Bias correction
Bias correction is performed on 0.2 deg grid, starting every 3 months and with the duration of 120 days to cover whole year 2011. SST bias fields are derived over the whole domain (EUR) and remapped to the original dataset resolution (0.05 deg for SEVIRI and 0.02 deg for Metop-A/AVHRR). Validation was performed also on the whole domain (EUR), but also on all defined sub-domains allowing inter-regional and inter-sensor analysis.
Following bias fields were derived: a) Metop-A corrected with AATSR, b) SEVIRI corrected with AATSR, c) SEVIRI corrected with Metop-A. All bias corrections are performed with two different approaches, based on DINEOF reconstruction of difference (RECDIFF) and based on differences of DINEOF reconstructions (DIFFREC). Here we show results divided in three groups:
- Val - time-series and spatial plots of results (besst adjusted, cms adjusted and without adjustment) based on validation with AATSR and buoy data
- Biases - time-series and spatial plots of derived biases (cms and besst based)
- Stat - table of overall statistics
| DIFFREC | RECDIFF | |||||
|---|---|---|---|---|---|---|
| Val | Biases | Stat | Val | Biases | Stat | |
| Metop-A vs. AATSR | O | O | O | O | O | O |
| SEVIRI vs. AATSR | O | O | O | O | O | O |
| SEVIRI vs. Metop-A | O | O | O | O | O | O |
Note: Links (O) open in a new window/tab.
