![]() Lire aussi Les eaux superficielles et souterraines de Sidi Bel Abbès 3.3 Monthly simulationsĪs described in section 2.2, to generate monthly precipitation, daily rainfall amounts are aggregated into monthly totals and they are modified using a lag one vector autoregressive process. rainfall at all stations and all months.įigure 3.9: Comparison of generated and observed monthly median rainfall at all stations and all months.įigure 3.10: Box plot of the observed and generated daily precipitation at all 12 months, at station 6 for the top panel and station 5 for the bottom panel. The variability is also well reproduced by the model, which confirms the results in figure 3.8.įigure 3.3: Comparison of empirical and theoretical densities.įigure 3.4: Comparison of empirical and theoretical CDFs.įigure 3.6: Comparison of generated and observed spatial correlations of daily rainfall amounts.įigure 3.7: Comparison of generated and observed mean monthly rainfall at all stations and all months.įigure 3.8: Comparison of generated and observed monthly s.d. The skewness and quartiles are well represented by the two part model. The statistics of daily rainfalls on long time scales are therefore well reproduced by the two part model.įigure 3.10 shows the box plot for observed and generated daily precipitation at station 5 and 6 at all 12 months. Similarly, figure 3.8, and 3.8 show respectively, the standard deviations of monthly precipitation (characterizing the interannual variation in total monthly precipitation) in the observations versus generated series and the observed median of monthly precipitation versus the simulated ones. Therefore, various monthly statistics are investigated.įigure 3.7 shows the relationship between generated and observed mean precipitation, for all 6 stations and all 12 months. It is of interest to examine the degree to which the daily precipitation occurrence and amounts models reproduce the statistics of observed precipitation climate on longer time scales. The generated spatial correlations are compared with the observed ones in figure 3.6, and they are well represented by the model. Thus, the Gamma distribution fit well the nonzero amounts. ![]() ![]() The empirical and theoretical densities, CDFs, quantiles, and probabilities for the same time series are compared in figure 3.3, 3.4, and 3.5. The Gamma distribution is fitted to nonzero amounts of station 1 in January, and the P-value resulted from the Kolmogorov Smirnov test is 0.7.įigure 3.2: Comparison of generated and observed spatial correlations of daily rainfall occurrences. 3.2.2 Daily amounts processĪs described in section 1.2, nonzero precipitation amounts are modeled using the Gamma distribution, whose parameters are estimated, for each site at each calendar month, using the maximum likelihood estimation. The spatial correlations are well represented by the model with some errors resulted during the simulation process. In figure 3.2, the spatial correlations of the observed daily occurrences are compared with the generated ones. In this section, the occurrence process model is used to generate the daily occurrences at all locations. To preserve the seasonal characteristics of the observed time series, each calendar month is considered separately in the model.Īt first the daily occurrences are generated using the multisite occurrence process at all stations, then the daily amounts are generated using the Gamma distribution. The method described in section 1.2 is used to simultaneously simulate daily sequences of precipitation at the 6 locations. Table 3.1: Location and catchement area of stations 3.2 Daily simulations įigure 3.1: Stations numbers and their locations on map. ![]() The data is available in the National River Flow Archive website. ![]() This data is used to simultaneously simulate daily, monthly and annual precipitation at the 6 locations. The data used in this study, is the daily rainfalls at 6 stations in Scotland (table 3.1), from 1 January 1961 to 31 December 2015. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |