Famine in Ethiopia

MONOGRAPH NO. 48

Famine in Ethiopia Editorial (Bo Vahlquist) I.

Study of shelter population in the Wollo region (Shewandagn Belete, Mehari Gebre-Medhin, Bantirgu Hailemariam, Mario Maffi, Bo Vahlquist and Zewdie Wolde-Gebriel)

II. Sequence of events. Model of a Consolidated Food and Nutrition Information Sysfem ("Early Warning System") (Mehari Gebre-Medhin) III. Initial experience of a Consolidated Food and Nutrition Information System. Analysis of data from the Ogaden area (Mehari Gebre-Medhin, Roger Hay, Yaicob Licke, and Mario Maffi)

Correspondence and request for reprints: Dr. Mehari Gebre-Medhin, Institute of Nutrition, Uppsala University, Box 551, S-751 22, Uppsala.

Environmental Child Health, February 1977

13

BOVAHLQUIST

Editorial At a time when knowledge about means of making the production, harvesting, processing and transport of food more efficient than ever is at our disposal and when family planning measures have caused a virtual standstill of the population growth in quite a few countries, it is an absurd situation that scarcity of food is a chronic problem for the majority of mankind. Over and again in recent years the situation has deteriorated to frank and downright famine that hits the headlines of the newspapers (though very rarely those of articles in scientific journals!). The causes of this appalling and intolerable situation are as manifold and complex as are the causes of poverty and indeed the two — famine and poverty — are closely interrelated. In this issue of the Journal, a Monograph appears which is devoted to Famine in Ethiopia. There are several reasons why the situation in Ethiopia has come so much to the foreground in the 1970s. Firstly, the extent and severity of human suffering has been colossal; secondly, the effects of the famine on the local and the national scene have been far-reaching and penetrating; -and thirdly, the new political system has allowed an insight into prevailing and underlying conditions that was previously lacking. Ethiopia, this vast and mountainous country with an area more than double the size of France and a population coming close to 30 million, has abundant agricultural resources. So far, however, these resources have only been partly utilized. Ninety percent or more of the total population is rural. Until recently large groups — perhaps half or more of the population — were tenants or farm labourers. Their life was one of marginal subsistence with recurring pre-harvest "hungry seasons" which, especially in the Northern part of the country, meant real hardship even in "normal" times. At about the same time as the rains began to fail in the Sahel area of Western and Central Africa, i.e. in the late 1960s drought problems accumulated increasingly also in parts of Ethiopia. At first it was • mainly the Northern regions that were afflicted, particularly Wollo and Tigrai. In the late fall of 1972 a large number of people started to migrate from the interior of these areas in search of food. Special shelters had to be erected which in Wollo alone, during 1973/74 at the peak of their activities, harboured some 30,000 destitutes, with probably as many living outside but nearby. No one knows exactly the number of lives lost, but it has been stated that out of the estimated 2.4 million inhabitants of Wollo the excess loss of lives due to famine already in August 1973 may have been in the order of 50-100,000. When the situation in Wollo started to improve in 1975 other regions of Ethiopia had already begun to suffer severly, this time primarily the Eastern and Southern lowlands. The demands on the new 14

Government were heavy. In the spring of 1974 the Relief and Rehabilitation Commission (RRC) was given wide powers and became the focal point for national relief activities as well as for the coordination of national and international efforts. The Commission cooperated closely with various international agencies, especially UNICEF, and was also able to draw extensively from the resources of a well established national centre, the Ethiopian Nutrition Institute (ENI) — in part supported by the Swedish government. It soon became clear that a system for monitoring the food situation in different areas, which at the same time could serve as an "early warning system" was urgent. Late in 1974 the Commissioner appointed an Interministerial Working Group comprising members of all relevant ministries and agencies (some 15 in all) with the Director of ENI as chairman. In September 1975 the report of the group ("Consolidated Food and Nutrition Information System") and the proposals for action which it contained, were endorsed by representatives of participating ministries and agencies. Special funds have been allocated to make action possible without delay. Parallel with the desk work on the report and on the basis of experience from Wollo, systematic data collection and relief programs had already started in the field. In the semi-arid lowlands of the Hararghe region, which form part of the vast Ogaden area on either side of the Ethio-Somali border, the effects of persistent severe drought had become more and more apparent. After earlier ad hoc action by voluntary organizations systematic screening of relief needs started in May-June 1974. The situation deteriorated markedly between.November 1974 and March 1975. At this time the ENI, operating on behalf of the RRC and through generous UNICEF support, established a number of shelters and feeding sites. Within a few months the total number of beneficiaries exceeded 80,000. By September 1975 the situation had improved but the risk of a recurrence of severe food shortage in a population which to a considerable extent had been deprived of its cattle was still imminent. The famine has also struck hard in other regions of Ethiopia, such as Bale, Gemu Gofa and S.W. Shoa ("the Kembata area"). The report to be found in paper No. Ill (in the Monograph) deals only with the Ogaden observations. We believe that in many ways they are typical for areas inhabited by nomads. Both the plan for an Early Warning System as described in part II of the Monograph and the experience gained in Ogaden as reported in part HI deserve attention on an international as well as a national level. Bo Vahlquist (Guest Editor) Department of Paediatrics, University Hospital, Uppsala, Sweden Environmental Child Health, February 1977

