Observations
The observations are input to numerical models, to national analyses and as the basis for a local very short range forecast (up to 12 hours) as well as for nowcasting (up to 6 hours). The difficulty lies in the relatively high space resolution and of course in the measurement errors as they appear for instance with precipitation.
The characteristics for an observation network (based on the weather service of Austria) are:
- Space resolution: depends on the network but usually the distance between stations is in the order of some tens to some hundred kilometers; in mountaineous areas the space resolution of mountain stations is very bad
- Time resolution: with manual observations: 1 hour; with automatic information: continuosly but visualisation interval is usually 5 to 10 minutes
- Forecast period: observations, therefore no forecast data; but obeservations are the input for forecasts based on persistence and the nowcasting for the very next hours. They are of course also inut data for all numerical models
"If there is such a bad spacial resoltion of observation stations how do you get good values for locations in between? Can you merely interpolate?"
"Yes, you found exactly the weakness. A mere interpolation will in most cases not be correct. You must take into account orography and possibly other relevant data sources. So parameters like radar and satellites but also the model analyses have to be combined to get an optimum analysis and we think that we can get close to the (not known) reality. But you never will arrive at a complete correct analysis."
"The next set of data is Radar. Follow
this link
to learn more!"