Clemens, Marco (2002) Machbarkeitsstudie zur räumlichen Niederschlagsanalyse aus Schiffsmessungen über der Ostsee. (PhD/ Doctoral thesis), Christian-Albrechts-Universität, Kiel, Germany, 178 pp. . Berichte aus dem Institut für Meereskunde an der Christian-Albrechts-Universität Kiel, 327 . DOI 10.3289/ifm_ber_327.

Text MClemens_Diss_2002.pdf - Published Version Restricted to Registered users only Available under License German copyright act UrhG. Download (93MB) | Contact

Abstract

In the frame of this work measurements of precipitation over sea are investigated and analysed. The measurements result from two new designed instruments: a ship rain gauge and an optical disdrometer. The statistical evaluation of the measured rain rates, made with the disdrometer, results from the count method. A new relationship between liquid water content and rainrate, based on integral rainfall parameters, is derived from the disdrometer measurments. Likewise, a new relationship between radar reflectivity and rain rate is derived. The large number of rainfall data measured by the optical disdrometer serves as a data set for calibration of the ship rain gauge. The calibration takes the temporal delay of the ship rain gauge and the different accuracy of both instruments into account. It is based on an algorithm used for the correction of ordinary rain gauges. In comparison to other reference rain gauges, operated at synoptic stations, the new derived calibration gives quite good results. The distribution characteristics of the measurements by the ship rain gauge correspond in general to the properties of a lognormal probability distribution. Considering the distribution characteristics, statistical moments of the ship rain gauge measurements are derived on a monthly time scale. The highest correlations between the different moments are obtained for the mean rates and the frecquency of precipitation. The main goal of this study, however, is the spatial analysis of precipitation measurements performed on merchant ships over the Baltic Sea. The analysis scheme makes use of the Kriging method. This method minimises statistical errors under consideration of the spatial structure of rainfall. The main error is the sampling error due to the low data density in time and space. An analysis of the standard error by applying a Monte-Carlo model allows the estimation of the sampling error, needed as an input parameter for the Kriging method. The spatial structure functions in terms of the autocorrelation or the autocovariance respectively give seasonal dependencies of the precipitation processes on short timescales. The decorrelationlength is shortest in autumn and longest in summer, indicating a predominance of convective precipitation events in autumn. The spatial analyses are made on a geographical 1◦ × 1◦ grid. The temporal resolution of the spatial estimates corresponds to a seasonal respectively annual timescale. Whereas the raw data fields are dominated by the sampling errors resulting in high spatial variances, the new developed analysis scheme based on the Kriging method reduces the sampling errors significantly. Comparisons of the analysed precipitation fields to synoptic observations at coastal stations show a good agreement. The generated precipitation fields over the Baltic Sea represents an unique data set, which is only based on in-situ measurements over sea in contrast to other existing estimates. Thus, these new derived precipitation fields are an useful tool for the validation of numerical model output and precipitation estimates derived from satellite or radar measurements.

Actions (login required)

View Item