Tunisia - Sfax Flood Protection Project

Documnt of The World Bank FOR OMcIAL USE ONLY Repot No. 12919 PERFORMANCE AUDIT REPORT TUNISIA SFAX FLOOD PROTECTION PROJECT (LOAN 2289-TUN) MARCH 30, 1994 Operations Evaluation Department This document has a restricted distribution and may be used by recipients only i the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization. EXCHANGE RATES Appraisal Year Average (1983) USD1 = TND .63 Intervening Years Average USD1 = TND .82 Completion Year Average (1989) USD1 = TND .92 GLOSSARY AfDB - African Development Bank CSM - National Tender Board (Commission Supdrieure des Marchds) DHU - Urban Hydraulics Directorate (Direction de l'Hydraulique Urbaine) DRME - Regional Directorate of the Ministry of Equipment (Director R6gionale du Ministbre de l'Equipement) ERR - Economic Rate of Return ICB - International Competitive Bidding OED - Operations Evaluation Department ONAS - Office National de l'Assainissement PAR - Performance Audit Report PCR - Project Completion Report FISCAL YEAR January 1 - December 31 FOR OFFICIAL USE ONLY THE WORLD BANK Washington, D.C. 20433 U.S.A. Office of Director-General Operations Evaluation March 30, 1994 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Performance Audit Report on Tunisia Sfax Flood Protection Project (Loan 2289-TUN) Attached is the "Performance Audit Report on Tunisia - Sfax Flood Protection Project (Loan 2289-TUN)" prepared by the Operations Evaluation Department. The project was a rapid Bank response to the urgent request of the Tunisian government for assistance in the design and construction of flood protection works for the city of Sfax, Tunisia's second largest city, following devastating floods in October 1982 in which 70 lives were lost and heavy property damage was sustained. The expeditious processing (government request was received in November 1982 and appraisal took place in January 1983) and subsequent implementation was only marred by a one year delay in effectiveness due to slow legal clearances by the government. By their nature, the effectiveness of the flood protection works can not be tested at will (they were designed for a flood of 100 year occurrence) but the well constructed and maintained facilities clearly increased the safety and confidence of the city's population as evidenced by the new developments in the previously flood damaged areas. The performance of the project is rated as satisfactory and its sustainability as likely. Institutional development was substantial. Attachment This document has a restricted distribution and may be used by recipients only in the performance of their officiaL duties. Its contents may not otherwise be discLosed without WorLd Bank authorization.  FOR OFFICIAL USE ONLY PERFORMANCE AUDIT REPORT TUNISIA SFAX FLOOD PROTECTION PROJECT (LOAN 2289-TUN) TABLE OF CONTENTS Page No. PREFACE ........................................ i BASIC DATA SHEET ................................ ii EVALUATION SUMMARY ............................ iv I. BACKGROUND ................................ 1 Flood Protection and Storm Drain ...................... 1 General ....................................... 1 The Sfax Region ................................. 1 Sectoral Responsibilities ............................. 2 II. THE PROJECT ................................. 2 Design and Objectives .............................. 2 Implementation .................................. 3 III. THE RESULTS ................................. 4 IV. ENVIRONMENTAL ISSUES ........................ 6 V. CONCLUSIONS AND LESSONS ..................... 7 Conclusions .................................... 7 Lessons Learned ................................. 8 MAP - IBRD 17067R This report was prepared by Messrs. Jozsef B. Bilky (Lead Evaluation Officer) and Jean- Francois Landeau (Senior Evaluation Officer), who audited the project in April 1993. This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization. 1 i PERFORMANCE AUDIT REPORT TUNISIA SFAX FLOOD PROTECTION PROJECT (LOAN 2289-TUN) PREFACE 1. This report presents the results of the Performance Audit of the Sfax Flood Protection project for which Loan 2289-TUN for US$25 million was approved on May 24, 1983. The loan was made to the Government of Tunisia. 2. The Loan Agreement was signed on June 6, 1983 and the loan became effective, with a one year delay, on June 8, 1984. The loan was closed on schedule on June 30, 1989 but the account was kept open until January 17, 1990 at which time US$2.7 million of the loan amount was canceled. 