Getachew Berhan Demisse, a visiting scholar at the National Drought Mitigation Center since August 2011, will return to Ethiopia in February, where he will demonstrate his method for predicting drought up to four months ahead.
Getachew is using his interdisciplinary background in forestry, information technology and geographic information systems to create an object-oriented approach to modeling drought, which is the subject of his dissertation. He tested his model on historic data and found that it correctly anticipated severe droughts in Ethiopia in 1984 and in 2002.
Getachew said he has found four key attributes that explain 85 percent of the variability between dry and wet years, and adds seven more to the mix to explain up to 90 percent. The four main attributes include dynamic data -- the three-month Standardized Precipitation Index (SPI) and the Normalized Differentiated Vegetation Index (NDVI) -- and static data -- a digital elevation model and the water holding capacity of the soil. The other seven are five oceanic indices, land cover, and ecological zones. Oceanic indices are long-term fluctuations in sea surface temperatures that scientists have connected to rainfall patterns over land.
The model is designed to run each month to produce predictions for the growing season ahead, at one-month intervals into the future, up to four months out.
Getachew, a doctoral student in IT at Addis Ababa University, is working with Dr. Tsegaye Tadesse, NDMC climatologist, who is developing the Vegetation Drought Outlook (VegOut). Getachew said he has also appreciated the chance to work with Karin Callahan, NDMC GIS/Remote Sensing specialist, and others. Getachew has a bachelor's degree in forestry from the Alemaya University of Agriculture, a certificate in computer science from Ethio Computer Training, and a master's degree in GIS and Earth Observation from the International Institute for Geo-information Science and Earth Observation in The Netherlands. He is a GIS lecturer at Addis Ababa University and at HilcoE School of Computer Science and Technology.
During his time in the U.S., Getachew has been working with NDVI data from the National Oceanic and Atmospheric Administration and with data from USAID's Famine Early Warning System Network (FEWS NET). He said that when he returns to Ethiopia, instead of NOAA data, he will substitute data from a Meteosat Second Generation feed, available via a receiving station recently installed at his university by the European Union.
The new data feed was part of the inspiration for Getachew's research. With the meteorological satellite data coming from the feed, "It's possible to look at your environment every 15 minutes," he said. "We thought, why not use this data for drought monitoring? That's a top agenda item for us."
Development and aid agencies use drought predictions to try to reduce the effects of food insecurity in areas where people depend on subsistence agriculture for their livelihoods.
Although the data isn't new, Getachew believes he has found the signal in the noise, having identified which data is meaningful, and he is bringing it together in a way that literally presents a clearer picture of areas affected by drought. Instead of the pixelated images that come from satellite data, he uses a technique called anisotropic diffusion to show patterns more clearly.
Getachew hopes to complete his doctorate in June. He looks forward to showing the product of his research – drought prediction maps -- to FEWS NET and to the Ethiopian Disaster Prevention and Preparedness Agency, and to seeing the central component of his research validated through the peer-review process.
"This is practical, problem-solving research," Getachew said. "It's on the top of the agenda in our country."
Getachew Berhan Demisse CV
-- Kelly Helm Smith, National Drought Mitigation Center