) #Make it into an array gfs.array <- ModelGrid(model.data, c(0.5, 0.5)) #What variables and levels we have print(gfs.array$levels) print(gfs.array$variables) #Find minimum temperature at the ground surface, and where it is min.temp <- min(gfs.array$z[2, 1,,] - 273.15) sprintf("%.1f", min.temp) #in Celsius ti <- which(gfs.array$z[2, 1,,] == min.temp + 273.15, arr.ind = TRUE) lat <- gfs.array$y[ti[1,2]] #Lat of minimum temp lon <- gfs.array$x[ti[1,1]] #Lon of minimum temp #Find maximum temperature at 100 mb atmospheric pressure max.temp <- max(gfs.array$z[1, 1,,]) - 273.15 sprintf("%.1f", max.temp) #Brrr! ## End(Not run)
A list of abbreviations, names and URLs for the NOMADS models archived on the NCDC web site.. Users can refer to this list to find out more information about the available models, and rNOMADS uses the abbreviations to determine how to access the archives.
NOMADSArchiveList(url.type, abbrev = NULL)NOMADSArchiveList(url.type, abbrev = NULL)
url.type |
Determine whether to return a URL for extracting GRIB files ( |
abbrev |
Return information about the model that this abbreviation refers to.
Defaults to |
abbrevs |
An abbreviation for each model |
names |
A full name for each model |
urls |
The web address of the download page for each model |
Daniel C. Bowman [email protected]
#The archived model list in rNOMADS archived.model.list <- NOMADSArchiveList("grib")#The archived model list in rNOMADS archived.model.list <- NOMADSArchiveList("grib")
A list of abbreviations, names and URLs for the NOMADS models. Users can refer to this list to find out more information about the available models, and rNOMADS uses the abbreviations to determine which URLs to scan and download.
NOMADSRealTimeList(url.type, abbrev = NULL)NOMADSRealTimeList(url.type, abbrev = NULL)
url.type |
Determine whether to return a URL for extracting GRIB files ( |
abbrev |
Return information about the model that this abbreviation refers to.
Defaults to |
abbrevs |
An abbreviation for each model |
names |
A full name for each model |
urls |
The web address of the download page for each model |
Daniel C. Bowman [email protected]
WebCrawler, ParseModelPage, NOMADSArchiveList, GribGrab, DODSGrab
#The full model list in rNOMADS model.list <- NOMADSRealTimeList("dods")#The full model list in rNOMADS model.list <- NOMADSRealTimeList("dods")
This function parses the model download pages on NOMADS, and extracts information on predictions, levels, and variables available for each.
ParseModelPage(model.url)ParseModelPage(model.url)
model.url |
The URL of the model to extract information from, probably returned by |
This function scrapes the web page for a given model and determines which predictions, levels, and variables are present for each. Predictions are instances returned by each model (for example, the GFS model produces 3 hour predictions up to 192 hours from the model run). Levels are regions of the atmosphere, surface of the Earth, or subsurface that the model produces output for (for example the GFS model has a “2 m above ground” level that has data for temperature, etc, at that height across the Earth). Variables are types of data (temperature, for example).
pred |
Model predictions |
levels |
Locations of data points |
variables |
Data types |
Many of the names for predictions, levels, and variables are somewhat cryptic.
Future versions of rNOMADS may have a reference function similar to NOMADSRealTimeList to help users with this issue.
Daniel C. Bowman [email protected]
WebCrawler, ParseModelPage, GribGrab
#An example for the Global Forecast System 0.5 degree model #Get the latest model url ## Not run: urls.out <- CrawlModels(abbrev = "gfs_hd", depth = 1) #Get a list of forecasts, variables and levels ## Not run: model.parameters <- ParseModelPage(urls.out[1])#An example for the Global Forecast System 0.5 degree model #Get the latest model url ## Not run: urls.out <- CrawlModels(abbrev = "gfs_hd", depth = 1) #Get a list of forecasts, variables and levels ## Not run: model.parameters <- ParseModelPage(urls.out[1])
This function wraps wgrib2 and wgrib, external grib file readers provided by the National Weather Service Climate Prediction Center (see http://www.cpc.ncep.noaa.gov/products/wesley/wgrib2/ and http://www.cpc.ncep.noaa.gov/products/wesley/wgrib.html).
ReadGrib extracts forecast data into R.
It does this by building an argument string, executing a system call to the appropriate external grib file reader, and extracting the result.
Note that wgrib2 must be installed for ReadGrib to work for current grib files, and wgrib may need to be installed when looking at archived data.
ReadGrib(file.name, levels, variables, file.type = "grib2")ReadGrib(file.name, levels, variables, file.type = "grib2")
file.name |
The path and file name of the grib file to read. |
levels |
The levels to extract. |
variables |
The variables to extract. |
file.type |
Whether the file is in GRIB ( |
This function constructs system calls to wgrib and wgrib2.
Therefore, you must have installed these programs and made it available on the system path.
Unless you are interested in accessing archive data that's more than a few years old, you can install wgrib2 only.
