Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Spatial and temporal drought incidence analysis in the northeastern highlands of Ethiopia

Spatial and temporal drought incidence analysis in the northeastern highlands of Ethiopia This study investigated the space-time drought incidence in the northeastern highlands of Ethiopia using monthly rainfall data. It also aims to predict drought events for 100 years. The Standardized Precipitation Index (SPI) was used to compute the drought severity classes of rainy months and seasons at 1-, 4- and 8-months timesteps. The Mann-Kendall’s test and Sen’s slope estimator were used to analyze the trends of drought events and to determine the magnitude of change. Inverse Distance Weighted spatial analysis tool was used to illustrate the spatial patterns of the drought risk events. The study detected extreme severe droughts in the belg rainy months in March 2008 and April 1984. However, during the belg season, the year 1999 was the driest for the recorded periods. On the other hand, the extremely severe droughts were observed during the kiremt rainy months of July 1987 and 2015, August 1984, and September 2009. In general, 1984, 1987 and 2015 were the driest years recorded in the kiremt season. The study noted that the drought risk events of months in the belg season were threefold greater than that of the months in kiremt season under moderate drought intensity class. Equally, the drought risk events of months in kiremt season were threefold greater than that of the belg season under extreme drought intensity class. Complex spatial variations of drought risk events were also observed in 1-, 4- and 8-months timesteps. During the belg seasons, the southern half was subjected to more frequent drought risk events while the northern half experienced more frequent drought risk events during kiremt season. Almost the eastern half of the livelihood zones experienced higher drought frequency events than the other parts in the 8-month timestep. The observed space-time drought risk event analysis has shown a potential threat to the rainfed agricultural practices that have a great influence on the livelihoods of smallholder farmers. Hence, documentation and assessment of drought risk events based on the livelihood zones are essential for drought risk management, early warning responses, local- scale planning and food security management. Finally, the study recommended further research on additional indices of climatic variables such as evapotranspiration and soil water content. Keywords: Livelihood zone, Belg and kiremt, Drought, Standardized precipitation index, South Wollo, Ethiopia Introduction average usually for a season or more (AghaKouchak and Drought is a commonly used term, but the most com- Nakhjiri 2012; Guha-Sapir et al. 2012; Dai 2011; Ashraf plex and least realized of all the natural hazards affecting and Routray 2013; WFP 2014). Many scholars, for in- more people than any other hazards (Wilhite 2000; Ash- stance, AghaKouchak and Nakhjiri (2012), Jeyaseelan raf and Routray 2013). As opposed to other extreme (2003) and Wilhite and Glantz (1985) classified drought events, (like floods, tornadoes, and hurricanes), drought as meteorological, hydrological, agricultural, and socio- develops slowly and steadily, making it difficult to deter- economic based on the causative factors. These drought mine the onset and end (WMO 2016). Generally, types share the deficiency of precipitation as a common drought is related to a continual deficit of water below phenomenon. Among these, the meteorological drought is the main driver; leading to the more likely occurrences of agricultural and hydrological droughts (WMO 2016; * Correspondence: abebemekkon@gmail.com Department of Geography and Environmental Studies, Bahir Dar University, Zhan et al. 2016; Mpelasoka et al. 2008). P.O. Box 79, Bahir Dar, Ethiopia © The Author(s). 2020 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 2 of 17 Since drought is a recurrent climatic phenomenon, its in great human sufferings and losses of life (Seleshi and outcome is manifested through immense damage to Camberlin 2006). Precipitation exhibits great spatial agriculture production, reduction of water supply and variation both in annual average and inter-annually that energy production, mass migration and loss of life could be taken as the major cause for the occurrence of (Masih et al. 2014). Writers such as –Sheffield and the droughts in the country (Viste et al. 2013; World Wood (2007), Bannayan et al. (2010), and Degefu and Bank 2006a, 2006b). In this regard, Viste et al. (2013) Bewket (2015) differently showed that the causes and in- conclude that there are no years without drought in fluences of drought in space-time on the environment many parts of Ethiopia. Consequently, environmental are determined by the attributes of frequency, magni- and socioeconomic deprivation in Ethiopia happened tude, intensity, and spatial extent. Likewise, Trenberth due to the cumulative induced drought over several cri- et al. (2011) reported that the magnitude, intensity, fre- sis seasons (Endale 1993). In the future, due to the in- quency and spatial coverage of the dry area has been creasing resource depletion, single-year droughts may be doubled since the 1970s in Africa, East and South Asia, sufficient to cause severe hardship among vulnerable eastern Australia, Southern Europe, Northern and households (Webb and Von Braun 1990; Webb and Southern America, most of Alaska, and Western Canada. Braun 1994). Within the period 1960–2016, there were about 669 Although the name “Water Tower of North-East Af- drought events reported across the world, and hence it rica” has been given to Ethiopia, the country is one of has killed about 2.2 million people and affected over 2.6 the Horn countries highly vulnerable to drought (Gebre- billion people and an estimated economic damage of hiwot et al. 2011). The history of drought and associated 146 billion US Dolar Centre for Research on the Epi- disasters in Ethiopia traced back to 250 B.C. with a total demiology of Disasters/CRED (2016). of 39 periods of food shortage and excess mortality rate The vulnerability of developing countries to drought have been identified (Block 2008; Webb and Von Braun has risen by a factor of three in the decade of the 1990s 1990; Webb and Braun 1994; Webb et al. 1992). Impacts compared with the 1960s (Mirza 2003). Much of the of droughts in Ethiopia are perhaps the most widely northern, southern, eastern and Sahelian African coun- publicized given their severity and frequency (Gautam tries experienced the worst droughts in terms of the 2006). As recent literature ratifies, Ethiopia’s history has number of people killed and affected in 1972/73 and been periodically marked by major drought incidences 1984/85 (Gommes and Petrassi 1996). For instance, of for the period 1950–2016 (Table 1). all the natural disasters, droughts account for only 8% of Drought frequency in Ethiopia has been increasing. the global phenomenon. It also accounts for 25% in Af- For example, in the 1970s and 1980s drought usually oc- rica between 1960 and 2006 (Gautam 2006). In the curred once per decade (Block 2008; United Nations 1980s drought killed more than half of a million people Environment Program/UNEP 2011). It was also about in Africa (Dai 2011). In Sub Saharan African countries, twice every 3 years between 1980 and 2004 (World Bank crop production and food security are mainly threatened 2006a, 2006b) and once a year between the 1990s and by drought where the effects are inter-temporal and 2011 (Viste et al. 2013). The estimated human cost long-lasting (Shiferaw et al. 2014). Africa stands first caused by drought in Ethiopia between 1972 and 1984 with a total of 289 reported the number of drought were about 1. 2 million people (United Nations Environ- events that killed almost 700 thousand people and af- ment Program/UNEP 2002). These evidences showed fected above 414 million people between 1960 and 2016 that Ethiopia is often stricken by droughts in the 1970s due to the induced drought frequency (Centre for Re- and 1980s resulted in widespread poverty, economic search on the Epidemiology of Disasters/CRED 2016). stagnation, depletion of household assets and savings, Drought has a severe negative impact on key socioeco- and excess mortality (Dorosh and Rashid 2013). South nomic sectors of most Eastern African countries (Butterfield Wollo, located in the northeastern highlands of Ethiopia 2011). Its incidence has increased steadily in East Africa for is an epicenter for drought and famine for many years. the last five decades (Gautam 2006). For instance, The zone is a drought prone area and faces the horrible deprivation associated with drought and food crises disaster combination of both high risk and low potential as well victims for East African countries (Burundi, Djibouti, as chronic food insecurity (World Bank 2006a, 2006b). Ethiopia, Kenya, Somalia, and Uganda) increased from one Of the total recorded drought years (1953–2016) in million people in the 1980s to 21 million in the Ethiopia, Wollo experienced about 72% of the drought 2010s (Hao et al. 2014;Guha-Sapiretal. 2012). events (Little et al. 2006). In fact, the existence of diverse In Ethiopia, extreme weather events owing to insuffi- agroecology enables the zone to raise different types of cient total rainfall amount and long dry periods affected livestock and to produce cash and food crops (ANRSPC the agro-socioeconomic environment in the northern, 2017; 2018). But, the agricultural activities are entirely southern and eastern parts of the country and resulted rainfed except in a few localities where there are small- Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 3 of 17 Table 1 Reported drought years, regions affected, causes and severities since the 1960s in Ethiopia Year Region Affected Causes and severity description 1957/58 Tigray and Wollo Rain failure in 1957; locusts and epidemic in 1958 and about 100,000 people died 1962/63 Western Ethiopia Very severe 1964–66 Tigray and Wollo About 1.5 million people were affected and about 300, 000 livestock died 1971–75 Ethiopia Sequence of rain failures; estimated 250,000 dead; 50% of livestock lost in Tigray and Wollo 1978/79 Southern Ethiopia Failure of Belg rain & 1.4 million people were affected 1982 Northern Ethiopia Late Meher rains & 2 million people were affected 1984/85 Ethiopia Sequential rain failure; 8 million people affected; 1987/88 Ethiopia 7 million people were affected 1990–92 Northern, Eastern, and SW Ethiopia Rain failure and regional conflicts; estimated 4 million people suffering food shortage 1993/94 Tigray and Wollo Widespread food insecurity (7.6 million people were affected), but few deaths or cases of displacement were reported. 1997 Borena, Bale, Omo, Somali Almost 986, 000 people affected 1999 N. & S. Wollo, Wag, Himra; Tigray; B. Gumu, Almost 5 million people affected Gambela, Oromia, SNNPR, Somali 2003/4 All Regions Over 13 million people affected, but the response mitigated the worst potential outcomes. 2005 Somali, Oromia Almost 2.6 million drought disaster affected people 2008/9 All regions Almost 12.6 million People were affected. 2011 S & E. Oromia, Somali Severe food insecurity and 4 million people affected 2015/16 N., E, & SW Ethiopia About 10.2 million people were affected Source: Degefu 1987; Webb and Von Braun 1990; Webb and Braun 1994; Webb et al. 1992; Philip et al. 2017 scale irrigation and water harvesting practices (SWAD SPI-values of 3-, 6-, 9-, 12-, and 24-months timesteps. 2018). Dominant cereal crops produced in the region in- However, the SPI can be computed at any timescales for clude teff (Eragrostis tef), sorghum (Sorghum bicolor), any set of successive months of 1-month up to 72- maize (Zea mays), wheat (Triticum aestivum), barley months. The spatial and temporal drought event analysis (Hordeum vulgare) and pulses (Leguminosae/ Fabaceae). for SPI 1-, 4- and 8-months timesteps are not included in Nevertheless, the larger reliance on agriculture-based in- their studies. Therefore, the analysis of the space-time pat- come, which is under the sensitivity of nature exposed tern of drought risk events of this study is based on the the people to food deficiency (Little et al. 2006; Kahsay rainy months and seasons of 1-, 4- and 8-months time- 2013). All these attributes contributed to the reduction steps,. The space-time drought risk events of the small of productivity in the zone. This might be the reason and big rainy months as well as seasons (SPI: 1-, 4- and 8- that the zone is exposed to food shortage and dependent months timesteps) are appropriate for planning and man- on safety nets for many years. Therefore, an assessment agement of drought and related risks at the local-scale of the spatio-temporal drought risk events is imperative level. This study, therefore, aims to characterize the spatial for policy makers and practitioners in the area under and temporal drought incidences as a function of fre- study. quency, magnitude, intensity and severity status using the The classification and representation of observed SPI model over the northeastern highlands of Ethiopia. spatio-temporal drought risk events using SPI are vital in operational monitoring for policy decisions. Accord- Drought quantification: the standardized precipitation ingly, many studies, but different in study periods and index spatial coverage related to drought were documented in Drought indices are essential to characterize several Ethiopia using SPI (e.g. Suryabhagavan 2017; Degefu and drought features such as the onset and cease time of Bewket 2015; Viste et al. 2013; Gebrehiwot et al. 2011; drought, drought duration, areal extent, severity and fre- Williams and Funk 2011; Edossa et al. 2010; Korecha quency at global, regional and local level (Piechota and and Barnston 2007; Segele and Lamb 2005; Seleshi and Dracup 1996). Owing to the complexity of drought, nu- Zanke 2004). Many of these studies documented either merous operational drought indices have been developed temporal or spatial or both to show the most common and used by meteorologists and climatologists around Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 4 of 17 the world (WMO 2016; Dai 2011). In the international SPI might be more effective in highlighting available works of literature, a large number of studies acknow- moisture conditions than the slow-responding Palmer ledge numerous indices in the detection, analysis, and Index or other currently available indices (WMO 2012). monitoring of drought events for recent and projected As compared to other indices, SPI is an appropriate changes (Loukas and Vasiliades 2004). Among the nu- index for drought analysis to East African countries indi- merous indices, the most commonly used drought index cated by various studies. For instance, Ntale and Gan with its strengths and weaknesses is the Standardized (2003) suggested that SPI is more suitable for monitor- Precipitation Index (SPI) (McKee et al. 1993). ing droughts in East Africa because (i) it is easily The SPI is relatively new application, gaining world- adapted to the local climate, (ii) it needs a modest data wide acceptance and widely used at present because of requirement, (iii) it produces spatially consistent results, its powerfulness, effectiveness, flexibility, versatile and and (iv) it can be computed almost at any time scale. standardized nature (Beth and Brown 2003;WMO 2012; Tsakiris and Vangelis 2004; Mashari Eshghabad et al. Materials and methods 2014). Hence, choosing a precipitation-based drought Description of the study site measure (SPI model) is preferred for its low data re- South Wollo is one of the eleven administrative zones of quirement and computational simplicity for developing the Amhara National Regional State which is located be- 0 0 0 countries like Ethiopia where the access to data is lim- tween 10 10′Nand 11 41′Nlatitudes and38 28′Eand ited (Viste et al. 2013; Degefu and Bewket 2015). 40 5′E longitudes (Fig. 1). The total area of the Zone is SPI values in one region can be directly compared to about 18,157.48 km , which is divided into 19 rural dis- SPI values in a completely different climate zone. It can tricts and four administrative towns (Dessie, Kombolcha, monitor the onset, intensity, and duration of drought Haik, and Mekaneselam) (ANRSPC 2017). Of the total (Hayes et al. 1999; Loukas and Vasiliades 2004). The SPI area, 36.3%, 13.5%, 18.3% and 31.9% are covered by arable is a very suitable index to study the geospatial and tem- land, forest and bush lands, grazing land, and others (e.g. poral variation of drought with spatially invariable re- bare land, buildings, water bodies), respectively (SWAD sults for historic time series analysis (Masih et al. 2014; 2018). The landscape of the South Wollo zone comprises Guttman 1998). The ideal strength of SPI is that precipi- highly diversified and dissected topography (rugged terrain tation anomalies can be consistently calculated over flex- with very steep slopes). Elevation ranges from 927 m (over ible time scales in a consistent fashion. Similarly, the dry plain/Arabat) in the east to 4261 m above sea level drought information can be provided promptly for oper- (Mount Ambaferit) in the west (SWAD 2018;Littleet al. ational drought-monitoring applications (AghaKouchak 2006). It has six livelihood zones, namely, Abay-Beshilo and Nakhjiri 2012). Even, following the discussion at Basin (ABB), Chefa Valley (CHV), Meher-Belg zone,Belg (WMO 2009), drought experts made a consensus agree- zone,Meher zone and eastern lowland sorghum and cattle ment to recommend the SPI for the characterization of (SWS) (USAID 2009). meteorological droughts (Hayes et al. 2011). The SPI The Zone is characterized by a distinctive bi-modal measures moisture to quantify the precipitation deficit rainfall regime, locally known as kiremt and belg seasons. for multiple time scales at any one location and reflects Kiremt is the big rainy season usually extending from the impact of drought on a range of meteorological, agri- June-to-September, and belg is the small rainy season cultural and hydrological applications (Marimon 2016; extending from February-to-May (National Meteoro- Degefu and Bewket 2015; Loukas and Vasiliades 2004). logical Services Agency of Ethiopia/NMSA 1996; Degefu The SPI can be calculated at various timescales for any 1987; Korecha and Barnston 2007; Conway 2000a, set of successive months of 1-month up to 72-months 2000b). Crop production in the entire livelihood zones that one wishes to detect and to illustrate the effect of follow these rainfall regimes but varies among the liveli- drought (Viste et al. 2013; Stagge et al. 2015; Padhee hood zones. Accordingly, the livelihood zones of ABB, et al. 2014; Beth and Brown 2003; AghaKouchak and SWS Meher and CHV are dependent on big kiremt rain- Nakhjiri 2012). But statistically, 1–24 months is the best fall for the cultivation of both long and short cycle crop practice range of application (Guttman 1999). The short production. Conversely, Belg and Meher-Belg crop pro- period time scales SPI-values (1-, 2-, 3-, 4-, 5- and 6- ductions follow the bimodal rainfall regimes (the small months) are appropriate and paramount for measuring belg and big kiremt rains) leading to two harvest periods. the effect of drought on agriculture, soil moisture and The mean annual rainfall ranges between 500 and 900 crop yield reduction (Bussay et al. 1998; Morid et al. mm in the Kolla (arid and semi-arid) area and 950– 2006; Szalai and Szinell 2000). 