S. BELETE, M. GEBRE-MEDHIN, B. HAILEMARIAM, M. MAFFI, B. VAHLQUIST, Z. WOLDE-GEBRIEL

From the Ethiopian Nutrition Institute, Addis Ababa, Ethiopia

I. Study of shelter population in the Wollo region* by SHEWANDAGN BELETE.f MEHARI GEBRE-MEDHIN.f BANTIRGU HAILEMARIAM.f MARIO MAFFI.t BO VAHLQUISTJ and ZEWDIE WOLDE-GEBRIELf Ethiopia is one of several African countries in the Sahel (the semi-arid region south of the Sahara u ) which have suffered severely from the effects of prolonged drought during the 1970s. Large parts of the country have successively been declared famine areas and numerous activities, both national and international, have been generated in an attempt to relieve the situation.*12 One of the most seriously affected parts of the country has been the administrative region of Wollo (Fig. 1). At about the end of 1972 large numbers of people migrated from the interior of this area to roadside towns in search of food and labour. During the spring of 1973 grain distribution started on an irregular basis. At this time most of the migrants were severely malnourished and ill and were in need of immediate care. The first systematic emergency relief measures, initiated by local authorities in July 1973, were to gather the destitutes from streets and Fig. 1

SOMALIA UGAN

The Wollo administrative region. Cross-hatched lines indicate arid and semi-arid areas. *This paper is based on one of a series of mimeographed reports from surveys conducted by the Ethiopian Nutrition Institute and presented to the Relief and Rehabilitation Commission for the purpose of action in famine stricken areas. It is republished with permission from the Editor, "Courrier". t Ethiopian Nutrition Institute. ^Department of Pediatrics, Uppsala University, Uppsala, Sweden. Environmental Child Health, February 1977

courtyards and house them in provisional shelters where they could obtain food and medical care. During the autumn of 1973 the organization became more stable, and eventually comprised 13 shelters. At the peak of the activities these shelters harboured in all some 30,000 people. After fluctuations in utilization, due to government policy as much as to food situation, successive "demobilization" started towards the end of 1974. In the summer of 1975 the shelter relief activities had diminished very markedly, only a few of the. shelters still being operative. The time of active shelter involvement averaged 18 months, but there was considerable variation from one shelter to another. Along with many other organizations the Ethiopian Nutrition Institute (ENI) in Addis Ababa, a collaborative Ethio-Swedish project, was involved in relief work, particularly in relation to food and nutrition. For the purpose of rapid assessment of human involvement in the famine, the requirement of rehabilitative foods, the organization and supervision of feeding programs in shelters, and the training of field personnel, the Institute established three teams, each consisting of one health officer and two home economic assistants. The group started its work in Wollo in October 1973. The frame of the ENI team activities was later extended, to include permanent supervision of feeding programs in every shelter. At the same time all necessary kitchen utensils were procured. Soon after its involvement in the relief activities the Institute felt that a systematic study of the camps was necessary not only for continued relief work in these shelters but also for the purpose of future national planning8 and as an important source of information on the situation in the interior of the region from which the people had migrated. The administrative region of Wollo* The topography of Wollo is characterized by three distinctive physical features: the western highlands, the eastern lowlands and the central plateau. The western highlands constitute a chain of rugged mountains reaching peaks above 4,000 metres. The eastern lowlands fall within the western catchment of the Awash River and extend down to the dry plains of the Awash Valley and the Danakil Depression. The *Vne numerical information in this section mainly derives from a report by the Central Statistical Office, Addis Ababa.5 15

