Eastern Africa is susceptible to recurrent severe droughts which threaten millions of people throughout the sub-region with famine and starvation. Drought is not simply a result of climatic change, but a phenomenon in which human activities play a growing role. Deforestation, inappropriate agricultural and pastoral policies and practices, civil wars and physical insecurity increasingly threaten food security in Africa, and throughout the world.

The shared concern of Djibouti, Ethiopia, Kenya, Somalia, the Sudan and Uganda on the effects of such disasters led to the establishment in 1986 of the Intergovernmental Authority on Drought and Development (IGADD). This body coordinated sub-regional efforts to combat the effects of drought and other natural disasters. Eritrea joined IGADD in 1993 and, in 1995, a revitalized and expanded authority was renamed the Intergovernmental Authority on Development (IGAD).

The original IGADD Plan of Action recommended the establishment or strengthening of national early warning systems for food security and the establishment of a sub-regional food security information system to coordinate national, sub-regional, regional and international efforts in this field.

Although some IGADD countries had already established National Early Warning Units (NEWUs), they suffered from a number of weaknesses and constraints. Valuable information generated by different agencies was not always utilized effectively, while other information was incomplete or unavailable at the right time. The lack of constant, guaranteed and adequate funding, resulting in unsatisfactory equipment and staffing levels, threatened the sustainability of the units, while the officers assigned to the NEWUs frequently lacked the knowledge and skills required to contribute to a functioning Early Warning and Food Information System (EWFIS). Furthermore, methodologies, tools and techniques were often inadequate or poorly integrated, threatening the reliability and timeliness of food assessments.

Awareness of these weakness led to the proposal, formulation and eventual implementation of the project GCPS/RAF/256/ITA, "IGADD Early Warning and Food Information System for Food Security". The project was designed to provide each Member State with the opportunity to strengthen its national EWFIS by introducing further refinements. These would include the interpretation and incorporation of remote sensing data, the development of pre-famine nutritional indicators, the examination of new techniques for early warning activities and the development of methodologies, including simple computer-based models, to identify areas likely to suffer crop failure and any other factors that might threaten food security at the household level.

1.2 OUTLINE OF OFFICIAL ARRANGEMENTS

The agreement for project GCPS/RAF/256/ITA "IGAD Early Warning and Food Information System for Food Security" was signed on 23 January 1989 by IGADD, the Government of Italy and FAO. The project had an initial duration of three years and a budget of $US 6 023 352. Field activity began with a consultancy mission in September 1989, but owing to initial funding constraints, the project became fully operational in January 1991. In 1992 it was agreed to extend the project until the end of 1993. A further extension of six months, until June 1994, was made on the basis of a recommendation of the third Tripartite Review. This extension would serve as a bridge to a Phase II project. Several other short extensions were approved to ensure continuity, bringing the termination of the project to February 1996 and the final budget to $US 6 590 841.

1.3. PROJECT OBJECTIVES

The original long-term objective of the project was to strengthen food security in the IGADD countries by the establishment of a food information system incorporating existing and relevant institutions and mechanisms. The project was expected to provide advance information on crop and animal production and food supply and to alert those concerned well in advance of impending food shortage and surplus so that timely and remedial action could be taken as appropriate.

Six immediate objectives were defined. It was decided to create an IGADD sub-regional EWFIS within the IGADD Secretariat and to provide information concerning the sub-regional food security situation to the IGADD Secretariat and Member States in the form of regular and ad hoc bulletins and reports. The project would organize the reception of a steady flow of information from national early warning systems, which would serve as a basis for the regular assessment of the situation. The project would also advise and assist the national units as and when required. It would help to evaluate the food supply situation in Djibouti and make recommendations on the need for a national unit. Finally, it would promote the concept of food security in each member country and the need for sub-regional cooperation.

The second Tripartite Review meeting in July 1992 established a new long-term objective: to achieve food security in the IGADD countries by ensuring that practically all people have physical and financial access to adequate food of appropriate nutritional quality throughout the year. The meeting also established an intermediate objective: that national and regional early warning and food information systems operate on a self-sustained basis in the IGADD sub-region by 1997.

The original six immediate objectives were reduced to two and reformulated. The first new objective was to increase significantly the capacity of the NEWUs and some related national institutions to collate, process, analyse and disseminate early warning data and information The second was to strengthen the capacity of the IGADD Secretariat to carry out regional EWFIS activities for food security. Both these objectives were to be achieved by the end of 1993, allowing the intermediate objective to be met within the time limit established.