1100 mm in the Woina Dega (semi-humid) and Dega A 3-month SPI reflects short- and medium-term mois- (cool and humid) areas. The average annual temperature ture conditions and provides a seasonal estimation of also ranges from15°c to 20°c (SWAD 2018). Between precipitation. In primary agricultural regions, a 3-month 1901 to 2016 the minimum temperature ranges were Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 5 of 17 Fig. 1 The geographic location of livelihood zones of South Wollo and Meteorological stations. Source: Extracted from USAID 2009 0 0 6.43 c in December and 11.82 c in June while the max- of the study area (ANRSPC 2017;SWAD 2018). Even imum ranges were 22.93 c in August and December and though the small and large-scale irrigation practices 29.13 c in June. are very minimal, the zone is known by many large Based on the population statistics of ANRSPC and small rivers categorized into two major drainage (2017), theSouth Wollo zonehas atotal population basins (Abay and Awash accounting for about 82% of nearly 3.1 million of whom 49.5% were males and and 18% of the area, respectively) (SWAD 2018). Cat- 50.5% females. The zone is one of the densely popu- tle, goats, sheep, and equines are the major livestock lated areas of the country with 171 persons per km . in the area. However, the contribution of livestock to The old-style subsistence farming (a mix of both the livelihood of the people is constrained by the grain production and rearing of livestock) is the pri- prevalence of livestock diseases (USAID 2009;Little mary form of economic activity in all livelihood zone et al. 2006). Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 6 of 17 Data sources and data analysis techniques (1993) and WMO (2012), have been applied in the quan- For the characterization of spatial and temporal drought tification and classification of SPI drought severity status incidences as a function of frequency, magnitude, inten- of the study. In the computation of drought status classi- sity, and severity status, long-term meteorological data fication of this study, only months with SPI ≤− 1.00 were were acquired from the Ethiopian National Meteorological considered to be as drought initiation and cessation Agency (ENMA-Kombolcha Regional Meteorological Sta- since mild drought (SPI between − 0.99 and 0.99) tion). Since rainfall is not normally distributed, a gamma showed only a slight variation from the normal rainfall probability distribution has been found fitting to the pre- distribution (Łabędzki 2007; Sternberg et al. 2011; cipitation time series frequency distribution for the com- Degefu and Bewket 2015). Therefore, using the SPI_SL_ putation of SPI (Khan and Gadiwala 2013). The gamma 6.exe file programming, moderate drought (SPI between probability density function is expected to fit with the fre- − 1.00 and − 1.49), severe drought (SPI between − 1.50 quency distribution of precipitation over the period and − 1.99) and extreme drought (SPI ≤− 2.00) were gen- (1981–2017). For the time scales of 1-, 4- and 8- months, erated and presented in the form of time series graphs alpha (α) and beta (β) parameters of the gamma probabil- and maps. ity density function were estimated. As a result, the cumu- Temporal drought event characteristics (frequency, lative probability distribution function is approximated magnitude, intensity and severity status) were computed into the standard normal cumulative distribution prob- using one of the most common models, SPI analysis pa- ability function with a mean of zero and a standard devi- rameters. Drought frequency is the return period be- ation of one to determine the probability distribution tween drought events that have severity threshold values function. Conceptually, the SPI is equivalent to Z- score, of SPI ≤− 1. The magnitude of drought event corre- the number of standard deviations that the observed value sponds to the cumulative water deficit during the would deviate from the long-term mean. The statistical drought period below some threshold (SPI-values ≥− 1) procedure is: (Thompson 1999) and drought intensity (D ) is the ratio between the drought magnitude and the duration of the X X event computed as: i− SPI ¼ ð1Þ D ¼− SPI ð2Þ M ij Where: xi = observed precipitation value of the selected i¼1 period during the i th year; Xand σ = the mean and the standard deviation re- M D ¼ ð3Þ spectively over the selected period. In this study, SPI was computed for shorter-term and Where: D drought magnitude, M= intermediate periods to quantify drought events at 1-, 4- n = number of months with drought event at j and 8-months timesteps to represent agricultural/me- timestep; teorological droughts. The 1-month SPI (analogous to D = drought duration. the percent of normal precipitation) reflected short-term However, drought severity (D ) should not be mis- conditions of soil moisture and crop stress during the guided for intensity, which is generally signified to the growing season while the 4-months SPI represented the lowest SPI-value of the drought event (Spinoni et al. short and medium-term moisture conditions applied for 2014). The number of droughts per 100 years for differ- the seasonal estimation of rainfall. The 4-months accu- ent timesteps of the rainy months and seasons (SPI:1-, mulation of SPI was calculated for May and (September) 4- and 8-months) of the study as used by (Łabędzki to assess the seasonal drought of belg and (kiremt). This 2007; Sternberg et al. 2011; Degefu and Bewket 2015; is because of the 4-month definition of seasons in Ghosh 2019) which was estimated as: Ethiopia are the most commonly used (Viste et al. 2013). The 8-months timestep accumulation of SPI was com- N ¼ X100 ð4Þ puted for September (February-to-September), which i;100 i Xn was also a good indicator to assess the drought condi- tion of slowly maturing long-cycle crops like sorghum Where: and maize that typically planted during the belg season Ni,100 = the number of droughts for a timestep i in and harvested following the kiremt season. 100 years; In the computation of SPI, the SPI_SL_6.exe file pro- Ni = the number of years with droughts for a timestep gramming which was available at http://drought.unl.edu/ i in the n-year set; MonitoringTools/DownloadableSPIProgram.aspx has i = the timesteps (1-, 4- and 8- months) and. been employed. The criteria employed by McKee et al. n = the number of years in the data set (37 years). Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 7 of 17 Furthermore, the MK (Mann Kendall) trend test (Ken- Results dall 1975; Mann 1945) has been computed to check the Characteristics of drought events: frequency, magnitude, statistical significance (increasing and decreasing trends) intensity and severity for rainy months and seasons of SPI: 1-, 4- and 8- Table 2 shows the observed drought frequency, magni- months timesteps. The relative strength of MK trend tude, intensity and severity of the small, big, and long- test in time series analysis has been quantified using term rainy months and seasons as computed using SPI. Sen’s(1968) slope estimator. As indicated by Chattopad- The highest frequency of drought years for the small hyay and Edwards (2016), Sen’s slope estimator has been rainy months within the moderately to extremely dry the commonly used estimator due to its relative insensi- (SPI ≤− 1.00) found to be 9 years; in February and May tivity and robustness to extreme values. The MK statistic account for about 24%of the total years over the study S of the series x is calculated as: period. The drought magnitudes of the small rainy months were ranged from − 9.45 in March to − 11.83 in N−1 N May. Nevertheless, the drought intensity found to be be- X X S¼ sgn X X ð5Þ j− i tween the moderately dry in May with SPI = − 1.23 and se- i¼1 j¼iþ1 verely dry in March with SPI = − 1.58. The extreme severe droughts were observed in March 2008 and April 1984 Where: with the severity of peak SPI-values of − 2.49 and − 2.23, N = the number of data points; respectively. Likewise, the highest drought frequency in X and X = the time series observations. Assuming (x the big rainy months was 6 years in June that account for i j j – x )= θ, the value of sgn (θ) is computed from: 15% of the total drought risk years. The observed magni- tude of drought events ranged from − 9.7 in July to − 5.69 in September. The moderately dry drought intensity was observed in June, July, and September while the extremely þ1………θ > 0 dry drought intensity was observed in August. The worst sgnðÞ θ ¼ 0………θ ¼ 0 ð6Þ droughts were recorded in July 2015, August 1984, and −1………θ < 0 September 2009 with the severity peak of SPI values of − 2.82, − 3.61, and 2.31, respectively. While Positive values of S indicate increasing trends, Seasonally, the drought frequency for SPI 4-months of negative S indicates decreasing trends. Under the hy- both belg and kiremt as well as long-term season of SPI pothesis of independent and randomly distributed ran- 8-months at the end of May and September account for dom variables, for large samples, when n ≥ 10 (in some about 4 years (9%) of the total drought years under papers n ≥ 8), the σ statistic is approximately normally study. A slight difference in the observed drought risk distributed, with zero mean and variance as follows: magnitude was shown among the SPI4-months (belg and kiremt) and SPI8-months (long-term) seasons. Nonethe- nnðÞ −1ðÞ 2nþ5 σ ¼ ð7Þ less, the drought intensity was varied from severely dry in SPI4-months (belg) and SPI8-months (long-term) sea- sons with SPI = − 1.92 and − 1.79 respectively to ex- As a consequence, the standardized normal deviate (Z- tremely dry in SPI4-months (kiremt) season with SPI = statistics) distribution has been then calculated as: − 2.0. The SPI 4- and 8-months experienced severe drought events in 1984, 1999 and 2015 with severity S−1 peak SPI-values of − 2.83, − 2.84 and − 2.32, respectively. > if S > 0 Z ¼ ð8Þ 0if S ¼ 0 S þ 1 > Temporal trends of drought events if S < 0 Using the SPI, the drought condition of the study for rainy months (SPI-1), belg and kiremt (SPI-4), and long- While Positive values of Z signify increasing trends, term (SPI-8) rainy seasons covering the period of 1981– negative Z signifies decreasing trends. In the spatial ana- 2017 were examined. A mix of dry and wet years have lysis of drought severity status, Inverse Distance been observed in February, March, April, and May as Weighted, IDW average interpolation method was used well as belg season but with temporal variation in sever- in the ArcGIS. The SPI values generated from SPI_SL_ ity, magnitude and intensity (Fig. 2a-d and Fig. 4a). Sev- 6.exe file programming was used as an input for map- eral drought risk incidences with SPI ≤− 1.00 were ping the spatial distribution of drought severity status. detected. It was observed that March 2008, April 1984, As a consequence, the 1-, 4- and 8-months timesteps and the belg season of 1999 were characterized by ex- spatial distributions of drought incidences were mapped. treme drought events with SPI ≤− 2 in the past 37 years. Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 8 of 17 Table 2 Drought risk characteristics of the rainy months and seasons for SPI ≤ 1.0 values Month Observed Drought risk Frequency Magnitude Intensity Severity (years) Peak SPI-Value Year/s February 9 −10.92 −1.23 −1.32 1999, 2000, 2012 March 6 −9.45 −1.58 −2.49 2008 April 7 −10.58 −1.5 −2.23 1984 May 9 −11.83 −1.37 − 1.52 1981 Feb-to-May (Belg: SPI-4) 4 −7.66 −1.92 −2.84 1999 June 6 −7.3 −1.22 −1.78 1981 July 5 −9.7 −1.94 −2.82 2015 August 3 −7.1 −2.37 −3.61 1984 September 3 −5.69 −1.9 −2.31 2009 June-to-Sept (Kiremt: SPI-4) 4 −8.03 −2 −2.32 2015 Feb-to-Sept (SPI-8) 4 −7.14 −1.79 −2.83 1984 Severe drought events with SPI values between − 1.5 Similarly, Fig. 3a-d and Fig. 4b signify a mix of drought and − 1.99 were also observed in March (1999 and 2000) and wet frequencies in the timesteps of SPI1-month ob- and April (1991, 1999 and 2001). Moreover, the largest servations (June, July, August and September) and SPI4- drought risk events (about 23 years) for February, March, months of the current season. A mix of dry and wet April, and May were shown under the moderate drought years have been observed in June, July, August, and Sep- intensity class. Nevertheless, 6 and (2) years were under tember as well as kiremt season but with temporal vari- the severe and (extreme) drought intensity classes. ation in severity, duration, magnitude, and intensity. It Drought ends when the SPI value becomes positive was witnessed from Fig. 3a-d and Fig. 4b extreme while drought begins when SPI value becomes negative. drought events with SPI ≤− 2 were investigated in July Since drought ends when the SPI value becomes posi- 1987 and 2015, August 1984 and 1993, September 2009 tive, the increasing drought frequency of February found and kiremt season of 1984, 1987 and 2015. Severe to be by far the highest (Table 2). This increasing drought also was detected in June 1981, July 1981 and drought frequency of February was substantiated by a 1984, and September 1984. Moreover, moderate drought significant decline of rainfall as revealed in the MK trend events were detected in June 1982, 1985, 1987, 1988 and test (Table 3) and regression trend line (Fig. 2a). 2017, July 1992, August 1990 and September 1997. It Fig. 2 SPI 1-month anomalies and trends of belg rainy months: (a) SPI-1: February; (b) SPI-1: March; (c) SPI-1: April; (d) SPI-1: May Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 9 of 17 Table 3 Mann-Kendall’s trend analysis and the probability of drought events per 100-year for the rainy months and seasons at 1-, 4- and 8-months timesteps SPI timesteps Months/Season MK test P-values Slope Probability of drought events per a 100-year 1-month February −0.2539** 0.0305 −0.0366 24 March −0.1388 0.2337 −0.0189 14 April −0.1508 0.1952 −0.02 19 May 0.0693 0.556 0.0149 24 June 0.2493** 0.0327 0.0311 16 July 0.1851 0.1105 0.0291 14 August 0.1458 0.2092 0.0188 8 September −0.0532 0.6559 −0.0077 8 4-month Feb-to-May −0.1583 0.1736 −0.0248 11 June-to-Sept 0.2122 0.067 0.0281 11 8-month Feb-to-Sept 0.1414 0.2238 0.0161 11 Note: ** is statistically significant at P < 0.05 was also shown in Fig. 4c that moderate drought event Incidences of drought trends and its probability per was detected in 2009 while extreme droughts were de- 100-years over the small and big rainy months (SPI-1) tected in 1984 and 2015 during the long-term rainy sea- and seasons (SPI-4 and SPI-8) timesteps were shown son of SPI 8-month timestep. Furthermore, the largest (Table 3). The Mann-Kendall trend test of SPI-values drought event years (about 8 years) for June, July, Au- showed statistically significant decreasing and (increas- gust and September were shown under the moderate ing) trends for small and (big) rainy months of February drought intensity class while three and (six) years were and (June) at p < 0.05. However, the 1-month timestep under the severe and (extreme) drought intensity classes. values of March, April, September and 4-month time- Seasonally, belg and (kiremt) were under severe and (ex- steps of February-to-May showed statistically non- treme) drought intensity classes for 3 years while 2 years significant decreasing rends. The 1-month timestep of the long-term season (SPI8-months) were under ex- values of May, July, August, and 4-month timestep treme drought intensity class. However, values of SPI1- (June-to-September) on the other hand showed statisti- (June, July, August) months and SPI 4- months (kiremt cally non-significant increasing trends. The 8-month season) showed increasing trends except SPI1-month of long-term seasonal timestep (Feb-to-Sept) value also September. showed a statistically non-significant increasing trend. Fig. 3 SPI 1-month anomalies and trends of kiremt rainy months: (a) SPI-1: June; (b) SPI-1: July; (c) SPI-1: August; (d) SPI-1: September Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 10 of 17 Fig. 4 The SPI anomalies and trends of the rainy seasons. a SPI-4: belg season (February-to-May); b SPI-4: kiremt season (June-to-September) and c SPI-8 long-term seasons (February-to-September)] The predicted number of droughts for 100 years rainy western and southern parts of the ABB, Meher, Belg, months and seasons timesteps of 1-, 4- and 8-months Meher-Belg livelihood zones that ranged from 6 years based on eq. (4) is also presented (Table 3). The