S. BELETE, M. GEBRE-MEDHIN, B. HAILEMARIAM, M. MAFFI, B. VAHLQUIST, Z. WOLDE-GEBRIEL

central plateau falls within the catchment areas of the Blue Nile and Tekeze Rivers. The population of Wollo is estimated to be 2.4 million. In normal times only 5.3% live in urban areas. The population density in general varies from 20-80 inhabitants per sq km in the highlands and in the central plateau to less than 10 in the eastern lowlands. Some parts of the highlands and the central plateau are among the most densely populated segments of the country, with densities even exceeding 160 per sq km. Inhabitants of the highlands and central plateau are settled farmers, whereas the eastern lowlands are inhabited by nomadic herdsmen. The average annual rainfall is 1,000 to 2,000mm in the highlands, 800 to 1,200mm in the central plateau and 50 to 400mm in the eastern lowlands. The highlands and the central plateau have two distinct cropping seasons: the 'meker' (big rains) crops sown in June-July and harvested in November-December and the 'belg' (small rains) crops sown in February-March and harvested in June. There is only one harvest per year in the lowlands, occurring in NovemberDecember. Tef, barley and sorghum are the main crops, while maize, wheat/ noug (Guizotia abyssinica Cass.) and horse beans are only grown to a limited extent. The noma'dic herdsmen roam with their cattle which they barter for food from the highlands. In the agricultural zone, more than 50% of the farmers used to rent the land they tilled. The share of crops paid to the owners ranged from 50-75% of the total produce. The average holding per family was about two acres.* Communication in the region is very difficult, with few roads branching from the main highway into the interior. This is particularly true of the western section, which is virtually inaccessible. The present famine Severe food shortages occurring in Wollo along with other regions in Ethiopia have been described in a number of early records. 1516 In the most recent disaster, which affected Wollo less than a decade ago, there was severe famine in the highland sub-regions of Lasta and Wag. The vulnerability of Wollo, together with several other regions of Ethiopia stems from its archaic agricultural and land tenure system, its dense population, continuous deterioration of the soil, lack of infrastructure, and heavy bureaucracy. These factors have always contributed to the maintenance of a marginal subsistence level10 with the ever present risk of famine whenever climatic conditions become unfavourable. Such unfavourable conditions prevailed in the early 1970s. Erratic precipitation in certain areas and general inadequacy of rainfall throughout the region continuously and drastically reduced the 'meker' harvest and also caused a heavy loss of •According to the Land reform proclamation of March 1975, all land has been distributed among the peasants with the aim that each family should cultivate 10 acres of land. See also Discussion. 16

livestock. The situation was further aggravated when the 1973 'belg' rains failed to come, causing a serious reduction of yields in the highlands. Soon food reserves in many hamlets were completely depleted, famine became rampant and migration in search of food was the last resort.1 Population studied, methods and procedures used The present report is based on a study carried out in the 13 shelters of the region of Wollo during the period May-June 1974. All heads of households, numbering in all 6,490, were interviewed, with use of a special questionnaire. Information was obtained regarding the origin of the family, the age and sex distribution of the family members, the reason for migration and the fate of property, etc. The majority of the inmates of the shelters, numbering about 29,000, were screened, by means of standard anthropometric measurements. Only for the children, however, was the screening sufficiently systematic to allow a meaningful analysis of the figures. Results Background Fig. 2 gives some background data, including the character of the staples in different areas of Wollo and the outcome of the crop production in 1972-73. It also indicates the areas of most intense population migration. Fig. 2

STAPLE CROPS OTHER SORGHUM BARLEY TEFF

POPULATION TERRAIN A80VE

MIGRATION

ELEVATION 1000 m

CROP PRODUCTION 1972-73 ' NORMAL I

I BELOW NORMAL

The Wollo administrative region. Staple crops in different areas, crop production in 1972-73, and population migration (dots indicate major areas involved). Alphabetical letters (A-L) indicate sub-regions ("awrajas") — see also Table I. The figures (1-13) atfilledtriangles indicate shelters — see also Table I. Origin of shelter population The majority of the migrants originated from districts in the eastern lowlands of the region (Fig. 3). Three of the sub-regions ("awrajas"), viz. Yeju, Ambassel and Raya/Kobo, account for over 75% of the migration to the shelters (Table I). Size of shelter population The size of the population varied considerably in individual shelters. Thus the number of heads of Environmental Child Health, February 1977

3?

n E.

o

IE

I

8

1

4 Gobiye

345

156

320

9

F , Raya& Kobo

TOTAL

13 Serdo

12 Kemisse

11 Bati

10 Komboltcha

169

182

18

2

22

1

15

14

1

1,350

98

3

—

—

1,994

278

16

285

441

21

9 Dessie 16

384

63

—

603

362

9

*Yeju

—

G

8 Haik

26

—

—

—

—

—

E Borena

7 Worgessa —

—

—

—

—

D Wore Himenu

9 —

—

—

—

—

C Wadla Delanta

6 Mersa

45

80

—

3 Kobo

4

27

49

. 2 Alamata

5 Woldiya

—

B Lasta

73

A Wag

1 Korem

No

Shelter Name

Sub-region

1,639

20

643

419

—

542

9

—

—

H Ambassel

110

2

104

1

3

I Dessie Zuria

Table I Relative size of individual shelters as indicated by number of heads of households. Origin of inmates from different subregions.