3. The Performance Audit Report (PAR) was prepared by the Operations Evaluation Department. Parts I and III of the Project Completion Report (PCR) were prepared by the then Europe, Middle East and North Africa Regional Office and Part II by the Borrower. The PCR was forwarded to the Board for information on October 15, 1991. OED has reviewed the PCR, the Staff Appraisal and Presidents Report, the legal documents and Bank files and the transcript of the meeting of the Executive Directors at which the loan was approved and discussed the project with the Bank staff involved. An OED mission visited Tunisia in April 1993 and discussed the project experience with officials of the borrower, other involved agencies and beneficiaries. The cooperation and assistance of these officials is gratefully acknowledged. 4. The Audit found that the PCR gave a concise and adequate account of the project experience. The PAR further analyzed cetain aspects of the PCR's findings and provides independent comments and conclusions on various aspects of the project. 5. The draft PAR was sent to the borrower for comments but none were received.  PERFORMANCE AUDIT REPORT TUNISIA SFAX FLOOD PROTECTION PROJECT (LOAN 2289-TUN) BASIC DATA SHEET KEY PROJECT DATA Appraisal Actual or Actual as % Item Expectation Current Estimate of Appraisal Estimate Total Project Costs (US$ million) 47.89 40.33 22 Loan Amount (US$ million) 25.00 22.30 89 Date Physical Components Completed 6/88 6/88 Proportion completed by that date (%) 100 Economic Rate of Return 18 23 CUMULATIVE ESTIMATED AND ACIUAL DISBURSEMENTS FYM FY EYM EYV2 EY EY2 EY Appraisal Estimate (US$ million) 4.87 19.75 42.00 63.50 81.87 96.01 Actual (US$ million) 0.06 1.71 8.18 37.38 68.43 86.51 66.6 Actual as % of Appraisal (%) 1 19 19 59 84 90 66 Date of Final Disbursement: January 17, 1990 PROJECT DATES Original Plan Revisions Actual First Mention in Files Negotiations 04/20/83 04/20/83 Board Approval 05/31/83 05/24/83 05/24/83 Signing 06/06/83 06/06/83 Effectiveness 10/17/83 01/17/84 06/08/84 Closing Date 06/30/89 06/30/89 111 STAFF INPUTS (Staff weeks) EY82 FY83 EYM EYM EYM FY87 [YM FY89 [Y% FY91 Total Preparation 2.8 0.2 0.2 3.2 Appraisal 30.6 30.6 Negotiations 2.3 2.3 Loan Processing 5.5 5.5 Supervision 5.4 4.6 0.7 3.6 3.0 1.8 24.8 PCR 2.6 2.6 Total 41.2 5.4 4.8 0.2 0.7 3.6 3.0 1.8 2.6 69.0 MISSION DATA Month/ No. of Days in Specializations Year Staff Field Represented Ratings Identification 11/82 1 n.a. Division Chief Appraisal 01/83 4 11 Sanitary Engineer 1 Financial Analyst Supervision 10/83 2 13 Sanitary Engineer 1 Financial Analyst Supervision 05/84 1 15 Sanitary Engineer 1 Supervision 12/84 2 15 Sanitary Engineer 1 Financial Analyst Supervision 05/85 3 10 Sanitary Engineer 1 Financial Analyst Supervision 05/86 2 10 Sanitary Engineer 1 Financial Analyst Supervision 09/87 1 13 Sanitary Engineer 1 Supervision 02/88 1 2 Sanitary Engineer 1 Supervision 05/88 2 12 Sanitary Engineer 1 Financial Analyst Supervision 01/88 2 15 Sanitary Engineer 1 Financial Analyst Supervision 11/89 1 15 Sanitary Engineer 1 iv PERFORMANCE AUDIT REPORT TUNISIA SFAX FLOOD PROTECTION PROJECT (LOAN 2289-TUN) EVALUATION SUMMARY BACKGROUND The Project 1. At the time of appraisal (1983) about half of Tunisia's 6.4 million population 3. The Government of Tunisia reacted resided in urban areas. Sfax, Tunisia's quickly to the flood disaster in Sfax and second largest city, located some 310 km requested Bank assistance for the design and south of Tunis had a 1983 population of construction of flood protection facilities. about 313,00. This port city lies at the foot Bank reaction was equally quick and the of a gently sloping plane with sandy soil and project was appraised in January 1983. sparse vegetation. Annual rainfall records range from 37 mm to 650 mm with 4. The objective of the project was to torrential intensities occurring in October, provide flood protection facilities to Sfax the wettest month. The high water table, and neighboring communities which would extensive urbanization and the sparse minimize potential damage at economically vegetation contribute to intense flash floods acceptable cost. The analysis of the cost of of, sometimes, catastrophic proportions. alternative measures and the projected residual damage resulted in the choice of a 2. Since 1934 five major floods have design flood of 100 years frequency. caused severe damage in the city and surrounding areas. In October 1982 a flood 5. The project was approved by the of estimated 