Autumn PNA gave no clear added winter PDO forecast skill in 1981-2025

We tested whether autumn Pacific–North American (PNA) circulation improves winter Pacific Decadal Oscillation (PDO) forecasts beyond autumn PDO and Nino 3.4. We used three fixed NOAA index tables. Each of 45 winter forecasts from 1981 to 2025 used only earlier winter targets for fitting. Adding autumn PNA increased mean squared error from 0.328746 to 0.349191, a 6.22% increase. The conditional 95% five-year block interval for error reduction was -19.81% to +4.57%. Trend, rolling training, separate monthly PDO controls and November PNA checks all increased point error. The separate-month control interval was below zero. The primary test gives no clear added forecast skill. It does not establish that atmospheric forcing is absent. This extends prior campaign work on target-month PNA information to a forecast with preceding-autumn inputs. Revised indices, 45 dependent test winters, simple models and uncertainty conditional on saved forecasts limit inference. Code, exact source paths and hashes, retrieval instructions, all predictions, score draws, a figure and a separate numerical audit are included. Raw source tables remain outside the package.

PDO · PNA · ENSO · seasonal forecast · atmospheric forcing · null result