World Bank SDI Report - Learning from Experience

This article is a part of World Bank SDI Report.

Introduction
Having detailed a range of SDI implementation issues in the previous chapter, this section of the report aims to further share the experience of earlier SDI developers through a series of 'lessons learned', i.e. what has worked and what has not worked and also through other potential challenges for consideration. Firstly, the chapter reflects on the experience of SDI implementation in the two best practice case study countries detailed in the previous chapter (Korea and Brazil) as a means of further guidance to developing countries in their SDI implementation process. Secondly, as an additional source of reference the experience of several additional World-Bank supported SDI-related projects follows. Thirdly, following on from these experiences, as a means of providing a comprehensive set of guidance for future SDI developers, this chapter then outlines several additional challenges which deserve consideration throughout the development process.

The Korean Experience in SDI
The lessons learned from the Korean NSDI are divided into 7 components. These are: data, law/institutions [ legal and institutional frameworks ], standards, access and metadata, technology, partnerships and GIS applications.

Data
Korea has excellent topographic data which underpins and makes a major contribution to the successful implementation of its NSDI. NGII's digital topographical map at 1:1,000 covering most all of the country is the core base data set. Use of this data has improved the function of the geospatial civil service and added value to business in geospatial industry sectors. The Korean underground facilities map based on the 1:1000 digital topographical maps is another example of a key core data set.

At the beginning of the NGIS project, a large amount of the budget was invested in building geospatial data. With the evolution of NGIS projects, it has also become clear that that not only building geospatial data, but also its maintenance for sharing and further use is extremely important.

Lessons Learned

 * To prevent duplicated investment for revision and data updating, strategies for data updating should be considered at the beginning of data building.


 * In terms of digital topographical maps, on completion of their creation in the 1st phase, continued revision and updating must follow in the 2nd phase.


 * UFID for efficient and real-time updating is useful.of cost with quality is recommended.


 * Digital ortho-photo, satellite imagery or new surveying technologies are preferred over traditional surveying methods.


 * For rapid return on investment, proper selection of map scale and decision on data building method will need to be reviewed by efficient methodologies. These include: an estimation of cost based on the analysis of various alternatives with multiple scenarios based on combination of each production methods considering different areas and different scales, a survey of experts and users' need and CBA (Cost-Benefit Analysis).


 * Different map scales are appropriate for different areas (e.g. 1:2500 in urban areas, 1:10000 in rural areas)Efforts to bridge the gap between assessing formal institutional data quality and informal real usage for business are recommended.


 * An over-emphasis on data accuracy has retarded the development of SDI in South Korea. According to the recent survey in Korea, both at a central and local level, digital topographical maps at 1:5,000 are most widely used for business across the various domains. It also estimates that 1:1,000 for the higher data accuracy is not used more widely than 1:5,000, but is more popular than 1:25,000. On the other hand, standard estimation of digital topographical map production costs per unit indicates that it could be different per scale. If the same production methods using aerial photogrammetry were used, cost would be rapidly increased for higher data accuracy.

Law/Institutions
The Korean SDI has been driven by a top-down rather than a bottom-up approach:

As a legal basis upon which to drive SDI and promote geospatial industries, two main legislations for NGIS at national government level have been established. An interagency committee to coordinate matters concerning national spatial data policy, the National Spatial Data Committee, has been established under this legislation. Its members include representatives from both national and local government and from public and private sectors with the Minister for Land, Transport and Maritime Affair as its chairperson.

There are also many GIS regulations, implementation plans and organizational teams at the local government level, which additionally provide the legal and organizational basis for local SDI.

Local legislation is established in accordance with the national legislation and based on specific articles. Since the autonomous local government system was launched in 1995, Korea has 15 years of local government experience. During that time, the underground facility management projects have strived to make legal regulations at a local level a necessity. In this context, many local governments established GIS regulations concerning building, using and managing GIS data and security. Local legislations such as Gyeonggi GIS Ordinance and regulations for spatial data security in Daegu, Dajeon, Busan, Incheon, Gwnagju and others have been made, reflecting the unique requirements of each local government.

Lessons Learned

 * Legislation for the NSDI at a national and local level are required.


 * legislation for SDI must be introduced in a timely manner.


 * A clear definition of role, based on legislation is critical.


 * A Strategic and sustainable Master Plan is needed for SDI from the beginning


 * The NGIS Master Plan proposed the vision, tasks, project and budgets for the NSDI, for every phase according to the Master Plan An implementation Plan is also needed.