150 years frequency caused the Board on May 24, 1983, the loan signed on loss of 70 lives, and destroyed 700 and June 6, 1983 and became effective on June damaged over 8000 houses and disrupted 8, 1984, after a one year delay due to communications. Existing flood protection lengthy legal procedures. It consisted of the measures were minimal and although plans rehabilitation, and raised dike protection of for more systematic protection existed since two natural water courses (Oued Ezzit and the 1950s, they were not constructed due to Oued Maou) bracketing the city north and lack of financing. Flood protection for south, the construction of a semi-circular urban areas is the responsibility of the "belt canal" some 4 km from the city center, Ministry of Equipment (PAR paras. 1-7). minor protection works in adjoining V municipalities, the establishment and materialize, they will remain untested. Nor equipping of a maintenance unit in the can the PCR's recalculated ERR of 23% be Ministry of Equipment's Regional Office verified as it is based on the estimation of and consulting services. the damage avoided during a design flood. The Audit was a.lso unable to verify claims 6. The estimated cost of the project was regarding "multifold" increase in US$47.89 million equivalent, to be financed land/property values as there is no property by a US$25 million Bank loan, disaster taxation in Sfax and the authorities keep no relief grants from the Kingdom of Saudi records of property transactions. Arabia and others and special taxes levied by the government. 10. Nevertheless, visual inspection of previously damaged areas did confirm the 7. The works were substantially increased confidence of developers as completed in mid 1988 and the loan closed demonstrated by substantial new on schedule on June 30, 1989. Final construction in the area. Other visible disbursement took place on January 17, benefits of the project include the new roads 1990 with US$2.7 million of the loan constructed along the belt canal which canceled (PAR paras. 8-10). provide for better traffic circulation (all previous roads are radial) as well as acting Implementation as a "by-pass" for the city center. Flood protection works in outlying areas now 8. Implementation was efficient with permit the re-zoning of protected only (apart from late effectiveness) minor agricultural land for urban development. delays resulting from the necessary realignment of railway lines and water and 11. Negative effects of the project sewerage services. There were no include the accumulation of contaminated procurement problems and consultants and ground water at the northern outlet of the contractors performed well. The final cost belt canal, resulting in unpleasant odors and of the project was US$40.33 million mosquito breeding. The contamination is equivalent, about 16% below estimates in the result of a now removed olive oil US dollar terms. Due to the devaluation of factory, the waste products of which the Tunisian currency against the US dollar, saturated the ground in the area. Further the dinar cost of the project exceeded problems are created by insufficient roadside estimates by about 22%. The Audit mission drainage systems in the city so that heavy confirms the high quality of construction and rains can not be readily drained into the the effective maintenance procedures (PAR canals and standing water remains on the paras. 11-13). roads and housing areas. This problem is clearly the result of insufficient cooperation The Results between the city and the government during the design of the project. 9. The effectiveness of the flood protection works constructed under the 12. The negative effects of the project project can not be tested at will and, should mentioned above should be regarded as the design flood (hopefully) fail to temporary as appropriate remedial measures vi can be found. The resolution of these 15. Lessons from the project confirm problems, at present, are hampered by a that: division of responsibility. The city authorities regard all real or perceived 1. The Bank's efficiency in problems relating to the canals as processing disaster relief type government responsibility. In the Audit's of operations exceeds that in view the maintenance of the flood protection "routine" projects. works and further development of the overall drainage system should be the city's 2. The Bank tends to be responsibility. preoccupied with the implementing agency at the 13. Overall, there is no doubt that the expense of the beneficiaries. project resulted in substantial benefits in 3. In "fast track" operations enhancing urban development in Sfax and important issues can be contributing to the feeling of increased overlooked such as, in this security by the population. The case, the integration of the performance of the project is rated by the flood control works with the Audit as satisfactory and its sustainability as city's drainage system. likely. Institutional development, in the establishment of the efficient maintenance 4. Future projects of this nature unit, has been substantial (PAR paras. 