max- (15%) to 8 years (21%) of the total observation periods. imum predicted drought frequency events (except for Except for the areas around the stations of Kelala and August and September) were expected to be in the Wereillu, which experienced the highest drought risks shorter timesteps of 1-month as compared to 4- month (8–9 years) in the south, the remaining livelihood zones’ and 8-month timesteps. The 1-month predicted drought drought frequency during April varied from 6 years frequency events expected to be between 24 and (8) (15%) to 7 years (18%) of the total study periods. Simi- years for February and May and (for August and Sep- larly, the eastern part including ABB, Meher-Belg,SWS, tember). The seasonal predicted drought frequency Meher, CHV, and Belg livelihood zones and the western events of all the 4-month (Feb-to-May and June-to-Sept) part of ABB, Meher experienced frequent drought risks and 8-month (Feb-to-Sept) expected to be 11 years. ranged from 9 (24%) to 11 years (30%) of the total study years during May. Spatial patterns of drought incidence The spatial pattern of drought frequency maps in the The spatial patterns of drought frequency incidence big rainy months (June, July, August, and September) maps for 1-, 4- and 8-months timesteps for moderately also revealed the complex and high local-scale variation dry-to- extremely dry intensity classes of drought events as indicated in Fig. 6. It was found that all the livelihood were depicted in Figs. 5, 6 and 7. The spatial pattern of zones except the eastern and southeastern periphery drought risk frequency maps over the livelihood zones (ABB, Belg, and CHV) experienced more frequent in the small rainy months (February, March, April, and drought risks ranged from 7 years (18%) to 9 years (24%) May) was exhibited complex and high local-scale vari- of the observation periods during June. More frequent ation as indicated in Fig. 5. All the livelihood zones drought events ranged from 6 to 7 years, which accounts (ABB, Meher, Belg, CHV, Meher-Belg, and SWS) except between 15 and 18% experienced during July in the around the station of Amba Mariam exposed to more northwest, southern and northern parts of Meher, ABB, frequent drought risks ranged from 9 years (24%) to 12 Meher-Belg, SWS, and Belg livelihood zones. The inci- years (33%) of the total observation years. The drought dences of drought frequency during August were found frequency of March found to be higher in the northern, to be ranged from 3 years (6%) to 6 years (15%) in the Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 11 of 17 Fig. 5 The 1-month timesteps spatial pattern of drought frequency events (SPI≤−1.0) for the small rainy months of February, March, April, and May eastern half including all the livelihood zones while the Figure 7 presents the spatial pattern of drought fre- western half, including ABB and Meher livelihood zones quency maps for 4- and 8-months timesteps of belg and covered about 7 years (18%) and 9 years (24%) of the kiremt, and the long-term (both belg and kiremt). The study period. During September, the incidence of SPI4-months of belg season the northern half, including drought frequency was found to be more pronounced in Belg, Meher-Belg, CHV, and SWS livelihood zones the northeastern and southwestern parts including ABB, (where belg crop harvesting dominate) experienced Meher, Belg, and Meher-Belg livelihood zones that varied drought frequency events varied from 4 to 6 years (9– from 7 years (18%) to 9 years (24%). However, inci- 15%) while the southern half, including ABB and Meher dences of drought frequency varied between 4 years livelihood zones experienced 7–9 years (18–24%) of (9%) and 6 years (15%) in the northwestern, southern drought frequency events of the total observation pe- and southeastern parts including ABB, Meher, Belg, riods. Conversely, during SPI:4-months of kiremt season, Meher-Belg, CHV, and SWS livelihood zones of the the southern half comprising part of ABB, Meher, Belg, study area. and CHV livelihood zones (where kiremt season crop Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 12 of 17 Fig. 6 The 1-month timesteps spatial pattern of drought frequency events (SPI≤− 1.0) for the big rainy months of June, July, August, and September harvesting dominate) experienced 5–7 years (12–18%) Discussion drought frequency incidences while the northern cover- The rainfed agricultural activities of the study area de- ing parts of ABB, Meher, Belg, Meher-Belg, SWS, and pend on two rainy seasons: belg and kiremt. However, CHV were experienced 8-11 years (21–30%) of drought the incidence of drought disaster due to the shortfall of frequency events of the total observation periods. The rainfall was the most important concern of this study. In long-term spatial tendency of drought incidences (Feb- this study, the temporal and spatial spread of drought ruary-to-September) were also observed in SPI:8- frequency, magnitude, intensity and severity have been months. Almost the eastern half except around the sta- examined. Of the total observation years, the highest tions of Wuchalle and Kombolcha, experienced 8–9 drought frequency was detected for February. The study years (21–24%) drought frequency incidences. The results indicated that March 2008 and April 1984 were southern and western parts on the other hand experi- the most severe drought months. However, the belg sea- enced 5–7 years (12–18%) drought frequency occur- son in 1999 was the driest year during the record. In rences of the total study periods. agreement with this, Conway (2000a, 2000b) and Viste Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 13 of 17 Fig. 7 The spatial pattern of drought frequency events (SPI≤− 1.0) for SPI: 4- month (Feb-to-May, belg); SPI: 4- month (June-to-Sept, kiremt); and SPI: 8-month (Feb-to-Sept) timesteps et al. (2013) reported that the belg season in 1999 was frequency of drought incidences for March-to-May the driest year during the record over Ethiopia. A gen- months. The increasing of drought incidences in the 4- eral tendency of increasing drought risk frequency has months belg season was detected due to a rainfall deficit been observed in the study period of belg rainy months in the rainy months of February-to-May. The relative and season. The MK test for belg rainy months and sea- abrupt decline was seen since 1996, which is concurrent son (1-and 4-months) timesteps also verifies the general to Funk et al. (2005) at the national level. Multiple study decline of rainfall. Lyon and DeWitt (2012) and Lyon reports in agreement with this study revealed that the (2014) also showed the decline of rainfall or increase greatest continual decline of rainfall has occurred during Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 14 of 17 the belg season in the post-1980s (Tierney and Ummen- years on the record (Fig. 4c). Similarly, Funk et al. (2005) hofer 2015;Funk et al. 2008;Funket al. 2005). For ex- reported that 1984 stands the main horrible year on the ample, in 1984, 72% of Ethiopia’s regions were affected by record although the 1997, 2002 and 2004 undergone drought in the belg season (Shanko and Camberlin 1998). poor rainfall in the long-term (March-to-Sept), belg and Likewise, the highest drought frequency was observed kiremt seasons. March-to-Sept rainfall totals have also in June during the big rainy months. However, extremely exhibited dry periods in the mid-1980s and mid-2000s. severe droughts were detected in August 1984, July 1987 The drought risk events of months in the belg season and 2015, and September 2009. Thus, 1984, 1987 and were threefold greater than the months in kiremt season 2015 were the three driest years in the record. Drought under moderate drought intensity class. Conversely, the risk frequency and magnitude were more pronounced drought risk events of months in kiremt season were during the months of belg as compared to kiremt. None- threefold greater than that of the belg season under ex- theless, the drought risk intensity was more weighted treme drought intensity class. during the months of kiremt. In agreement of this, Viste In the 100-years, the predicted drought frequency et al. (2013), Segele and Lamb (2005) reported that the events for 1-month timesteps were relatively more fre- kiremt season of 1982, 1984 and 1987 was severely dry quent than the 4-and 8-months seasonal timesteps ex- over Ethiopia, particularly in the northeastern half of the cept for August and September. In agreement with this, country, which primarily caused by the missing rain in Łabędzki (2007) and Degefu and Bewket (2015) reported July and August. Lyon (2014) also revealed that the kir- that drought events of shorter timesteps occur more fre- emt season (June, July, and August) rainfall during 1982 quently and last a shorter time than drought events of and 1984 was below-average rainfall for prolonged longer timesteps, which are more concentrated and last drought events that lead to the most devastating impact longer (changes slowly). However, Sternberg et al. (2011) for 1950–2010 in northern Ethiopia. Further, Jjemba reported contrasting results that frequency of drought et al. (2017) and Philip et al. (2017) reported that the events have been the lowest at 1-month timesteps and worst drought of 2015–16 in the northern and central highest for the longer 17-months timestep. part of Ethiopia because of belg rains had failed and soon The spatial distribution of drought frequencies was ex- after, kiremt rains were severely delayed, erratic and hibited by the presence of complex patterns for 1-, 4-, below the long-term average (deficit of 167 mm). Segele and 8-months (Figs. 5, 6 and 7). The northern and and Lamb (2005) also revealed that the greatest dam- (southern) half during 4-month belg and (kiremt) sea- aging droughts in Ethiopia are connected with the failure sons experienced less frequent drought risk events. Con- of kiremt rains. Specifically, Suryabhagavan (2017), testi- versely, the southern and (northern) half during 4- fied that South Wollo as one of the highest peaks of month belg and (kiremt) seasons experienced a higher drought recorded area in 1984. A general decreasing frequency of drought events. Almost the eastern half ex- trend tendency of drought events has been observed cept around the stations of Wuchalle and Kombolcha, during 1-, 4- and 8-months timesteps of kiremt season. experienced the most drought frequency events while The MK test also confirms increasing trend tendencies, the southern and western parts experienced less drought but it was only statistically significant for June. frequency occurrences during the long-term 8-month A cyclical oscillation of wetter and drier events has timestep. In this regard, belg, and long-term (in both shown in 1-, 4- and 8-months timesteps. Wetter condi- belg and kiremt) seasons crop harvesting areas were tions were illustrated in the period of the 1990s. In more exposed to frequent drought risks. The 4-month contrast, increasing drought risk event frequencies were of kiremt season drought frequency pattern influenced illustrated in the 1980s and 2000s (Fig. 3b-c) and and manifested over the 8-months drought frequency (Fig. 4b-c). In corroboration of this, for the 1970s on- spatial pattern. wards, drought has occurred during kiremt season in 1982, 1984, 1897, 1990, 1991, 1995, 1997, and 2002 Potential implications of drought incidences due to the deficit of rainfall in June, July, August, and The SPI values of the rainy months of 1-month as well September (Conway 2000a, 2000b; Korecha and as 4- and 8-months timesteps had an illustrious impact Barnston 2007). The influence of rainfall deficit dur- on the rainfall of the study periods that reflects the ing the kiremt season affects a significant portion of prompt occurrence of agricultural drought. In this re- the county. For instance, in 1984 about 53% of the gard, Łabędzki (2007) confirmed that the 1- to 6-months regions of Ethiopia were affected by drought in the timesteps of SPI values soundly reflects well the quick kiremt season (Shanko and Camberlin 1998). development of agricultural drought, the exacerbating of The 8-month long-term time step (Feb-to-Sept) dem- the actual state of water conditions and the negative ef- onstrated the drought events for the 1980s and 2000– fects of rainfall deficit in agriculture (soil moisture, 2015. The 1984 and 2015 were the foremost terrible groundwater table depth, crop yield). The recurrent Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 15 of 17 incidence of droughts for 1-, 4-, and 8-months has im- shocks that increase vulnerability and then step up to a portant implications on the socio-economy of rainfed cycle of poverty (Funk et al. 2012). Mekonen and Berlie agricultural practices and productions for smallholder (2019) and McSweeney et al. (2008) also reported that farmers. The incidences of drought owing to the short- the respective decreasing and increasing trends of rain- fall of rainfall during the months of belg (Feb-to-April), fall and temperature were the potential threats to the kiremt (June-to-Sept) and long-term rain (Feb-to-Sept) agricultural sector through pest intensification. seasons were influencing the agricultural production. The greatest drought risk incidence of months in the Conclusions belg and (kiremt) seasons under moderate and (extreme) This study presented the spatiotemporal drought risk drought intensity classes highly impacted the agricultural events in the northeastern highlands of Ethiopia using practices. The influence of moderate and extreme Standardized Precipitation Index, SPI. The results of the drought risk expected to be via the timely deficit of rain study confirm that highly complex and localized tem- at planting and critical stage of crop growing that finally poral and spatial patterns of drought risk events were exposed the society to shift to food aid. For example, the detected. This helped to identify and characterize local reports of Viste et al. (2013), Block (2008), Prasad and droughts. The majority of 1–4- and 8-months timestep Staggenborg (2009), World Bank (2006a, 2006b) and underwent severe and extreme (SPI ≤− 1.50) drought United Nations Environment Program/UNEP (2011) re- events. However, the detected frequency and magnitude vealed that from the 1970s onwards, the decline of rain- of drought risk events were more pronounced during fall amount and the increasing frequency of drought the rainy months of the belg season. directly influenced water shortages and limited crop The 1-month belg and (kiremt) rainy months experi- growth and development, and reduced yield. This fur- enced the largest drought risk events under the moderate ther threatened millions of people in the country that and (extreme) drought intensity classes. Moreover, the 4- demanded emergency humanitarian assistance. In rela- month belg and (the 4- month kiremt and 8-month long- tion to this, Webb et al. (1992), Von Braun (1991), and term) seasons experienced the largest drought risk events Webb and Von Braun (1990) also indicated that the pri- under the severe and (extreme) drought intensity classes. mary effects of drought on crop production and yields Increased drought risk frequency events were observed in were quite strong. For instance, a 10% decline in rainfall the 1980s and 2000s onwards for kiremtand long-term results in an average reduction in cereal yields of about seasons. A steady increased drought risk event frequency 4.2%. was detected since 1980s during belg season. In the spatial The deficit of February-to-May rainfall is also be pattern of drought, the southern part during belg,the threatening those slowly maturing long-cycle crop pro- northern and eastern during kiremt and the eastern half duction like sorghum and maize in the long-cycle crop during the long-term rainy seasons experienced more fre- growing areas of the study planted during belg and har- quent drought risk events. vested after kiremt seasons. Degefu (1987) and Funk The observed temporal and spatial drought risk events et al. (2005) report that poor belg season rainfall per- indicate a potential threat to the rainfed agricultural formance not only affecting belg crops but also adversely practices that gradually jeopardized the smallholder impacting the long-season crop production such as sor- farmers to food insecurity and socio-economic vulner- ghum and maize, the country’s major food crops. The ability. The findings of this study also could be a rudi- detected incidence of drought risks of 8-months (Feb-to- mentary stage to enhance drought risk management Sept) was also influencing the long-cycle crop produc- strategy through the appraisal of the rainfed agricultural tion such as Sorghum, Maize, and Millet growing areas practices in the study area. Therefore, the documenta- (Meher-Belg livelihood zones) owing to deficit of rainfall tion and informed assessment of drought frequency, in belg and kiremt seasons. In this regard, Funk et al. magnitude, intensity and severity based on the livelihood (2012) revealed that insufficient amount of belg and kir- zones are essential for drought risk management, early emt rainfall likely have negative impacts on agricultural warning responses, local-scale planning, and food secur- production of slowly maturing varieties of crops (Maize, ity management. Finally, the study recommended further Sorghum, and Millet) and food availability. Mahoo et al. research on additional indices of climatic variables such (2013) also reported that recurrent drought, changes in as evapotranspiration and soil water content. the consistent crop planting and shift in crop types, as well as rainfall variability related with the amount, tim- Abbreviations ing, and intensity, are impacting the agricultural prac- ANRSPC: Amhara National Regional State Plan Commission; CRED: Center for Research on the Epidemiology of Disaster; IPCC: Intergovernmental Panel on tices leading to frequent crop failures, losses of life and Climate Change; MK: Mann