PM2.5 Prediction on EEA ground monitoring stations Flat Stations T+1 2022 (σz < 50 m, N = 2096)
8.94RMSECRAN-PM
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| CRAN-PMModel Type=Neural model, Resolution=1 km, Number of Parameters=96M2026.03 | 8.94 | 4.86 | 80 | -0.75 | |
| SimVPModel Type=Neural model, Resolution=25 km, Number of Parameters=4.9M2026.03 | 9.14 | 4.97 | 73 | -1.05 | |
| EarthformerModel Type=Neural model, Resolution=25 km, Number of Parameters=61M2026.03 | 9.16 | 5.03 | 72 | -1.12 | |
| ConvLSTMModel Type=Neural model, Resolution=25 km, Number of Parameters=8.2M2026.03 | 9.18 | 4.96 | 72 | -1.04 | |
| TopoFlowModel Type=Neural model, Resolution=25 km, Number of Parameters=61M2026.03 | 9.21 | 4.95 | 74 | -1.38 | |
| ClimaXModel Type=Neural model, Resolution=25 km, Number of Parameters=57M2026.03 | 9.22 | 5.01 | 71 | -1 | |
| CAMSModel Type=Physical model2026.03 | 10.83 | 6.26 | 58 | -3.81 |