Kernel density estimation on maps

Hi everyone!

I am in need of your help the following challenge. I have a 3D data (83 rows X 92 columns X 480 months), which are time-series maps of temperature. I need to produce maps of hotspot areas by considering different number of grids for all 480 maps. How can I apply Kernel density estimation on each of these maps?

Kindly,

Chane