Evaluating shifts in rainfall patterns using multi-test change point detection across several rainfall stations in semi-arid climate

The rainfall records from twelve stations in Algeria were examined to assess the consistency of the time series and to identify any potential breaks or inhomogeneities. Instead of relying on a single method, four different statistical tests were applied: Pettitt, standard normal homogeneity test (SNHT), Buishand, and the Buishand range test. The analysis focused on annual totals. The objective was to identify periods when major changes might have happened and to judge how assess the strength of statistical evidence. For the longer records, such as St40405, St40406, St40414, St40415, St40418, and St40419, breaks were clearly detected. Test values were high and the p-values were very small, indicating stronger confidence in the results. In contrast, for the shorter series, St40423 and St40438, the signal was weak and less stable. This pattern is typical when the data record is short, as it is harder to draw firm conclusions. In a few cases, the break appeared at the very end of the series. This may reflect recent climate shifts, although it could also be the result of noise, and the cause is not always clear. When the results of all four tests were compared, at least two of them identified the same breakpoint for every station. This agreement strengthens the reliability of the detections and is consistent with earlier studies. Overall, the point is clear: rainfall series should be tested for consistency before trends are studied. Otherwise, the results of regional climate studies may be unreliable.