Nothing is wrong with the plots, I would caution against assuming something must be wrong just because you don't quite understand what it shows. It generates the wrong mental state and approach to solving the problem.

In this case, you should learn about dotplots and what they show, then the image would make more sense.

That being said, let's analyze the situation. The dot plots ought to be similar, right? After all, you align the same sequences first in nucleotide and then in peptide space. It would be a much bigger concern if the alignments were wildly different.

Then just as expected, the protein alignments show much longer sections of similarity than the nucleotide plots. The distance between two dots is the region of similarity. Note how many more dots the nucleotide-based alignments have. So it all behaves as expected. Protein alignments find longer regions of similarity.

Google the concept "dot plot alignment" and read up on what dotplots show.

As with everything that has to do with alignments, dot plots can be very simple yet may also show surprisingly complex patterns that are not so easy to explain fully.