A description of wgrib2 and installation links are available at http://www.cpc.ncep.noaa.gov/products/wesley/wgrib2/ and http://www.cpc.ncep.noaa.gov/products/wesley/wgrib.html.
Note that Windows users will likely have to use a UNIX emulator like Cygwin to install these programs.
Also, rNOMADS is focused towards GRIB2 files; I have included GRIB1 format support as a convenience.
model.data |
A structure with a series of elements containing data extracted from the grib file. |
Daniel C. Bowman [email protected]
Ebisuzaki, W, Bokhorst, R., Hyvatti, J., Jovic, D., Nilssen, K, Pfeiffer, K., Romero, P., Schwarb, M., da Silva, A., Sondell, N., and Varlamov, S. (2011). wgrib2: read and write GRIB2 files. National Weather Service Climate Prediction Center, http://www.cpc.ncep.noaa.gov/products/wesley/wgrib2/
GribGrab, ArchiveGribGrab, ModelGrid
#Operational Forecast Data Extraction #NCEP output is always in GRIB2 format - this makes things easy for us #An example for the Global Forecast System 0.5 degree model #Get the latest model url ## Not run: urls.out <- CrawlModels(abbrev = "gfs_hd", depth = 1) #Get a list of forecasts, variables and levels ## Not run: model.parameters <- ParseModelPage(urls.out[1]) #Figure out which one is the 6 hour forecast #provided by the latest model run #(will be the forecast from 6-12 hours from the current date) ## Not run: my.pred <- model.parameters$pred[grep("06$", model.parameters$pred)] #What region of the atmosphere to get data for levels <- c("2 m above ground", "800 mb") #What data to return variables <- c("TMP", "RH") #Temperature and relative humidity #Get the data ## Not run: model.info <- GribGrab(urls.out[1], my.pred, levels, variables) #Extract the data ## Not run: model.data <- ReadGrib(model.info$file.name, levels, variables) #Reformat it ## Not run: model.grid <- ModelGrid(model.data, c(0.5, 0.5)) #Show an image of world temperature at ground level ## Not run: image(model.grid$z[2, 1,,]) #Archived Data Extraction #This is sometimes in GRIB1 format #This example is in GRIB1 abbrev <- "gfsanl" model.date <- 20040302 #March 2, 2004 model.run <- 18 #1800 GMT model run pred <- 0 #Analysis ## Not run: grib.info <- ArchiveGribGrab(abbrev, model.date, model.run, pred, file.type = "grib1") ## End(Not run) ## Not run: model.data <- ReadGrib(grib.info$file.name, c("1000 mb"), c("TMP"), file.type = "grib1") ## End(Not run)#Operational Forecast Data Extraction #NCEP output is always in GRIB2 format - this makes things easy for us #An example for the Global Forecast System 0.5 degree model #Get the latest model url ## Not run: urls.out <- CrawlModels(abbrev = "gfs_hd", depth = 1) #Get a list of forecasts, variables and levels ## Not run: model.parameters <- ParseModelPage(urls.out[1]) #Figure out which one is the 6 hour forecast #provided by the latest model run #(will be the forecast from 6-12 hours from the current date) ## Not run: my.pred <- model.parameters$pred[grep("06$", model.parameters$pred)] #What region of the atmosphere to get data for levels <- c("2 m above ground", "800 mb") #What data to return variables <- c("TMP", "RH") #Temperature and relative humidity #Get the data ## Not run: model.info <- GribGrab(urls.out[1], my.pred, levels, variables) #Extract the data ## Not run: model.data <- ReadGrib(model.info$file.name, levels, variables) #Reformat it ## Not run: model.grid <- ModelGrid(model.data, c(0.5, 0.5)) #Show an image of world temperature at ground level ## Not run: image(model.grid$z[2, 1,,]) #Archived Data Extraction #This is sometimes in GRIB1 format #This example is in GRIB1 abbrev <- "gfsanl" model.date <- 20040302 #March 2, 2004 model.run <- 18 #1800 GMT model run pred <- 0 #Analysis ## Not run: grib.info <- ArchiveGribGrab(abbrev, model.date, model.run, pred, file.type = "grib1") ## End(Not run) ## Not run: model.data <- ReadGrib(grib.info$file.name, c("1000 mb"), c("TMP"), file.type = "grib1") ## End(Not run)
This routine simplifies the rapid generation of data for specific points on the Earth's surface.
RTModelProfile(model.url, pred, levels, variables, lon, lat, resolution, grid.type, model.domain = NULL, spatial.average = FALSE, verbose = TRUE)RTModelProfile(model.url, pred, levels, variables, lon, lat, resolution, grid.type, model.domain = NULL, spatial.average = FALSE, verbose = TRUE)
model.url |
The address of a model download page, probably from |
pred |
The requested model prediction. |
levels |
A list of model levels to get for the profile. |
variables |
A list of model variables to download. |
lon |
Longitudes of points of interest. |
lat |
Latitudes of points of interest. |
resolution |
Resolution of model, in degrees if Lat/Lon, in kilometers if cartesian, as a 2 element vector (ZONAL, MERIDIONAL) |
grid.type |
If the model is gridded in Lat/Lon or cartesian units. Use |
model.domain |
A four element vector of latitudes and longitudes that defines a rectangular area to get data for.
