Vízügyi Közlemények, 1995 (77. évfolyam)
2. füzet - Rátky István: A turbulens áramlás matematikai alapjai
212 Gyarmati Györgyi Utilization of precipitation measurements in the correction of precipitation measurement by radar by Györgyi GYARMATI, meteorologist Errors of radar measurement necessitate the correction of the precipitation data measured by radar stations by utilizing the data of earth-based precipitation monitoring stations. Records of two Hungarian radar stations, the Farkasfa and the Nyíregyháza-Napkor stations have been utilized. The essence of the correction procedure is to generate a matrix of correction factors by jointly utilizing the traditional and radar precipitation data for the modification of the radar data. Л correction factor is calculated as the ratio of earth-based precipitation data to radar data for each monitoring points and the matrix of correction factors is generated by an interpolation method. Next the matrix (field) of radar data are multiplied with this matrix of correction factors. A simple interpolation technique, widely used in hydrology, is used for the generation of the matrix of correction factors. In a nodal point the correction factor is approximated with the weighed average of the values of the monitoring points. The weighing factors are determined as the inverse of the squared distance between the nodal point and the station. Precipitation sums of the rainfall events of 24. 08. 1994, 03. 09. 1994 and 15. 09. 1994 were utilized for this study. Patterns of processed daily precipitation data arc shown in Table I. The table shows the date of measurements, the radar station considered, the method of operation and the wave length utilized. Measured values of reflectivity were tranfonned to rainfall intensity by utilizing the Marshall-Palmer formula (Z - A* P; B), where Z is the reflectivity, P; is the rainfall intensity and A and В are experimental constants (Table I.). Different constants were used for high intensity rainfall, rainstroins and lasting continuous rainfall. A traditional technique was used for data evaluation. The available stations were grouped into two sets. The first set contained those stations the data of which were used for the development of correction factors. The other set of stations were considered the independent sample to which the original and corrected radar data were compared. Results were tested in two ways. Correlations of radar and pulviometer data were analyzed before and after the correction of radar data. These correlations characterize the relationship between pulviometer data and the areal distribution of radar data and the presence of noise in the records. Further, the radar data (P r) were plotted in function of the pulviometer data along with a linear regression analysis. The slope of the straight line characterizes the extent of under- or over estimation on the basis of the radar data. Table II shows the slopes and the correlation coefficients before and after the correction. Figure 1. shows the relationship between radar data and pulviometer data. Figure 2. shows the logarithm of the correction factors in function of the distance (D) of the radar station from the pulviometer station. The relatively low correlation values of non-corrected patterns are due to the well known error sources (decreasing intensity of rainfall, geometric faults of the radar rays and the permanent echoes). It can be seen that the procedure presented can substantially improve the correctness of radar-based estimation of precipitation parameters. This is indicated by the increase of correlation coefficients and by the decreasing dependency of the absolute value of In (P/Pr*) on the distance. * * *
