COMPSTAT 2016: Start Registration
View Submission - COMPSTAT
A0525 - Status: Accepted
Type of publication: Only Abstract Type of presentation: Poster Presentation Title: Generation of monthly precipitation series and the performance of homogeneity tests Authors:  Elif Akca - Middle East Technical University (Turkey) [presenting]
Ceylan Yozgatligil - Middle East Technical University (Turkey)
Ceyda Yazici - Middle East Technical University (Turkey)
Keywords: computational methods,structural breaks,time series Abstract: Climate studies have gained importance due to the significant effect of climate change. The extreme meteorological events can cause floods, droughts, sudden change in the temperature or change in the climate trends. Since they have an important effect on human beings and the environment, these meteorological variables should be predicted and some precautions should be taken if possible. In order to conduct any kind of statistical analysis, the nonclimatic effects should be determined and corrected or removed. Otherwise, any kind of meteorological effect should be kept to reflect the change in the variable. If there is any nonclimatic effect in the variable, the series is defined as inhomogeneous. The most important climatic variables are temperature and precipitation. The precipitation variable cannot be modeled efficiently due to its nonsymmetric shape and high variability. First, monthly total precipitation amount is simulated based on real station records. Then, in each simulation, artificial breakpoints are created and the homogeneity analysis are conducted to detect the inhomogeneity and the results are compared to decide on the best performed homogeneity test for precipitation. Comments: