How can I find the nearest distance between points in a 2D array and a 3D matrix grid?

Question

I have buoy data as an array of longitudes and latitudes for 30 days, however I would like to find the closest distance between the buoys' locations and the 0% sea ice concentration for each day. The sea ice concentration data is a 3D matrix, however the space dimensions are in x y coordinates not latitudes and longitudes. I have converted all the concentrations above 0% to Nan. I am not to sure now how to locate the closest latitude+longitude points of the sea ice to each point along the buoy trajectory.

This is my ice dataset:

Dimensions: (time: 363, x: 2528, y: 2656)

Coordinates:

• y (y) int16 1 2 3 4 5 6 ... 2652 2653 2654 2655 2656

• x (x) int16 1 2 3 4 5 6 ... 2524 2525 2526 2527 2528

• time (time) datetime64[ns] 2017-01-01T12:00:00 ... 2017-12-30T12:00:00

  longitude (time, y, x) float32 dask.array

  latitude (time, y, x) float32 dask.array

Data variables:

sea_ice_concentration  (time, y, x) float32 dask.array<shape=(363, 2656, 2528), chunksize=(1, 2656, 2528)>

land                   (time, y, x) int8 dask.array<shape=(363, 2656, 2528), chunksize=(1, 2656, 2528)>

Answer 1

You could try using Djikstras shortest path algorithm. The link contains an example in python which could be your starting point.

You will then need to convert your results, which will be in x,y dimensions to latitude/longitude.