Are negative values in deseasonalized monthly rainfall acceptable?

I used the Stable Seasonal Filter to deseasonalized monthly rainfall data for 41 years using this method. However, there are a considerable amount of negative values in the resulting deseasonalized data, which is probably the effect of zero-rainfall/very-low-rainfall months. My question is, is this acceptable to be reported or should I adjust all negative values to become zero instead?

I'm going to use the resulting deseasonalized data to analyzed monthly and yearly trends. Hence, retiming it to total yearly rainfall would be affected by my decision to either use the negatives or adjust them to a realistic zero.

Thanks in advance.

[Rik]

This isn't strictly a Matlab question. Either choice can be supported, but I don't know what is most common in this field.