14- should ensure the recording 20). These ratings confirm the ratings based of baseline data to permit on the review of the PCR. impact measurements at a later stage. Conclusions and Lessons 5. In a water scarce country, flood protection measures 14. The project met its objectives in should consider conserving reducing the risk of periodic flood damage flood waters and not just in the city of Sfax and neighboring evacuating. municipalities. In order to better integrate the flood control facilities with the city's overall infrastructure, the city should assume ownership of the facilities and their maintenance. The benefits of the project accrued to rich and poor alike but, as the poor would find it more difficult to recover from flood damage losses, the benefit to them is greater. The negative effects of the project emerged over time and could not have been noted in the PCR (PAR paras. 23 to 28).  1 PERFORMANCE AUDIT REPORT TUNISIA SFAX FLOOD PROTECTION PROJECT (LOAN 2289-TUN) I. BACKGROUND Flood Protection and Storm Drainage General 1.1 Tunisia, located on the southern coast of the Mediterranean sea, occupies about 160,000 sq.kms. The Atlas mountain chain, running east to west, divides the country into two distinct regions. The northern part, a semi-arid area with average annual precipitation of about 600 mm contains most of the arable land. Major urban centers and industries are concentrated in the northern coastal areas. The area, south of the Atlas mountains is arid to desertic with annual precipitation less than 200 mm. At the time of appraisal (1983) more than half the 6.4 million population resided in urban areas with the rural population widely dispersed. Some two third of the urban population was concentrated in Tunis and Sfax, the two largest cities. Localized torrential rainfalls are common occurrence in the country, often causing widespread damage and heavy erosion. The Sfax Region 1.2 Sfax, Tunisia's second largest city is a major port, some 310 km south of Tunis had a 1982 population of 313,000. The growth of the city is above the Tunisian average due to its strategic position and good employment opportunities which presents an attraction to migrants from the hinterland. From its dense coastal center and port area and surrounding low income housing the city is rapidly spreading, in low density housing, into the agricultural land to the west. The surrounding area consists, largely, of mature olive plantations. 1.3 Topographically, Sfax lies at the coastal end of a gently sloping plain with gradients not exceeding 3%, the city itself being virtually flat. The area is semi-arid with sparse vegetation and sandy soil. Although the soil permeability is good, the high water table limits infiltration of flood waters. Exploitation of ground water, near the city, is limited by saline intrusion from the sea. 1.4 The average annual rainfall in Sfax is about 200 mm with extremes of 37 mm and 650 mm having been recorded. October is the wettest month with a recorded maximum of 463 mm. Rainstorms are, generally, localized with intensities up to 80 mm per hour. Rainfall records span about a 100 years. 2 1.5 The intense rainstorms, combined with the sparse vegetation, high water table and the spread of urbanization - resulting in paved roads, destruction of water retaining hedgerows built around large garden plots and the blocking of natural water courses - regularly result in intense flash floods which can reach catastrophic proportions. Five severe floods have been recorded in Sfax between 1934 and 1982 causing serious damage. The flood of October 1982, as a result of an estimated 150 year frequency rainfall, with peak intensity of 45 mm/hour, caused the loss of 70 lives, destroyed 700 houses and damaged 8000 others as well as roads, railways, bridges, power and telecommunication facilities. Excluding loss of life, the damage has been estimated at about US$75 million. 