RasterDataSources.jl

RasterDataSourcesModule

RasterDataSources.jl

CI codecov.io

RasterDataSources downloads raster data for local use or for integration into other spatial data packages, like Rasters.jl. The collection is largely focussed on datasets relevant to ecology, but will have a lot of crossover with other sciences.

Currently sources include:

SourceURLStatus
CHELSAhttps://chelsa-climate.orgBioClim, BioClimPlus, and Climate
WorldClimhttps://www.worldclim.orgClimate, Weather, BioClim, and Elevation
EarthEnvhttp://www.earthenv.orgLandCover and HabitatHeterogeneity
AWAPhttp://www.bom.gov.au/jsp/awap/index.jspComplete
ALWBhttp://www.bom.gov.au/water/landscape/Complete
SRTMhttps://www2.jpl.nasa.gov/srtm/Complete
MODIShttps://modis.ornl.govComplete (beta)

Please open an issue if you need more datasets added, or (even better) open a pull request following the form of the other datasets where possible.

Retrieving data

Usage is generally via the getraster method - which will download the raster data source if it isn't available locally, or simply return the path/s of the raster file/s:

julia> using RasterDataSources

julia> getraster(WorldClim{Climate}, :wind; month=1:12)
12-element Array{String,1}:
 "/home/user/Data/WorldClim/Climate/wind/wc2.1_10m_wind_01.tif"
 "/home/user/Data/WorldClim/Climate/wind/wc2.1_10m_wind_02.tif"
 "/home/user/Data/WorldClim/Climate/wind/wc2.1_10m_wind_03.tif"
 "/home/user/Data/WorldClim/Climate/wind/wc2.1_10m_wind_04.tif"
 "/home/user/Data/WorldClim/Climate/wind/wc2.1_10m_wind_05.tif"
 "/home/user/Data/WorldClim/Climate/wind/wc2.1_10m_wind_06.tif"
 "/home/user/Data/WorldClim/Climate/wind/wc2.1_10m_wind_07.tif"
 "/home/user/Data/WorldClim/Climate/wind/wc2.1_10m_wind_08.tif"
 "/home/user/Data/WorldClim/Climate/wind/wc2.1_10m_wind_09.tif"
 "/home/user/Data/WorldClim/Climate/wind/wc2.1_10m_wind_10.tif"
 "/home/user/Data/WorldClim/Climate/wind/wc2.1_10m_wind_11.tif"
 "/home/user/Data/WorldClim/Climate/wind/wc2.1_10m_wind_12.tif"

Installation and setup

Install as usual with:

] add RasterDataSources

To download data you will need to specify a folder to put it in. You can do this by assigning the environment variable RASTERDATASOURCES_PATH:

ENV["RASTERDATASOURCES_PATH"] = "/home/user/Data/"

This can be put in your startup.jl file or the system environment.

RasterDataSources was based on code from the SimpleSDMDataSoures.jl package by Timothée Poisot.

source

getraster

RasterDataSources.jl only exports a single function, getraster.

Specific implementations are included with each source, below.

Data sources

ALWB

RasterDataSources.ALWBType
ALWB{Union{Deciles,Values},Union{Day,Month,Year}} <: RasterDataSource

Data from the Australian Landscape Water Balance (ALWB) data source.

See: www.bom.gov.au/water/landscape

The dataset contains NetCDF files. They have a time dimension so that multiple dates are stored in each file.

The available layers are: (:rain_day, :s0_pct, :ss_pct, :sd_pct, :sm_pct, :qtot, :etot, :e0, :ma_wet, :pen_pet, :fao_pet, :asce_pet, :msl_wet, :dd), available in daily, monthly and annual resolutions, and as Values or relative Deciles.

getraster for ALWB must use a date keyword to specify the date to download.

See the getraster docs for implementation details.

source
RasterDataSources.getrasterMethod
getraster(source::Type{<:ALWB{Union{Deciles,Values},Union{Day,Month,Year}}}, [layer]; date)

Download ALWB weather data from www.bom.gov.au/water/landscape as values or deciles with timesteps of Day, Month or Year.

Arguments

  • layer: Symbol or Tuple of Symbol from (:rain_day, :s0_pct, :ss_pct, :sd_pct, :sm_pct, :qtot, :etot, :e0, :ma_wet, :pen_pet, :fao_pet, :asce_pet, :msl_wet, :dd). Without a layer argument, all layers will be downloaded, and a NamedTuple of paths returned.

