How best to handle this data?

Hi,

I have a large amount of data that I need to work with and I'm looking for some advice on the best way to go about it...

I have 5 [130536 x 8] double class variables, each containing meterological data for certain times of the day. Columns 1-5 of each variable are the same, containing the index, pressure, altitude, latitude and longitude of the point in 3D space, followed by 3 unique columns in each variable, containing temperature, wind_u and wind_v components. E.g:

>> metdata_0909(1:5,:)

ans =

               0        850.00       4781.00         47.97         -9.50        269.03         14.74        -15.42
            1.00        850.00       4781.00         47.97         -9.39        269.10         14.83        -15.72
            2.00        850.00       4781.00         47.97         -9.28        269.17         14.94        -16.00
            3.00        850.00       4781.00         47.97         -9.17        269.24         15.10        -16.28
            4.00        850.00       4781.00         47.97         -9.06        269.25         15.25        -16.55

Each variable is only valid between a certain time of day.

I need to be able to input an altitude, latitude, longitude and time and whatever code I write needs to return an interpolated temperature, u and v components of the wind for the input location.

How/what is the best way of going about this? I have a fair few toolboxes so feel free to suggest anything that may be included in them.

J

[Jan]

Fortunately arrays with only some 100 thousands of elements are not considered "big" on modern computers.

The expression "valid between a certain time" sounds very funny.

Answer 1

[Jan]

At first avoid magic information in the names of the variables. Tricks like the "0909" makes the life so much harder without any benefit. So create a 3D array and keep the information hidden in the "0909" separately.

Then find the wanted altitude, latitude, longitude and time inside this array and the interpolation can be done by interp1 directly, as far as I understand. But it is not exactly clear to me, what you want to interpolate.