1.6 Prior to 1982, flood protection measures in the Sfax area were minimal. Two dry river beds or "oueds" bracketed the town on the north and the south providing channels for flood waters. The Oued Maou in the south had earth dike levees for added protection but the Oued Ezzit was in a natural state and, in fact, substantially blocked by buildings, trees and debris. The olive groves in the outlying areas had minor earth dikes in a random pattern to retain run- off for irrigation and, incidentally, reduce flood intensity downstream. More systematic flood protection measures have been considered since the 1950s and, in 1978, a sewerage and storm drainage master plan was prepared by consultants but no works were implemented due to lack of financing. Sectoral Responsibilities 1.7 Flood protection for urban areas is the responsibility of the Directorate of Urban Hydraulics (DHU) of the Ministry of Equipment and the maintenance of such works is entrusted to the Ministry's Regional Directorates. The "Office National de l'Assainissement" (ONAS) is the agency responsible for sewerage and storm drainage and the Directorate of National Land Planning prepares national and urban plans. The Ministry of Agriculture manages all water resources. II. THE PROJECT Design and Objectives 2.1 The government of Tunisia reacted quickly to the October 1982 flood of Sfax and, in November 1982, it requested assistance from the Bank for the design and execution of appropriate flood protection facilities. A feasibility study, based on the updated 1978 flood protection master plan, was prepared in January 1983, with Bank assistance, and formed the basis for the appraisal of the Sfax Flood Protection project, in the same month. 2.2 The objective of the project was to provide a degree of protection to Sfax and neighboring communities that would minimize potential damage at economically acceptable cost (full protection is neither physically nor economically feasible). The analysis of the cost of 3 alternative measures and the projected residual damage resulted in the selection of a design flood of 100 year frequency as providing the most economical alternative. 2.3 The Project was appraised in January 1983, approved by the Executive Directors on May 24, 1983, the loan signed on June 6, 1983 and became effective on June 8, 1984. The one year delay in effectiveness, due to lengthy legal clearances of the loan documents negated much of the otherwise efficient processing of this "emergency" project. The project consisted of: (a) rehabilitation of the water courses of the Oued Ezzit and the Oued Maou, north and south of Sfax respectively; (b) construction of a "belt canal", 4 km from the city center; (c) establishment, staffing and equipping a Flood Protection Section of the Ministry of Equipment in Sfax for the maintenance of the flood protection works; and (d) consulting services for design and supervision of construction. The total estimated cost of the project was US$47.89 million equivalent. An integral part of the design, but financed outside the project, was the construction of three storm water collectors, to evacuate flood waters from the center of the city. These works were financed under the Second Urban Sewerage Project (Loan 1675-TUN). The financing plan for the project included a US$25 million Bank loan, disaster relief donations from the Kingdom of Saudi Arabia (US$10.5 million) and the Islamic Conference Organizations and exceptional national taxes levied for flood prevention and reconstruction. The project was substantially completed in June 1988 and the Bank loan closed, as scheduled, on June 30, 1989. Final disbursements were made on January 17, 1990 at which time US$2.7 million of the loan was canceled. Implementation 2.4 Except for the unreasonable delay in effectiveness, implementation proceeded smoothly with only minor delays resulting from design changes required at some railway and road crossings of the belt canal and from the realignments of water supply and sewer pipes. Expropriation of land for the canals and adjacent roads proceeded expeditiously. By the first scheduled effectiveness date 94% of the land was acquired, a notable contrast to "normal" projects where this matter is often a major bottleneck. Procurement presented no problems. The Audit mission verified the high quality of construction of the works and their current effective maintenance. 