Index

maximum(A::ClimGrid)

Compute the maximum value of ClimGrid A

mean(A::ClimGrid)

Compute the mean of ClimGrid A

merge(A::ClimGrid, B::ClimGrid)

Combines two ClimGrid. Based on the AxisArrays method.

minimum(A::ClimGrid)

Compute the minimum value of ClimGrid A

std(A::ClimGrid)

Compute the standard deviation of ClimGrid A

var(A::ClimGrid)

Compute the variance of ClimGrid A

annualmax(C::ClimGrid)

Annual maximum of array data.

Let data[i,j] be daily time serie on day i in year j. Extract the highest value for year j.

annualmean(C::ClimGrid)

Annual mean of array data.

Let data[i,j] be daily time serie on day i in year j. Calculate the mean value for year j.

annualmin(C::ClimGrid)

Annual minimum of array data.

Let data[i,j] be daily time serie on day i in year j. Extract the lowest value for year j.

annualsum(C::ClimGrid)

Annual sum of array data.

Let data[i,j] be daily time serie on day i in year j. Sums daily values for year j.

applymask(A::AbstractArray{N, n}, mask::AbstractArray{N, n})

This function applies a mask on the array A. Return an AbstractArray{N, n}.

  approx_surfacepressure(sealevel_pressure::ClimGrid, orography::ClimGrid, daily_temperature::ClimGrid)

Returns the approximated surface pressure (sp) (Pa) using sea level pressure (psl) (Pa), orography (orog) (m), and daily mean temperature (tas) (K).

$sp = psl * 10^{x}$

where $x = \frac{-orog}{18400 * tas / 273.15}$

customthresover(C::ClimGrid)

customthresover, annual number of days over a specified threshold.

Let TS[i,j] be a daily time serie value on day i in year j. Count the number of days where:

TS[i,j] > thres.

customthresunder(C::ClimGrid)

customthresover, annual number of days under a specified threshold.

Let TS[i,j] be a daily time serie value on day i in year j. Count the number of days where:

TS[i,j] < thres.

daysabove10(C::ClimGrid)

Annual number of days with temperature >= 10 Celsius. This function returns a ClimGrid.

diurnaltemperature(temperatureminimum::ClimGrid, temperaturemaximum::ClimGrid, α::Float64)

Returns an estimation of the diurnal temperature (temperature between 7:00 (7am) and 17:00 (5pm)). The estimation is a linear combination of the daily minimum temperature (temperatureminimum) and daily maximum temperature (temperaturemaximum). The value of α has to be estimated seperatly from observations and depends on the location. The daily max and min must be in the same unit and in Celsius or Kelvin The diurnal temperature returned is in the same units as the daily minimum temperature and daily maximum temperature.

$Tdiu = α * Tmin + (1 - α) * Tmax$

extractpoly(file::String, n::Int)

Returns the n-th polygon contained in file.

frostdays(C::ClimGrid)

FD, Number of frost days: Annual count of days when TN (daily minimum temperature) < 0 Celsius.

Let TN[i,j] be daily minimum temperature on day i in year j. Count the number of days where:

TN[i,j] < 0 Celsius.

icingdays(C::ClimGrid)

ID, Number of summer days: Annual count of days when TX (daily maximum temperature) < 0 degree Celsius.

Let TX[i,j] be daily maximum temperature on day i in year j. Count the number of days where:

TX[i,j] < 0 Celsius.

inpoly(p, poly::Matrix)

Determines if a point is inside a polygon.

• p – point (x,y) or [x,y]

• poly – polygon vertices [x1 x2 ... xn x1 y1 y2 ... yn y1]

Returns true if point has an odd winding number. This should label points as exterior which are inside outcrops. See test for a test.

Author: Github "Mauro3" / "Mauro"

inpolygrid(lon, lat, poly::AbstractArray{N,2} where N)

Used to test a grid of points. Returns a mask of ones and NaNs of the same size as lon and lat.

load(files::Array{String,1}, variable::String; poly = ([]), start_date::Date = Date(-4000), end_date::Date = Date(-4000), data_units::String = "")

Loads and merge the files contained in the arrar files.

load(file::String, variable::String; poly = Array{Float64}([]), start_date::Tuple, end_date::Tuple, data_units::String = "")

Returns a ClimGrid type with the data in file of variable var inside the polygon poly. Metadata is built-in the ClimGrid type, from the netCDF attributes.