Kendal; NMSA: National Meteorological Service property. This means that more frequent droughts, and Agency; PDSI: Palmer Drought Severity Index; SPI: Standardized Precipitation drier climate in general, maybe producing repeated Index; SWAD: South Wollo Agriculture Department; UNEP: United Nations Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 16 of 17 Environmental Program; WFP: World Food Program; WMO: World Edossa DC, Babel MS, Gupta AD (2010) Drought analysis in the awash river basin, Meteorological Organization Ethiopia. Water Resour Manag 24(7):1441–1460. https://doi.org/10.1007/ s11269-009-9508-0 Endale D (1993) The Ethiopian famines, entitlements, and governance. World Acknowledgments Institute for Development Economics Research of the United Nations We are grateful to the Ethiopian National Meteorological Agency-East Amhara University, Annankatu 42 C, 00100 Helsinki, Finland. Working papers No. 102 Meteorological service center (Kombolcha) for providing us the monthly rainfall Funk C, Dettinger MD, Michaelsen JC, Verdin JP, Brown ME, Barlow M, Hoell A data. We are also indebted to the two anonymous reviewers for their insightful (2008) Warming of the Indian Ocean threatens eastern and southern African and thoughtful feedback for further improvement of the manuscript. food security but could be mitigated by agricultural development. Proc Natl Acad Sci 105(32):11081–11086. https://doi.org/10.1073/pnas.0708196105 Authors’ contributions Funk C, Rowland J, Eilerts G, Kebebe E, Biru N, White L, Galu G (2012) A climate The first author collects and analyzes the data, and wrote the manuscript. All trend analysis of Ethiopia. In: US Geol Surv, Fact Sheet, p 3053 authors read, edit and approved the manuscript. Funk C, Senay G, Asfaw A, Verdin J, Rowland J, Michaelson J et al (2005) Recent drought tendencies in Ethiopia and equatorial-subtropical eastern Africa. Funding FEWS-NET, Washington DC The first author is grateful to Bahir Dar and Debremarkos Universities for Gautam M (2006) Managing drought in sub-Saharan Africa: policy perspectives. funding support as part of his PhD study. Invited paper prepared for presentation at the International Association of Agricultural Economists, Gold Coast, Australia Availability of data and materials Gebrehiwot T, van der Veen A, Maathuis B (2011) Spatial and temporal The datasets used and/or analyzed during the current study are available assessment of drought in the northern highlands of Ethiopia. Int J Appl Earth from the corresponding author on reasonable request. Obs Geoinf 13(3):309–321. https://doi.org/10.1016/j.jag.2010.12.002 Ghosh KG (2019) Spatial and temporal appraisal of drought jeopardy over the Competing interests Gangetic West Bengal, eastern India. Geoenviron Disas 6(1):1. https://doi.org/ The authors declare that they have no competing interests. 10.1186/s40677-018-0117-1 Gommes R, Petrassi F (1996) Rainfall variability and drought in sub-Saharan Africa Received: 9 August 2019 Accepted: 28 January 2020 SD dimensions, FAO Guha-Sapir D, Vos F, Below R, Ponserre S (2012) Annual disaster statistical review 2011: the numbers and trends Centre for Research on the Epidemiology of References Disasters (CRED) AghaKouchak A, Nakhjiri N (2012) A near real-time satellite-based global drought Guttman NB (1998) Comparing the palmer drought index and the standardized climate data record. Environ Res Lett 7(4):044037. https://doi.org/10.1088/ precipitation index. JAWRA J Am Water Resour Assoc 34(1):113–121. https:// 1748-9326/7/4/044037 doi.org/10.1111/j.1752-1688.1998.tb05964.x ANRSPC (2017) Amhara National Regional State Plan Commission 2015/2016 Guttman NB (1999) Accepting the standardized precipitation index: a calculation budget year annual statistical bulletin, Ethiopia algorithm 1. JAWRA J Am Water Resour Assoc 35(2):311–322 Ashraf M, Routray JK (2013) Perception and understanding of drought and Hao Z, AghaKouchak A, Nakhjiri N, Farahmand A (2014) Global integrated coping strategies of farming households in north-West Balochistan. Int J Dis drought monitoring and prediction system. Scientific data 1:140001. https:// Risk Reduct 5:49–60. https://doi.org/10.1016/j.ijdrr.2013.05.002 doi.org/10.1038/sdata.2014.1 Bannayan M, Sanjani S, Alizadeh A, Lotfabadi SS, Mohamadian A (2010) Hayes M, Svoboda M, Wall N, Widhalm M (2011) The Lincoln declaration on Association between climate indices, aridity index, and rainfed crop yield in drought indices: universal meteorological drought index recommended. Bull northeast of Iran. Field crops research 118(2):105–114. https://doi.org/10. Am Meteorol Soc 92(4):485–488. https://doi.org/10.1175/2010BAMS3103.1 1016/j.fcr.2010.04.011 Hayes MJ, Svoboda MD, Wiihite DA, Vanyarkho OV (1999) Monitoring the 1996 Beth L, Brown TJ (2003) A comparison of precipitation and drought indices related drought using the standardized precipitation index. Bull Am Meteorol Soc to fire activity in the US. In: 5th symposium on fire and Forest meteorology 80(3):429–438 Block PJ (2008) An assessment of investments in agricultural and transportation Jeyaseelan AT (2003) Droughts and floods assessment and monitoring using remote infrastructure, energy, and hydroclimatic forecasting to mitigate the effects sensing and GIS. In: Satellite remote sensing and GIS applications in agricultural of hydrologic variability in Ethiopia. In: CPWF working paper 01 the CGIAR meteorology (Vol. 291). World Meteorol. Org. Dehra Dun, India Geneva challenge program on water and food, Colombo, Sri Lanka, p 53 Jjemba E, Singh R, Arrighi J (2017) Extreme drought in Ethiopia stretches drought Bussay AM, Szinell C, Hayes M, Svoboda M (1998) Monitoring drought in Hungary management systems using the standardized precipitation index. In: Annales Geophysicae, Kahsay B (2013) Diagnosis and intervention plans for south Wollo zone, Amhara Supplement (Vol. 11) region: livestock and irrigation value chains for Ethiopian smallholders (LIVES) Butterfield R (2011) Climate changes in East Africa https://www.weadapt.org/ Google Scholar knowledge-base/national-adaptation-planning/climate-changes-in-east-africa Kendall MG (1975) Rank correlation methods, 4th edn. Griffin, London Centre for Research on the Epidemiology of Disasters/CRED (2016) International Khan MA, Gadiwala MS (2013) A study of drought over Sindh (Pakistan) using disaster database http://www.emdat.be/disaster_profiles/index.html standardized precipitation index (SPI) 1951 to 2010. Pakistan J Meteorol 9(18):15–22 Chattopadhyay S, Edwards DR (2016) Long-term trend analysis of precipitation Korecha D, Barnston AG (2007) Predictability of June–September rainfall in Ethiopia. and air temperature for Kentucky, United States. Climate 4(1):10. https://doi. Mon Weather Rev 135(2):628–650. https://doi.org/10.1175/MWR3304.1 org/10.3390/cli4010010 Łabędzki L (2007) Estimation of local drought frequency in Central Poland using Conway D (2000a) Some aspects of climate variability in the northeast Ethiopian the standardized precipitation index SPI. Irrigation and Drainage J Int highlands-Wollo and Tigray. Sinet: Ethiopian J Sci 23(2):139–161. https://doi. Commission Irrigation Drain 56(1):67–77. https://doi.org/10.1002/ird.285 org/10.4314/Sinet. v23i2.18163 Little PD, Stone MP, Mogues T, Castro AP, Negatu W (2006) ‘Moving in place’: Conway D (2000b) The climate and hydrology of the upper Blue Nile River. drought and poverty dynamics in south Wollo, Ethiopia. J Dev Stud 42(2): Geogr J 166(1):49–62. https://doi.org/10.1111/j.1475-4959.2000.tb00006.x 200–225. https://doi.org/10.1080/00220380500405287 Dai A (2011) Drought under global warming: a review. Wiley Interdiscip Rev Clim Loukas A, Vasiliades L (2004) Probabilistic analysis of drought spatiotemporal Chang 2(1):45–65. https://doi.org/10.1002/wcc.81 characteristics inThessaly region, Greece. Nat Hazards Earth Syst Sci 4(5/6):719–731 Degefu MA, Bewket W (2015) Trends and spatial patterns of drought incidence in Lyon B (2014) Seasonal drought in the greater horn of Africa and its recent the Omo-Ghibe River basin, Ethiopia. GeografiskaAnnaler: Ser A, Phys increase during the march–may long rains. J Climate 27(21):7953–7975. Geograph 97(2):395–414. https://doi.org/10.1111/geoa.12080 Degefu W (1987) Some aspects of meteorological drought in Ethiopia. https://doi.org/10.1175/JCLI-D-13-00459.1 Cambridge University Press, Cambridge, pp 23–36 Lyon B, DeWitt DG (2012) A recent and abrupt decline in the east African long Dorosh P, Rashid S (2013) Food and agriculture in Ethiopia: Progress and policy rains. Geophys Res Lett 39(2). https://doi.org/10.1029/2011GL050337 challenges. University of Pennsylvania Press, Philadelphia. International MahooH, RadenyM,KinyangiJ,Cramer L (eds)(2013) Climatechange Journal of the Commons, 7(2):573–574. https://doi.org/10.18352/ijc.448 vulnerability and risk assessment of agriculture and food security in Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 17 of 17 Ethiopia: which way forward? CCAFS working paper no. 59. CGIAR Stagge JH, Tallaksen LM, Gudmundsson L, Van Loon AF, Stahl K (2015) Candidate Research Program on Climate Change, Agriculture and Food Security distributions for climatological drought indices (SPI and SPEI). Int J Climatol (CCAFS), Copenhagen 35(13):4027–4040 Mann HB (1945) Nonparametric tests against trend. Econometrica 13:245–259 Sternberg T, Thomas D, Middleton N (2011) Drought dynamics on the Mongolian Marimon DC (2016) Contributions to the knowledge of the multitemporal spatial steppe, 1970–2006. Int J Climatol 31(12):1823–1830. https://doi.org/10.1002/ patterns of the Iberan peninsula droughts from a geographic information joc.2195 service perspective, Universitat Autònoma de Barcelona, PhD thesis. Suryabhagavan KV (2017) GIS-based climate variability and drought Accessed from https://www.tdx.cat/handle/10803/379829#page=1. characterization in Ethiopia over three decades. Weather Clim Extremes 15: 11–23. https://doi.org/10.1016/j.wace.2016.11.005 Mashari Eshghabad S, Omidvar E, Solaimani K (2014) Efficiency of some SWAD (2018) South Wollo Zone Agricultural Department 2016/17 budget year meteorological drought indices in different time scales (case study: Tajan annual report. SWAD, Dessie Basin, Iran). ECOPERSIA 2(1):441–453 Szalai S, Szinell CS (2000) Comparison of two drought indices for drought Masih I, Maskey S, Mussá FEF, Trambauer P (2014) A review of droughts on the monitoring in Hungary—a case study. In: Drought and drought mitigation in African continent: a geospatial and long-term perspective. Hydrol Earth Syst Europe, pp 161–166 Sci 18(9):3635. https://doi.org/10.5194/hess-18-3635-2014 Thompson SA (1999) Hydrology for Water Management. CRC Press, London, pp McKee, T. B., Doesken, N. J., & Kleist, J. 1993. The relationship of drought frequency 286-317. https://doi.org/10.1201/9780203751435 and duration to time scales. In Proceedings of the 8th conference on applied Tierney JE, Ummenhofer CC (2015) Past and future rainfall in the horn of Africa. climatology. Boston, American Meteorological Society, (17)22, 179–183 Sci Adv 1(9):e1500682. https://doi.org/10.1126/sciadv.1500682 McSweeney C, New M, Lizcano G (2008) UNDP climate change country profiles: Trenberth K, Meehl J, Masters J, Somerville R (2011) Current extreme weather and Ethiopia Available from http://country-profiles.geog.ox.ac.uk climate change. Climate communication science and outreach Mekonen AA, Berlie AB (2019) Spatiotemporal variability and trends of rainfall Tsakiris G, Vangelis H (2004) Towards a drought watch system based on spatial and temperature in the northeastern highlands of Ethiopia. Modeling Earth SPI. Water Resour Manag 18(1):1–12. https://doi.org/10.1023/B:WARM. Syst Environ:1–16. https://doi.org/10.1007/s40808-019-00678-9 0000015410.47014.a4 Mirza MMQ (2003) Climate change and extreme weather events: can developing United Nations Environment Program/UNEP (2002) Global environment outlook countries adapt? Clim Pol 3(3):233–248. https://doi.org/10.1016/s1469- 3: past, present and future perspectives. Environ Manag Health 13(5):560–561 3062(03)00052-4 United Nations Environment Program/UNEP (2011) Food security in the horn of Morid S, Smakhtin V, Moghaddasi M (2006) Comparison of seven meteorological Africa: the Implicationsof a drier, hotter and more crowded future Ecosystem indices for drought monitoring in Iran. Int J Climatol 26(7):971–985. https:// Management, Climate Change doi.org/10.1002/joc.1264 USAID (2009) Ethiopia livelihood zones Accessed from http://www.fews.net/east- Mpelasoka F, Hennessy K, Jones R, Bates B (2008) Comparison of suitable drought africa/ethiopia/.livelihood-zone November, 2017 indices for climate change impacts assessment over Australia towards resource Viste E, Korecha D, Sorteberg A (2013) Recent drought and precipitation management. Int J Climatol 28(10):1283–1292. https://doi.org/10.1002/joc.1649 tendencies in Ethiopia. Theor Appl Climatol 112(3–4):535–551. https://doi. National Meteorological Services Agency of Ethiopia/NMSA, 1996. Climatic and org/10.1007/s00704-012-0746-3 agro-ecology resources of Ethiopia, Addis Ababa, Ethiopia, 1(1):0–137 Von Braun J (1991) A policy agenda for famine prevention in Africa (no. 1) Ntale HK, Gan TY (2003) Drought indices and their application to East Africa. Int J International food policy research institute (IFPRI) Climatol 23(11):1335–1357. https://doi.org/10.1002/joc.931x Webb P, Braun JV (1994) Famine and food security in Ethiopia: lessons for Africa. Wiley, Padhee SK, Nikam BR, Aggarwal SP, Garg V (2014) Integrating effective drought Chichester index (EDI) and remote sensing derived parameters for agricultural drought Webb P, Braun JV, Yohannes Y (1992) Famine in Ethiopia: policy implications of assessment and prediction in Bundelkhand region of India. Int Arch coping failure at national and household levels. IFPRI research reports Photogram Remote Sensing Spatial Inform Sci 40(8):89. https://doi.org/10. Webb P, Von Braun JV (1990) Drought and food shortages in Ethiopia: a 5194/isprsarchives-XL-8-89-2014 preliminary review of effects and policy implications International Food Philip S, Kew SF, Jan van Oldenborgh G, Otto F, O’Keefe S, Haustein K et al (2017) Policy Research Institute The drought in Ethiopia, 2015 climate and development knowledge network WFP (2014) Resilience measurement principles: toward an agenda for and world weather attribution initiative Raising Risk Awareness measurement design. Technical series no.1 Piechota TC, Dracup JA (1996) Drought and regional hydrologic variation in the Wilhite DA (2000) Drought as a natural hazard: concepts and definitionsDrought United States: associations with the El Niño-southern oscillation. Water Mitigation Center Faculty Publications, pp 69.3–69.18 Resour Res 32(5):1359–1373. https://doi.org/10.1029/96WR00353 Wilhite DA, Glantz MH (1985) Understanding: the drought phenomenon: the role Prasad PV, Staggenborg SA (2009) Growth and production of sorghum and of definitions. Water Int 10(3):111–120. https://doi.org/10.1080/ millets. soils, plant growth and crop production, 2. In: Encyclopedia of life support systems (EOLSS), developed under the auspices of the UNESCO Williams AP, Funk C (2011) A westward extension of the warm pool leads to a Segele ZT, Lamb PJ (2005) Characterization and variability of Kiremt rainy season westward extension of the Walker circulation, drying eastern Africa. Climate over Ethiopia. Meteorol Atmos Phys 89(1–4):153–180. https://doi.org/10.1007/ Dynam 37(11–12):2417–2435. https://doi.org/10.1007/s00382-010-0984-y s00703-005-0127-x WMO (2009) Guide to hydrological practices, volume II, management of water Seleshi Y, Camberlin P (2006) Recent changes in dry spell and extreme rainfall resources and application of hydrological practices Extreme value analysis, events in Ethiopia. Theor Appl Climatol 83(1):181–191. https://doi.org/10. WMO-No. 168 1007/s00704-005-0134-3 WMO (2012) Standardized Precipitation Index User Guide (M. Svoboda, M. Hayes Seleshi Y, Zanke U (2004) Recent changes in rainfall and rainy days in Ethiopia. Int J and D. Wood). (WMO-No. 1090), Geneva Climatol A J R Meteorol Soc 24(8):973–983. https://doi.org/10.1002/joc.1052 WMO (2016) Guidelines on the definition and monitoring of extreme weather Sen PK (1968) Estimates of the regression coefficient based on Kendall's tau. J and climate events Am Stat Assoc 63(324):1379–1389 World Bank (2006a) Ethiopia: managing water resources to maximize sustainable Shanko D, Camberlin P (1998) The effects of the Southwest Indian Ocean tropical growth. World Bank, Washington, D.C. cyclones on Ethiopian drought. Int J Climatol 18(12):1373–1388 World Bank (2006b) IDA countries and exogenous shocks. International Sheffield J, Wood EF (2007) Projected changes in drought occurrence under future development association resource mobilization. World Bank, Washington, DC global warming from multi-model, multi-scenario, IPCC AR4 simulations. Zhan W, Guan K, Sheffield J, Wood EF (2016) Depiction of drought over sub- Climate dynamics 31(1):79–105. https://doi.org/10.1007/s00382-007-0340-z Saharan Africa using reanalysis precipitation data sets. J Geophys Res Atmos Shiferaw B, Tesfaye K, Kassie M, Abate T, Prasanna BM, Menkir A (2014) Managing 121(18):10,555–10,574. https://doi.org/10.1002/2016JD024858 vulnerability to drought and enhancing livelihood resilience in sub-Saharan Africa: technological, institutional and policy options. Weather and Climate Extremes 3:67–79. https://doi.org/10.1016/j.wace.2014.04.004 Publisher’sNote Spinoni J, Naumann G, Carrao H, Barbosa P, Vogt J (2014) World drought Springer Nature remains neutral with regard to jurisdictional claims in frequency, duration, and severity for 1951–2010. Int J Climatol 34(8):2792– published maps and institutional affiliations. 2804. https://doi.org/10.1002/joc.3875 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Geoenvironmental Disasters Springer Journals

Spatial and temporal drought incidence analysis in the northeastern highlands of Ethiopia

Download PDF
17 pages

Loading next page...