2.5 The PCR reports the final cost of the project as US$40.33 million equivalent or some 16% below appraisal estimates. The summarized cost of the major project components is shown in Table 2.1 below. It should be noted that the substantial reduction of the maintenance operations and equipment was not due to the scaling down of the relevant plans but to the fact that during the project period the contractor was responsible for maintaining the facilities. The current staffing and equipment of the Ministry's maintenance unit is adequate. However, the presentation of project costs in US dollar terms only appears to overstate the cost savings. The Audit was unable to obtain the year by year project costs but the 46% devaluation of the Dinar (vis-a-vis the dollar) between appraisal and completion suggests that, in dinar terms, there was, in fact an appreciable cost overrun. Converting at the rate ruling at the time of completion, this 4 would amount to 22.7%. Future PCRs should ensure that adequate information is available to permit analysis of the effect of currency depreciation (if any) on project costs. 2.6 Disbursements were slow for the first two years (the authorized US$600,000 retroactive financing was not utilized), accelerated in 1986 and reached projected levels in mid-1988. As the loan was made to the Republic of Tunisia, the exchange risk was borne by the government. As of April 30, 1993 this, just due to the currency pooling system, amounted to US$2.19 million or 18.5% of the outstanding debt. Due to recent devaluation of the dinar against the US dollar debt servicing has been increased by a further 46%. The performance of consultants and contractors was satisfactory. Table 2.1 - Project Costs (US$ million) Components Appraisal Estimate Actual Change Local For. Total Local For. Total Belt Canal and Dikes 5.37 9.46 14.84 5.28 9.56 14.84 0 along Oued Khalifa Rehab. of Oued Ezzit 6.54 11.53 18.07 5.88 10.72 16.60 -8 Jebeniana, El Amra & 0.48 0.87 1.35 0.61 1.10 1.71 +26 Sidi Mansour Rehab. of Gued Maou 0.22 0.40 0.62 0.12 0.21 0.33 -47 Additional 0 0 0 1.04 1.87 2.92 Infrastructure Land & Relocation 5.53 0.95 6.42 2.30 0.00 2.30 -64 Maintenance Equipm. 1.14 2.67 3.81 0.20 0.55 0.75 -81 Operation & Maint. 0.43 0.07 0.50 0.00 0.02 0.02 -95 Consulting Services 0.46 1.83 2.28 0.22 0.65 0.87 -62 TOTALS 20.17 27.78 47.89 15.63 24.70 40.34 -16 III. THE RESULTS 3.1 By their nature, the effective performance of the flood protection works constructed under the project cannot be tested at will. And, it is possible (and indeed hoped), that floods of the design frequency and lower will never be experienced in Sfax, leaving the design parameters (and the cost) technically unjustified. Indeed, during the Audit mission's discussions with local officials such comments have been heard. Theoretical justification on economic grounds such as the verification of the Appraisal estimate of the Economic Rate of Return (ERR) of 18% and the PCR's recalculated value of 23%, is equally difficult as the reduction of potential damage have to remain an estimate. 5 3.2 The Audit mission also found it impossible to verify the claims of the appraisal report and the PCR that property values, in the previously flood prone areas, have increased "multifold". As there are no capital gain or formal property taxes levied and the city administration keeps no records of transactions, only a detailed survey of recent property transactions could verify such claims. The Audit mission was in no position to conduct such a survey. Anecdotal evidence suggested that no substantial land/property price increases occurred in Sfax itself, possibly (partly) due to the fact that change of ownership itself has been very limited. In the small town of El Amra, outside Sfax, slight better proof emerged in that an area, now protected by a dike constructed under the project, has been re-zoned from agricultural to urban land, making some four fold land price increase possible. 3.3 A similar claim for reduced flood insurance premiums, as a result of the protective works, also proved unverifiable as the administrative official informed the mission that "hardly anyone has flood insurance". 3.4 Nevertheless, visual inspection did confirm the increased confidence of developers (both public and private) in the form of substantial developments in the flood damaged areas of the city. The new roads, constructed on both sides of the belt canal provided the city with both better internal circulation and a "by-pass" system which diverts through traffic from the city center. At the same time several complaints have been heard that the new roads and bridges across the canals are too narrow. 