Skip navigation
 

Please use this identifier to cite or link to this item: http://hdl.handle.net/123456789/13872
Title: Knowledge Discovery From Satellite Images For Drought Monitoring
???metadata.dc.contributor.*???: Tsegaye Tadesse(Dr.)
Solomon Atnafu (Dr.)
. Shawndra Hil (lDr)
Getachew, Berhan
Keywords: Drought Monitoring;Drought Prediction;Geo-spatial Information;Knowledge Discovery;Satellite Image
Issue Date: Jun-2013
Publisher: Addis Ababa University
Abstract: Drought is one of the most important challenges facing the planet. When it happens, it usually results in serious economic, environmental, and social crises. Despite the growing number of freely available biophysical, climate, and satellite data for characterizing and modeling drought, research efforts have been constrained to using only meteorological point data, such as the amount of rainfall, for drought monitoring information. This point data is insufficient for representing diversified ecosystems, and the data has coarse resolution levels (limited spatial coverage). Researchers also have limited tools for data retrieval and integration for improved drought identification and modeling, which usually results in a time delay for information to reach decision makers. Taking this into account, this dissertation research has three objectives: 1) identify the most relevant attributes for efficiently implementing drought monitoring, 2) develop a new approach for extracting knowledge from satellite imageries for improved identification and prediction of drought, and 3) evaluate the new approach for national and regional drought prediction applications. Using an exploratory research approach and modeling research method, different data collection and analysis techniques were executed using knowledge discovery in a database approach. The data mining models developed using artificial neural network and regression tree models were able to predict DroughtObject with accuracy of 0.70 – 0.95 correlation coefficients, in oneto four months’ time lag. The developed DroughtObject model was evaluated for its application in showing drought severity and food deficit status. There were positive relationships between DroughtObject products and crop yield data up to 0.91 R2 values. The results confirmed that the model can directly be used by those who are currently responsible for drought monitoring and risk management. The new concept developed in this research was prototyped and demonstrated in an easy-touse approach, with a focus on demonstrating the concept of DroughtObject characterization and identification from a group of pixels. This demonstration also revealed possible future system developments. This dissertation research could help decision makers use advanced satellite technology for effective drought monitoring and early warning systems in various regions. Combined with proper policies, these systems can help to prevent famine and starvation in food-insecure regions. Up to now, satellite technologies have been used primarily in areas of meteorological applications. In this research, the main emphasis is on mining knowledge from satellite images for drought risk assessment and saving the lives of individuals who are affected by recurring droughts. The findings of this research can help decision makers take timely and appropriate actions to save lives in drought-affected areas using advanced satellite technology.
URI: http://hdl.handle.net/123456789/13872
Appears in Collections:Thesis - Information Science

Files in This Item:
File Description SizeFormat 
Getachew Berhan.pdf10.33 MBAdobe PDFView/Open
Show full item record


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.