The expected outputs of the project were as follows. It hoped to test and adopt a fully documented early warning methodology at regional and national level and to train staff to run the NEWUs and the Regional Early Warning Unit (REWU). It also planned to strengthen the EWFIS communication network among Member States and to install the first level of a modular NEWU-REWU communication system. It intended to produce databases on agricultural statistics, agrometeorology, prices, socio-economic aspects, nutrition and remote sensing in IGADD and Member States and to publish early warning newsletters and ad hoc food security reports. Finally, it aimed at producing recommendations for further work on the EWFIS in IGADD countries, to be carried out during a second phase project.

2. RESULTS AND CONCLUSIONS

2.1 DEVELOPMENT OF AN EARLY WARNING METHODOLOGY

Developing, testing and installing early warning methodologies, tools and techniques, including computerized data systems and personnel training, at both national and regional levels was a major activity of the project. The methodologies and software developed, known as the ADDATI package, have been widely adopted at regional level and to varying degrees at national level.

The project contributed significantly to developing early warning methodologies throughout the region. The systematic comparison of actual and expected values of key indicators, the use of statistical clustering techniques to classify land units on the basis of these indicators and the merging of classifications for a range of indicators to produce vulnerability and risk maps are valuable diagnostic techniques in early warning methodologies.

Since the methodologies and software tools were developed concurrently, it was impossible to explain the underlying methodological concepts independently of the software itself. As a result, the methodology was often neglected in favour of aspects of the software. This had a negative effect on the adoption of the methodology in some Member States. Where the methodology was understood, however, it was generally well received.

Many senior NEWU technical staff were computer-illiterate and tended to leave computing tasks to subordinate staff, thus hampering full understanding of project methodologies. In some NEWUs, staff were not encouraged to adopt the new methodologies in their daily activities. Some staff, however, have acquired a good understanding of the package.

2.2 REGIONAL AND NATIONAL STAFF TRAINING

Training was regarded as a continuous process, owing to the need to introduce new technologies and to the very high turnover of national staff. However, the project had neither the time nor the resources to provide training to everyone involved in early warning systems in all Member States. As a result, the project input was principally that of training future trainers, enabling those being trained to transfer their newly acquired knowledge and skills to others.

The training provided was practice-oriented and varied according to the subject, the number and level of participants and the teaching aids available. Most training was in-house training at national and regional level, although a few selected officers were trained at the project office.

Three levels of training were adopted. The first, at national level, provided basic computer skills. The second offered specialized training on the different early warning components through in-country workshops and backstopping from Djibouti, while the third provided training on the interpretation of different types of data.

In all cases, training was related to the specific job the trainee was doing or was expected to do in the EWFIS, although all activities also contained a general introduction component to EWFIS and IGADD as a whole.

The project provided background, basic and high-level training and backstopping to more than 140 people in the region. It monitored trainee performance and adjusted training activities on the basis of modifications in NEWU staff composition. The project also produced a substantial number of training manuals and background documentation (see Appendix 4).

Despite being appreciated by the Member States, the results of training were disappointing. Four main factors impeded the full achievement of this output. These were the high turnover of personnel trained by the project, the frequent lack of collaboration from NEWU heads in transferring knowledge from trained personnel to other staff, the constant loss of training manuals and technical documentation provided by the project and, in particular, the appropriation by NEWU heads of laptops intended to facilitate the daily training and activities of personnel.

Training of regional staff took place in Djibouti. By the end of the project a core of competent regional experts, able to provide effective support for the development and operation of the IGADD regional EWFIS, was in place.

2.3 DEVELOPING A COMMUNICATION NETWORK FOR EWFIS

The establishment and improvement of national EWFIS communications networks was regarded as crucial. The project established a simple, free, computer-assisted communication system linking NEWUs with the REWU. Fully operational internal national communication networks were improved or established in all Member States except Somalia and Djibouti.

In order to conserve scarce resources, access to these networks was given to institutions concerned with early warning and food security. Member States were encouraged to monitor and maintain the networks to ensure the availability of timely information.

A short wave computer-assisted communication system was chosen to link IGADD headquarters with the IGADD focal points in Member States. The project successfully installed short wave computer-assisted links in Djibouti, Khartoum, Addis Ababa, Kampala, Asmara and Nairobi. The Communication Authority refused authorization in Ethiopia and the situation in Somalia remains unpredictable. However, although the system was highly appreciated by the Secretariat and Member States, the recent advent of electronic mail is likely to render it obsolete in most circumstances.

