From point observations to regionally comprehensive climate grids
Gridded data are produced using regional interpolation. It generates an estimate on the regional distribution of a weather variable. Interpolation produces information on climate also for locations where no meteorological observations have been made.
Meteorological observations on grid
Gridded data are commonly used in climate research. Gridded data refers to a uniformly graduated grid. For each grid point, the value of the required variable, e.g. temperature, is calculated using the regional interpolation method. The grid most used in Finland's climate studies is 10 x 10 kilometres (Figure 1). The interpolation method is called 'kriging'. Maps for instance on temperature distribution in Finland can be produced with the help of gridded data.
Figure 1. 10 x 10 kilometre grid square in Finland.
© Ilmatieteen laitos
The value of a variable for each grid point is calculated on the basis of measurements by nearby observation stations. This calculation method takes account of the impact of geography and coast and water bodies. The assumption is that points located close to each other are more likely to have similar climate than points further away. Climate grids provide assessments on the regional distribution of weather variables even for locations on which no meteorological observations are available.
The number of observation stations as a quality indicator
How well the grid actually describes the actual distribution of temperature or rain depends, above all, on the number of observation stations available. A large number of observation stations produce more accurate estimates, whereas a result calculated on the basis of a sparse network of stations describes regional characteristics less accurately.
Figure 2. Finland's annual mean temperature (map to the left) and precipitation sum illustrated on the basis of grid material.
© Ilmatieteen laitos
The variable to be analysed is significant, too. In comparison with precipitation, temperature is fairly evenly distributed and a sparser network of stations will suffice to facilitate its interpolation. However, precipitation may vary considerably from one location to another. Therefore, it is challenging to include high and local accumulated precipitations resulting from intense rain showers in the grid. Therefore, interpolated rain distribution attempts to balance high amounts of precipitation.