# 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 non-cyclic coordinates, extrapolation is used so that the first and last cells have their coordinates in their centres.

For instance, the Voronoi cells created for non-cyclic coordinates 0, 40, 90, 180 have bounds (-20, 20), (20, 65), (65, 135), (135, 225). If those coordinates were instead cyclic with a period of 360, then the bounds would be (-90, 20), (20, 65), (65, 135), (135, 270).

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, setting bound=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 of bound=-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.

• cf.TimeDuration

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) (see cf.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 setting flt=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, setting cellsize=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, or max=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, or min=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 co-located 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 non-contiguous 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 date-time coordinates can use cf.TimeDuration cell sizes:

>>> d = cf.DimensionCoordinate(
...   data=cf.Data(['1984-12-01', '1984-12-02'], 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(['1985-01-16: 12:00', '1985-02-15'], 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(['1984-06-01', '1985-06-01'], 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)]]