Strengths: A well-written investigation. Good use of references to support an original inquiry.
Points for consideration:
Be careful saying “there is significant evidence” without a test. Stationarity is a slippery thing (e.g., Notes Q3.5).
Trend: model is missing error term. More importantly, if the AIC is based on a white noise error model (which seems to be the case) that is problematic.
Several roots are very close to the unit circle.