Your expanded Hlog plot looks quite satisfactorily smooth and without any major deformities.
Considering where the road-works have occurred in relationship to the PCP and its fibre twin, I would agree with your suggestion. It does appear that the elevated noise level, thus depressed SNR and synchronisation speed, is a secondary effect arising from those circuits carrying xDSL services of exchange origin.
You mention "exposing the cables". From my understanding of your description, I would be very surprised to know that the E-side cables were not ducted. Was there any evidence of ducting that had been broken away? Perhaps those exposed cables were DIG (direct in ground) D-side cables, as there is no logical reason why D-side cables should not travel in the direction towards the exchange. Considering the PCP to be just a marker on a map, D-side cables should be expected to radiate in a full 360 degrees from that point.