2.4 DEVELOPMENT OF DATABASES

EWFIS monitoring depends on good-quality databases. The project made considerable progress in developing computerized databases for both historical and current data, agricultural statistics, agrometeorology, remote sensing, food prices, nutrition and other socio-economic parameters for each Member State. It also provided assistance in data entry, cleaning and updating sectoral databases to meet the needs of NEWUs and related institutions. These databases are frequently used by NEWUs for early warning monitoring.

Socio-economic district profiles, covering the smallest administrative areas of the region, were set up and regularly updated by the REWU in order to assist NEWU assessments.

A new regional database of crop production system zones was established. This covered more than 1 200 cropping units and contained information on physical and agronomic factors, livestock, climatic hazard and pests and diseases. It represents a basic element for any future crop-related analysis in the sub-region, not necessarily connected with EWFIS activities.

2.5 EARLY WARNING NEWSLETTERS AND REPORTS

The project began to produce regular Early Warning Newsletters and Food Security Reports in September 1991. Since July 1992, this activity has been the responsibility of the IGADD. However, owing to data acquisition problems, the production of these documents continues to be problematic and their coverage only partial. In addition, there is no evidence that the bulletins are being used by Member State Governments, which have provided virtually no feedback.

2.6. FUTURE SUPPORT FOR EARLY WARNING ACTIVITIES

The project assisted the IGADD Secretariat by providing suggestions and technical assistance for follow-up activities. This included the formulation of full project documents to support the sub-regional EWFIS and separate documents for each national EWFIS with the exception of Somalia, which would be operated from the REWU in Djibouti. These documents were intended to facilitate discussions with interested donors.

2.7. CONSTRAINTS TO PROJECT IMPLEMENTATION

Some time after the project started, inconsistencies in the original Project Document were noted. These had a negative effect on activities and, although a new framework was approved by the Tripartite Review Meeting in 1992, it came too late to allow the activities originally envisaged to be accomplished.

Despite sensitization activities at different levels, some countries remained reluctant to accept regional coordination and standardization procedures and to exchange information with other Member States. Although the original IGADD decision to implement a regional system was appropriate, the situation at national level made it far too optimistic.

The project was hampered by the lack of a clearly defined institutional framework at the IGADD Secretariat and of a written set of working guidelines and procedures. By the end of the project, the Secretariat had still not fully developed its internal technical capability to carry out EWFIS activities at a regional level. The contract between FAO and the IGADD to provide the services of three regional experts also constrained the smooth day-to-day running of the project.

3. RECOMMENDATIONS

The IGADD Secretariat should develop a wider understanding of regional EWFIS activities and increase its technical authority in the sub-region.

During the project, IGADD EWFIS emphasis was on early warning. Future efforts should be expanded to include the broader range of food security. IGADD EWFIS staff should work to develop the broader food security approach in full collaboration with the environmental monitoring activities developed at IGADD Headquarters.

The IGADD Secretariat should create and fill permanent IGADD EWFIS posts as soon as possible, in order to transfer responsibilities from the project to the Secretariat. In order to preserve the knowledge gained during the project, these posts should ideally be filled by the project's regional experts.

The regional aspects of EWFIS activities should be maintained and strengthened. In order to do this, the IGADD Secretariat must build its own EWFIS technical capability, by developing and implementing a plan to reduce its dependence on donor assistance.

The strength of the regional EWFIS largely depends on the relative strengths of the national EWFISs. Although the IGADD Secretariat has made a commendable effort to secure future support, most Member States still need direct technical and financial support for their national systems. The IGADD should continue to seek such assistance in order to help ensure food security and the future viability of the regional EWFIS.

