Why aren't my variables the same dimension?

Hi All,

I was hoping a new pair of eyes could find what is wrong with the following code:

    v = landPercent(x1:x2,y1:y2)>80;

    SWNA_S = zeros(length(time),1);

    for i = 1:length(time);

    nonZero = nonzeros(v.*pr(x1:x2,y1:y2,i));

    Area = nonzeros(v.*areacell(x1:x2,y1:y2));

    sumArea = sum(Area);

    prArea = Area .* nonZero;

    sum_prArea = sum(prArea);

    SWNA_S(i) = sum_prArea/sumArea;

    end

For some reason the nonZero and Area variables are different dimensions at i=425. I tried putting an index for Area: Area = nonzeros(v.*areacell(x1:x2,y1:y2,i)); But then it says that the index exceeds the matrix dimension. I think I'm on the right path with that approach, but wonder if anyone could help me figure out how to put the correct index in that doesn't exceed the matrix dimensions?

Thanks.

[Walter Roberson]

We do not know size(pr) or size(areacell) nor do we know if they are functions.

[Shannon]

Good point. pr is a 192x145x1872 (1872 dimension is time) array of precipitation values and area cell is a 192x145 array of grid cell area values. I'm trying to calculate the values of each variable across an averaging region called SWNA_S. I am then weighting the pr value in each grid cell within SWNA_S according to each grid cell area, which is why I multiply Area * nonZero, which need to be the same size. I'm not sure why they aren't the same size in some time steps, because I am extracting those values from (x1:x2,y1:y2) for both pr and areacell. Thanks!

[Walter Roberson]

The nonzeros() selects for nonzero values. Your code is assuming that the zeros come only from v, but if that is not true, if there is a 0 in pr, then there would be fewer nonzeros() for that layer than you expect. The code I suggest handles it a different way, by using v as a logical index.

Answer 1

[Walter Roberson]

pr(x1:x2,y1:y2,425) contains a 0 in a position where v is true. In the iterations before that, that was not true.

With the extra 0, the nonzeros of the multiplication are different than the nonzero of the areacell.

To repair this:

before the loop calculate

areaxy = areacell(x1:x2,y1:y2);
Area = areaxy(v);
sumArea = sum(Area);

and inside the loop,

pri = pr(x1:x2,y1:y2,i);
nonZero = pri(v);

There is no need to recalculate Area or sumArea inside the loop as you do not change the array and you do not change the indices.

[Shannon]

Hi Walter, that worked like a charm. Thanks!!