Table 3 Statistical results for regression models’ estimations of cloudless DLR and the statistically significant models are in italics
From: Deep learning and regression modelling of cloudless downward longwave radiation
Model | Training phase (n = 143) | Testing phase (n = 62) | ||||||
|---|---|---|---|---|---|---|---|---|
r | RMSE | MBE | ts | r | RMSE | MBE | ts | |
Brunt [17] | 0.7985 | 13.2107 | − 4.4736 | 4.2887 | 0.7330 | 13.5892 | − 4.1961 | 2.5356 |
Efimova [35] | 0.8086 | 28.0357 | 25.1381 | 20.1708 | 0.7496 | 28.6559 | 25.5969 | 15.5186 |
Swinbank [111] | 0.2981 | 20.6139 | 1.2239 | 0.7088 | 0.1184 | 21.0120 | 1.3252 | 0.4935 |
Idso and Jackson [49] | 0.2986 | 21.4172 | 5.4152 | 3.1251 | 0.1183 | 21.9029 | 5.5732 | 2.0550 |
Maykut and Church [69] | 0.2994 | 35.5477 | − 25.8874 | 12.6628 | 0.1182 | 32.6729 | − 26.2765 | 10.5688 |
Brutsaert [18] | 0.8047 | 13.7358 | 5.5704 | 5.3054 | 0.7374 | 14.4938 | 5.9371 | 3.5071 |
Satterlund [99] | 0.2998 | 23.8285 | 12.8603 | 7.6663 | 0.1181 | 23.8656 | 12.7461 | 4.9339 |
Idso [48] | 0.3013 | 41.4920 | 36.3682 | 21.7740 | 0.1179 | 40.3991 | 36.2112 | 14.0946 |
Idso [48] | 0.8260 | 30.0430 | 26.9126 | 24.1021 | 0.7749 | 30.7321 | 27.4740 | 15.5819 |
Prata [89] | 0.7983 | 13.1831 | 4.3305 | 4.1589 | 0.7316 | 13.7783 | 4.6229 | 2.7818 |
Guest [42] | 0.2994 | 23.6458 | − 12.6583 | 7.5292 | 0.1182 | 23.7629 | − 12.8264 | 5.0079 |
Iziomon et al. [51] | 0.8066 | 12.9193 | − 3.9510 | 2.8919 | 0.7431 | 13.3529 | − 3.6223 | 2.8919 |
Present model | 0.8275 | 11.5847 | 0.2412 | 0.2490 | 0.7785 | 11.8883 | 0.3393 | 0.2230 |