That's a pretty good summary but add in the sealing off of the Lugg: a 1,083km2 tributary supporting a large salmon population with a poorly designed flood defense scheme and the concreting in of a fish pass in the early 80's.

The Wye has 1,564km of potential juvenile nursery stream. In 1995, 767km of this was behind impassable barriers and farming and forestry was doing its best to extinguish salmon populations from the rest. Only 12% (190km) was working well!

Combined with a fall in marine survival, that especially affecting the spring salmon the Wye run was composed off and you have the perfect storm.

After 20 years of ever increasing effort the freshwater is in a much better state. When last assessed at end of 2014:
43% (669km) was working well (good or excellent juvenile densities)
25% (391km) was in recovery (limiting factor dealt with and populations building)
32% remains impacted or severely impacted

Now that acid waters, fish passage, Sheep dip pollution and most of the degraded habitat has been dealt with and a new more sympathetic agreement for public water abstraction in place. The remaining major pressures come from agriculture, particularly the massive soil loss in Herefordshire and Monmouthshire and the explosion of poultry and maize plus to a lesser extent, forestry with the deep drainage and afforestation of the headwater blanket bogs which has a large part to play in reduced summer flows.

As for why the Wye never switched to a late run. At the time of the switch from spring to summer/autumn salmon runs the part of the river dominated by late running fish on similar large rivers (the lower third and main stem) was rendered fishless by farming practices. We lost salmon production in the main stem below Ross in the 1980's, below Hereford in the 90's and below Hay in the 00's as the abuse of the catchment's soils cranked up.

A lot of work with farmers and regulators is now turning this around but there remains still some way to go.