—

—

—

—

—

J Were Ilu

—

—

7

—

515

_

67

3

438

•

——

—

—

—

K Kalu

•

184

79

7

94

—

4

—

—

—

——

—

—

—

L Awsa

1

1

292

11

9

147

—

48

—

6,490

177

403

765

1,393

981

590

644

372

363

247

467

62 11

84

TOTAL

2

Others/ unknown

O r O to

i

I

2 >

>

O

O m

S. BELETE, M. GEBRE-MEDHIN, B. HAILEMARIAM, M. MAFFI, B. VAHLQUIST, Z. WOLDEGEBRIEL

Fig. 3 WOLLO: ORIGIN OF SHELTER

POPULATION

Intensity of migration In Table II the intensity of migration from different sub-regions has been calculated. It is evident that also with respect to these relative figures the sub-regions of Raya/Kobo, Yeju and Ambassel contribute more than any of the others. Age and sex distribution The results are given in Fig. 4. Fig. 4 AGE PYRAMID

OF SHELTER ^ POPULATION

AND PRE-DROUGHT |""j

IN WOLLO.

AGE FEMALE

MALE TERRAIN ELEVATION ABOVE 1000 m

The Wollo administrative region. Origin of shelter population. For explanation of alphabetical letters and figures, see Table I.

households at the'time of the study varied from 84 to 1,393 (cf. Table I). The number of members per household, on a gross average, was 4.5 (this figure is calculated on the ratio pf heads of households to aggregate shelter population). Using this figure as a basis the-population of individual shelters varied from ca. 400 to 6,300. In some of the shelters the number increased considerably during feeding hours, since also people living outside the shelters were granted meals. By far the largest groups of migrants were harboured in the two neighboring centers of Komboltcha and Dessie (the capital of the region) located close to the southern border of Wollo. Table II Intensity of migration from different sub-regions.

Sub-region ("awraja") Wag Lusta Wadla Delanta Wore Himenu Borena Raya & Kobo Yeju Ambassel Dessie Zuria Were Ilu Kalu Awsa

Population of Sub-region x 1000

Shelter inmates in 0/00 of population*

142 292 147 225 265 54 159 261

1.8 2.8 0.6 0.4

•

179

137 126 73

0.3 112.5 56.4 28.3 2.8 0 18.3 11.4

•Number of shelter inmates calculated as Heads of Households x 4.5 (see text). 18

40

30

20

10

10

20

30

No per year of age per 1000 individuals

Age pyramid (males and females) of aggregate shelter population as compared with pre-drought situation.

With respect to age the composition of the shelter population shows a striking "deficit" at the base and top of the pyramid when comparisons are made with pre-drought times. Children 0-4 years of age and adults 45 years or over are markedly underrepresented. In contrast to this there is a marked "excess" in the proportion of children aged 5-14 years. For the group 15-44 years the total representation of subjects tallies fairly well with pre-drought figures (the sex ratio is changed, however; see below). The observations with respect to sex ratio in individual shelters are presented in Figs. 5a-d. It is obvious that the variations between different shelters as well as between different age groups in one and the same shelter are very great indeed. The most striking observation is that in the age group 15-44 years the male/female ratio is considerably below par. Incidentally the male/female ratio in the pre-drought Wollo population shows an unexpected feature in the form of a larger than one ratio in the two child groups, especially the group of 5-14-year-olds. Anthropometry The results of weight for height measurements are presented in Table III. In the two child groups 0-4 years and 5-14 years of age, 19.2% and 10.3%, Environmental Child Health, February 1977

S. BELETE, M. GEBREMEDHIN, B. HAILEMARIAM, M. MAFFI, B. VAHLQUIST, Z. WOLDE-GEBRIEL

Fig. 5a SEX RATIO OF SHELTER

Fig. 5d

POPULATION: WOLLO 0-4years

Mole per 100 females

200

4,

SEX RATIO OF SHELTER POPULATION: WOLLO 45 yeors end above Mall [»r 100 f«males 1600 , t 7 150 JL X

150

100

50 IA

iM

Sex ratio of shelter population in four different age groups (a. 0-4; b. 5-14; c. 15-44 and d. > 45 years). The figures indicate shelter numbers (see Table I).

