Efficient Method for Multiple Gridded Interpolation

I have scattered points of rain guage stations with random missing values. I also have NDVI values at these stations (complete records). The data is from January -Decemeber, 2016.

I want to interpolate both variables indepedently at daily time scale excluding any NaN stations for BOTH variables (i.e. exclude Nan from rainfall and corresponding NDVI). They will be saved as a Geotiff. This is my code for a single grid.

I want to do this in a loop with NaN excluded at each time step.

%%% load data
load data_now.dat
[lon,lat] = ndgrid(lon,lat); 
lonobs = data_now(:,1);
latobs = data_now(:,2);
J=  data_now(:,3);  
G=data_now(:,4) 
% lon and lat are read from a netcdf file
lon = ncread('important_gh.nc','lon');
lat = ncread('important_gh.nc','lat');
%Interpolate
Rainq = griddata(latobs,lonobs,J,lat,lon);
NDVIq = griddata(latobs,lonobs,G,lat,lon);

Thanks

Answer 1

[Sindar]

I don't know how fast it will be, but I think this'll work

% lon and lat are read from a netcdf file
lon = ncread('important_gh.nc','lon');
lat = ncread('important_gh.nc','lat');
[lon,lat] = ndgrid(dl.lon,dl.lat);

data_files = ...
for ind=1:length(data_files)
    % load into a variable
    data_now = load(data_files{ind});
    
    lonobs = data_now(:,1);
    latobs = data_now(:,2);
    J= data_now(:,3);
    G= data_now(:,4);
    % find valid locations 
    idx = (~isnan(J)) & (~isnan(G));
    
    %Interpolate, dropping nan locations
    Rainq = griddata(latobs(idx),lonobs(idx),J(idx),lat,lon);
    NDVIq = griddata(latobs(idx),lonobs(idx),G(idx),lat,lon);
    % save
end