Inside the ClimgGrid type, the data is stored into an AxisArray data type, with time, longitude/x and latitude/y dimensions.

The polygon provided should be in the -180, +180 longitude format. If the polygon crosses the International Date Line, the polygon should be splitted in multiple parts (i.e. multi-polygons).

Options for data_units are for precipitation : "mm", which converts the usual "kg m-2 s-1" unit found in netCDF files. For temperature : "Celsius", which converts the usual "Kelvin" unit.

Temporal subsetting can be done by providing start_date and end-date Tuples of length 1 (year), length 3 (year, month, day) or 6 (hour, minute, second).

Note: load uses CF conventions. If you are unable to read the netCDF file with load, the user will need to read it with low-level functions available in the NetCDF.jl package.

load2D(file::String, variable::String; poly=[], data_units::String="")

Returns a 2D array. Should be used for fixed data, such as orography

mapclimgrid(C::ClimGrid; region::String="auto", poly, level, mask, caxis, start_date::Tuple, end_date::Tuple, titlestr::String, surface::Symbol, ncolors::Int, center_cs::Bool, filename::String, cs_label::String)

Maps the time-mean average of ClimGrid C. If a filename is provided, the figure is saved in a png format.

Optional keyworkd includes precribed regions (keyword region, see list below), spatial clipping by polygon (keyword poly) or mask (keyword mask, an array of NaNs and 1.0 of the same dimension as the data in ClimGrid C), start_date and end_date. For 4D data, keyword level is used to map a given level (defaults to 1). caxis is used to limit the colorscale. ncolors is used to set the number of color classes (defaults to 12). Set center_cs to true to center the colorscale (useful for divergent results, such as anomalies, positive/negative temprature). cs_label is used for custom colorscale label.

Arguments for keyword region (and shortcuts)

• Europe ("EU")

• NorthAmerica ("NA")

• Canada ("CA")

• Quebec, QuebecNSP ("QC", "QCNSP")

• Americas ("Ams")

• World, WorldAz, WorldEck4 ("W", "Waz", "Weck4")

• Greenwich ("Gr")

Arguments for keyword surface

• :contour

• :contourf

• :pcolormesh

mapclimgrid(; region::String="auto", poly, level, mask, caxis, start_date::Date, end_date::Date)

Empty map generator, when called without a ClimGrid as the positional argument.

X, Y = meshgrid{T}(vx::AbstractVector{T}, vy::AbstractVector{T})

This function creates a 2-D mesh-grid in a format consistent with Matlab's function meshgrid(). XV and YV are vectors.

X, Y = ndgrid(XV, YV)

This function creates a 2-D mesh-grid in a format consistent with Matlab's function ndgrid(). XV and YV are vectors.

prcp1(C::ClimGrid)

Annual number with preciptation >= 1 mm. This function returns a ClimGrid.

qqmap(obs::Array{N, 1} where N, ref::Array{N, 1} where N, fut::Array{N, 1} where N; method="Additive", detrend=true, window=15, rankn=50, thresnan=0.1, keep_original=false, interp::Function = Linear(), extrap::Function = Flat())

Quantile-Quantile mapping bias correction for single vector. This is a low level function used by qqmap(A::ClimGrid ..), but can work independently.

qqmap(obs::ClimGrid, ref::ClimGrid, fut::ClimGrid; method="Additive", detrend=true, window::Int=15, rankn::Int=50, thresnan::Float64=0.1, keep_original::Bool=false, interp::Function = Linear(), extrap::Function = Flat())

Quantile-Quantile mapping bias correction. For each julian day of the year (+/- window size), a transfer function is estimated through an empirical quantile-quantile mapping.

The quantile-quantile transfer function between ref and obs is etimated on a julian day (and grid-point) basis with a moving window around the julian day. Hence, for a given julian day, the transfer function is then applied to the fut dataset for a given julian day.

