MOESM4 of Assessing Lagrangian inverse modelling of urban anthropogenic CO2 fluxes using in situ aircraft and ground-based measurements in the Tokyo area

Additional file 4: Figure S4. S4a) Source receptor relationship matrix calculated with different winds. Left: panel ERA Interim. Center panel: WRF. Right panel: difference. Rows (vertical axis) represent measurements grouped by site and release time (D = Dodaira, K = Kisai, T1 = Tsukuba 25 m, T2 = Tsukuba 100 m, T3 = Tsukuba 200 m, C = CONTRAIL). Columns (horizontal axis) represent the Tokyo Bay Area spatiotemporal surface fluxes between 2007-01-11 00:00 and 2007-01-13 24:00 JST. This corresponds to trajectory ensembles released during 2007-01-13 and integrated 48 h backwards in time. Source regions are aggregated by prefecture in the Kanto area for this particular case totaling 9 regiones (7 prefectures, rest of the land and sea) to improve the visualization as higher resolution SRRs are usually sparser. The gaps (matrix entries with SRR = 0) correspond to source regions not reached by the backward trajectories, i.e. for which the measurements provide no constraint. Time resolution of the fluxes is 3 hours here, but can change between 1 hour and static. Color scale represents the source-receptor relationship value in hours (i.e., the residence time: a factor that depends on the footprint layer height gives the sensitivity in e.g. mg CO2 m2s−1)−1. The integrated difference is of the order of 15 % of the source-receptor relationship calculated either with WRF of with ERA Interim winds. S4b) Three hourly footprints corresponding to the SRR described above. First columns: ECMWF winds. Second column: WRF winds. Third column: Difference ECMWF minus WRF. The shift North East– South West is apparent. The wind situation depicted is not uncommon. Compare with Fig. 6, where the misalignment is artificially produced by shifting the SRR directly.