Keywords

  • date: a DateTime, AbstractVector of DateTime or a Tuple of start and end dates. For multiple dates, a Vector of multiple filenames will be returned. ALWB is available with a daily, monthly, and yearly, timestep.

Example

This will return the file containing annual averages, including your date:

julia> getraster(ALWB{Values,Year}, :ss_pct; date=Date(2001, 2))
"/your/RASTERDATASOURCES_PATH/ALWB/values/month/ss_pct.nc"

Returns the filepath/s of the downloaded or pre-existing files.

source

AWAP

RasterDataSources.AWAPType

AWAP <: RasterDataSource

Daily weather data from the Australian Water Availability Project, developed by CSIRO.

See: www.csiro.au/awap

The available layers are: (:solar, :rainfall, :vprpress09, :vprpress15, :tmin, :tmax).

source
RasterDataSources.getrasterMethod
getraster(source::Type{AWAP}, [layer]; date)

Download data from the AWAP weather dataset, from www.csiro.au/awap.

The AWAP dataset contains ASCII .grid files.

Arguments

  • layer Symbol or Tuple of Symbol for layers in (:solar, :rainfall, :vprpress09, :vprpress15, :tmin, :tmax). Without a layer argument, all layers will be downloaded, and a NamedTuple of paths returned.

Keywords

  • date: a DateTime, AbstractVector of DateTime or a Tuple of start and end dates. For multiple dates, A Vector of multiple filenames will be returned. AWAP is available with a daily timestep.

Example

Download rainfall for the first month of 2001:

julia> getraster(AWAP, :rainfall; date=Date(2001, 1, 1):Day(1):Date(2001, 1, 31))

31-element Vector{String}:
 "/your/path/AWAP/rainfall/totals/20010101.grid"
 "/your/path/AWAP/rainfall/totals/20010102.grid"
 ...
 "/your/path/AWAP/rainfall/totals/20010131.grid"

Returns the filepath/s of the downloaded or pre-existing files.

source

CHELSA

RasterDataSources.CHELSAType
CHELSA{Union{BioClim,BioClimPlus,Climate,<:Future}} <: RasterDataSource

Data from CHELSA, currently implements the BioClim BioClimPlus, and Climate variables for current and future conditions.

See: chelsa-climate.org for the dataset, and the getraster docs for implementation details.

source
RasterDataSources.getrasterMethod
getraster(source::Type{CHELSA{BioClim}}, [layer]; version = 2, [patch]) => Union{Tuple,String}

Download CHELSA BioClim data from chelsa-climate.org.

Arguments

  • layer: Integer or tuple/range of Integer from (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19), or Symbols form (:bio1, :bio2, :bio3, :bio4, :bio5, :bio6, :bio7, :bio8, :bio9, :bio10, :bio11, :bio12, :bio13, :bio14, :bio15, :bio16, :bio17, :bio18, :bio19). Without a layer argument, all layers will be downloaded, and a NamedTuple of paths returned.

Keyword arguments

  • version: Integer indicating the CHELSA version, currently either 1 or 2.
  • patch: Integer indicating the CHELSA patch number. Defaults to the latest patch (V1.2 and V2.1)

Returns the filepath/s of the downloaded or pre-existing files.

source
RasterDataSources.getrasterMethod
getraster(T::Type{CHELSA{Future{Climate}}}, [layer]; date, month) => String

Download CHELSA Climate data, choosing layers from: (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19).

See the docs for Future for model choices.

Without a layer argument, all layers will be downloaded, and a NamedTuple of paths returned.

Keywords

  • date: a Date or DateTime object, a Vector, or Tuple of start/end dates. Note that CHELSA CMIP5 only has two datasets, for the periods 2041-2060 and 2061-2080. CMIP6 has datasets for the periods 2011-2040, 2041-2070, and 2071-2100. Dates must fall within these ranges.
  • month: the month of the year, from 1 to 12, or a array or range of months like 1:12.

Example

using Dates, RasterDataSources
getraster(CHELSA{Future{Climate, CMIP6, GFDLESM4, SSP370}}, :prec; date = Date(2050), month = 1)
source

EarthEnv

RasterDataSources.getrasterMethod
getraster(source::Type{EarthEnv{HabitatHeterogeneity}}, [layer]; res="25km")

Download EarthEnv habitat heterogeneity data.

