MODIS NDVI satellite data for assessing drought in Somalia during the period 2000-2011

Abstract

This thesis studied the drought monitoring and assessing in Somalia environment during the period 2000-2011 with more details about the drought in 2010- 2011. Time series analysis of remote sensing data as well as rainfall data were used to achieve the aim of the study. Most of Somalia is located in dry zone with only some grass and pasture valid only pastoralism which is practices by half of the population. The country is frequently exposed to drought, occurring moderately every 3-4 years and severely every 7-9 years. Drought is characterized as a slow process, can happen everywhere and anytime and prolongs from months to years. The advance in remote sensing data both spatially and temporally makes it possible to use this data for drought monitoring regionally and globally.
MODIS NDVI satellite data was used mainly to achieve the purpose of the study. The analysis was done at pixel-scale in seven locations as well as at regional-scale. To monitor and assess drought, three components were studied: onset, duration, and severity or intensity.
The first important result is that among all drought years during the study period, the drought in 2010-2011 was the worst as it continued for longer time than previous droughts during the study period (around 12 months). In most parts in southern Somalia, there was extreme drought where the percent of negative change reached up to 30-50% below the normal. The impact of this drought was catastrophe because of existence of three factors which are the conflict in the country, nature of livelihood (pastoralism), and the population density. Drought often starts in Deyr season and continues to the Gu season in the next year e. g. 2003-2004, 2005-2006 and 2010-2011 and it is often followed by above normal rainfall. Also, droughts with long duration are lower frequent than that with short duration. The monthly correlation between rainfall and NDVI was significant in most parts of the country and surrounding areas, though not very strong (r ~ 0.40). This relationship got higher with time lag one month for NDVI. The correlation in the second rainy months was found to be higher than the first rainy months with almost no correlation. Annual correlations were very low and even negative. Finally, remote sensing data proved to be a significant tool in monitoring and detecting drought components. In the thesis’s case where rainfall data and other meteorological data were absence or unreliable, remote sensing data was the only practical choice.

Abstract

Popular science
Using remote sensing to assess drought in Somalia

Somalia, the country in East Africa in a region referred to as Horn of Africa is in a civil war, conflict and unrest since 1991. The prolonged conflict increases the impact of droughts in the country which is already prone to drought continuously. The country is frequently exposed to drought, occurring moderately every 3-4 years and severely every 7-9 years. This study focused on assessing and monitoring drought in Somalia during the study period 2000-2011, especially drought of 2010-2011 which became a disaster. Drought is characterized as a slow process, can happen everywhere and anytime and prolongs from months to years. Satellite remote sensing data, especially vegetation index, was mainly used to achieve the purpose of the study. Satellite remote sensing uses the reflectance of the light bands from ground objects that is received. Vegetation index used the difference in red and Near-infrared (NIR) reflectance. Healthy vegetation absorbs radiation relatively high in the visible red band and reflects significantly high in the NIR band. The other data used is rainfall data from Climatic Research Unit (CRU) 3.1 and Rainfall Estimation (RE) 2.0.
The most important result is that among all drought years during the study period, drought of 2010-11 was the most influence one in its duration (more than a year), intensity (30%-50% negative change) and extension to most parts of southwest Somalia and north and northeast of Kenya. The nature of livelihood (pastoralism); the absence of stability and peace; and the high population density in south of Somalia made the impact of drought catastrophe (crossing the famine threshold
During the study period, most dry spells started in the second season and continued to the next season e. g. 2003-2004, 2005-2006 and 2010-2011 and was often followed by above normal rainfall. Also, droughts with long duration are less frequent than that with short duration. The correlation between monthly rainfall and vegetation showed to be significant in most of the locations though the dependency of vegetation on rainfall not more than 26%. This relationship becomes stronger with time lag of one month for vegetation. It is the time needed for vegetation to grow or respond to change in rainfall. In a country like Somalia where reliable data is difficult to obtain continuously, remote sensing data proved to be a significant tool in monitoring and detecting drought components.

• author: Alwesabi, Mohammed
• supervisor: Lars Eklundh
• organization: Dept of Physical Geography and Ecosystem Science
• course: NGEM01 20121
• year: 2012
• type: H2 - Master's Degree (Two Years)
• subject: Earth and Environmental Sciences
• keywords: remote sensing, vegetation index, drought, Somalia, geography, physical geography
• publication/series: Student thesis series INES
• report number: 257
• funder: European Commission, Erasmus/Socrates Program
• language: English
• id: 3044245
• date added to LUP: 2012-09-04 15:45:10
• date last changed: 2012-09-04 15:45:10