Is it possible to "raster paint" outlier  data (ie  data that you do not believe) then convert back to vector?

Example:  2009 wet holes that have brought down the typically highest yielding portion of a field. Suppose it is then drained.   I need to fix that management zone.

  I think you could build robust enough logic in the equation - but I would also use "raster paint" to build in some "omit plots" or negative controls in anything I was doing.

So, can you touch up these management zones or the final Rx?

Thanks.