3.5 The mission has also confirmed two significant negative effects of the project, both, essentially, design problems. The first is the accumulation of stagnant, contaminated ground water at the northern outlet of the belt canal resulting in strong odors and mosquito breading (a distinct health hazard). This is due to the canal invert being well below ground water level and that the ground water, in that area, is still saturated by the residue of a now removed olive oil factory. The well functioning maintenance crews which remove rubbish from the canal on a regular basis, can not cope with this problem. The second problem is that the design of the protective works obviously focused on the evacuation of flood waters approaching the city from the hinterland but paid little or no attention to disposal of storm water falling on the city itself - apart from the storm drains constructed under the urban project. As a result, it is claimed, that substantial parts of the city remain under some water, after the rains, as no drainage system exists to divert the water into the main canals. Although the weather was dry during the mission's visit, the total lack of roadside drains in the very flat areas toured was a clear evidence that this complaint was justified. It is evident that insufficient consultation has taken place between the city and the government designers of the project with regard to local flooding issues. This matter should be taken up by the Bank Regional Office in the course of the current considerations of improving the sewer system of Sfax. 3.6 The resolution of this problem, as well as that of the odor and contamination, is not made easier by the fact that the city's current administration does not have "ownership" of the project facilities. They regard the canals, protective works and their maintenance as well as any problems arising therefrom, an entirely government affair. In the Audit's view, while the Bank 6 was entirely correct in insisting on the creation of adequate maintenance facilities for the works, on which the sustainability of the effectiveness of the works depends, entrusting these to the government may have been incorrect. The city should have the responsibility both for the maintenance of the facilities in its territory as well as for their further development and incorporation of the canal functions in the city's overall drainage system. As it is, no one seems to be ready to take any initiative and while the city is, understandably, happy to have the government bear the cost of maintenance, it also expects the government to carry the responsibility for future developments which is not really appropriate. 3.7 In summary, certain beneficial effects of the flood protection works are evident. Recent heavy rainfalls (not of flood proportion) have, apparently, been evacuated at greater ease then before. The population does have a greater sense of security and development work, in the area of "usual flood damage" has accelerated. Some of the "undesirable" effects such as the "separation of the city" by the belt canal are permanent; others, such as the odor and health hazards, should be possible to solve with the cooperation of the city and the government. Overall there is no doubt that the project resulted in substantial benefits even if the magnitude of these may never be ascertained. As a final note, the mission feels that it is worth recording an observation made in the Ministry of Equipment, namely that in the design of future flood protection projects, in water scarce countries like Tunisia, more attention should be paid to trying to "keep the water on the land" and not just to its quick evacuation to the sea. IV. ENVIRONMENTAL ISSUES 4.1 The overall environmental effects of the project are, clearly, beneficial. The realignment, leveeing and lining of the main canals, by confining flood waters, will have reduced the danger of not only property damage and threat to life but of the extensive soil erosion which accompanied each flood. This, in turn, will permit the growth of vegetation assisting in the retention and percolation of runoff. 