W0L10 P R E - DROUGHT MEAN FOR ALL SHELTERS

o OO

Fig. 5b SEX

RATIO OF SHELTER

POPULATION: WOLLO 5-14 yeors

Mole per 100 femoles

respectively, scored below 60% of the WHO standard" and were thus severely malnourished. Only 9.5% of the children 0-4 years old and 8.2% of the 5-14 year-olds were above 90% weight for age of the standard. Table III Nutritional status. Frequency distribution of % weight for height. Children 0-4 and 5-14 years.

150

Age group years

100

m

[7]

0-4 5-14

50

O OO

WOLLO PRE-OROUGHT MEAN FOR ALL SHELTERS

Fig. 5c SEX RATIO OF SHELTER Male per 100 females 340

POPULATION: WOLLO 15-44 yeors

o W0L10 PRE-DROUGHT 0 0 MEAN FOR AU. SHELTERS


Frequency distribution of weight for height > 90

81-90

71-80

61-70

9.5 8.2

19.5 19.4

23.3 34.4

28.5 27.7

O)

3

ss i

§ K

1

3 3 2

Fig. 7 Average values for N. and S. Ogaden.

Table II Mean market prices by type of commodities. Eth.S/quintal or head: 1 Eth.S = c:a 0.50 USS 1 quintal = 100kg. NORTHERN OGADEN Time May-June '74 Nov-Dec '74 Feb-March '75 June '75 Pre-drought Year 32

Sorghum

Maize

Cattle

42 49.8 36.3 31.3 17

36 42.1 35.3 28.5" 14

62 83.0 32.2 146.3 87

SOUTHERN OGADEN

Sheep & Goats Camels Sorghum 18 16.0 13.1 19.1 21

193 132.5 77.9 168.3 238

51 48.0 36.3 35.5 16

Maize

Cattle

48 43.8 35.3 32.1 14

77 46.3 32.2 104.4 86

Sheep & Goats Camels 19 14.5 13.1 20.7 18

223 201.7 77.9 247.7 222


MEHARI GEBRE-MEDHIN, ROGER HAY, YAICOB LICKE A N D MARIO MAFFI

Nutritional status A significant deterioration in the nutritional condition of the population took place in the second half of 1974. This trend continues to reach the lowest levels in the period February to March 1975. Fig. 9. The proportion of moderately and severely malnourished children (below 80% weight/height standard) doubled between May-June and NovemberDecember 1974. Table III. A degree of recovery, more marked in the north than in the south, is observed in June-August 1975. In the south the proportion of children with severe malnutrition was still twice as great in June-August 1975 as one year earlier. The mode point for both north and south in June-August 1975 still lies between 80 and 89% of standard.

PRICE INDEX FOR ANIMALS AND FOOD COMMODITIES

J3U-

•-~

•

300250-

MAIZE m

>

\

•^

SORGHUM

-

\

SHEEP i ' GOATS

200-

CATTLE 150100-

CAMELS .•

/

y

—• - -T*^~*\— *''

50-

a I % Human mortality Mortality figures were obtained by interviewing heads of households about deaths having occurred during the preceding 12 month's. In plotting against time (Fig. 10) it seemed reasonable to antedate the mortality figures thus obtained six months earlier than the actual time of interviewing. Our observations would indicate that the overall mortality rate was at its highest during the latter part of 1974 (and, indeed, probably remained high during the early part of 1975). For the most vulnerable age group, the under five, the mortality rate in Southern Ogaden remained high throughout the whole observation period. In Northern Ogaden there was a steady deterioration during the second half of 1974. For the 5-15 year old the trends were similar, although the mortality rate was, of course, on a much lower level. For those above 15 years the figures do not lend themselves to any definite conclusions in this respect.

Fig. 8 Price index for animals and food commodities. Mean values for N. and S. Ogaden. MEAN WEIGHT FOR HEIGHT AS V. NORTHERN OGADEN OF STANDARD 100-1

SOUTHERN OGADEN

w Fig. 9 Human nutritional status. Weight for height as % of standard over time. Mean value and standard errors of the mean for N. and S. Ogaden.

