I'm trying to forecast future sales via SARIMA model. Since I have multiple time series, I have used a grid search technique to find the best parameters for the SARIMA model. But when looking at the fitted values the fit shows a huge dip at September. Details: Model fitted on $2$ years of weekly sales. SARIMA: $(0, 1, 0) \times (1, 1, 0, 52)$ AIC=$4.000$
Can someone explain why is the SARIMA model fit shows such dips. Ideally it show consider it as an outlier.
