Variability and Trend of Annual Maximum Daily Rainfall in Northern Algeria

The daily rainfall dataset of 35 weather stations covering the north of Algeria was studied for a period up to 43 years, recorded after 1970s. The variability and trends in annual maximum daily rainfall (AMDR) time series and their contributions in annual rainfall (AR) were investigated. The analysis of the series was based on statistical characteristics, Burn's seasonality procedure, Mann-Kendall test, and linear regression technique. The contribution of the AMDR to AR analysis was subjected to both the Buishand test and the double mass curve technique. The AMDR characteristics reveal a strong temporal irregularity and have a wide frequency of occurrence in the months of November and December while the maximum intensity occurred in October. The observed phenomenon was so irregular that there was no dominant season and the occurrence of extreme event can arrive at any time of the year. The AMDR trends showed that only six of 35 stations have significant trend. For other stations, no clear trend was highlighted. This result was confirmed by the linear regression procedure. On the contrary, the contribution of AMDR in annual totals exhibited a significant increasing trend for 57% of the sites studied with a growth rate of up to 50%.