Appendix 1

PROJECT STAFF

Dates of Service

Name Function Starting Date Concluding Date

International

P. Santacroce Chief Technical Adviser 24 Nov. 1989 31 Dec. 1993

A.M. Comin Administrative Officer 22 June 1990 31 Dec. 1993

A. Castermans Associate Professional Officer 24 Jan. 1993 23 Jan. 1994

L. Verelst Agrometeorologist 24 April 1993 23 April 1994

Regional

J. Mwikya Agrometeorologist 9 July 1990 28 Feb. 1993

S.M. Giama Nutritionist 15 Nov. 1990 14 Nov. 1991

S. Zziwa Agricultural Economist 10 Jan. 1991 28 Feb. 1995

E. Tekle Nutritionist 28 Jan. 1992 31 Oct. 1995

A. Zahran Agrometeorologist 1 Nov. 1993 31 June 1995

Appendix 2

PROJECT TRAINING ACTIVITIES

National workshops

No. of

Focus Participants Location Date

EWFIS concepts 7 Entebbe July 1991

3 Khartoum Aug. 1991

2 Addis Ababa Sept. 1991

ADDATI 1 4 Entebbe Sept-Oct. 1991

Project software 3 Entebbe Oct. 1991

ADDATI 2 3 Entebbe Jan. 1992

ADDATI 1 and 2 11 Addis Ababa Feb. 1992

12 Khartoum Feb. 1992

Project software 1 13 Khartoum March 1992

Data entry 2 Entebbe March 1992

Project software 10 Khartoum June-July 1992

ADDATI 2 8 Nairobi Oct. 1992

EWFIS orientation 6 Entebbe Oct. 1992

ADDATI 1 and 2 5 Entebbe Nov.-Dec. 1992

ADATTI 3 5 Nairobi Nov. 1992

Agrometeorology 2 8 Khartoum Jan. 1993

Bulletin preparation 1 4 Nairobi Feb. 1993

Bulletin preparation 2 4 Nairobi May 1993

No. of

Focus Participants Location Date

Project software 4 Khartoum May 1993

Agrometeorology 9 Asmara May-June 1993

EWFIS orientation 10 Asmara June-July 1993

Nutrition 26 Kisumu Aug. 1993

34 Nakuru Aug. 1993

Regional workshops

EWFIS methodology 3 Djibouti Nov. 1990

Nutrition in EWFIS 11 Nairobi Oct. 1991

Orientation to EWFIS 6 Djibouti Dec. 1991

Market price data 1 12 Addis Ababa March-April 1992

Agrometeorology 1 10 Kampala June 1992

CCD remote sensing 7 Khartoum Sept. 1992

EWFIS methodology

evaluation 8 Kampala Dec. 1992

Agrometeorology 2 5 Khartoum Jan. 1993

Market price data 2 11 Kampala June 1993

Nutrition and socio-

economic data 19 Kampala June 1993

Crop monitoring 14 Kampala Sept. 1993

Appendix 3

MAJOR ITEMS OF EQUIPMENT PROVIDED

Cost

Quantity Item ($ US)

2 Photocopier, Canon NP3825 16 187

2 Computer, Zenith 286 7 561

3 Laser printer, Bull 10 803

3 Communication system, Nestor 56 310

3 Comunication system, Mestor 58 195

1 Fax machine, Murata 1 384

12 Vehicle, Toyota Hilux 159 796

1 Photocopier, Canon NP 1010 2 401

2 Computer table 1 341

1 Computer, Matco BM 75 9 758

1 Stabilizer, Bull 1 902

1 Computer, Matco BM 200 6 321

5 Computer, Matco BM 286 22 430

4 Printer, Bull 4202 4 747

1 Vehicle, Toyota Land Cruiser 25 326

1 Vehicle, Suzuki SJ431 11 661

1 Computer, EXDR BM 286 3 736

1 Stabilizer, Merlin Gerin 7 327

1 Optical disk 3 203

2 Telefax, Canon 120 2 825

1 Typewriter, ET 2400 1 416

1 Colour monitor, Zenith HR 1 064

2 Printer, portable, BJ-10E 1 039

1 Projector, Datashaw 1 700

1 Computer, SMJ 286 3 512

1 Computer, portable, Toshiba T1200 3 175

1 Printer, HP Paintjet 2 198

3 Air-conditioner, Amana 1 390

1 Computer, portable, Toshiba T2000 3 629

3 Printer, HP Laserjet III 5 776

1 Computer, Zenith Z386 3 375

1 Air-conditioner, split 1 553

3 Computer, T4400SX Notebook 10 273

19 Codan SSB X-2 and accessories 54 269

1 Computer, T1900 Notebook 2 538

Cost

Quantity Item ($ US)

