It is a different model from the full model that took in account I believe 111 monthly data points between 2009 and 2019 (r2 is also different).
Okay, then you have a good case about this model's ability to forecast.
However, careful about "overfitting" and let's see about it in the next few years
