Analyses of Alpine High-Resolution Rain-Gauge Observations for the MAP Special Observing Period

   Attention: the data from Croatia seems to be anticipated by 1 day. For day n, data from Croatia are for day n-1 (Ahrens and Beck, July 2003). Investigations are on the way.

This site provides digital access to an objective precipitation analysis for the Special Observing Period of the Mesoscale Alpine Program (MAP SOP, September 7 - November 15, 1999). The analysis is based on observations at about 6500 rain-gauge stations, covers the entire Alpine region (2.1-18.9W and 42.18-49.0N), including the ridge and the foreland, has a time resolution of one day and a space resolution of about 25 km. This data product was produced in a joint effort of the Institute for Atmospheric and Climate Science at ETH Zurich and the MAP Data Centre.

Some of the institutions providing the data have extensively checked the quality of the data supplied to the MDC. For the purpose of overall consistency, the dataset has been quality checked again, using automated tests and manual/visual proofing. The latter was based on the resulting analysis. The quality testing led to the exclusion of some of the original data. A list of the flagged stations and reports is provided in the quality test log.

The analysis was undertaken onto a regular longitude x latitude grid with a resolution of 0.3 x 0.22 degrees, corresponding to a grid-spacing of about 25 km. Grid-point values represent regional area mean values around each grid-point. A modified version of the SYMAP algorythm was adopted, similar to the previous Alpine precipitation analysis conducted for a climatological period. Details of the method and original references can be found in Frei and Schär (1998).

• Domain-mean precipitation in the LMTA region for IOP2B is larger by nearly 20%. (See the comparison of domain mean values for version 2.0 (red) and 1.5 (green)). LMTA domain is defined as (8.1-9.6E,45.48-46.58N).
• SOP mean precipitation of version 2.0 and 1.5 exhibits differences in fine-scale patterns especially over Italy. Larger values are found in the version 2.0 over the Bergamask Alps and parts of the Po valley, while lower values are found at the Ligurian coast. (See figures of SOP mean precipitation in version 2.0 and version 1.5 and the difference.)

Data version 2.0    Stations per grid-point    Search radius

When using the products of the Alpine precipitation climatology in publications (any media, scientific literature, the Internet or others) please refer to Frei and Häller (2001) which describes the details of the dataset and to Frei and Schär (1998) which contains an extensive description of the method of analysis.

• Users should carefully take into consideration the error sources discussed in Frei and Schär (1998), in Frei and Häller (2001) and in the Interpretation Hints on this site.

File Format:

• heading lines:

  head: data source and version number
  title: field (parameter) and unit
  time: period of the analysis
  dimensions: dimensions of the analysis grid in longitude (x), latitude (y) and time (t),
  loc of (1,1): lower left corner in degrees lat/lon,
  loc of (nx,ny): upper right corner in degrees lat/lon.
  missing data flag: indication used to demark missing analysis (insufficient observations)
  

• data block:

  The data block follows the heading lines and comprises the two-dimensional analyses grids sequentially in time. Each of the fields is preceded by a line of the form nt=time, where time is number of the day starting at September 7, 1999. Each of the 2D-fields is stored row-by-row from the south-western to the north-eastern corner of the analysis domain.

The nature of rain-gauge data and the variable density both in time and space require some care in the interpretation of the results of the Alpine precipitation analysis. Here we list some of the most important points, a detailed discussion can be found in Frei and Schär (1998).

• No correction was made of the systematic measurement bias of the rain-gauge measurements. Underestimates are in the order of 4-10% for rain and between 10-50 % for snowfall.

• The spatial analysis does not compensate for the unrepresentative distribution of the observing network with respect to topographic height.

• Random errors due to limited spatial sampling add up to the systematic errors.

• The grid-resolution is 25 km, features below this scale are not resolved. In particular, the analysis peak values are lower than the peak values observed at individual stations.

• The resolution of the analyses can regionally be coarser than the grid-resolution, depending on the density of the available station network.

• Frei, C. and Schär, C., 1998: A precipitation climatology of the Alps from high-resolution rain-gauge observations. Int. J. Climatol., 18, 873-900. [pdf-file]

• Frei, C. and Häller, E., 2001: Mesoscale precipitation analysis from MAP SOP rain-gauge data. MAP Newsletter, 15, 257-260. [pdf-file]

Acknowledgements

We are indebted to the following institutes for providing access to daily precipitation data: Deutscher Wetterdienst, Offenbach a.M.; Hydrographisches Zentralbüro des Bundesamtes für Land- und Forstwirtschaft, Wien; Zentralanstalt für Meteorologie und Geophysik, Wien; Tiroler Wasserkraftwerke AG, Austria; Météo France, Toulouse; MeteoSwiss, Zurich; Servizio Idrografico e Mareografico Nazionale, Roma; Ufficio Centrale di Ecologia Agraria, Roma; the Italian MAP working group on climatology; Croatian Meteorological and Hydrological Service, Zagreb; Hydrometeorological Institute of Slovenia, Ljubljana; Czech Hydrometeorological Institute, Prag; Instituto Nacional de Meteorologia, Spain.