If |
spatial.average |
If |
verbose |
If |
It is much more efficient to download a large chunk of data and extract profile points from that as opposed to downloading individual small model chunks in the vicinity of each point of interest. That is why I developed this function.
profile$profile.data |
Table of requested values, with rows corresponding to the requested levels. |
profile$spatial.averaging |
What kind of spatial interpolation was used, if any, for the profile calculations. |
profile$pred |
The model prediction used for generating the profile. |
profile$model.date |
When the model was run. |
profile$variables |
Model variables, in the order presented in |
profile$levels |
Model levels, in the order presented in |
Daniel C. Bowman [email protected]
GetClosestGFSForecasts, BuildProfile
#Get temperature profiles in Pantego, Chapel Hill, and Asheville, NC #First, define each location lon <- c(-76.662819, -79.052094, -82.550011) lat <- c(35.589446, 35.907562, 35.591994) #Get latest GFS 0.5 model, use analysis forecast ## Not run: model.url <- CrawlModels(abbrev = "gfs_hd", depth = 1)[1] pred <- ParseModelPage(model.url)$pred[1] ## End(Not run) #Get levels pressure <- c(1, 2, 3, 5, 7, 10, 20, 30, 50, 70, seq(100, 1000, by = 25)) levels <- paste(pressure, " mb", sep = "") #Variables - temperature and height only variables <- c("TMP", "HGT") #Resolution of GFS is 0.5 x 0.5 degree resolution <- c(0.5, 0.5) grid.type <- "latlon" #Get data ## Not run: profile <- RTModelProfile(model.url, pred, levels, variables, lon, lat, resolution, grid.type, spatial.average = TRUE) #Plot it plot(c(-100, 50), c(0, 50000), type = "n", xlab = "Temperature (C)", ylab = "Height (m)", main = paste("GFS", profile$model.date, "GMT Analysis Forecast")) for(k in seq_len(3)) { points(profile$profile.data[[k]][,2] - 273.15, profile$profile.data[[k]][,1], pch = k, col = k) } legend("topright", pch = 1:3, col = 1:3, legend = c( "Pantego, NC", "Chapel Hill, NC", "Asheville, NC")) ## End(Not run)#Get temperature profiles in Pantego, Chapel Hill, and Asheville, NC #First, define each location lon <- c(-76.662819, -79.052094, -82.550011) lat <- c(35.589446, 35.907562, 35.591994) #Get latest GFS 0.5 model, use analysis forecast ## Not run: model.url <- CrawlModels(abbrev = "gfs_hd", depth = 1)[1] pred <- ParseModelPage(model.url)$pred[1] ## End(Not run) #Get levels pressure <- c(1, 2, 3, 5, 7, 10, 20, 30, 50, 70, seq(100, 1000, by = 25)) levels <- paste(pressure, " mb", sep = "") #Variables - temperature and height only variables <- c("TMP", "HGT") #Resolution of GFS is 0.5 x 0.5 degree resolution <- c(0.5, 0.5) grid.type <- "latlon" #Get data ## Not run: profile <- RTModelProfile(model.url, pred, levels, variables, lon, lat, resolution, grid.type, spatial.average = TRUE) #Plot it plot(c(-100, 50), c(0, 50000), type = "n", xlab = "Temperature (C)", ylab = "Height (m)", main = paste("GFS", profile$model.date, "GMT Analysis Forecast")) for(k in seq_len(3)) { points(profile$profile.data[[k]][,2] - 273.15, profile$profile.data[[k]][,1], pch = k, col = k) } legend("topright", pch = 1:3, col = 1:3, legend = c( "Pantego, NC", "Chapel Hill, NC", "Asheville, NC")) ## End(Not run)
Discover all links on a given web page, follow each one, and recursively scan every link found. Return a list of web addresses whose pages contain no links.
WebCrawler(url, depth = NULL, verbose = TRUE)WebCrawler(url, depth = NULL, verbose = TRUE)
url |
A URL to scan for links. |
depth |
How many links to return.
This avoids having to recursively scan hundreds of links.
Defaults to |
verbose |
Print out each link as it is discovered.
Defaults to |
CrawlModels uses this function to get all links present on a model page.
urls.out |
A list of web page addresses, each of which corresponds to a model instance. |
While it might be fun to try WebCrawler on a large website such as Google, the results will be unpredictable and perhaps disastrous if depth is not set.
This is because there is no protection against infinite recursion.
Daniel C. Bowman [email protected]
#Find the first 10 model runs for the #GFS 0.5x0.5 model ## Not run: urls.out <- WebCrawler( "http://nomads.ncep.noaa.gov/cgi-bin/filter_gfs_hd.pl", depth = 10) ## End(Not run)#Find the first 10 model runs for the #GFS 0.5x0.5 model ## Not run: urls.out <- WebCrawler( "http://nomads.ncep.noaa.gov/cgi-bin/filter_gfs_hd.pl", depth = 10) ## End(Not run)