Arguments

  • layer: Symbol or Tuple of Symbol from (:cv, :evenness, :range, :shannon, :simpson, :std, :Contrast, :Correlation, :Dissimilarity, :Entropy, :Homogeneity, :Maximum, :Uniformity, :Variance). Without a layer argument, all layers will be downloaded, and a NamedTuple of paths returned.

Keywords

  • res: String chosen from ("1km", "5km", "25km"), defaulting to "25km".

Returns the filepath/s of the downloaded or pre-existing files.

source
RasterDataSources.getrasterMethod
getraster(T::Type{EarthEnv{LandCover}}, [layer]; discover=false) => Union{Tuple,String}

Download EarthEnv landcover data.

Arguments

  • layer: Integer or tuple/range of Integer from (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12), or Symbols from (:needleleaf_trees, :evergreen_broadleaf_trees, :deciduous_broadleaf_trees, :other_trees, :shrubs, :herbaceous, :cultivated_and_managed, :regularly_flooded, :urban_builtup, :snow_ice, :barren, :open_water). Without a layer argument, all layers will be downloaded, and a NamedTuple of paths returned.

Keywords

  • discover::Bool: whether to download the dataset that integrates the DISCover model.

Returns the filepath/s of the downloaded or pre-existing files.

source

WorldClim

RasterDataSources.getrasterMethod
getraster(T::Type{WorldClim{BioClim}}, [layer::Union{Tuple,AbstractVector,Integer}]; res::String="10m") => Union{Tuple,AbstractVector,String}

Download WorldClim BioClim data.

Arguments

  • layer: Integer or tuple/range of Integer from (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19). or Symbols from (:bio1, :bio2, :bio3, :bio4, :bio5, :bio6, :bio7, :bio8, :bio9, :bio10, :bio11, :bio12, :bio13, :bio14, :bio15, :bio16, :bio17, :bio18, :bio19). Without a layer argument, all layers will be downloaded, and a NamedTuple of paths returned.

Keywords

  • res: String chosen from ("30s", "2.5m", "5m", "10m"), "10m" by default.

Returns the filepath/s of the downloaded or pre-existing files.

source
RasterDataSources.getrasterMethod
getraster(T::Type{WorldClim{Weather}}, [layer::Union{Tuple,Symbol}]; date) => Union{String,Tuple{String},Vector{String}}

Download WorldClim Weather data, for layer/s in: (:tmin, :tmax, :prec). Without a layer argument, all layers will be downloaded, and a NamedTuple of paths returned.

Keywords

  • date: a Date or DateTime object, a Vector of dates, or Tuple of start/end dates. WorldClim Weather is available with a daily timestep.

Returns the filepath/s of the downloaded or pre-existing files.

source
RasterDataSources.getrasterMethod
getraster(T::Type{WorldClim{Climate}}, [layer::Union{Tuple,Symbol}]; month, res::String="10m") => Vector{String}

Download WorldClim Climate data.

Arguments

  • layer Symbol or Tuple of Symbol from (:tmin, :tmax, :tavg, :prec, :srad, :wind, :vapr). Without a layer argument, all layers will be downloaded, and a NamedTuple of paths returned.

Keywords

  • month: Integer or AbstractArray of Integer. Chosen from 1:12.
  • res: String chosen from ("30s", "2.5m", "5m", "10m"), "10m" by default.

Returns the filepath/s of the downloaded or pre-existing files.

source
RasterDataSources.getrasterMethod
getraster(T::Type{WorldClim{Elevation}}, [layer::Union{Tuple,Symbol}]; res::String="10m") => Union{Tuple,AbstractVector,String}

Download WorldClim Elevation data.

Arguments

  • layer: Symbol or Tuple of Symbol from (:elev,). Without a layer argument, all layers will be downloaded, and a NamedTuple of paths returned.

Keywords

  • res: String chosen from ("30s", "2.5m", "5m", "10m"), "10m" by default.

Returns the filepath/s of the downloaded or pre-existing files.

source

MODIS

RasterDataSources.getrasterMethod
getraster(T::Union{Type{<:ModisProduct}, Type{MODIS{X}}}, [layer::Union{Tuple,AbstractVector,Integer, Symbol}]; kwargs...) => Union{String, AbstractVector, NamedTuple}

Download MODIS data for a given ModisProduct as ASCII raster(s).