 * After implementation, evaluation is essential for a successful NSDI.


 * More participation and determination of GIS experts in the policy making process is needed.


 * An imbalance between administrative experts and GIS experts as decision makers will result in excessive trial and error at the early stages of NSDI development.


 * Conflicts between GIS and surveying communities are also a problem. Surveyor's laws have been more substantial than GIS laws. Accuracy has prevailed over fitness for use, which has retarded the development of GIS applications.


 * Korea has experienced that SDI policy focusing on a top-down approach at the beginning leads to some problems. Now, harmonization with a bottom-up approach has emphasized a move towards a strategy that is demand-driven, cooperative, with an open data policy, and an emphasis on linkage and integration, sharing and participation. Because the development of SDIs is a long-term process, in the opinion of Korean GIS expert's, a top-down approach is somewhat necessary to drive SDI development at the beginning. Over time, the harmonization with the bottom-up approach is however more efficient. Therefore, GIS education, to include the concept of "Neogeography", should be prioritised.

Standards
Due to a lack of standardization, data sharing and cooperation among agencies has been difficult. To-date, Korea has developed a range of standards (see Annex C). However, according to NGIS evaluation literature and results from Korean GIS experts' investigations, a time lag, lack of availability and no obligation to conform to GIS standards for NGIS has resulted. Regarding standards for current Korean NSDI, the Subcommittee on Standardization and Technical Standards (a sub-committee of the National Spatial Data Committee) is responsible to develop and manage standards for spatial data by the provision of the National Spatial Data Infrastructure Act, the Framework Act on National Standards, and the Industrial Standardization Act. According to the Framework Act on National Standards, KS and TTA standards for NSDI are not mandatory while technical standards are. However, as yet there are no technical standards in the field of GIS domain (see Annex C).

To solve the problem of standards and develop mandatory standards for the NGIS projects (such as FGDC standards in USA) a project for developing GIS technical standards is currently ongoing in Korea.

In order to integrate and share data more effectively and easily and to provide interoperability to geospatial data and services, standards are of the utmost importance.

Lessons Learned

 * Strategic standardization based on consensus among GIS communities for SDI is more  effective than dictating standards


 * Both de facto (industry based and informal) standards and formal de jure standards exist. In Korea, for national SDI two kinds of de jure standards have developed.


 * To develop standards while obtaining consensus, is a long process. Thus, adoption of international standards can be a solution. Nonetheless, it is not the best way for developing countries to adopt all of them without a strategic GIS standard plan.


 * Standards such as the FGDC standards and USGS geospatial bluebook can be considered as a consistent data model and for specification guidelines.

Access & Metadata
'Open data' policies are becoming more prevalent. Data.gov in the USA, Data.gov.uk in the UK, and also Data.gov.kr in Korea are good examples. However, in Korea, open geospatial data policy still remains a sensitive and controversial issue. Until now, due to the political and military relationship with North Korea, data security is more important than perhaps in most other countries. Thus, access to geospatial data in some sensitive areas for military purposes is limited and permission is required. The NSDI ACT (Act No. 9705) includes provisions on both the establishment of a National Spatial Data Center and the protection of national spatial data.

In spite of data security issues mentioned above, Korea opened a national clearing house in 2000. Now, various kinds of spatial data including digital topographical maps and many thematic maps are available free or for cost price.

In addition, there are 3 metadata standards containing 2 TTA standards and 1KS standard in Korea (see Annex E). One of the TTA metadata standards is for distribution, the other for management. The KS metadata standard is adopted from the ISO 19115 metadata standard.

Lessons Learned

 * Efficient online or distribution networks are necessary for successful SDI


 * One Stop Portals for access services are beneficial


 * For data sharing and distribution via the One Stop Portal, standards for metadata and data catalogues are required.


 * Linkage of e-government portal with the One Stop Portal is desirable.


 * Incorporation of VGI data should be encouraged to increase data availability and relevance.

Technology
Availability of advanced information and communication infrastructure has been a major contributor to the success of the Korean SDI implementation. However, the separation of IT and GIS communities hindered progress: The two domains should have had a shared perspective from the beginning for a better NSDI. Integration and linkage of e-government and NSDI should also have been achieved from the beginning to avoid duplication of efforts and for a better citizen spatial service. NGIS projects have provided a number of opportunities for developing GIS technologies, from GIS components technology in the beginning through 3D GIS technology, Mobile GIS technologies, Geoweb technologies in the mid phase to ubiquitous geospatial technologies currently under development. Although Korea lacked international competitiveness in GIS S/W technology at the beginning, more research to develop GIS components, platform interface technologies and ubiquitous geospatial technologies has contributed to open new geospatial markets and encouraged SDI development.

