SARIMA model predicted values are not on the actual time series (level too low)

[Phil Decker]

Hey, we want to use the SARIMA model to clean up the time series forecast more values.
The code represents what we have tried. Unfortunately, the trend line is represented graphically lower than the SARIMA model.
hat can we do to get the graphs to the same height. Is there a better alternative?

Thanks for any help.

gen month=tm(2011m1) +_n-1
format month%tm
tsset month
keep if tin(2011m1,2019m12)

gen Dpassenger= passengernumbers- L12.passengernumbers
pac Dpassenger
ac Dpassenger

arima Dpassenger, arima (2,0,0) sarima (1,0,0,12)
estat ic

tsappend, add(60)
predict f_Dpassenger, xb dynamic(108)
predict mse_Dpassengerr, mse dynamic(108)
predict f_passengernumbers, y dynamic(tm(2019m12))
predict mse_passengernumbers, mse dynamic(tm(2019m12))
tsline passengernumbers f_passengernumbers