CDO

I would find it very helpful to have an enhanced cdo regres function. If a second output file could be generated with the standard error of b, that would allow easy significance testing of the slope.