Representativity of in situ precipitation measurements - A case study for the LITFASS area in North-Eastern Germany

In situ precipitation measurements can extremely differ in space and time. Taking into account the limited spatial-temporal representativity and the uncertainty of a single station is important for validating mesoscale numerical model results as well as for interpreting remote sensing data. In situ precipitation data from a high resolution network in North-Eastern Germany are analysed to determine their temporal and spatial representativity. For the dry year 2003 precipitation amounts were available with 10 min resolution from 14 rain gauges distributed in an area of 25 km x 25 km around the Meteorological Observatory Lindenberg (Richard-Assmann Observatory). Our analysis reveals that short-term (up to 6 h) precipitation events dominate (94% of all events) and that the distribution is skewed with a high frequency of very low precipitation amounts. Long-lasting precipitation events are rare (6% of all precipitation events), but account for nearly 50% of the annual precipitation. The spatial representativity of a single-site measurement increases slightly for longer measurement intervals and the variability decreases. Hourly precipitation amounts are representative for an area of 11 km x 11 km. Daily precipitation amounts appear to be reliable with an uncertainty factor of 3.3 for an area of 25 km x 25 km, and weekly and monthly precipitation amounts have uncertainties of a factor of 2 and 1.4 when compared to 25 km x 25 km mean values. (C) 2011 Elsevier B.V. All rights reserved.