4.2 Currently, negative environmental effects are confined to the accumulation of noxious ground water at the northern outlet of the belt canal. This sea level exit, being below ground water level, has a perforated floor to relieve uplift pressures which permits the rise of ground water above invert level which, in spite of some tidal flushing effects, remains in the canal as stagnant water. The site of this exit, in the past, was occupied by an olive oil factory which, over many years, discharged olive wastes onto the adjoining ground. The residual olive oil, which saturates the ground to an unknown extent, contaminated the ground water in the area and the oil's breakdown gives rise to strong and unpleasant odors. Apart from its overall unpleasantness, this, to a great extent negates the city's efforts of beautifying and turning into a promenade the nearby shoreline. Due to the divided responsibilities noted above, there are no current plans to solve this problem although it offers itself as an interesting environmental rehabilitation project. Also, the problem of olive oil contamination is not unique to Sfax, the mission encountered a similar problem in Morocco where olive oil contamination of a river 7 requires the shutting down of a nearby water abstraction/treatment plant during the fall harvest period, depriving a number of towns of regular water supply. The design of future water supply projects and water resource management in general needs to pay careful attention to this particular form of contamination in the areas of olive oil production. V. CONCLUSIONS AND LESSONS Conclusions 5.1 The project met its objectives in reducing the rik of flood damage in the city and adjoining areas of Sfax and its performance is rated as satisfactory. The well managed and equipped maintenance unit is likely to ensure the sustainability of the project facilities and their effectiveness, provided it is maintained at its present level of efficiency. As mentioned above, the Audit believes, that the maintenance function should be vested in the city to ensure a better sense of ownership and to facilitate better integration of the flood control canals in the city's overall drainage system. Institutional development, in the form of creating an efficient maintenance unit for flood protection works--a first in the Ministry--was substantial. The Audit was informed that, modeled on the Sfax example, three similar units have been established in other parts of the country. 5.2 Cost effectiveness of the project can only be demonstrated in theoretical projections of flood damage prevented but the implementation of the project was efficient. Cost increase, in dinar terms, seem to have occurred in the foreign exchange components of the works, due to the appreciation of the US dollar against the dinar. 5.3 The nature of the project does not lend itself to direct cost recovery except in the form of general or earmarked taxation. 5.4 The benefits of the project accrued (or will accrue) to rich and poor alike but, in the sense that the poor would find it more difficult to recover from loss of property, the proportional benefit to them is greater. 5.5 The Audit is largely in accord with the findings of the PCR and, in particular, reconfirms (as suggested in para. 8. of the PCR) the physical sustainability of the flood protection works on the basis of the current level and standard of maintenance. The negative effects of the project on Sfax, which are noted in the Audit, have only become evident with the passage of time and do not represent shortcomings in the PCR. However, further Bank action to encourage remedial measures and enhance environmental benefits and sustainability appears warranted. 8 Lessons Learned 5.6 The project offers the following principal lessons: (1) The project confirmed, once again, that Bank response to disaster relief type of requests is considerably faster in processing and, in many ways, more efficient than a routine lending operation. (2) The apparent gaps in cooperation between the government and the city of Sfax highlights the danger of the Bank being preoccupied with the main implementing agency without adequately involving local authorities. (3) The speed and efficiency of preparing and implementing these type of projects often poses the danger of overlooking some important issues, in this case, the study of retaining more water on the land and the integration of the city's drainage system with main flood protection facilities. (4) When the economic justification of a project (of this nature) rests on assumptions which (i) may never be verifiable (reduction of flood damage) or (ii) relate to future prospects such as property price increases, it is desirable that base-line data be gathered and recorded at appraisal to permit future impact measurements. NOTES  г о � � ш и а а �  »s S.kii Ed Dair A i .4 !4 .9 n » A LGERIA