Table 5.

Stage 4 model performance (predicting daily 200 m residuals with an ensemble): 10-fold cross-validated performance across all years (2000 – 2016), overall and by month and season (residual scale).

	R_Tmin			R_Tmean			R_Tmax
	R²	RMSE	MAE	R²	RMSE	MAE	R²	RMSE	MAE
Overall	0.79	0.63	0.39	0.79	0.41	0.26	0.85	0.51	0.31

Jan	0.84	0.56	0.34	0.82	0.40	0.24	0.85	0.48	0.27
Feb	0.82	0.59	0.36	0.81	0.39	0.24	0.84	0.49	0.28
Mar	0.80	0.63	0.39	0.79	0.40	0.26	0.83	0.50	0.30
Apr	0.77	0.63	0.40	0.76	0.39	0.25	0.83	0.51	0.31
May	0.77	0.60	0.37	0.76	0.38	0.24	0.84	0.51	0.31
Jun	0.77	0.62	0.40	0.79	0.39	0.25	0.87	0.52	0.33
Jul	0.76	0.66	0.43	0.77	0.42	0.28	0.86	0.55	0.35
Aug	0.77	0.67	0.44	0.78	0.41	0.28	0.87	0.54	0.34
Sep	0.77	0.69	0.46	0.75	0.42	0.29	0.84	0.54	0.34
Oct	0.78	0.65	0.41	0.76	0.42	0.27	0.82	0.52	0.32
Nov	0.80	0.61	0.37	0.79	0.41	0.25	0.81	0.50	0.29
Dec	0.83	0.60	0.37	0.84	0.43	0.27	0.84	0.52	0.31

Winter	0.83	0.58	0.36	0.83	0.41	0.25	0.84	0.50	0.28
Spring	0.78	0.62	0.39	0.77	0.39	0.25	0.84	0.51	0.31
Summer	0.76	0.65	0.42	0.78	0.41	0.27	0.86	0.54	0.34
Fall	0.78	0.65	0.41	0.77	0.42	0.27	0.82	0.52	0.32