If you look at the documentation for readtable you'll see that you can read a sepcific sheet name from an excel spreadsheet.
To put together a string that you'll pass to this function, there are lot's of options, and it probably depends a little bit on how your data is stored in matlab. For instance:
a='Jan';
b=10;
c=50;
join([string(a) string(b) string(c)],'_')
sprintf('%s_%d_%d', a, b, c)
If your question is more about how to rethink your excel organization - you can import a lot of data pretty quickly. So is it maybe more convenient to keep month, wind speed, humudity as columns of one big sheet and just import it all at once to MATLAB? I generally think it's more easy to get subsets and move data around in MATLAB than excel, because you can automate it so easily with code, so if I can get all my data there early it's easier than multiple rounds of importing. But that may be because I have lots of experience. Either way, if you can separate out file IO from data processing that seems good!