3 Printer, HP Laserjet 6 080

1 Computer, IBM PS/VP638 5 418

9 Fax, Panasonic UF-127 5 995

2 Printer, HP Laserjet 3 582

9 Computer, portable, Zenith 21 474

1 Computer, IBM PS/2 55SX 3 601

2 Printer, HP Laserjet TV 3 620

15 SG-712EX SSB and accessories 35 295

2 Photocopier, Canon NP 3325 6 330

8 Telcop RX 707 9 750

1 Computer, IBM PS/2 3 429

1 Computer, Zenith 2-320SX 3 512

18 Motorcycle, Yamaha AG100 23 799

16 Tansc TW2000 SSB 36 960

1 Photocopier, RX 5030 6 778

8 Motorcycle, Yamaha 12 453

1 Photocopier, Rank Xerox 4 295

6 Computer, Notebook T1800 10 160

1 Duplicator, Gestet 431 1 640

2 Scanner, Gestet 3 193

2 Air-conditioner, Westhouse 1 793

1 Photocopier, Sharp SF 2027 2 972

3 Computer, Winner 386SX 18 030

1 Photocopier, Sharp SF-7850 1 932

1 Photocopier, Sharp SF-7370 1 728

2 PK 2075 UPS 1 016

1 Computer, AST 386SX 4 200

20 Codan FC SSB and accessories 72 183

1 Printer, laser, Wang 1 620

3 Printer, Epson FX-1170 2 122

3 Computer, ADS 386X/25 11 310

1 Computer, Wang 250/16 4 560

1 Typewriter, electric, IBM 1 000

1 Photocopier, Canon NP1215 2 005

7 Fax, Canon 80 3 553

1 Computer, Wang 80386SX 3 620

1 Photocopier, RX 5014 4 023

Appendix 4

DOCUMENTS PREPARED DURING THE PROJECT

A4.1 MANUALS AND METHODOLOGIES

A short guide to software and standard procedures to IGADD EWFIS publications. Revised February 1991.

A guide to the use of CARTE. Revised November 1993.

A quick reference to project filenames convention. Revised December 1993.

User manual on Lotus macro routines for EWFIS. February 1991.

District profiles files, description of contents. Revised December 1993.

FAOINDEX, crop specific soil water balance, version 2.1 adapted for IGADD. March 1992.

A proposal for a technique of area averaging. Revised December 1993.

ADDATI, early warning version (3.1), tutorial. April 1993.

ADDATI, early warning version (3.1), reference manual. November 1992.

Introductory guide to ADDATI, early warning version (3.1). November 1992.

A short training guide to the use of SURFER. Revised December 1993.

A short guide to packages and procedures for Agromet analysis. Revised December 1993.

ABC, early warning version (1.2). December 1993.

Training manual on use of price data for early warning, version 1. December 1993.

A4.2 WORKING PAPERS

NDVI processing methodology used by the project, an application to Ethiopia during the last three years. February 1991.

Description of standard procedures for detecting probable stress areas. April 1991 (under revision).

A4.3 STATISTICAL PAPERS

IGADD long-term demographic estimates and projections. October 1991.

IGADD, crops, livestock and fishery products. October 1991.

IGADD, food aid by product and donors. October 1991.

A4.4 MAPS AND GRAPHICS

Overall NDVI average of IGADD countries. September 1991.

Definition of growing periods using NDVI information (IGADD countries at district level). December 1991.

A4.5 PROCEEDINGS

A role for nutritional and socio-economic information in early warning activities. February 1992.

A role for nutritional and socio-economic information in early warning activities. Background papers. February 1992.

Collection and utilization of nutritional and socio-economic data for early warning and food information. July 1993.

A4.6 OTHER PUBLICATIONS

Project exhibition, contents of panels. September 1992.

Technical documentation on regional databases. May 1993.

Crop production system zones (CPSZ), national matrices, Vols I-VIII.

My country. Case study. November 1992.

Introduction to the IGADD EW methodology. November 1993.

Catalogue of documents. Revised October 1993.

EWFIS newsletters (September 1991-October 1993).

EWFIS - Bulletin d'information (in French). September and December 1991 and April 1992.

Food situation reports (September 1992 and March 1993).

Agromet bulletin. June 1993.

A4.7 SOFTWARE

CARTE, version 3.1, distribution diskette.

FAOINDEX, version 2.1F, distribution diskette.

Area averaging, example diskette.

ADDATI, early warning version 3.1, tutorial, sample file diskette and installation diskette.

Introductory guide to ADDATI, early warning version 3.1, case studies, application diskette.

A guide to the use of Surfer, distribution diskette.

Software for Agromet analysis, 2 distribution diskettes.

ABC, distribution diskette.

Price data for early warning, distribution diskette.

IGAD EARLY WARNING AND FOOD INFORMATION SYSTEM FOR FOOD SECURITY PROJECT FINDINGS AND RECOMMENDATIONS

Table of Contents

LIST OF ABBREVIATIONS

1. INTRODUCTION

1.1 PROJECT BACKGROUND