 
/lp/springer-journals/spatial-and-temporal-drought-incidence-analysis-in-the-northeastern-NCIc6uLo0h

References (91)

Publisher
Springer Journals
Copyright
Copyright © The Author(s). 2020
eISSN
2197-8670
DOI
10.1186/s40677-020-0146-4
Publisher site
See Article on Publisher Site

Abstract

This study investigated the space-time drought incidence in the northeastern highlands of Ethiopia using monthly rainfall data. It also aims to predict drought events for 100 years. The Standardized Precipitation Index (SPI) was used to compute the drought severity classes of rainy months and seasons at 1-, 4- and 8-months timesteps. The Mann-Kendall’s test and Sen’s slope estimator were used to analyze the trends of drought events and to determine the magnitude of change. Inverse Distance Weighted spatial analysis tool was used to illustrate the spatial patterns of the drought risk events. The study detected extreme severe droughts in the belg rainy months in March 2008 and April 1984. However, during the belg season, the year 1999 was the driest for the recorded periods. On the other hand, the extremely severe droughts were observed during the kiremt rainy months of July 1987 and 2015, August 1984, and September 2009. In general, 1984, 1987 and 2015 were the driest years recorded in the kiremt season. The study noted that the drought risk events of months in the belg season were threefold greater than that of the months in kiremt season under moderate drought intensity class. Equally, the drought risk events of months in kiremt season were threefold greater than that of the belg season under extreme drought intensity class. Complex spatial variations of drought risk events were also observed in 1-, 4- and 8-months timesteps. During the belg seasons, the southern half was subjected to more frequent drought risk events while the northern half experienced more frequent drought risk events during kiremt season. Almost the eastern half of the livelihood zones experienced higher drought frequency events than the other parts in the 8-month timestep. The observed space-time drought risk event analysis has shown a potential threat to the rainfed agricultural practices that have a great influence on the livelihoods of smallholder farmers. Hence, documentation and assessment of drought risk events based on the livelihood zones are essential for drought risk management, early warning responses, local- scale planning and food security management. Finally, the study recommended further research on additional indices of climatic variables such as evapotranspiration and soil water content. Keywords: Livelihood zone, Belg and kiremt, Drought, Standardized precipitation index, South Wollo, Ethiopia Introduction average usually for a season or more (AghaKouchak and Drought is a commonly used term, but the most com- Nakhjiri 2012; Guha-Sapir et al. 2012; Dai 2011; Ashraf plex and least realized of all the natural hazards affecting and Routray 2013; WFP 2014). Many scholars, for in- more people than any other hazards (Wilhite 2000; Ash- stance, AghaKouchak and Nakhjiri (2012), Jeyaseelan raf and Routray 2013). As opposed to other extreme (2003) and Wilhite and Glantz (1985) classified drought events, (like floods, tornadoes, and hurricanes), drought as meteorological, hydrological, agricultural, and socio- develops slowly and steadily, making it difficult to deter- economic based on the causative factors. These drought mine the onset and end (WMO 2016). Generally, types share the deficiency of precipitation as a common drought is related to a continual deficit of water below phenomenon. Among these, the meteorological drought is the main driver; leading to the more likely occurrences of agricultural and hydrological droughts (WMO 2016; * Correspondence: abebemekkon@gmail.com Department of Geography and Environmental Studies, Bahir Dar University, Zhan et al. 2016; Mpelasoka et al. 2008). P.O. Box 79, Bahir Dar, Ethiopia © The Author(s). 2020 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 2 of 17 Since drought is a recurrent climatic phenomenon, its in great human sufferings and losses of life (Seleshi and outcome is manifested through immense damage to Camberlin 2006). Precipitation exhibits great spatial agriculture production, reduction of water supply and variation both in annual average and inter-annually that energy production, mass migration and loss of life could be taken as the major cause for the occurrence of (Masih et al. 2014). Writers such as –Sheffield and the droughts in the country (Viste et al. 2013; World Wood (2007), Bannayan et al. (2010), and Degefu and Bank 2006a, 2006b). In this regard, Viste et al. (2013) Bewket (2015) differently showed that the causes and in- conclude that there are no years without drought in fluences of drought in space-time on the environment many parts of Ethiopia. Consequently, environmental are determined by the attributes of frequency, magni- and socioeconomic deprivation in Ethiopia happened tude, intensity, and spatial extent. Likewise, Trenberth due to the cumulative induced drought over several cri- et al. (2011) reported that the magnitude, intensity, fre- sis seasons (Endale 1993). In the future, due to the in- quency and spatial coverage of the dry area has been creasing resource depletion, single-year droughts may be doubled since the 1970s in Africa, East and South Asia, sufficient to cause severe hardship among vulnerable eastern Australia, Southern Europe, Northern and households (Webb and Von Braun 1990; Webb and Southern America, most of Alaska, and Western Canada. Braun 1994). Within the period 1960–2016, there were about 669 Although the name “Water Tower of North-East Af- drought events reported across the world, and hence it rica” has been given to Ethiopia, the country is one of has killed about 2.2 million people and affected over 2.6 the Horn countries highly vulnerable to drought (Gebre- billion people and an estimated economic damage of hiwot et al. 2011). The history of drought and associated 146 billion US Dolar Centre for Research on the Epi- disasters in Ethiopia traced back to 250 B.C. with a total demiology of Disasters/CRED (2016). of 39 periods of food shortage and excess mortality rate The vulnerability of developing countries to drought have been identified (Block 2008; Webb and Von Braun has risen by a factor of three in the decade of the 1990s 1990; Webb and Braun 1994; Webb et al. 1992). Impacts compared with the 1960s (Mirza 2003). Much of the of droughts in Ethiopia are perhaps the most widely northern, southern, eastern and Sahelian African coun- publicized given their severity and frequency (Gautam tries experienced the worst droughts in terms of the 2006). As recent literature ratifies, Ethiopia’s history has number of people killed and affected in 1972/73 and been periodically marked by major drought incidences 1984/85 (Gommes and Petrassi 1996). For instance, of for the period 1950–2016 (Table 1). all the natural disasters, droughts account for only 8% of Drought frequency in Ethiopia has been increasing. the global phenomenon. It also accounts for 25% in Af- For example, in the 1970s and 1980s drought usually oc- rica between 1960 and 2006 (Gautam 2006). In the curred once per decade (Block 2008; United Nations 1980s drought killed more than half of a million people Environment Program/UNEP 2011). It was also about in Africa (Dai 2011). In Sub Saharan African countries, twice every 3 years between 1980 and 2004 (World Bank crop production and food security are mainly threatened 2006a, 2006b) and once a year between the 1990s and by drought where the effects are inter-temporal and 2011 (Viste et al. 2013). The estimated human cost long-lasting (Shiferaw et al. 2014). Africa stands first caused by drought in Ethiopia between 1972 and 1984 with a total of 289 reported the number of drought were about 1. 2 million people (United Nations Environ- events that killed almost 700 thousand people and af- ment Program/UNEP 2002). These evidences showed fected above 414 million people between 1960 and 2016 that Ethiopia is often stricken by droughts in the 1970s due to the induced drought frequency (Centre for Re- and 1980s resulted in widespread poverty, economic search on the Epidemiology of Disasters/CRED 2016). stagnation, depletion of household assets and savings, Drought has a severe negative impact on key socioeco- and excess mortality (Dorosh and Rashid 2013). South nomic sectors of most Eastern African countries (Butterfield Wollo, located in the northeastern highlands of Ethiopia 2011). Its incidence has increased steadily in East Africa for is an epicenter for drought and famine for many years. the last five decades (Gautam 2006). For instance, The zone is a drought prone area and faces the horrible deprivation associated with drought and food crises disaster combination of both high risk and low potential as well victims for East African countries (Burundi, Djibouti, as chronic food insecurity (World Bank 2006a, 2006b). Ethiopia, Kenya, Somalia, and Uganda) increased from one Of the total recorded drought years (1953–2016) in million people in the 1980s to 21 million in the Ethiopia, Wollo experienced about 72% of the drought 2010s (Hao et al. 2014;Guha-Sapiretal. 2012). events (Little et al. 2006). In fact, the existence of diverse In Ethiopia, extreme weather events owing to insuffi- agroecology enables the zone to raise different types of cient total rainfall amount and long dry periods affected livestock and to produce cash and food crops (ANRSPC the agro-socioeconomic environment in the northern, 2017; 2018). But, the agricultural activities are entirely southern and eastern parts of the country and resulted rainfed except in a few localities where there are small- Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 3 of 17 Table 1 Reported drought years, regions affected, causes and severities since the 1960s in Ethiopia Year Region Affected Causes and severity description 1957/58 Tigray and Wollo Rain failure in 1957; locusts and epidemic in 1958 and about 100,000 people died 1962/63 Western Ethiopia Very severe 1964–66 Tigray and Wollo About 1.5 million people were affected and about 300, 000 livestock died 1971–75 Ethiopia Sequence of rain failures; estimated 250,000 dead; 50% of livestock lost in Tigray and Wollo 1978/79 Southern Ethiopia Failure of Belg rain & 1.4 million people were affected 1982 Northern Ethiopia Late Meher rains & 2 million people were affected 1984/85 Ethiopia Sequential rain failure; 8 million people affected; 1987/88 Ethiopia 7 million people were affected 1990–92 Northern, Eastern, and SW Ethiopia Rain failure and regional conflicts; estimated 4 million people suffering food shortage 1993/94 Tigray and Wollo Widespread food insecurity (7.6 million people were affected), but few deaths or cases of displacement were reported. 1997 Borena, Bale, Omo, Somali Almost 986, 000 people affected 1999 N. & S. Wollo, Wag, Himra; Tigray; B. Gumu, Almost 5 million people affected Gambela, Oromia, SNNPR, Somali 2003/4 All Regions Over 13 million people affected, but the response mitigated the worst potential outcomes. 2005 Somali, Oromia Almost 2.6 million drought disaster affected people 2008/9 All regions Almost 12.6 million People were affected. 2011 S & E. Oromia, Somali Severe food insecurity and 4 million people affected 2015/16 N., E, & SW Ethiopia About 10.2 million people were affected Source: Degefu 1987; Webb and Von Braun 1990; Webb and Braun 1994; Webb et al. 1992; Philip et al. 2017 scale irrigation and water harvesting practices (SWAD SPI-values of 3-, 6-, 9-, 12-, and 24-months timesteps. 2018). Dominant cereal crops produced in the region in- However, the SPI can be computed at any timescales for clude teff (Eragrostis tef), sorghum (Sorghum bicolor), any set of successive months of 1-month up to 72- maize (Zea mays), wheat (Triticum aestivum), barley months. The spatial and temporal drought event analysis (Hordeum vulgare) and pulses (Leguminosae/ Fabaceae). for SPI 1-, 4- and 8-months timesteps are not included in Nevertheless, the larger reliance on agriculture-based in- their studies. Therefore, the analysis of the space-time pat- come, which is under the sensitivity of nature exposed tern of drought risk events of this study is based on the the people to food deficiency (Little et al. 2006; Kahsay rainy months and seasons of 1-, 4- and 8-months time- 2013). All these attributes contributed to the reduction steps,. The space-time drought risk events of the small of productivity in the zone. This might be the reason and big rainy months as well as seasons (SPI: 1-, 4- and 8- that the zone is exposed to food shortage and dependent months timesteps) are appropriate for planning and man- on safety nets for many years. Therefore, an assessment agement of drought and related risks at the local-scale of the spatio-temporal drought risk events is imperative level. This study, therefore, aims to characterize the spatial for policy makers and practitioners in the area under and temporal drought incidences as a function of fre- study. quency, magnitude, intensity and severity status using the The classification and representation of observed SPI model over the northeastern highlands of Ethiopia. spatio-temporal drought risk events using SPI are vital in operational monitoring for policy decisions. Accord- Drought quantification: the standardized precipitation ingly, many studies, but different in study periods and index spatial coverage related to drought were documented in Drought indices are essential to characterize several Ethiopia using SPI (e.g. Suryabhagavan 2017; Degefu and drought features such as the onset and cease time of Bewket 2015; Viste et al. 2013; Gebrehiwot et al. 2011; drought, drought duration, areal extent, severity and fre- Williams and Funk 2011; Edossa et al. 2010; Korecha quency at global, regional and local level (Piechota and and Barnston 2007; Segele and Lamb 2005; Seleshi and Dracup 1996). Owing to the complexity of drought, nu- Zanke 2004). Many of these studies documented either merous operational drought indices have been developed temporal or spatial or both to show the most common and used by meteorologists and climatologists around Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 4 of 17 the world (WMO 2016; Dai 2011). In the international SPI might be more effective in highlighting available works of literature, a large number of studies acknow- moisture conditions than the slow-responding Palmer ledge numerous indices in the detection, analysis, and Index or other currently available indices (WMO 2012). monitoring of drought events for recent and projected As compared to other indices, SPI is an appropriate changes (Loukas and Vasiliades 2004). Among the nu- index for drought analysis to East African countries indi- merous indices, the most commonly used drought index cated by various studies. For instance, Ntale and Gan with its strengths and weaknesses is the Standardized (2003) suggested that SPI is more suitable for monitor- Precipitation Index (SPI) (McKee et al. 1993). ing droughts in East Africa because (i) it is easily The SPI is relatively new application, gaining world- adapted to the local climate, (ii) it needs a modest data wide acceptance and widely used at present because of requirement, (iii) it produces spatially consistent results, its powerfulness, effectiveness, flexibility, versatile and and (iv) it can be computed almost at any time scale. standardized nature (Beth and Brown 2003;WMO 2012; Tsakiris and Vangelis 2004; Mashari Eshghabad et al. Materials and methods 2014). Hence, choosing a precipitation-based drought Description of the study site measure (SPI model) is preferred for its low data re- South Wollo is one of the eleven administrative zones of quirement and computational simplicity for developing the Amhara National Regional State which is located be- 0 0 0 countries like Ethiopia where the access to data is lim- tween 10 10′Nand 11 41′Nlatitudes and38 28′Eand ited (Viste et al. 2013; Degefu and Bewket 2015). 40 5′E longitudes (Fig. 1). The total area of the Zone is SPI values in one region can be directly compared to about 18,157.48 km , which is divided into 19 rural dis- SPI values in a completely different climate zone. It can tricts and four administrative towns (Dessie, Kombolcha, monitor the onset, intensity, and duration of drought Haik, and Mekaneselam) (ANRSPC 2017). Of the total (Hayes et al. 1999; Loukas and Vasiliades 2004). The SPI area, 36.3%, 13.5%, 18.3% and 31.9% are covered by arable is a very suitable index to study the geospatial and tem- land, forest and bush lands, grazing land, and others (e.g. poral variation of drought with spatially invariable re- bare land, buildings, water bodies), respectively (SWAD sults for historic time series analysis (Masih et al. 2014; 2018). The landscape of the South Wollo zone comprises Guttman 1998). The ideal strength of SPI is that precipi- highly diversified and dissected topography (rugged terrain tation anomalies can be consistently calculated over flex- with very steep slopes). Elevation ranges from 927 m (over ible time scales in a consistent fashion. Similarly, the dry plain/Arabat) in the east to 4261 m above sea level drought information can be provided promptly for oper- (Mount Ambaferit) in the west (SWAD 2018;Littleet al. ational drought-monitoring applications (AghaKouchak 2006). It has six livelihood zones, namely, Abay-Beshilo and Nakhjiri 2012). Even, following the discussion at Basin (ABB), Chefa Valley (CHV), Meher-Belg zone,Belg (WMO 2009), drought experts made a consensus agree- zone,Meher zone and eastern lowland sorghum and cattle ment to recommend the SPI for the characterization of (SWS) (USAID 2009). meteorological droughts (Hayes et al. 2011). The SPI The Zone is characterized by a distinctive bi-modal measures moisture to quantify the precipitation deficit rainfall regime, locally known as kiremt and belg seasons. for multiple time scales at any one location and reflects Kiremt is the big rainy season usually extending from the impact of drought on a range of meteorological, agri- June-to-September, and belg is the small rainy season cultural and hydrological applications (Marimon 2016; extending from February-to-May (National Meteoro- Degefu and Bewket 2015; Loukas and Vasiliades 2004). logical Services Agency of Ethiopia/NMSA 1996; Degefu The SPI can be calculated at various timescales for any 1987; Korecha and Barnston 2007; Conway 2000a, set of successive months of 1-month up to 72-months 2000b). Crop production in the entire livelihood zones that one wishes to detect and to illustrate the effect of follow these rainfall regimes but varies among the liveli- drought (Viste et al. 2013; Stagge et al. 2015; Padhee hood zones. Accordingly, the livelihood zones of ABB, et al. 2014; Beth and Brown 2003; AghaKouchak and SWS Meher and CHV are dependent on big kiremt rain- Nakhjiri 2012). But statistically, 1–24 months is the best fall for the cultivation of both long and short cycle crop practice range of application (Guttman 1999). The short production. Conversely, Belg and Meher-Belg crop pro- period time scales SPI-values (1-, 2-, 3-, 4-, 5- and 6- ductions follow the bimodal rainfall regimes (the small months) are appropriate and paramount for measuring belg and big kiremt rains) leading to two harvest periods. the effect of drought on agriculture, soil moisture and The mean annual rainfall ranges between 500 and 900 crop yield reduction (Bussay et al. 1998; Morid et al. mm in the Kolla (arid and semi-arid) area and 950– 2006; Szalai and Szinell 2000). 