Options

method::String = "Additive" (default) or "Multiplicative". Additive is used for most climate variables. Multiplicative is usually bounded variables such as precipitation and humidity.

detrend::Bool = true (default). A 4th order polynomial is adjusted to the time series and the residuals are corrected with the quantile-quantile mapping.

window::Int = 15 (default). The size of the window used to extract the statistical characteristics around a given julian day.

rankn::Int = 50 (default). The number of bins used for the quantile estimations. The quantiles uses by default 50 bins between 0.01 and 0.99. The bahavior between the bins is controlled by the interp keyword argument. The behaviour of the quantile-quantile estimation outside the 0.01 and 0.99 range is controlled by the extrap keyword argument.

thresnan::Float64 = 0.1 (default). The fraction is missing values authorized for the estimation of the quantile-quantile mapping for a given julian days. If there is more than treshnan missing values, the output for this given julian days returns NaNs.

keep_original::Bool = false (default). If keep_original is set to true, the values are set to the original values in presence of too many NaNs.

interp = Interpolations.Linear() (default). When the data to be corrected lies between 2 quantile bins, the value of the transfer function is linearly interpolated between the 2 closest quantile estimation. The argument is from Interpolations.jl package.

extrap = Interpolations.Flat() (default). The bahavior of the quantile-quantile transfer function outside the 0.01-0.99 range. Setting it to Flat() ensures that there is no "inflation problem" with the bias correction. The argument is from Interpolation.jl package.

qqmap(fut::ClimGrid, ITP::TransferFunction)

Quantile-Quantile mapping bias correction with a known transfert function. For each julian day of the year, use the right transfert function to correct model values.

Options

method::String = "Additive" (default) or "Multiplicative". Additive is used for most climate variables. Multiplicative is usually bounded variables such as precipitation and humidity.

qqmaptf(obs::ClimGrid, ref::ClimGrid; partition::Float64 = 1.0, detrend::Bool=true, window::Int64=15, rankn::Int64=50, thresnan::Float64=0.1, keep_original::Bool=false, interp = Linear(), extrap = Flat())

Transfer function based on quantile-quantile mapping bias correction. For each julian day, a transfer function is estimated through an empirical quantile-quantile mapping for the entire obs' ClimGrid extent. The quantile-quantile transfer function between ref and obs is etimated on a julian day basis with a moving window around the julian day. The transfer function can then be used to correct another dataset.

Options partition::Float64 = 1.0. The proportion of grid-points (chosen randomly) used for the estimation of the transfer function. A transfer function is estimated for every chosen grid-points (and julian day) and averaged for the entire obs ClimGrid extent.

method::String = "Additive" (default) or "Multiplicative". Additive is used for most climate variables. Multiplicative is usually bounded variables such as precipitation and humidity.

window::Int = 15 (default). The size of the window used to extract the statistical characteristics around a given julian day.

rankn::Int = 50 (default). The number of bins used for the quantile estimations. The quantiles uses by default 50 bins between 0.01 and 0.99. The bahavior between the bins is controlled by the interp keyword argument. The behaviour of the quantile-quantile estimation outside the 0.01 and 0.99 range is controlled by the extrap keyword argument.

interp = Interpolations.Linear() (default). When the data to be corrected lies between 2 quantile bins, the value of the transfer function is linearly interpolated between the 2 closest quantile estimation. The argument is from Interpolations.jl package.

extrap = Interpolations.Flat() (default). The bahavior of the quantile-quantile transfer function outside the 0.01-0.99 range. Setting it to Flat() ensures that there is no "inflation problem" with the bias correction. The argument is from Interpolation.jl package.

C = regrid(A::ClimGrid, londest::AbstractArray{N, 1} where N, latdest::AbstractArray{N, 1} where N)A

Interpolate ClimGrid A onto lat-lon grid defined by londest and latdest vector or array. If an array is provided, it is assumed that the grid is curvilinear (not a regular lon-lat grid) and the user needs to provide the dimension vector ("x" and "y") for such a grid.