Table III Human nutritional status. Frequency distribution of % weight for height. Mean values and standard deviations. N. and S. Ogaden. NORTHERN

SOUTHERN

OGADEN

OGADEN

% Standard May-June '74 Nov-Dec '74 Feb-Mar '75 July-Aug '75 May-June '74 Nov-Dec '74 Feb-Mar '75 July-Aug '75 = 100 90— 99 80— 89 70— 79 = 70

10.5 23.0 41.3 21.9 3.3

5.2 18.5 . 37.3 30.7 7.9

1.6 11.9 41.0 38.0 7.5

X = 86.54 XO = 82.70 X = 80.91 SD = 9.71 SD = 9.99 SD = 8.58 n = 3O5 n = 563 Environmental Child Health, February 1977

18.9 20.0 35.8 19.4 5.9

24.0 33.9 29.5 9.9 2.7

4.4 18.6 44.7 26.8 5.5

X = 88.94 X = 92.15 XO = 83.50 SD = 14.03 SD= 11.54 SD = 9.05 n = 292 n = 609

2.6 16.1 38.4 35.4 7.5 X = 82.15 SD = 8.31

12.2 • 20.1 35.8 24.9 7.0 X = 86.27 SD = 13.36 n=1169 33

MEHARIGEBREMEDHIN, ROGER HAY, YAICOB LICKE A N D MARIO MAFFI

AGE SPECIFIC

HUMAN

MORTALITY

RATES PER 10OO POPULATION

NO. OF PERSONS PER 1000 SOUTHERN OGADEN

NORTHERN OGAOEN '

280-

—

^

.

^

< 5 yeors 5-15 •

200-

120-

40-

tz---:-.-Z— --—-

'J-

--.--'-•

Fig. 10 Human mortality rates per 1,000 population. Age/specific mean values for N. and S. Ogaden.

Discussion As evident from the foregging presentation the four different surveys are not strictly comparable in all respects. Thus, the first survey was carried out by one group of people, the following three surveys by another group. Also the techniques used were in certain respects different. In the first survey helicopter transportation could be used for the teams, in the others only landrovers were at disposal. Surveys among nomads always involve special problems.4 ' This is particularly true in periods of" food shortage. Migration patterns are changed, often drastically so. The creation of special camps for feeding programmes often means an abandoning of time honoured temporary settlements. It happened ever so frequently that the sampling made up in Addis Ababa on the basis of previous experience had little applicability in the field. The various factors just mentioned, with considerable risk of biased sampling, imply that only marked differences in results from one survey .to another should be looked upon as indicative of true change. Rainfall data were obtained from only two stations, Kebri Dehar and Gode (Fig. 5). Unfortunately no recordings are available from Northern Ogaden. The nearest rainfall measuring equipments were located in the town of Jijiga which is too close to the highland areas to provide a representative rainfall pattern. As rainfall normally varies considerably from month to month throughout the year, the observations made have been compared to what might reasonably be expected on the basis of average recordings over the 34

previous thirteen years (1957-70). From a technical view point these averages are not entirely reliable as a minimum recording period of 25 years is required for this purpose. Further, the variation in rainfall from area to area is not adequately reflected in the recordings from two stations. Nevertheless it has been possible to make a number of significant observations. It is evident that the disturbance of rains in the Ogaden date at least two years back (Fig. 5 and Table I) and that the deficit and/or delay in rainfall shown must have been a major contributory factor to the area's problems. Undoubtedly also a number of other factors such as overpopulation by both animals and humans compounded the tragedy but with the wisdom of hindsight the reduction in observed precipitation which even in normal years rarely exceeds 300mm should have been warning enough of what was to come. Animals are used for milk and meat but also as a very essential means to procure other foods. In Ogaden 50% of the energy foods are bought as grains in exchange for animals. Estimates of herd size are prone to many errors. Pastoralists anywhere are reluctant to report accurately. Where possible herd counts have been made but some clans practice split herd grazing so that only part of the herd can be counted near an encampment. In time of hardship further reporting bias is likely to be introduced. The most obvious and most serious losses have occurred among the cattle although sheep and goats have also been severely affected. If the FebruaryMarch 1975 level is ignored, the camel losses have been much less extensive; this, of course, is to be expected. The increase in animal numbers observed in the Northern Ogaden during the November-December 1974 survey may be explained by a traditional but exaggerated northward migration which occurred between May-June and November-December 1974. Regrettably many of these animals were lost due to a combination of over-grazing and a failure of the November-December 1974 rains, while congregated in river valleys just south of the highlands and also in the east. The changes in animal holdings observed since that time reflect, not only the effects of the drought, but probably other factors as well. It is clearly not possible that between February-March and June-August 1975 herd size could have increased to the extent indicated by normal regenerative processes. The low figure reported in February-March was probably a result of under-reporting, of migration by men and older boys out of the Ogaden with most of the remaining herd and of a poor selection of the sample. At best these figures should be interpreted as indicating the number of animals left with remnants of families in the Ogaden. Many sought and found food from the government relief programmes. Since February-March 1975 there has been an observable migration back into the area (as is normal after the main rains in April and May). Environmental Child Health, February 1977

