HOMOGENEITY ANALYSIS OF LONG-TERM MONTHLY PRECIPITATION DATA OF TURKEY

Availability of a long-term, continuous and homogeneous precipitation series is always essential for climate and hydrologic studies. Nevertheless, precipitation data may suffer from inhomogeneity, owing to various non-climatic factors. The observing network in Turkey is under continous development and but at the same time prone to effects of urbanization and changes in land-use conditions. This study aims to characterize homogeneity of the long-term Turkish precipitation data in order to ensure that they can be used reliably. Several parametric and non-parametric tests are applied to de-tect inhomogeneities with long-term monthly precipitation data of 211 stations scattered across Turkey for the period from 1973 to 2002. The homogeneity analysis was performed in two steps. In the first step, 4 parametric tests were applied to the data. Initially, the results indicated that 3 stations with the Kruskal-Wallis test, 13 stations with the Fried-man tests, 5 stations with the one-way ANOVA tests, and 73 stations with the Bartlett's tests were inhomogeneous. Further analyses indicated that only 2 of 211 stations resulted to be inhomogeneous under the all 4 tests. Again, 2 out of 211 stations were inhomogeneous according to the result of 3 tests. Finally, only 7 out of 211 stations demonstrated inhomogeneity according to 2 of the tests. In the second step, 27 pairs of stations which have closest proximity to each other are tested for homogeneity using the non-parametric Mann-Whitney and Wilcoxon signed-rank tests. A total of 10 pairs of stations exhibited inhomogeneity for the both tests. It was observed that most of the stations, which are characterized as inhomogeneous, are located in Eastern Anatolia, and 3 of them lie around Lake Van, which is the largest water-body inside Turkey. When the inhomogeneous stations were checked against their meta-data, nothing conclusive was found to justify the inhomogencities, except Gokceada station, which experienced considerable increase in plant canopy in its surrounding. It is expected that replacing the conventional network with the more up-to-date sensors (called AWOS) may create a great challenge in homogeneity of the Turkish precipitation series.