cf.DimensionCoordinate.create_bounds¶

DimensionCoordinate.
create_bounds
(bound=None, cellsize=None, flt=0.5, max=None, min=None)[source]¶ Create cell bounds.
Creates new cell bounds, irrespective of whether the cells already have cell bounds. The new bounds are not set on the dimension coordinate construct, but if that is desired they may always be added with the
set_bounds
method, for instance:>>> b = d.create_bounds() >>> d.set_bounds(b)
By default, Voronoi cells are created by defining cell bounds that are half way between adjacent coordinate values. For noncyclic coordinates, extrapolation is used so that the first and last cells have their coordinates in their centres.
For instance, the Voronoi cells created for noncyclic coordinates
0, 40, 90, 180
have bounds(20, 20), (20, 65), (65, 135), (135, 225)
. If those coordinates were instead cyclic with a period of360
, then the bounds would be(90, 20), (20, 65), (65, 135), (135, 270)
.See also
 Parameters
 bound: data_like scalar, optional
If set to a value larger (smaller) than the largest (smallest) coordinate value, then bounds are created which include this value and for which each coordinate is in the centre of its bounds. This can be useful for recreating the bounds of vertical coordinates from Earth system models.
 Parameter example:
For coordinates
1, 2, 10
, settingbound=0.5
will result in bounds of(0.5, 1.5), (1.5, 2.5), (2.5, 17.5)
. Parameter example:
Beware that if bound is too large or small, then some returned bounds values will by physically incorrect. For coordinates
1, 2, 10
, the setting ofbound=2
will result in bounds of(2, 4), (4, 0), (0, 20)
.
 cellsize: optional
Define the size of each cell that is created. Created cells are allowed to overlap and do not have to be contigious. The cellsize parameter may be one of:
data_like scalar
Defines the cell size, either in the same units as the coordinates or in the units provided. Note that in this case, the position of each coordinate within the its cell is determined by the flt parameter.

A time duration defining the cell size. Only applicable to reference time coordinates. It is possible to “anchor” the cell bounds via the
cf.TimeDuration
instance parameters. For example, to specify cell size of one calendar month, starting and ending on the 15th day:cellsize=cf.M(day=15)
(seecf.M
for details). Note that the flt parameter is ignored in this case.
 flt: array_like scalar, optional
When creating cells with sizes specified by the cellsize parameter, define the fraction of each cell which is less than its coordinate value. By default flt is 0.5, so that each cell has its coordinate at its centre. Ignored if cellsize is not set.
 Parameter example:
For coordinates
1, 2, 10
settingflt=0.25
will result in bounds of(0.75, 1.75), (1.75, 2.75), (9.75, 10.75)
. Parameter example:
For decreasing coordinates
2, 0, 12
, settingcellsize=6, flt=0.9
will result in bounds of(2.6, 3.4), (0.6, 5.4), (11.4, 17.4)
.
 max: data_like scalar, optional
Limit the created bounds to be no more than this number.
 Parameter example:
To ensure that all latitude bounds are at most
90
:max=90
, ormax=cf.Data(90, 'degrees_north')
.
 min: data_like scalar, optional
Limit the created bounds to be no less than this number.
 Parameter example:
To ensure that all latitude bounds are at least
90
:min=90
, ormin=cf.Data(90, 'degrees_north')
.
 Returns
Bounds
The new coordinate cell bounds.
Examples
Create bounds for Voronoi cells:
>>> d = cf.DimensionCoordinate(data=[0.0, 40, 90, 180]) >>> b = d.create_bounds() >>> print(b.array) [[20. 20.] [ 20. 65.] [ 65. 135.] [135. 225.]] >>> d = d[::1] >>> b = d.create_bounds() >>> print(b.array) [[225. 135.] [135. 65.] [ 65. 20.] [ 20. 20.]]
