Those look like rounding artifacts; the pic I posted just used some Gaussian smoothing, but there are (way more complex) methods for handling them correctly: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2847785/

The overall function is a p-curve, here's a paper on it: http://journals.sagepub.com/doi/pdf/10.1177/1745691614553988