MEHARI GEBRE-MEDHIN, ROGER HAY, YAICOB LICKE AND MARIO MAFFI

Losses may also be partly accounted for by off-take for marketing. Little is known about the changes in offtake rates during times of scarcity. It is probable that this is a very complex matter dictated by pastoralist expectations, the level of hardship, the fall-off in milk production, the rate of animal mortality and the prices being, offered. In June-August 1975 the average livestock holding per family was still well below the minimum required for survival, i.e. 7.5 units. Fig. 6. On the basis of this information a number of years will elapse before the herd size will attain proportions required to support the present population of the area, should they retain a pastoral way of life. Because exchange of animals for grain is an important part of the pastoralists economy the prices paid for grain and fetched by animals in local markets have been monitored during 12 months. Prices in these markets respond as usual to supply and demand but are also influenced by other factors. The supply of grain in the Ogaden is largely dependent on the adequacy of the harvest in the Hararghe highlands. This is, of course, affected by climatic and other conditions. Transport costs tend to add to grain prices the further south pne goes. The demand for grain, being a staple food, is probably fairly constant as long as purchasing power (i.e. animals) exists. The supply of animals is presumably directly related to holding size, fertility rattfs and to pasture condition. The demand for animals appears complex but is at least influenced to some extent by the condition of animals.' In times of severe scarcity, the market appears to become distorted by extraneous factors such as merchants capitalizing on an imbalance between supply and demand, pastoralists deciding not to sell animals at any price and merchants offering ludicrously low prices to pastoralists desperate for food. This complex of factors, added to a reluctance of both buyers and sellers to declare a true market value for commodities makes the interpretation of price changes somewhat difficult. Finally there comes a point of extreme scarcity where trading ceases. Close to this point prices become almost meaningless. • We have no information regarding the balance of effect on demand caused by reduced purchasing due to scarcity of animals and low price on one hand and shortage of animals products for food on the other. There is also no information available regarding the flow of animals through local markets but we would suspect that this in June-August 1975 has not yet achieved normal levels. It is probable, therefore, that the rise in price during the first six months of 1975 is as much due to excessive demand over supply as due to incomplete recovery of the animal economy. This indicates a disruption of food supply but in addition it should be noted that throughout, the pastoralists have operated in an adverse market situation — low animal prices, high grain prices — this adding to their hardship. Environmental Child Health, February 1977

The key issue in monitoring the situation during famine relates to the nutritional status of the population. For the purpose of this report we have chosen one indicator of nutritional status — "weight for height". We have done this for three reasons: it is a simple method, it is independent of age (which is difficult to determine accurately) and it expresses the process of body wasting seen in acute and extreme food shortage" such as has been seen in the Ogaden. It is also appropriate as the predominant type of malnutrition observed is not kwashiorkor' but marasmus. The first survey, in May-June 1974, indicated that the "nutritional status was poor but not exceptional for an Ethiopian population".5 A significant deterioration in nutritional status was observed in the second half of 1974. It is probable that this is partly accounted for by the lean time brought about by the long dry period between the main rains — an observation which has been shown to apply to many rural communities.10 In this case, however, the proportion of children below 80% weight/height standard doubled between MayJune and November-December 1974. In the south there was a further deterioration between NovemberDecember 1974 and February-March 1975. Fig. 9. This is more than can be attributed to seasonal variability characteristic of pre-drought times. The recovery shown in the June-August 1975 results could very well be an expression of increased animal production as a result of the April-May rains. The effect of the government relief operation may also have contributed but it should be noted that the measurements obtained for the June-August 1975 report did not come from relief centres but from settlements and encampments in the interior. The recovery in the south has been less dramatic than in the north where mean nutritional status in JuneAugust 1975 was significantly better than at the same time in 1974. Despite these encouraging results two points should be noted. Table III shows that in the south the proportion of children below 80% weight/height standard is twice as great as at the same time in 1974. In addition the mode point both in the north and the south still lies between 80 and 89% of standard which is too low for unqualified complacency. These data indicate that while there has been a general and encouraging improvement in nutritional status between February-March and June-August 1975 a significant proportion of the population at large still show signs of serious malnutrition. It should also be remembered that June is probably the most abundant time of the year in the Ogaden. The population of the south have not recovered their 1974 condition. There has been further depletion of animal resources in the entire area. A considerable number of people are now dependant on relief grain. Thus, if the population was vulnerable to disturbance of end-year 35

