Creating 2D Matrix from latitude/longitude values

I have three 1D arrays of the same dimensions (14114 x 1): latitude, longitude, and radar reflectivity. The 1D arrays are not sorted in any particular way (i.e., by latitude or longitude). Here are sample values for the first three elements:

latitude (1,1): 29.28

longitude (1,1): -92.5

radar (1,1): 25

latitude (2,1): 29.5

longitude (2,1): -88.48

radar (2,1): 22.5

latitude (3,1): 29.5

longitude (3,1): -88.06

radar (3,1): 20.5

I need to create a 2D lat/lon matrix containing the radar values from the 1D array that spans from longitude [-92.5:0.02:-87.5] in the X and latitude [27.2:0.02:32] in the Y, but has the appropriate radar reflectivity value from the 1D radar array. Lat/lon pairs not in the 1D radar array should be stored as 0 in the 2D matrix.

Appreciate any help you can provide.

Answer 1

[KSSV]

Let (x,y,r) be your lon,lat, radar reflectivity values in column.

xi = -92.5:0.02:-87.5 ; 
yi = 27.2:0.02:32 ; 
[X,Y] = meshgrid(xi,yi) ; 
F = scatteredInterpolant(x,y,r) ; 
R = F(X,Y) ; 
pcolor(X,Y,R) ;
shading interp ;
colorbar

[John Cruce]

The scatteredInterpolant interpolates between points. Is there a way to only project actual values from r onto the new meshgrid, but have missing r values (lat/lon pairs with no r value) set to 0? Thanks for the help.

[KSSV]

You can use knnsearch to find the nearest points, assign values to these and rest to zeros.

[John Cruce]

I did a help on knnsearch. I'm still unsure how to reflect those reflectivity values on to the meshgrid. What should I be looking to for my values of X and Y with knnsearch?