USE OF STATISTICAL METHODS TO DETECT

CLIMATE CHANGE AND/OR VARIABILITY IN

THE BLUE NILE RIVER

 

RIHAB HUSSEIN BAYOUMI AHMED

B.Sc (Hons) in Civil Engineering

 

 

 

A THESIS SUBMITTED TO THE UNESCO CHAIR IN WATER RESOURCES OF THE OMDURMAN ISLAMIC UNIVERSITY FOR THE DEGREE OF MASTER OF SCEINCE M.Sc. IN WATER RESOURCES DEVELOPMENT & MANAGEMENT

 

 

Supervisor

Dr. KAMAL ELDIN ELSIDDIG BASHAR

 

 

UNESCO Chair in Water Resources

 

OMDURMAN ISLAMIC UNIVERSITY

 

KHARTOUM, August 2005

 

 

I

ABSTRACT

 

Water is getting serious attention throughout the world. It has a vital role in sustaining life and development in arid and semi-arid regions. It could be argued that water resources in Sudan are rather limited, despite the fact that it comprises rainfall, surface flows and groundwater.

 

The major river (from the Ethiopian Plateau) contributes about 84 % of the Nile water, which is characterized by strong annual cycle and great inter-annual variability, with a dominant peak in August.

 

The objective of this research is to apply statistical techniques to detect evidence of climate variability or climate change in the time series of flow and rainfall of the Blue Nile River. The analyzed series includes rainfall series at (8) rain gauging stations and flow series at Eddeim.

 

The techniques used include trend analysis, Automatic Segmentation of Time Series, Split Sample Technique for detection of recent trends in a time series, Onset, Cessation, and length of the rainy season or Duration, the total rainfall within the rainy season, peak flow and time to peak.

 

The Annual Mean Areal Rainfall was found to be 1270 mm with standard deviation of 121.34 mm and the coefficient of variation 0.1. Trend analysis showed that there is a statistically significant declining trend. The average Onset of the Rainfall was found to be at dekad 11 (April 11-20) and the average Cessation date was found to be the 28th dekad (October 01–10).

 

 

The Mean Annual Flow was found to be 525079.4 m3/s with a standard deviation of 101711.2 m3/s and coefficient of variation 0.19. There is no significant trend in the annual Flow series. However, the Split Sampling and Automatic Segmentation revealed that the mean value of the flow dropped over 80s.

 

The average Onset of the Flow at Eddeim was found to be at dekad 19 (01-10 July) and the Cessation was found to be at dekad 31 (01-10 November), giving a Duration of 12 dekads of flow.

 

The Onset, Cessation and Duration analysis was redone for the periods of the Split Sampling and Automatic Segmentation. The results showed that the Onset systematically changed in the three periods and become one dekad earlier from the previous period. The Cessation remains the same. This makes Duration in each period to become longer by one dekad from the previous one.

 

There is no statistically significant trend in the Annual Peaks, and the average time to Peak was found to be 21st of August with standard deviations 13.54 and coefficient of variation 0.06.

 

The study found that, despite the statistically significant declining trends in the rainfall series and the change in mean flow series, the variations can be attributed to climate variability. There is no evidence of climate change detected within the study period as no systematic prevailing changes seen.

 

Abstract in Arabic