But when I look at various communities of Tulu region I really wonder if that was the case. Consider the case of Bunts, erstwhile landowning community. The majority of Bunts have surname "Shetty". Shetty or its variants Setty, Chetty in South Indian society is the title of merchants. However, Bunts were typically considered as cultivators and never merchant community. In my opinion, changes in socio-political situation might have led many members of this community to cease their trading activities.
The western coast had an active trading relationship with Romans and Semites(probably Chinese too) dating as back as 2000 years. It has been observed that in Kerala society generally merchants from East coast(Tamil Nadu) migrated to North Kerala region and engaged in trading activities. What about Tulu region just north to it?
Considering the position of Tulu language in Dravidian language tree, I would envisage a situation where merchants from Telugu region(Godavari region) migrating to coastal Karnataka region. As time passed and trading activities came down, these Shettys, remained only as landowners. The migration of Shettys(and other people) belonging to SD-II family probably occurred before interior Karnataka region was occupied or Proto-Kannada-Tamil region was still confined to Krishna river region. Of course, this requires a mature Proto-SD-II cultural region in Godavari basin long before Kaveri region (Tamil) . Probably, around 1000-1500 BCE. I don't think this situation impossible considering Proto-Dravidians were already a civilized community by the time they reached South India. Of course, South Indian culture was an amalgamation of Dravidian, IA, Semitic and Austro-Asiatic elements from the beginning.
Note: I believe both IA and Dravidian speakers, had they been only wheat eaters, might have faced tribalization in tropical South India. Michale Witzel et al. argue that the word for "rice" in Dravidian languages is a borrow from either IA or Austro-Asiatic. So, it is highly possible that Dravidian speakers might have been at a greater risk of tribalization as wheat does not grow in South India(if I go by Jared Diamond).However, this theory requires Tulu being part of SD-II family(which includes Telugu). Probably, it's a mistake including it in SD-I family. I think Tulu might have been part of Telugu family but for the influence of SD-I family it has become a separated, but curiously unrelated, branch in that family. I mean linguists start with Proto-Tamil-Kannada and not with Proto-Tulu-Kannada.
Probably, we can check this further by population genetics. If 30-40% of male lineages in this region belongs to Haplogroup R2(my clan) just like coastal Andhra(Godavari basin) then we can confidently say the male migration to this region was from SD-II family and Tulu should be part of Telugu family. Of course, we have to also look for all those Roman(R1b), Pre-Islamic Semitic (E3b, J1) lineages among Tulu castes.