Arguments

  • layer: Integer or tuple/range of Integer or Symbols. Without a layer argument, all layers will be downloaded, and a NamedTuple of paths returned.

Available layers for a given product can be looked up using RasterDataSources.layerkeys(T::Type{<:ModisProduct}).

Keywords

  • lat and lon: Coordinates in decimal degrees of the approximate center of the raster. The MODIS API will try to match its pixel grid system as close as possible to those coordinates.

  • km_ab and km_lr: Half-width and half-height of the raster in kilometers (kilometers above/below and left/right). Currently only Integer values are supported, up to 100.

  • date: String, Date, DateTime, AbstractVector of dates or Tuple of a start and end date for the request. Strings should be in format YYYY-MM-DD but can be in similar formats as long as they are comprehensible by Dates.Date. The available date interval for MODIS is 16 days, reset every first of January.

Example

Download 250m NDVI in the western part of Britanny, France, from winter to late spring, 2002:

julia> getraster(MOD13Q1, :NDVI; lat = 48.25, lon = -4, km_ab = 50, km_lr = 50, date = (Date(2002,1,1), Date(2002,6,1)))

    10-element Vector{String}:
    "/your/path/MODIS/MOD13Q1/250m_16_days_NDVI/47.8313_-4.5899_2002-01-01.asc"
    ...
    "/your/path/MODIS/MOD13Q1/250m_16_days_NDVI/47.8313_-4.5899_2002-05-25.asc"

Will attempt to download several files, one for each date and layer combination, and returns the filepath/s of the downloaded or pre-existing files. Coordinates in the file names correspond to the lower-left corner of the raster.

source
RasterDataSources.ModisProductType
ModisProduct <: RasterDataSet

Abstract supertype for MODIS/VIIRS products.

Usage

Some commonly used products are MOD13Q1 (250m resolution MODIS vegetation indices) and VNP13A1 (500m resolution VIIRS vegetation indices). Refer to the MODIS documentation for detailed product information.

source
RasterDataSources.layerkeysMethod
layerkeys(T::Type{<:ModisProduct}) => Tuple

Tuple of Symbols corresponding to the available layers for a given product. May issue a request to MODIS server to get the layers list, or might just read this information if the correctly named file is available.

source

Datasets

RasterDataSources.BioClimType
BioClim <: RasterDataSet

BioClim datasets. Usually containing layers from 1:19. These can also be accessed with :bioX, e.g. :bio5.

They do not usually use month or date keywords, but may use date in past/future scenarios.

Currently implemented for WorldClim and CHELSA as WorldClim{BioClim}, CHELSA{BioClim} and CHELSA{Future{BioClim, args..}}.

See the getraster docs for implementation details.

source
RasterDataSources.ClimateType
Climate <: RasterDataSet

Climate datasets. These are usually months of the year, not specific dates, and use a month keyword in getraster. They also use date in past/future scenarios.

Currently implemented for WorldClim and CHELSA as WorldClim{Climate}, CHELSA{Climate} and CHELSA{Future{Climate, args..}}.

See the getraster docs for implementation details.

source
RasterDataSources.WeatherType
Weather <: RasterDataSet

Weather datasets. These are usually large time-series of specific dates, and use a date keyword in getraster.

Currently implemented for WorldClim and CHELSA as WorldClim{Weather}, and CHELSA{Weather}

See the getraster docs for implementation details.

source
RasterDataSources.FutureType
Future{<:RasterDataSet,<:CMIPphase,<:ClimateModel,<:ClimateScenario}

Future climate datasets specified with a dataset, phase, model, and scenario.

Type Parameters

RasterDataSet

Currently BioClim and Climate are implemented for the CHELSA data source.

CMIPphase

Can be either CMIP5 or CMIP6.

ClimateModel

Climate models can be chosen from:

`forCMIP5`;

`forCMIP6`;"

ClimateScenario

CMIP5 Climate scenarios are all RepresentativeConcentrationPathway and can be chosen from: RCP26, RCP45, RCP60, RCP85

CMIP6 Climate scenarios are all SharedSocioeconomicPathway and can be chosen from: SSP126, SSP245, SSP370, SSP585

However, note that not all climate scenarios are available for all models.