1100 mm in the Woina Dega (semi-humid) and Dega A 3-month SPI reflects short- and medium-term mois- (cool and humid) areas. The average annual temperature ture conditions and provides a seasonal estimation of also ranges from15°c to 20°c (SWAD 2018). Between precipitation. In primary agricultural regions, a 3-month 1901 to 2016 the minimum temperature ranges were Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 5 of 17 Fig. 1 The geographic location of livelihood zones of South Wollo and Meteorological stations. Source: Extracted from USAID 2009 0 0 6.43 c in December and 11.82 c in June while the max- of the study area (ANRSPC 2017;SWAD 2018). Even imum ranges were 22.93 c in August and December and though the small and large-scale irrigation practices 29.13 c in June. are very minimal, the zone is known by many large Based on the population statistics of ANRSPC and small rivers categorized into two major drainage (2017), theSouth Wollo zonehas atotal population basins (Abay and Awash accounting for about 82% of nearly 3.1 million of whom 49.5% were males and and 18% of the area, respectively) (SWAD 2018). Cat- 50.5% females. The zone is one of the densely popu- tle, goats, sheep, and equines are the major livestock lated areas of the country with 171 persons per km . in the area. However, the contribution of livestock to The old-style subsistence farming (a mix of both the livelihood of the people is constrained by the grain production and rearing of livestock) is the pri- prevalence of livestock diseases (USAID 2009;Little mary form of economic activity in all livelihood zone et al. 2006). Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 6 of 17 Data sources and data analysis techniques (1993) and WMO (2012), have been applied in the quan- For the characterization of spatial and temporal drought tification and classification of SPI drought severity status incidences as a function of frequency, magnitude, inten- of the study. In the computation of drought status classi- sity, and severity status, long-term meteorological data fication of this study, only months with SPI ≤− 1.00 were were acquired from the Ethiopian National Meteorological considered to be as drought initiation and cessation Agency (ENMA-Kombolcha Regional Meteorological Sta- since mild drought (SPI between − 0.99 and 0.99) tion). Since rainfall is not normally distributed, a gamma showed only a slight variation from the normal rainfall probability distribution has been found fitting to the pre- distribution (Łabędzki 2007; Sternberg et al. 2011; cipitation time series frequency distribution for the com- Degefu and Bewket 2015). Therefore, using the SPI_SL_ putation of SPI (Khan and Gadiwala 2013). The gamma 6.exe file programming, moderate drought (SPI between probability density function is expected to fit with the fre- − 1.00 and − 1.49), severe drought (SPI between − 1.50 quency distribution of precipitation over the period and − 1.99) and extreme drought (SPI ≤− 2.00) were gen- (1981–2017). For the time scales of 1-, 4- and 8- months, erated and presented in the form of time series graphs alpha (α) and beta (β) parameters of the gamma probabil- and maps. ity density function were estimated. As a result, the cumu- Temporal drought event characteristics (frequency, lative probability distribution function is approximated magnitude, intensity and severity status) were computed into the standard normal cumulative distribution prob- using one of the most common models, SPI analysis pa- ability function with a mean of zero and a standard devi- rameters. Drought frequency is the return period be- ation of one to determine the probability distribution tween drought events that have severity threshold values function. Conceptually, the SPI is equivalent to Z- score, of SPI ≤− 1. The magnitude of drought event corre- the number of standard deviations that the observed value sponds to the cumulative water deficit during the would deviate from the long-term mean. The statistical drought period below some threshold (SPI-values ≥− 1) procedure is: (Thompson 1999) and drought intensity (D ) is the ratio between the drought magnitude and the duration of the X X event computed as: i− SPI ¼ ð1Þ D ¼− SPI ð2Þ M ij Where: xi = observed precipitation value of the selected i¼1 period during the i th year; Xand σ = the mean and the standard deviation re- M D ¼ ð3Þ spectively over the selected period. In this study, SPI was computed for shorter-term and Where: D drought magnitude, M= intermediate periods to quantify drought events at 1-, 4- n = number of months with drought event at j and 8-months timesteps to represent agricultural/me- timestep; teorological droughts. The 1-month SPI (analogous to D = drought duration. the percent of normal precipitation) reflected short-term However, drought severity (D ) should not be mis- conditions of soil moisture and crop stress during the guided for intensity, which is generally signified to the growing season while the 4-months SPI represented the lowest SPI-value of the drought event (Spinoni et al. short and medium-term moisture conditions applied for 2014). The number of droughts per 100 years for differ- the seasonal estimation of rainfall. The 4-months accu- ent timesteps of the rainy months and seasons (SPI:1-, mulation of SPI was calculated for May and (September) 4- and 8-months) of the study as used by (Łabędzki to assess the seasonal drought of belg and (kiremt). This 2007; Sternberg et al. 2011; Degefu and Bewket 2015; is because of the 4-month definition of seasons in Ghosh 2019) which was estimated as: Ethiopia are the most commonly used (Viste et al. 2013). The 8-months timestep accumulation of SPI was com- N ¼ X100 ð4Þ puted for September (February-to-September), which i;100 i Xn was also a good indicator to assess the drought condi- tion of slowly maturing long-cycle crops like sorghum Where: and maize that typically planted during the belg season Ni,100 = the number of droughts for a timestep i in and harvested following the kiremt season. 100 years; In the computation of SPI, the SPI_SL_6.exe file pro- Ni = the number of years with droughts for a timestep gramming which was available at http://drought.unl.edu/ i in the n-year set; MonitoringTools/DownloadableSPIProgram.aspx has i = the timesteps (1-, 4- and 8- months) and. been employed. The criteria employed by McKee et al. n = the number of years in the data set (37 years). Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 7 of 17 Furthermore, the MK (Mann Kendall) trend test (Ken- Results dall 1975; Mann 1945) has been computed to check the Characteristics of drought events: frequency, magnitude, statistical significance (increasing and decreasing trends) intensity and severity for rainy months and seasons of SPI: 1-, 4- and 8- Table 2 shows the observed drought frequency, magni- months timesteps. The relative strength of MK trend tude, intensity and severity of the small, big, and long- test in time series analysis has been quantified using term rainy months and seasons as computed using SPI. Sen’s(1968) slope estimator. As indicated by Chattopad- The highest frequency of drought years for the small hyay and Edwards (2016), Sen’s slope estimator has been rainy months within the moderately to extremely dry the commonly used estimator due to its relative insensi- (SPI ≤− 1.00) found to be 9 years; in February and May tivity and robustness to extreme values. The MK statistic account for about 24%of the total years over the study S of the series x is calculated as: period. The drought magnitudes of the small rainy months were ranged from − 9.45 in March to − 11.83 in N−1 N May. Nevertheless, the drought intensity found to be be- X X S¼ sgn X X ð5Þ j− i tween the moderately dry in May with SPI = − 1.23 and se- i¼1 j¼iþ1 verely dry in March with SPI = − 1.58. The extreme severe droughts were observed in March 2008 and April 1984 Where: with the severity of peak SPI-values of − 2.49 and − 2.23, N = the number of data points; respectively. Likewise, the highest drought frequency in X and X = the time series observations. Assuming (x the big rainy months was 6 years in June that account for i j j – x )= θ, the value of sgn (θ) is computed from: 15% of the total drought risk years. The observed magni- tude of drought events ranged from − 9.7 in July to − 5.69 in September. The moderately dry drought intensity was observed in June, July, and September while the extremely þ1………θ > 0 dry drought intensity was observed in August. The worst sgnðÞ θ ¼ 0………θ ¼ 0 ð6Þ droughts were recorded in July 2015, August 1984, and −1………θ < 0 September 2009 with the severity peak of SPI values of − 2.82, − 3.61, and 2.31, respectively. While Positive values of S indicate increasing trends, Seasonally, the drought frequency for SPI 4-months of negative S indicates decreasing trends. Under the hy- both belg and kiremt as well as long-term season of SPI pothesis of independent and randomly distributed ran- 8-months at the end of May and September account for dom variables, for large samples, when n ≥ 10 (in some about 4 years (9%) of the total drought years under papers n ≥ 8), the σ statistic is approximately normally study. A slight difference in the observed drought risk distributed, with zero mean and variance as follows: magnitude was shown among the SPI4-months (belg and kiremt) and SPI8-months (long-term) seasons. Nonethe- nnðÞ −1ðÞ 2nþ5 σ ¼ ð7Þ less, the drought intensity was varied from severely dry in SPI4-months (belg) and SPI8-months (long-term) sea- sons with SPI = − 1.92 and − 1.79 respectively to ex- As a consequence, the standardized normal deviate (Z- tremely dry in SPI4-months (kiremt) season with SPI = statistics) distribution has been then calculated as: − 2.0. The SPI 4- and 8-months experienced severe drought events in 1984, 1999 and 2015 with severity S−1 peak SPI-values of − 2.83, − 2.84 and − 2.32, respectively. > if S > 0 Z ¼ ð8Þ 0if S ¼ 0 S þ 1 > Temporal trends of drought events if S < 0 Using the SPI, the drought condition of the study for rainy months (SPI-1), belg and kiremt (SPI-4), and long- While Positive values of Z signify increasing trends, term (SPI-8) rainy seasons covering the period of 1981– negative Z signifies decreasing trends. In the spatial ana- 2017 were examined. A mix of dry and wet years have lysis of drought severity status, Inverse Distance been observed in February, March, April, and May as Weighted, IDW average interpolation method was used well as belg season but with temporal variation in sever- in the ArcGIS. The SPI values generated from SPI_SL_ ity, magnitude and intensity (Fig. 2a-d and Fig. 4a). Sev- 6.exe file programming was used as an input for map- eral drought risk incidences with SPI ≤− 1.00 were ping the spatial distribution of drought severity status. detected. It was observed that March 2008, April 1984, As a consequence, the 1-, 4- and 8-months timesteps and the belg season of 1999 were characterized by ex- spatial distributions of drought incidences were mapped. treme drought events with SPI ≤− 2 in the past 37 years. Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 8 of 17 Table 2 Drought risk characteristics of the rainy months and seasons for SPI ≤ 1.0 values Month Observed Drought risk Frequency Magnitude Intensity Severity (years) Peak SPI-Value Year/s February 9 −10.92 −1.23 −1.32 1999, 2000, 2012 March 6 −9.45 −1.58 −2.49 2008 April 7 −10.58 −1.5 −2.23 1984 May 9 −11.83 −1.37 − 1.52 1981 Feb-to-May (Belg: SPI-4) 4 −7.66 −1.92 −2.84 1999 June 6 −7.3 −1.22 −1.78 1981 July 5 −9.7 −1.94 −2.82 2015 August 3 −7.1 −2.37 −3.61 1984 September 3 −5.69 −1.9 −2.31 2009 June-to-Sept (Kiremt: SPI-4) 4 −8.03 −2 −2.32 2015 Feb-to-Sept (SPI-8) 4 −7.14 −1.79 −2.83 1984 Severe drought events with SPI values between − 1.5 Similarly, Fig. 3a-d and Fig. 4b signify a mix of drought and − 1.99 were also observed in March (1999 and 2000) and wet frequencies in the timesteps of SPI1-month ob- and April (1991, 1999 and 2001). Moreover, the largest servations (June, July, August and September) and SPI4- drought risk events (about 23 years) for February, March, months of the current season. A mix of dry and wet April, and May were shown under the moderate drought years have been observed in June, July, August, and Sep- intensity class. Nevertheless, 6 and (2) years were under tember as well as kiremt season but with temporal vari- the severe and (extreme) drought intensity classes. ation in severity, duration, magnitude, and intensity. It Drought ends when the SPI value becomes positive was witnessed from Fig. 3a-d and Fig. 4b extreme while drought begins when SPI value becomes negative. drought events with SPI ≤− 2 were investigated in July Since drought ends when the SPI value becomes posi- 1987 and 2015, August 1984 and 1993, September 2009 tive, the increasing drought frequency of February found and kiremt season of 1984, 1987 and 2015. Severe to be by far the highest (Table 2). This increasing drought also was detected in June 1981, July 1981 and drought frequency of February was substantiated by a 1984, and September 1984. Moreover, moderate drought significant decline of rainfall as revealed in the MK trend events were detected in June 1982, 1985, 1987, 1988 and test (Table 3) and regression trend line (Fig. 2a). 2017, July 1992, August 1990 and September 1997. It Fig. 2 SPI 1-month anomalies and trends of belg rainy months: (a) SPI-1: February; (b) SPI-1: March; (c) SPI-1: April; (d) SPI-1: May Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 9 of 17 Table 3 Mann-Kendall’s trend analysis and the probability of drought events per 100-year for the rainy months and seasons at 1-, 4- and 8-months timesteps SPI timesteps Months/Season MK test P-values Slope Probability of drought events per a 100-year 1-month February −0.2539** 0.0305 −0.0366 24 March −0.1388 0.2337 −0.0189 14 April −0.1508 0.1952 −0.02 19 May 0.0693 0.556 0.0149 24 June 0.2493** 0.0327 0.0311 16 July 0.1851 0.1105 0.0291 14 August 0.1458 0.2092 0.0188 8 September −0.0532 0.6559 −0.0077 8 4-month Feb-to-May −0.1583 0.1736 −0.0248 11 June-to-Sept 0.2122 0.067 0.0281 11 8-month Feb-to-Sept 0.1414 0.2238 0.0161 11 Note: ** is statistically significant at P < 0.05 was also shown in Fig. 4c that moderate drought event Incidences of drought trends and its probability per was detected in 2009 while extreme droughts were de- 100-years over the small and big rainy months (SPI-1) tected in 1984 and 2015 during the long-term rainy sea- and seasons (SPI-4 and SPI-8) timesteps were shown son of SPI 8-month timestep. Furthermore, the largest (Table 3). The Mann-Kendall trend test of SPI-values drought event years (about 8 years) for June, July, Au- showed statistically significant decreasing and (increas- gust and September were shown under the moderate ing) trends for small and (big) rainy months of February drought intensity class while three and (six) years were and (June) at p < 0.05. However, the 1-month timestep under the severe and (extreme) drought intensity classes. values of March, April, September and 4-month time- Seasonally, belg and (kiremt) were under severe and (ex- steps of February-to-May showed statistically non- treme) drought intensity classes for 3 years while 2 years significant decreasing rends. The 1-month timestep of the long-term season (SPI8-months) were under ex- values of May, July, August, and 4-month timestep treme drought intensity class. However, values of SPI1- (June-to-September) on the other hand showed statisti- (June, July, August) months and SPI 4- months (kiremt cally non-significant increasing trends. The 8-month season) showed increasing trends except SPI1-month of long-term seasonal timestep (Feb-to-Sept) value also September. showed a statistically non-significant increasing trend. Fig. 3 SPI 1-month anomalies and trends of kiremt rainy months: (a) SPI-1: June; (b) SPI-1: July; (c) SPI-1: August; (d) SPI-1: September Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 10 of 17 Fig. 4 The SPI anomalies and trends of the rainy seasons. a SPI-4: belg season (February-to-May); b SPI-4: kiremt season (June-to-September) and c SPI-8 long-term seasons (February-to-September)] The predicted number of droughts for 100 years rainy western and southern parts of the ABB, Meher, Belg, months and seasons timesteps of 1-, 4- and 8-months Meher-Belg livelihood zones