Mcmc convergence dignostics plot


I got fowllowing mcmc convergence dignostics plot (plz see attached file).
As you can see, red and bule lines begin to diverge from around 60,000 iterations.

I found log posterior kernel’s mode by mode_cpmpute=9, since Prof. Pfeifer recommended this for global mode finder in this site’s other posts.
Then, it looked considerably successful until around 60,000 iterations.

However, attached plot casts doubt on convergence.

Do I need more iteration or Can I use the results before 60,000 iterations, or do I need to seek another ways?

Thanks in advance.
allv_rot822_mdiag.pdf (5.8 KB)

Please take a look at the trace plots. You should be able to easily spot what is going on.