Example

using RasterDataSources
dataset = Future{BioClim, CMIP5, BNUESM, RCP45}
# output
Future{BioClim, CMIP5, BNUESM, RCP45}

Currently Future is only implented for CHELSA

datasource = CHELSA{Future{BioClim, CMIP5, BNUESM, RCP45}}
source

Models, phases and scenarios for Future data.

RasterDataSources.CMIPphaseType
CMIPphase

Abstract supertype for phases of the CMIP, the Coupled Model Intercomparison Project.

Subtypes are CMIP5 and CMIP6.

source

Other

Internal interface

These methods are not exported at this stage, but are for the most part internally consistent. Any new sources added to the package should use these methods in a consistent way for readability, consistency and the potential to use them for other things later.

RasterDataSources.layerkeysFunction
layerkeys(T::Type{<:ModisProduct}) => Tuple

Tuple of Symbols corresponding to the available layers for a given product. May issue a request to MODIS server to get the layers list, or might just read this information if the correctly named file is available.

source
RasterDataSources.rasternameFunction
rastername(source::Type, [layer]; kw...)

Returns the name of the file, without downloading it.

Arguments are the same as for getraster

Returns a String or multiple Strings.

source
RasterDataSources.rasterpathFunction
rasterpath(source::Type, [layer]; kw...)

Returns the name of the file, without downloading it.

Arguments are the same as for getraster

Returns a String or multiple Strings.

source
RasterDataSources.rasterurlFunction
rasterurl(source::Type, [layer]; kw...)

If the file has a single url, returns it without downloading.

Arguments are the same as for getraster.

Returns a URIs.jl URI or mulitiple URIs.

source
RasterDataSources.zipnameFunction
zipname(source::Type, [layer]; kw...)

If the url is a zipped file, returns its name.

Arguments are as the same for getraster where possible.

Returns a String or multiple Strings.

source
RasterDataSources.zippathFunction
zippath(source::Type, [layer]; kw...)

If the url is a zipped file, returns its path when downloaded. (This may not exist after extraction with getraster)

Arguments are the same as for getraster where possible.

Returns a String or multiple Strings.

source
RasterDataSources.zipurlFunction
zipurl(source::Type, [layer]; kw...)

If the url is a zipped file, returns its zip path without downloading.

Arguments are the same as for getraster where possible.

Returns a URIs.jl URI or mulitiple URIs.

source

Internal MODIS interface

Unlike all the other currently supported data sources, MODIS data is not available online in raster file format. Building rasters out of the available information therefore requires internal functions that are not exported. They might be extended as needed if other similar sources get supported.

Requesting to server and building raster files

RasterDataSources.modis_requestFunction
modis_request(T::Type{<:ModisProduct}, args...)

Lowest level function for requests to modis server. All arguments are assumed correct.

Arguments

  • layer: String matching the "exact" layer name (i.e. as it is written in the MODIS dataset itself) for the given product. e.g. "250m_16_days_EVI".

  • lat, lon, km_ab, km_lr in correct types

  • from, to: Strings of astronomical dates for start and end dates of downloaded data, e.g. "A2002033" for "2002-02-02"

Returns a NamedTuple of information relevant to build a raster header, and a Vector of Dicts containing raster data, directly downloaded from MODIS. Those will almost always directly be passed to RasterDataSources.process_subset

source
RasterDataSources.process_subsetFunction
process_subset(T::Type{<:ModisProduct}, subset::Vector{Any}, pars::NamedTuple)

Process a raw subset and argument parameters and create several raster files. Any already existing file is not overwritten.

For each band, a separate folder is created, containing a file for each of the required dates. This is inspired by the way WorldClim{Climate} treats the problem of possibly having to download several dates AND bands.

Can theoretically be used for MODIS data that does not directly come from RasterDataSources.modis_request, but caution is advised.

Returns the filepath/s of the created or pre-existing files.

source

Miscellaneous

RasterDataSources.sinusoidal_to_latlonFunction
sinusoidal_to_latlon(x::Real, y::Real)

Convert x and y in sinusoidal projection to lat and lon in dec. degrees

The EPSG.io API takes care of coordinate conversions. This is not ideal in terms of network use but guarantees that the coordinates are correct.

source