Cyclic coordinates have a colocated first and last bound:
>>> d.period(360) >>> d.cyclic(0) set() >>> b = d.create_bounds() >>> print(b.array) [[270. 135.] [135. 65.] [ 65. 20.] [ 20. 90.]]
Cellsize units must be equivalent to the coordinate units, or if the cell has no units then they are assumed to be the same as the coordinates:
>>> d = cf.DimensionCoordinate(data=cf.Data([0, 2, 4], 'km')) >>> b = d.create_bounds(cellsize=cf.Data(2000, 'm')) >>> print(b.array) [[1 1] [ 1 3] [ 3 5]] >>> b = d.create_bounds(cellsize=2) >>> print(b.array) [[1 1] [ 1 3] [ 3 5]]
Cells may be noncontiguous and may overlap:
>>> d = cf.DimensionCoordinate(data=[1.0, 2, 10]) >>> b = d.create_bounds(cellsize=1) >>> print(b.array) [[ 0.5 1.5] [ 1.5 2.5] [ 9.5 10.5]] >>> b = d.create_bounds(cellsize=5) >>> print(b.array) [[1.5 3.5] [0.5 4.5] [ 7.5 12.5]]
Reference datetime coordinates can use
cf.TimeDuration
cell sizes:>>> d = cf.DimensionCoordinate( ... data=cf.Data(['19841201', '19841202'], dt=True) ... ) >>> b = d.create_bounds(cellsize=cf.D()) >>> print(b.datetime_array) [[cftime.DatetimeGregorian(1984, 12, 1, 0, 0, 0, 0) cftime.DatetimeGregorian(1984, 12, 2, 0, 0, 0, 0)] [cftime.DatetimeGregorian(1984, 12, 2, 0, 0, 0, 0) cftime.DatetimeGregorian(1984, 12, 3, 0, 0, 0, 0)]] >>> b = d.create_bounds(cellsize=cf.D(hour=12)) >>> print(b.datetime_array) [[cftime.DatetimeGregorian(1984, 11, 30, 12, 0, 0, 0) cftime.DatetimeGregorian(1984, 12, 1, 12, 0, 0, 0)] [cftime.DatetimeGregorian(1984, 12, 1, 12, 0, 0, 0) cftime.DatetimeGregorian(1984, 12, 2, 12, 0, 0, 0)]]
Cell bounds defined by calendar months:
>>> d = cf.DimensionCoordinate( ... data=cf.Data(['19850116: 12:00', '19850215'], dt=True) ... ) >>> b = d.create_bounds(cellsize=cf.M()) >>> print(b.datetime_array) [[cftime.DatetimeGregorian(1985, 1, 1, 0, 0, 0, 0) cftime.DatetimeGregorian(1985, 2, 1, 0, 0, 0, 0)] [cftime.DatetimeGregorian(1985, 2, 1, 0, 0, 0, 0) cftime.DatetimeGregorian(1985, 3, 1, 0, 0, 0, 0)]] >>> b = d.create_bounds(cellsize=cf.M(day=20)) >>> print(b.datetime_array) [[cftime.DatetimeGregorian(1984, 12, 20, 0, 0, 0, 0 cftime.DatetimeGregorian(1985, 1, 20, 0, 0, 0, 0] [cftime.DatetimeGregorian(1985, 1, 20, 0, 0, 0, 0 cftime.DatetimeGregorian(1985, 2, 20, 0, 0, 0, 0]]
Cell bounds defined by calendar years:
>>> d = cf.DimensionCoordinate( ... data=cf.Data(['19840601', '19850601'], dt=True) ... ) >>> b = d.create_bounds(cellsize=cf.Y(month=12)) >>> print(b.datetime_array) [[cftime.DatetimeGregorian(1983, 12, 1, 0, 0, 0, 0) cftime.DatetimeGregorian(1984, 12, 1, 0, 0, 0, 0)] [cftime.DatetimeGregorian(1984, 12, 1, 0, 0, 0, 0) cftime.DatetimeGregorian(1985, 12, 1, 0, 0, 0, 0)]]