MEHARI GEBRE-MEDHIN, ROGER HAY, YAICOB LICKE A N D MARIO MAFFI

rains in 1974, it was equally or more so in 1975.* Reliable baseline figures for the Ogaden are lacking. Available information on age specific mortality rates for Harerghe are probably more relevant to the highland parts of the region. The mortality rates observed in the present study were excessively high as far as -under-five children are concerned, and still very high also in older children and, to a lesser extent, in. adults. There is little reason to doubt that during 1973-75 famine has taken a high death toll also in the whole of the Ogaden area. In evaluating the effect of famine it must always be kept in mind, however, that also in "normal" times the losses of human lives due to malnutrition and other preventable disease in a country like Ethiopia are enormous, a figure of 500,000 has been given.8 In this paper we have not discussed the food distribution system adopted for relief purposes, first in Wollo and then in the Ogaden area. In principle it follows similar lines as those used earlier in e.g. West Africa.1 Finally, a fundamental question: has the surveillance in Ogaden in 1974-75 permitted a prediction of forthcoming events and been able to "blow the Clarion "horn" in time for disaster relief when so needed? The answer seems to be in the positive. It is true that the results of the first survey may have conveyed an unduly optimistic picture. It seems quite clear, however, that the relief work instituted in the spring of 1975 would have been both later in appearance and less massive had not the deterioration from November-December 1974 to February-March 1975 given such a striking evidence of impending crisis. It seems therefore that the "early warning system" which is in the making has proven its value in the field. Summary Based on the principles worked out in Ethiopia by an Interministerial Technical Working Group concerning a Consolidated Food and Nutrition Information System repeated surveillance was carried out in the Ogaden area in SE Ethiopia. Data were obtained on 'Addendum Again in October-December 1975 there was in Ogaden severe rain failure, "by 75%". Thus the situation is once again deteriorating in an already severely affected population. And the influx to the camps has augmented anew.

36

four different occasions between June 1974 and JuneAugust 1975. They related to rainfall, livestock holdings, price indices, human nutritional status and mortality. Most of the data were collected separately for the northern and the sourthern part of Ogaden. The various recordings indicated a marked deterioration of the situation from June 1974 to February-March 1975. As a consequence coordinated and large-scale relief measures were inagurated. The considerable improvement noted with respect to all indicators in June-August 1975 could be interpreted as the combined effect of relief measures and delayed but quantitatively normal rainfall. Acknowledgements This survey was carried out at the suggestion of the Interministerial Technical Working Group (TWG). The authors are greatly indebted to all members of TWG for valuable criticism during the preparation and execution of the survey. The considerable amount of data presented here could not have been collected without the interest and collaboration of all members of our field staff. Their contribution is gratefully acknowledged. Special thanks are due to Mrs. Abeba Gobezie for useful comments and Mr. Leif Andersson for providing administrative support. We are grateful to Prof. Bo Vahlquist for positive criticism and for providing excellent working conditions during the preparation of this document. We also wish to thank Mrs. Karin Wennqvist who typed the manuscript.

References 1. Aall, C , and Helsing, E. (1974). / . Trop. Pediat. Environ. Child Health, 20,304. 2. Belete, S., Gebre-Medhin, M., Hailemariam, B., Mam, M., Vahlquist, B., and Wolde-Gebriel, Z. (1976). Courrier 26,113. 3. Gebre-Medhin, M. (1976). J. Trop. Pediat. Environ. Child Health 22,11. 4. Greene, M. H. (1974). Lancet. June 1, 1093. 5. Holt, J. F. J., Seaman, J., and Rivers, J. P. W. (1975). Proc. Nutr. Soc, p. 115A. 6. Holt, J., Seaman, J., and Rivers, J. (1975). Nature, 255, 180. 7. Jelliffe, D. B. (1966). The assessment of the nutritional status of the community. WHO Monogr. Ser. No. 53. 8. Miller, D. S., and Holt, J. F. J. (1975). Proc. Nutr. Soc, 34, 167. 9. Seaman, J., Holt, J., Rivers, J., and Murlis, J. (1973). Lancet, October 6, 774. 10. Selinus, R., Gobezie, A., Knutsson, K. E., and Vahlquist, B. (1971). Am.J. Clin. Nutr., 24,365. 11. Waterlow, J. C , and Rutishauser, I. H. E. (1974). In: Cravioto, J., Hambraeus, L., and Vahlquist, B. (Eds.): Early malnutrition .and mental development. Symposia of the Swedish Nutrition Foundation XII. Uppsala, Almqvist & Wiksell, p. 13.

"