that ranged from 6 years based on eq. (4) is also presented (Table 3). The max- (15%) to 8 years (21%) of the total observation periods. imum predicted drought frequency events (except for Except for the areas around the stations of Kelala and August and September) were expected to be in the Wereillu, which experienced the highest drought risks shorter timesteps of 1-month as compared to 4- month (8–9 years) in the south, the remaining livelihood zones’ and 8-month timesteps. The 1-month predicted drought drought frequency during April varied from 6 years frequency events expected to be between 24 and (8) (15%) to 7 years (18%) of the total study periods. Simi- years for February and May and (for August and Sep- larly, the eastern part including ABB, Meher-Belg,SWS, tember). The seasonal predicted drought frequency Meher, CHV, and Belg livelihood zones and the western events of all the 4-month (Feb-to-May and June-to-Sept) part of ABB, Meher experienced frequent drought risks and 8-month (Feb-to-Sept) expected to be 11 years. ranged from 9 (24%) to 11 years (30%) of the total study years during May. Spatial patterns of drought incidence The spatial pattern of drought frequency maps in the The spatial patterns of drought frequency incidence big rainy months (June, July, August, and September) maps for 1-, 4- and 8-months timesteps for moderately also revealed the complex and high local-scale variation dry-to- extremely dry intensity classes of drought events as indicated in Fig. 6. It was found that all the livelihood were depicted in Figs. 5, 6 and 7. The spatial pattern of zones except the eastern and southeastern periphery drought risk frequency maps over the livelihood zones (ABB, Belg, and CHV) experienced more frequent in the small rainy months (February, March, April, and drought risks ranged from 7 years (18%) to 9 years (24%) May) was exhibited complex and high local-scale vari- of the observation periods during June. More frequent ation as indicated in Fig. 5. All the livelihood zones drought events ranged from 6 to 7 years, which accounts (ABB, Meher, Belg, CHV, Meher-Belg, and SWS) except between 15 and 18% experienced during July in the around the station of Amba Mariam exposed to more northwest, southern and northern parts of Meher, ABB, frequent drought risks ranged from 9 years (24%) to 12 Meher-Belg, SWS, and Belg livelihood zones. The inci- years (33%) of the total observation years. The drought dences of drought frequency during August were found frequency of March found to be higher in the northern, to be ranged from 3 years (6%) to 6 years (15%) in the Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 11 of 17 Fig. 5 The 1-month timesteps spatial pattern of drought frequency events (SPI≤−1.0) for the small rainy months of February, March, April, and May eastern half including all the livelihood zones while the Figure 7 presents the spatial pattern of drought fre- western half, including ABB and Meher livelihood zones quency maps for 4- and 8-months timesteps of belg and covered about 7 years (18%) and 9 years (24%) of the kiremt, and the long-term (both belg and kiremt). The study period. During September, the incidence of SPI4-months of belg season the northern half, including drought frequency was found to be more pronounced in Belg, Meher-Belg, CHV, and SWS livelihood zones the northeastern and southwestern parts including ABB, (where belg crop harvesting dominate) experienced Meher, Belg, and Meher-Belg livelihood zones that varied drought frequency events varied from 4 to 6 years (9– from 7 years (18%) to 9 years (24%). However, inci- 15%) while the southern half, including ABB and Meher dences of drought frequency varied between 4 years livelihood zones experienced 7–9 years (18–24%) of (9%) and 6 years (15%) in the northwestern, southern drought frequency events of the total observation pe- and southeastern parts including ABB, Meher, Belg, riods. Conversely, during SPI:4-months of kiremt season, Meher-Belg, CHV, and SWS livelihood zones of the the southern half comprising part of ABB, Meher, Belg, study area. and CHV livelihood zones (where kiremt season crop Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 12 of 17 Fig. 6 The 1-month timesteps spatial pattern of drought frequency events (SPI≤− 1.0) for the big rainy months of June, July, August, and September harvesting dominate) experienced 5–7 years (12–18%) Discussion drought frequency incidences while the northern cover- The rainfed agricultural activities of the study area de- ing parts of ABB, Meher, Belg, Meher-Belg, SWS, and pend on two rainy seasons: belg and kiremt. However, CHV were experienced 8-11 years (21–30%) of drought the incidence of drought disaster due to the shortfall of frequency events of the total observation periods. The rainfall was the most important concern of this study. In long-term spatial tendency of drought incidences (Feb- this study, the temporal and spatial spread of drought ruary-to-September) were also observed in SPI:8- frequency, magnitude, intensity and severity have been months. Almost the eastern half except around the sta- examined. Of the total observation years, the highest tions of Wuchalle and Kombolcha, experienced 8–9 drought frequency was detected for February. The study years (21–24%) drought frequency incidences. The results indicated that March 2008 and April 1984 were southern and western parts on the other hand experi- the most severe drought months. However, the belg sea- enced 5–7 years (12–18%) drought frequency occur- son in 1999 was the driest year during the record. In rences of the total study periods. agreement with this, Conway (2000a, 2000b) and Viste Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 13 of 17 Fig. 7 The spatial pattern of drought frequency events (SPI≤− 1.0) for SPI: 4- month (Feb-to-May, belg); SPI: 4- month (June-to-Sept, kiremt); and SPI: 8-month (Feb-to-Sept) timesteps et al. (2013) reported that the belg season in 1999 was frequency of drought incidences for March-to-May the driest year during the record over Ethiopia. A gen- months. The increasing of drought incidences in the 4- eral tendency of increasing drought risk frequency has months belg season was detected due to a rainfall deficit been observed in the study period of belg rainy months in the rainy months of February-to-May. The relative and season. The MK test for belg rainy months and sea- abrupt decline was seen since 1996, which is concurrent son (1-and 4-months) timesteps also verifies the general to Funk et al. (2005) at the national level. Multiple study decline of rainfall. Lyon and DeWitt (2012) and Lyon reports in agreement with this study revealed that the (2014) also showed the decline of rainfall or increase greatest continual decline of rainfall has occurred during Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 14 of 17 the belg season in the post-1980s (Tierney and Ummen- years on the record (Fig. 4c). Similarly, Funk et al. (2005) hofer 2015;Funk et al. 2008;Funket al. 2005). For ex- reported that 1984 stands the main horrible year on the ample, in 1984, 72% of Ethiopia’s regions were affected by record although the 1997, 2002 and 2004 undergone drought in the belg season (Shanko and Camberlin 1998). poor rainfall in the long-term (March-to-Sept), belg and Likewise, the highest drought frequency was observed kiremt seasons. March-to-Sept rainfall totals have also in June during the big rainy months. However, extremely exhibited dry periods in the mid-1980s and mid-2000s. severe droughts were detected in August 1984, July 1987 The drought risk events of months in the belg season and 2015, and September 2009. Thus, 1984, 1987 and were threefold greater than the months in kiremt season 2015 were the three driest years in the record. Drought under moderate drought intensity class. Conversely, the risk frequency and magnitude were more pronounced drought risk events of months in kiremt season were during the months of belg as compared to kiremt. None- threefold greater than that of the belg season under ex- theless, the drought risk intensity was more weighted treme drought intensity class. during the months of kiremt. In agreement of this, Viste In the 100-years, the predicted drought frequency et al. (2013), Segele and Lamb (2005) reported that the events for 1-month timesteps were relatively more fre- kiremt season of 1982, 1984 and 1987 was severely dry quent than the 4-and 8-months seasonal timesteps ex- over Ethiopia, particularly in the northeastern half of the cept for August and September. In agreement with this, country, which primarily caused by the missing rain in Łabędzki (2007) and Degefu and Bewket (2015) reported July and August. Lyon (2014) also revealed that the kir- that drought events of shorter timesteps occur more fre- emt season (June, July, and August) rainfall during 1982 quently and last a shorter time than drought events of and 1984 was below-average rainfall for prolonged longer timesteps, which are more concentrated and last drought events that lead to the most devastating impact longer (changes slowly). However, Sternberg et al. (2011) for 1950–2010 in northern Ethiopia. Further, Jjemba reported contrasting results that frequency of drought et al. (2017) and Philip et al. (2017) reported that the events have been the lowest at 1-month timesteps and worst drought of 2015–16 in the northern and central highest for the longer 17-months timestep. part of Ethiopia because of belg rains had failed and soon The spatial distribution of drought frequencies was ex- after, kiremt rains were severely delayed, erratic and hibited by the presence of complex patterns for 1-, 4-, below the long-term average (deficit of 167 mm). Segele and 8-months (Figs. 5, 6 and 7). The northern and and Lamb (2005) also revealed that the greatest dam- (southern) half during 4-month belg and (kiremt) sea- aging droughts in Ethiopia are connected with the failure sons experienced less frequent drought risk events. Con- of kiremt rains. Specifically, Suryabhagavan (2017), testi- versely, the southern and (northern) half during 4- fied that South Wollo as one of the highest peaks of month belg and (kiremt) seasons experienced a higher drought recorded area in 1984. A general decreasing frequency of drought events. Almost the eastern half ex- trend tendency of drought events has been observed cept around the stations of Wuchalle and Kombolcha, during 1-, 4- and 8-months timesteps of kiremt season. experienced the most drought frequency events while The MK test also confirms increasing trend tendencies, the southern and western parts experienced less drought but it was only statistically significant for June. frequency occurrences during the long-term 8-month A cyclical oscillation of wetter and drier events has timestep. In this regard, belg, and long-term (in both shown in 1-, 4- and 8-months timesteps. Wetter condi- belg and kiremt) seasons crop harvesting areas were tions were illustrated in the period of the 1990s. In more exposed to frequent drought risks. The 4-month contrast, increasing drought risk event frequencies were of kiremt season drought frequency pattern influenced illustrated in the 1980s and 2000s (Fig. 3b-c) and and manifested over the 8-months drought frequency (Fig. 4b-c). In corroboration of this, for the 1970s on- spatial pattern. wards, drought has occurred during kiremt season in 1982, 1984, 1897, 1990, 1991, 1995, 1997, and 2002 Potential implications of drought incidences due to the deficit of rainfall in June, July, August, and The SPI values of the rainy months of 1-month as well September (Conway 2000a, 2000b; Korecha and as 4- and 8-months timesteps had an illustrious impact Barnston 2007). The influence of rainfall deficit dur- on the rainfall of the study periods that reflects the ing the kiremt season affects a significant portion of prompt occurrence of agricultural drought. In this re- the county. For instance, in 1984 about 53% of the gard, Łabędzki (2007) confirmed that the 1- to 6-months regions of Ethiopia were affected by drought in the timesteps of SPI values soundly reflects well the quick kiremt season (Shanko and Camberlin 1998). development of agricultural drought, the exacerbating of The 8-month long-term time step (Feb-to-Sept) dem- the actual state of water conditions and the negative ef- onstrated the drought events for the 1980s and 2000– fects of rainfall deficit in agriculture (soil moisture, 2015. The 1984 and 2015 were the foremost terrible groundwater table depth, crop yield). The recurrent Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 15 of 17 incidence of droughts for 1-, 4-, and 8-months has im- shocks that increase vulnerability and then step up to a portant implications on the socio-economy of rainfed cycle of poverty (Funk et al. 2012). Mekonen and Berlie agricultural practices and productions for smallholder (2019) and McSweeney et al. (2008) also reported that farmers. The incidences of drought owing to the short- the respective decreasing and increasing trends of rain- fall of rainfall during the months of belg (Feb-to-April), fall and temperature were the potential threats to the kiremt (June-to-Sept) and long-term rain (Feb-to-Sept) agricultural sector through pest intensification. seasons were influencing the agricultural production. The greatest drought risk incidence of months in the Conclusions belg and (kiremt) seasons under moderate and (extreme) This study presented the spatiotemporal drought risk drought intensity classes highly impacted the agricultural events in the northeastern highlands of Ethiopia using practices. The influence of moderate and extreme Standardized Precipitation Index, SPI. The results of the drought risk expected to be via the timely deficit of rain study confirm that highly complex and localized tem- at planting and critical stage of crop growing that finally poral and spatial patterns of drought risk events were exposed the society to shift to food aid. For example, the detected. This helped to identify and characterize local reports of Viste et al. (2013), Block (2008), Prasad and droughts. The majority of 1–4- and 8-months timestep Staggenborg (2009), World Bank (2006a, 2006b) and underwent severe and extreme (SPI ≤− 1.50) drought United Nations Environment Program/UNEP (2011) re- events. However, the detected frequency and magnitude vealed that from the 1970s onwards, the decline of rain- of drought risk events were more pronounced during fall amount and the increasing frequency of drought the rainy months of the belg season. directly influenced water shortages and limited crop The 1-month belg and (kiremt) rainy months experi- growth and development, and reduced yield. This fur- enced the largest drought risk events under the moderate ther threatened millions of people in the country that and (extreme) drought intensity classes. Moreover, the 4- demanded emergency humanitarian assistance. In rela- month belg and (the 4- month kiremt and 8-month long- tion to this, Webb et al. (1992), Von Braun (1991), and term) seasons experienced the largest drought risk events Webb and Von Braun (1990) also indicated that the pri- under the severe and (extreme) drought intensity classes. mary effects of drought on crop production and yields Increased drought risk frequency events were observed in were quite strong. For instance, a 10% decline in rainfall the 1980s and 2000s onwards for kiremtand long-term results in an average reduction in cereal yields of about seasons. A steady increased drought risk event frequency 4.2%. was detected since 1980s during belg season. In the spatial The deficit of February-to-May rainfall is also be pattern of drought, the southern part during belg,the threatening those slowly maturing long-cycle crop pro- northern and eastern during kiremt and the eastern half duction like sorghum and maize in the long-cycle crop during the long-term rainy seasons experienced more fre- growing areas of the study planted during belg and har- quent drought risk events. vested after kiremt seasons. Degefu (1987) and Funk The observed temporal and spatial drought risk events et al. (2005) report that poor belg season rainfall per- indicate a potential threat to the rainfed agricultural formance not only affecting belg crops but also adversely practices that gradually jeopardized the smallholder impacting the long-season crop production such as sor- farmers to food insecurity and socio-economic vulner- ghum and maize, the country’s major food crops. The ability. The findings of this study also could be a rudi- detected incidence of drought risks of 8-months (Feb-to- mentary stage to enhance drought risk management Sept) was also influencing the long-cycle crop produc- strategy through the appraisal of the rainfed agricultural tion such as Sorghum, Maize, and Millet growing areas practices in the study area. Therefore, the documenta- (Meher-Belg livelihood zones) owing to deficit of rainfall tion and informed assessment of drought frequency, in belg and kiremt seasons. In this regard, Funk et al. magnitude, intensity and severity based on the livelihood (2012) revealed that insufficient amount of belg and kir- zones are essential for drought risk management, early emt rainfall likely have negative impacts on agricultural warning responses, local-scale planning, and food secur- production of slowly maturing varieties of crops (Maize, ity management. Finally, the study recommended further Sorghum, and Millet) and food availability. Mahoo et al. research on additional indices of climatic variables such (2013) also reported that recurrent drought, changes in as evapotranspiration and soil water content. the consistent crop planting and shift in crop types, as well as rainfall variability related with the amount, tim- Abbreviations ing, and intensity, are impacting the agricultural prac- ANRSPC: Amhara National Regional State Plan Commission; CRED: Center for Research on the Epidemiology of Disaster; IPCC: Intergovernmental Panel