C = regrid(A::ClimGrid, B::ClimGrid; method="linear", min=[], max=[])

Interpolate ClimGrid A onto the lon-lat grid of ClimGrid B, where A and B are ClimGrid. Available methods for interpolation are "linear" (default), "nearest" and "cubic".

Min and max optional keyword are used to constraint the results of the interpolation. For example, interpolating bounded fields can lead to unrealilstic values, such as negative precipitation. In that case, one would use min=0.0 to convert negative precipitation to 0.0.

shapefile_coords(poly::Shapefile.Polygon)

This function return the polygons contained in shp.shapes[i]. It returns the x and y coordinates vectors.

See also shapefile_coords_poly, which returns a polygon that ca be used for data extraction of the load.

shapefile_coords_poly(poly::Shapefile.Polygon)

Return the polygons contained in shp.shapes[i]. It returns an array containing the polygons.

See also shapefile_coords, which returns vectors as opposed to array. Returned polygon is consistent with the data extraction of the load function.

spatialsubset(C::ClimGrid, poly::Array{N, 2})

Returns the spatial subset of ClimGrid C. The spatial subset is defined by the polygon poly, defined on a -180, +180 longitude reference.

summerdays(C::ClimGrid)

SD, Number of summer days: Annual count of days when TX (daily maximum temperature) > 25 degree Celsius.

Let TX[i,j] be daily maximum temperature on day i in year j. Count the number of days where:

TX[i,j] > 25 Celsius.

function temporalsubset(C::ClimGrid, startdate::Date, enddate::Date)

Returns the temporal subset of ClimGrid C. The temporal subset is defined by a start and end date.

tropicalnights(C::ClimGrid)

TropicalNights, Number of tropical nights: Annual count of days when TN (daily maximum temperature) > 20 degree Celsius.

Let TN[i,j] be daily minimum temperature on day i in year j. Count the number of days where:

TN[i,j] > 20 Celsius.

vaporpressure(specific_humidity::ClimGrid, sealevel_pressure::ClimGrid, orography::ClimGrid, daily_temperature::ClimGrid)

Returns the vapor pressure (vp) (Pa) estimated with the specific humidity (q), the sea level pressure (psl) (Pa), the orography (orog) (m) and the daily mean temperature (tas) (K). An approximation of the surface pressure is first computed by using the sea level pressure, orography and the daily mean temperature (see approx_surfacepressure). Then, vapor pressure is calculated by:

$vp = \frac{q * sp}{q+0.622}$

vaporpressure(surface_pressure::ClimGrid, specific_humidity::ClimGrid)

Returns the vapor pressure (vp) (Pa) based on the surface pressure (sp) (Pa) and the specific humidity (q).

$vp = \frac{q * sp}{q+0.622}$

wbgt(diurnal_temperature::ClimGrid, vapor_pressure::ClimGrid)

Returns the simplified wet-bulb global temperature (wbgt) (Celsius) calculated using the vapor pressure (Pa) of the day and the estimated mean diurnal temperature (Celsius; temperature between 7:00 (7am) and 17:00 (5pm)).

$wbgt = 0.567 * Tday + 0.00393 * vp + 3.94$

plot(C::ClimGrid, titlefig::String, gridfig::Bool, label::String, color, lw, linestyle)

Plots the spatial average timeserie of ClimGrid C.

ClimGrid{A <: AxisArray}

In-memory representation of Climate Forecast netCDF files.

ClimGrid(data; longrid=[], latgrid=[], msk=[], grid_mapping=Dict(), dimension_dict=Dict(), model="NA", frequency="NA", experiment="NA", run="NA", project="NA", institute="NA", filename="NA", dataunits="NA", latunits="NA", lonunits="NA", variable="NA", typeofvar="NA", typeofcal="NA", varattribs=Dict(), globalattribs=Dict())

Constructor of the ClimGrid function. Data is an AxisArray. Everything else is optional, but usually needed for further processing (mapping, interpolation, etc...).

TransferFunction

In-memory representation of the transfer function used during quantile-quantile mapping bias correction.

TransferFunction(itp, method)

Create a TransferFunction used during quantile-quantile mapping bias correction. Contains the ITP function and the methode used (either Additive or Multiplicative)