The grid values should not be considered as representing the weather conditions at the exact location of the grid point. They should be considered as a time-space average within a two- or three-dimensional grid box. The discrepancy between the forecast grid-point value and the verifying observed average value can be both systematic and non-systematic:
- systematic errors reflect the limitations of the NWP model’s ability to simulate the physical and dynamic properties of the system.
- non-systematic errors reflect synoptic phase and intensity errors (as indicated by the left hand green arrow in Fig3.2-1).
- systematic and non-systematic errors occur when the NWP output is verified against point observations. The NWP output may not be representative of the location, height, aspect of the observation or capture sub-grid scale variability.
Fig3.2-1: Comparison between NWP model output and observations ought ideally to follow a two-step procedure:
- first step: compare grid point average to observation area average. The systematic errors are then due to model shortcomings; the non-systematic errors stem from synoptic phase and intensity errors.
- second step: compare the systematic errors between observation average and point observation. The systematic errors come from station representativeness (i.e. the location, height and aspect of the observation). The non-systematic errors come from sub-grid scale variability.
Fig3.2-2: In reality, the comparison between NWP and observations must for simplicity bypass the area average stage. This results in the systematic and non-systematic errors arising from distinctly different sources. The effects related to the two green arrows in Fig3.2-1 are here combined into one.
Systematic errors maybe due to model deficiencies and/or observational representativeness. These can be partly corrected by statistical means (e.g. model output statistics MOS). A series of forecasts also helps with dealing with uncertainty.
Non-systematic synoptic errors can be dampened by different ensemble approaches (e.g. medium range ensemble, probability considerations, forecast error growth).
Sub-grid variability (notably for rainfall but other parameters too) can be addressed by downscaling.
Downscaling converts:
- the average grid box probability density functions from raw ENS into
- "point rainfall probability density functions" for points within the grid box.
New downscaling techniques are being developed accordingly (see for example the Point Rainfall product).