Mask 2D numpy array

Question

I want to apply mask on 2D numpy array. But it does not work correctly. Suppose I have

val(lat, lon)  ---> my 2D array (20, 30)


Mask_lat = np.ma.masked_array(lat, mask=latmask)  ---> masked lat (5,)


Mask_lon = np.ma.masked_array(lon, mask =lonmask)   ---> masked lon (8,)


Maks_val = np.ma.masked_array(val, mask=mask_lat_lon) ---> ?

I do not know how can I pass a correct mask_lat_lon to have masked val (5,8). I would appreciate if one guides me.

Thank you in advance.

Answer 1

If I understand your question correctly, you have two 1D arrays that represent y and x (lat and long) positions in a 2D array. You want to mask a region based on the x/y position in the 2D array.

The key part to understand is that mask for a 2D array is also 2D.

For example, let's mask a single element of a 2D array:

import numpy as np

z = np.arange(20).reshape(5, 4)
mask = np.zeros(z.shape, dtype=bool)

mask[3, 2] = True

print z
print np.ma.masked_array(z, mask)

This yields:

[[ 0  1  2  3]
 [ 4  5  6  7]
 [ 8  9 10 11]
 [12 13 14 15]
 [16 17 18 19]]

[[ 0  1  2  3]
 [ 4  5  6  7]
 [ 8  9 10 11]
 [12 13 -- 15]
 [16 17 18 19]]

In your case, you have two 1D x and y arrays that you need to create a 2D mask from. For example:

import numpy as np

x = np.linspace(-85, -78, 4)
y = np.linspace(32, 37, 5)
z = np.arange(20).reshape(5, 4)

xmask = (x > -82.6) & (x < -80)
ymask = (y > 33) & (y < 35.6)

print xmask
print ymask

We'd then need to combine them into a single 2D mask using broadcasting:

mask = xmask[np.newaxis, :] & ymask[:, np.newaxis]

Slicing with newaxis (or None, they're the same object) adds a new axis at that position, turning the 1D array into a 2D array. It you have seen this before, it's useful to take a quick look at what xmask[np.newaxis, :] and ymask[:, np.newaxis] look like:

In [14]: xmask
Out[14]: array([False, False,  True, False], dtype=bool)

In [15]: ymask
Out[15]: array([False,  True,  True, False, False], dtype=bool)

In [16]: xmask[np.newaxis, :]
Out[16]: array([[False, False,  True, False]], dtype=bool)

In [17]: ymask[:, np.newaxis]
Out[17]:
array([[False],
       [ True],
       [ True],
       [False],
       [False]], dtype=bool)

mask will then be (keep in mind that True elements are masked):

In [18]: xmask[np.newaxis, :] & ymask[:, np.newaxis]
Out[18]:
array([[False, False, False, False],
       [False, False,  True, False],
       [False, False,  True, False],
       [False, False, False, False],
       [False, False, False, False]], dtype=bool)

Finally, we can create a 2D masked array from z based on this mask:

arr = np.masked_array(z, mask)

Which gives us our final result:

 [[ 0  1  2  3]
  [ 4  5 --  7]
  [ 8  9 -- 11]
  [12 13 14 15]
  [16 17 18 19]]