on tices leading to frequent crop failures, losses of life and Climate Change; MK: Mann Kendal; NMSA: National Meteorological Service property. This means that more frequent droughts, and Agency; PDSI: Palmer Drought Severity Index; SPI: Standardized Precipitation drier climate in general, maybe producing repeated Index; SWAD: South Wollo Agriculture Department; UNEP: United Nations Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 16 of 17 Environmental Program; WFP: World Food Program; WMO: World Edossa DC, Babel MS, Gupta AD (2010) Drought analysis in the awash river basin, Meteorological Organization Ethiopia. Water Resour Manag 24(7):1441–1460. https://doi.org/10.1007/ s11269-009-9508-0 Endale D (1993) The Ethiopian famines, entitlements, and governance. World Acknowledgments Institute for Development Economics Research of the United Nations We are grateful to the Ethiopian National Meteorological Agency-East Amhara University, Annankatu 42 C, 00100 Helsinki, Finland. Working papers No. 102 Meteorological service center (Kombolcha) for providing us the monthly rainfall Funk C, Dettinger MD, Michaelsen JC, Verdin JP, Brown ME, Barlow M, Hoell A data. We are also indebted to the two anonymous reviewers for their insightful (2008) Warming of the Indian Ocean threatens eastern and southern African and thoughtful feedback for further improvement of the manuscript. food security but could be mitigated by agricultural development. Proc Natl Acad Sci 105(32):11081–11086. https://doi.org/10.1073/pnas.0708196105 Authors’ contributions Funk C, Rowland J, Eilerts G, Kebebe E, Biru N, White L, Galu G (2012) A climate The first author collects and analyzes the data, and wrote the manuscript. All trend analysis of Ethiopia. In: US Geol Surv, Fact Sheet, p 3053 authors read, edit and approved the manuscript. Funk C, Senay G, Asfaw A, Verdin J, Rowland J, Michaelson J et al (2005) Recent drought tendencies in Ethiopia and equatorial-subtropical eastern Africa. Funding FEWS-NET, Washington DC The first author is grateful to Bahir Dar and Debremarkos Universities for Gautam M (2006) Managing drought in sub-Saharan Africa: policy perspectives. funding support as part of his PhD study. Invited paper prepared for presentation at the International Association of Agricultural Economists, Gold Coast, Australia Availability of data and materials Gebrehiwot T, van der Veen A, Maathuis B (2011) Spatial and temporal The datasets used and/or analyzed during the current study are available assessment of drought in the northern highlands of Ethiopia. Int J Appl Earth from the corresponding author on reasonable request. Obs Geoinf 13(3):309–321. https://doi.org/10.1016/j.jag.2010.12.002 Ghosh KG (2019) Spatial and temporal appraisal of drought jeopardy over the Competing interests Gangetic West Bengal, eastern India. Geoenviron Disas 6(1):1. https://doi.org/ The authors declare that they have no competing interests. 10.1186/s40677-018-0117-1 Gommes R, Petrassi F (1996) Rainfall variability and drought in sub-Saharan Africa Received: 9 August 2019 Accepted: 28 January 2020 SD dimensions, FAO Guha-Sapir D, Vos F, Below R, Ponserre S (2012) Annual disaster statistical review 2011: the numbers and trends Centre for Research on the Epidemiology of References Disasters (CRED) AghaKouchak A, Nakhjiri N (2012) A near real-time satellite-based global drought Guttman NB (1998) Comparing the palmer drought index and the standardized climate data record. Environ Res Lett 7(4):044037. https://doi.org/10.1088/ precipitation index. JAWRA J Am Water Resour Assoc 34(1):113–121. https:// 1748-9326/7/4/044037 doi.org/10.1111/j.1752-1688.1998.tb05964.x ANRSPC (2017) Amhara National Regional State Plan Commission 2015/2016 Guttman NB (1999) Accepting the standardized precipitation index: a calculation budget year annual statistical bulletin, Ethiopia algorithm 1. JAWRA J Am Water Resour Assoc 35(2):311–322 Ashraf M, Routray JK (2013) Perception and understanding of drought and Hao Z, AghaKouchak A, Nakhjiri N, Farahmand A (2014) Global integrated coping strategies of farming households in north-West Balochistan. Int J Dis drought monitoring and prediction system. Scientific data 1:140001. https:// Risk Reduct 5:49–60. https://doi.org/10.1016/j.ijdrr.2013.05.002 doi.org/10.1038/sdata.2014.1 Bannayan M, Sanjani S, Alizadeh A, Lotfabadi SS, Mohamadian A (2010) Hayes M, Svoboda M, Wall N, Widhalm M (2011) The Lincoln declaration on Association between climate indices, aridity index, and rainfed crop yield in drought indices: universal meteorological drought index recommended. Bull northeast of Iran. Field crops research 118(2):105–114. https://doi.org/10. Am Meteorol Soc 92(4):485–488. https://doi.org/10.1175/2010BAMS3103.1 1016/j.fcr.2010.04.011 Hayes MJ, Svoboda MD, Wiihite DA, Vanyarkho OV (1999) Monitoring the 1996 Beth L, Brown TJ (2003) A comparison of precipitation and drought indices related drought using the standardized precipitation index. Bull Am Meteorol Soc to fire activity in the US. In: 5th symposium on fire and Forest meteorology 80(3):429–438 Block PJ (2008) An assessment of investments in agricultural and transportation Jeyaseelan AT (2003) Droughts and floods assessment and monitoring using remote infrastructure, energy, and hydroclimatic forecasting to mitigate the effects sensing and GIS. In: Satellite remote sensing and GIS applications in agricultural of hydrologic variability in Ethiopia. In: CPWF working paper 01 the CGIAR meteorology (Vol. 291). World Meteorol. Org. Dehra Dun, India Geneva challenge program on water and food, Colombo, Sri Lanka, p 53 Jjemba E, Singh R, Arrighi J (2017) Extreme drought in Ethiopia stretches drought Bussay AM, Szinell C, Hayes M, Svoboda M (1998) Monitoring drought in Hungary management systems using the standardized precipitation index. In: Annales Geophysicae, Kahsay B (2013) Diagnosis and intervention plans for south Wollo zone, Amhara Supplement (Vol. 11) region: livestock and irrigation value chains for Ethiopian smallholders (LIVES) Butterfield R (2011) Climate changes in East Africa https://www.weadapt.org/ Google Scholar knowledge-base/national-adaptation-planning/climate-changes-in-east-africa Kendall MG (1975) Rank correlation methods, 4th edn. Griffin, London Centre for Research on the Epidemiology of Disasters/CRED (2016) International Khan MA, Gadiwala MS (2013) A study of drought over Sindh (Pakistan) using disaster database http://www.emdat.be/disaster_profiles/index.html standardized precipitation index (SPI) 1951 to 2010. Pakistan J Meteorol 9(18):15–22 Chattopadhyay S, Edwards DR (2016) Long-term trend analysis of precipitation Korecha D, Barnston AG (2007) Predictability of June–September rainfall in Ethiopia. and air temperature for Kentucky, United States. Climate 4(1):10. https://doi. Mon Weather Rev 135(2):628–650. https://doi.org/10.1175/MWR3304.1 org/10.3390/cli4010010 Łabędzki L (2007) Estimation of local drought frequency in Central Poland using Conway D (2000a) Some aspects of climate variability in the northeast Ethiopian the standardized precipitation index SPI. Irrigation and Drainage J Int highlands-Wollo and Tigray. Sinet: Ethiopian J Sci 23(2):139–161. https://doi. Commission Irrigation Drain 56(1):67–77. https://doi.org/10.1002/ird.285 org/10.4314/Sinet. v23i2.18163 Little PD, Stone MP, Mogues T, Castro AP, Negatu W (2006) ‘Moving in place’: Conway D (2000b) The climate and hydrology of the upper Blue Nile River. drought and poverty dynamics in south Wollo, Ethiopia. J Dev Stud 42(2): Geogr J 166(1):49–62. https://doi.org/10.1111/j.1475-4959.2000.tb00006.x 200–225. https://doi.org/10.1080/00220380500405287 Dai A (2011) Drought under global warming: a review. Wiley Interdiscip Rev Clim Loukas A, Vasiliades L (2004) Probabilistic analysis of drought spatiotemporal Chang 2(1):45–65. https://doi.org/10.1002/wcc.81 characteristics inThessaly region, Greece. Nat Hazards Earth Syst Sci 4(5/6):719–731 Degefu MA, Bewket W (2015) Trends and spatial patterns of drought incidence in Lyon B (2014) Seasonal drought in the greater horn of Africa and its recent the Omo-Ghibe River basin, Ethiopia. GeografiskaAnnaler: Ser A, Phys increase during the march–may long rains. J Climate 27(21):7953–7975. Geograph 97(2):395–414. https://doi.org/10.1111/geoa.12080 Degefu W (1987) Some aspects of meteorological drought in Ethiopia. https://doi.org/10.1175/JCLI-D-13-00459.1 Cambridge University Press, Cambridge, pp 23–36 Lyon B, DeWitt DG (2012) A recent and abrupt decline in the east African long Dorosh P, Rashid S (2013) Food and agriculture in Ethiopia: Progress and policy rains. Geophys Res Lett 39(2). https://doi.org/10.1029/2011GL050337 challenges. University of Pennsylvania Press, Philadelphia. International MahooH, RadenyM,KinyangiJ,Cramer L (eds)(2013) Climatechange Journal of the Commons, 7(2):573–574. https://doi.org/10.18352/ijc.448 vulnerability and risk assessment of agriculture and food security in Mekonen et al. Geoenvironmental Disasters (2020) 7:10 Page 17 of 17 Ethiopia: which way forward? CCAFS working paper no. 59. CGIAR Stagge JH, Tallaksen LM, Gudmundsson L, Van Loon AF, Stahl K (2015) Candidate Research Program on Climate Change, Agriculture and Food Security distributions for climatological drought indices (SPI and SPEI). Int J Climatol (CCAFS), Copenhagen 35(13):4027–4040 Mann HB (1945) Nonparametric tests against trend. Econometrica 13:245–259 Sternberg T, Thomas D, Middleton N (2011) Drought dynamics on the Mongolian Marimon DC (2016) Contributions to the knowledge of the multitemporal spatial steppe, 1970–2006. Int J Climatol 31(12):1823–1830. https://doi.org/10.1002/ patterns of the Iberan peninsula droughts from a geographic information joc.2195 service perspective, Universitat Autònoma de Barcelona, PhD thesis. Suryabhagavan KV (2017) GIS-based climate variability and drought Accessed from https://www.tdx.cat/handle/10803/379829#page=1. characterization in Ethiopia over three decades. Weather Clim Extremes 15: 11–23. https://doi.org/10.1016/j.wace.2016.11.005 Mashari Eshghabad S, Omidvar E, Solaimani K (2014) Efficiency of some SWAD (2018) South Wollo Zone Agricultural Department 2016/17 budget year meteorological drought indices in different time scales (case study: Tajan annual report. SWAD, Dessie Basin, Iran). ECOPERSIA 2(1):441–453 Szalai S, Szinell CS (2000) Comparison of two drought indices for drought Masih I, Maskey S, Mussá FEF, Trambauer P (2014) A review of droughts on the monitoring in Hungary—a case study. In: Drought and drought mitigation in African continent: a geospatial and long-term perspective. Hydrol Earth Syst Europe, pp 161–166 Sci 18(9):3635. https://doi.org/10.5194/hess-18-3635-2014 Thompson SA (1999) Hydrology for Water Management. CRC Press, London, pp McKee, T. B., Doesken, N. J., & Kleist, J. 1993. The relationship of drought frequency 286-317. https://doi.org/10.1201/9780203751435 and duration to time scales. In Proceedings of the 8th conference on applied Tierney JE, Ummenhofer CC (2015) Past and future rainfall in the horn of Africa. climatology. Boston, American Meteorological Society, (17)22, 179–183 Sci Adv 1(9):e1500682. https://doi.org/10.1126/sciadv.1500682 McSweeney C, New M, Lizcano G (2008) UNDP climate change country profiles: Trenberth K, Meehl J, Masters J, Somerville R (2011) Current extreme weather and Ethiopia Available from http://country-profiles.geog.ox.ac.uk climate change. Climate communication science and outreach Mekonen AA, Berlie AB (2019) Spatiotemporal variability and trends of rainfall Tsakiris G, Vangelis H (2004) Towards a drought watch system based on spatial and temperature in the northeastern highlands of Ethiopia. Modeling Earth SPI. Water Resour Manag 18(1):1–12. https://doi.org/10.1023/B:WARM. Syst Environ:1–16. https://doi.org/10.1007/s40808-019-00678-9 0000015410.47014.a4 Mirza MMQ (2003) Climate change and extreme weather events: can developing United Nations Environment Program/UNEP (2002) Global environment outlook countries adapt? Clim Pol 3(3):233–248. https://doi.org/10.1016/s1469- 3: past, present and future perspectives. Environ Manag Health 13(5):560–561 3062(03)00052-4 United Nations Environment Program/UNEP (2011) Food security in the horn of Morid S, Smakhtin V, Moghaddasi M (2006) Comparison of seven meteorological Africa: the Implicationsof a drier, hotter and more crowded future Ecosystem indices for drought monitoring in Iran. Int J Climatol 26(7):971–985. https:// Management, Climate Change doi.org/10.1002/joc.1264 USAID (2009) Ethiopia livelihood zones Accessed from http://www.fews.net/east- Mpelasoka F, Hennessy K, Jones R, Bates B (2008) Comparison of suitable drought africa/ethiopia/.livelihood-zone November, 2017 indices for climate change impacts assessment over Australia towards resource Viste E, Korecha D, Sorteberg A (2013) Recent drought and precipitation management. Int J Climatol 28(10):1283–1292. https://doi.org/10.1002/joc.1649 tendencies in Ethiopia. Theor Appl Climatol 112(3–4):535–551. https://doi. National Meteorological Services Agency of Ethiopia/NMSA, 1996. Climatic and org/10.1007/s00704-012-0746-3 agro-ecology resources of Ethiopia, Addis Ababa, Ethiopia, 1(1):0–137 Von Braun J (1991) A policy agenda for famine prevention in Africa (no. 1) Ntale HK, Gan TY (2003) Drought indices and their application to East Africa. Int J International food policy research institute (IFPRI) Climatol 23(11):1335–1357. https://doi.org/10.1002/joc.931x Webb P, Braun JV (1994) Famine and food security in Ethiopia: lessons for Africa. Wiley, Padhee SK, Nikam BR, Aggarwal SP, Garg V (2014) Integrating effective drought Chichester index (EDI) and remote sensing derived parameters for agricultural drought Webb P, Braun JV, Yohannes Y (1992) Famine in Ethiopia: policy implications of assessment and prediction in Bundelkhand region of India. Int Arch coping failure at national and household levels. IFPRI research reports Photogram Remote Sensing Spatial Inform Sci 40(8):89. https://doi.org/10. Webb P, Von Braun JV (1990) Drought and food shortages in Ethiopia: a 5194/isprsarchives-XL-8-89-2014 preliminary review of effects and policy implications International Food Philip S, Kew SF, Jan van Oldenborgh G, Otto F, O’Keefe S, Haustein K et al (2017) Policy Research Institute The drought in Ethiopia, 2015 climate and development knowledge network WFP (2014) Resilience measurement principles: toward an agenda for and world weather attribution initiative Raising Risk Awareness measurement design. Technical series no.1 Piechota TC, Dracup JA (1996) Drought and regional hydrologic variation in the Wilhite DA (2000) Drought as a natural hazard: concepts and definitionsDrought United States: associations with the El Niño-southern oscillation. Water Mitigation Center Faculty Publications, pp 69.3–69.18 Resour Res 32(5):1359–1373. https://doi.org/10.1029/96WR00353 Wilhite DA, Glantz MH (1985) Understanding: the drought phenomenon: the role Prasad PV, Staggenborg SA (2009) Growth and production of sorghum and of definitions. Water Int 10(3):111–120. https://doi.org/10.1080/ millets. soils, plant growth and crop production, 2. In: Encyclopedia of life support systems (EOLSS), developed under the auspices of the UNESCO Williams AP, Funk C (2011) A westward extension of the warm pool leads to a Segele ZT, Lamb PJ (2005) Characterization and variability of Kiremt rainy season westward extension of the Walker circulation, drying eastern Africa. Climate over Ethiopia. Meteorol Atmos Phys 89(1–4):153–180. https://doi.org/10.1007/ Dynam 37(11–12):2417–2435. https://doi.org/10.1007/s00382-010-0984-y s00703-005-0127-x WMO (2009) Guide to hydrological practices, volume II, management of water Seleshi Y, Camberlin P (2006) Recent changes in dry spell and extreme rainfall resources and application of hydrological practices Extreme value analysis, events in Ethiopia. Theor Appl Climatol 83(1):181–191. https://doi.org/10. WMO-No. 168 1007/s00704-005-0134-3 WMO (2012) Standardized Precipitation Index User Guide (M. Svoboda, M. Hayes Seleshi Y, Zanke U (2004) Recent changes in rainfall and rainy days in Ethiopia. Int J and D. Wood). (WMO-No. 1090), Geneva Climatol A J R Meteorol Soc 24(8):973–983. https://doi.org/10.1002/joc.1052 WMO (2016) Guidelines on the definition and monitoring of extreme weather Sen PK (1968) Estimates of the regression coefficient based on Kendall's tau. J and climate events Am Stat Assoc 63(324):1379–1389 World Bank (2006a) Ethiopia: managing water resources to maximize sustainable Shanko D, Camberlin P (1998) The effects of the Southwest Indian Ocean tropical growth. World Bank, Washington, D.C. cyclones on Ethiopian drought. Int J Climatol 18(12):1373–1388 World Bank (2006b) IDA countries and exogenous shocks. International Sheffield J, Wood EF (2007) Projected changes in drought occurrence under future development association resource mobilization. World Bank, Washington, DC global warming from multi-model, multi-scenario, IPCC AR4 simulations. Zhan W, Guan K, Sheffield J, Wood EF (2016) Depiction of drought over sub- Climate dynamics 31(1):79–105. https://doi.org/10.1007/s00382-007-0340-z Saharan Africa using reanalysis precipitation data sets. J Geophys Res Atmos Shiferaw B, Tesfaye K, Kassie M, Abate T, Prasanna BM, Menkir A (2014) Managing 121(18):10,555–10,574. https://doi.org/10.1002/2016JD024858 vulnerability to drought and enhancing livelihood resilience in sub-Saharan Africa: technological, institutional and policy options. Weather and Climate Extremes 3:67–79. https://doi.org/10.1016/j.wace.2014.04.004 Publisher’sNote Spinoni J, Naumann G, Carrao H, Barbosa P, Vogt J (2014) World drought Springer Nature remains neutral with regard to jurisdictional claims in frequency, duration, and severity for 1951–2010. Int J Climatol 34(8):2792– published maps and institutional affiliations. 2804. https://doi.org/10.1002/joc.3875

Journal

Geoenvironmental Disasters – Springer Journals

Published: Mar 3, 2020

There are no references for this article.