Lessons Learned

 * Strategic and rapid development of selected technologies for a developing countries' SDI is recommended.


 * The import of advanced foreign technologies is needed.


 * Due to a lack of best practice in the use of technologies in the beginning, education programs for GIS technologists and experts are a high priority.


 * Wider strategic use of open source technologies is recommended.

Partnership
Partnership can be defined as relationships and agreements across relevant actors and organizations that reduce duplication and collection costs and leverage local, national and global technology and skills. Subsequent NGIS projects have shown that the main cause of inefficiency and inconsistency has been a lack of partnership. Project success depends on coordinating many stakeholders and many organizations and on cooperation between these agencies. Therefore, a great deal of attention should be paid to both vertical and horizontal partnership in the SDI hierarchy. Also, international partnership can be a very important component for SDI of developing countries.

Lessons Learned
For more powerful organizational partnerships, more efficient institutional reorganization and regulations are required. Insufficient leadership leads to, government organizational conflicts, reduced focus and a lack of cooperation with other governmental organizations and agencies. A lack of organisational partnership also served to delay driving the organization of NGIS standards in the 2nd phase of NGIS.

GIS Applications
The 'Best Practice' examples from Korea (see 3.4.1) demonstrated that GIS applications can be developed in multiple domains, for example in economic developments environmental monitoring and for monitoring the MDGs.

From Korean cases, interoperability among GIS applications would contribute to the avoidance of duplicated efforts and inefficiency. Integration of GIS applications can also be critical, as exemplified by the National Integrated Spatial Information system (NIIS).

Lessons Learned

 * A geoweb platform assists interoperability of GIS applications.


 * The recent trend of using Open API's through which users can make their own GIS applications, has significant potential for developing countries.


 * For 15 years (up to the 3rd phase of NGIS in Korea) many GIS applications have been developed and used for efficient resource management. Considering often limited resources in developing countries budget allocation priorities need to be established in the master plan.


 * Rapid return on investment of GIS applications also must be prioritised.


 * Interoperability of GIS applications is key and the priorities of GIS applications should be established.

The Brazilian experience in SDI
As to human capacity, it has been demonstrated that Brazil has a diversified and capable academic sector, with a worthy presence in world science and is therefore capable of educating people in all required SDI-related subjects. There is, nevertheless, a strong market pressure on highly specialized professionals resulting from the recent increase in economic growth.

The remaining part of the equation, namely policies, is the most interesting. Coherent and effective policies are only possible in a context of institutional maturity. Taking Brazil as an example, we observe that the creation of INDE was possible because various conditions were met. First, a national geography/cartography/statistics bureau (IBGE) is in place, with a long term involvement in geographic information and a vision on the demands for GI from the government and the society in general. IBGE in turn is part of a broader arrangement of cartographic institutions, all of which have undergone a transition from conventional mapping to GIS technologies in the last two decades. When CONCAR, led by IBGE, set forth the initiative of creating INDE, provisions were made to allow the participation of the most important institutions in the process, and the resulting action plan has arisen from consensus, rather than from a centralized planning decision.

In a sense, one can say that institutional maturity led to the kind of cooperation and agreement necessary to create something like INDE. Of course, the initiative is in its initial steps and much remains to be seen as to the reality of funding INDE's operation and to the actual usage of the information provided in the infrastructure. IBGE is currently undertaking an initiative to disseminate INDE in GIS-related events countrywide and is offering support for the establishment of local or thematic SDIs that are to become part of the national infrastructure. CONCAR and IBGE are working with the objective of encouraging GIS-enabled organizations to move towards SDI and to join INDE. If the installation and configuration of INDE-compatible servers becomes simple enough, many more geographic information sources can join the bandwagon and thus increase the importance of INDE to wider society. It is important to ensure that not only governmental or "official" data sources are allowed in; there must be openings for other types of organizations to join, making available then for instance, research data, community-oriented points of interest, commercial locations and others.

As in the case of GIS projects from the 1990s, current SDI initiatives appear to be growing out of the already existing infrastructure and funding embedded in data collection and updating efforts. In some cases, special projects have been developed, seeking funding from organizations such as the World Bank, but such projects are presented as a small IT part of a much larger physical infrastructure initiative. In this sense, a relatively small percentage of a large effort is being redirected towards improving information availability and dissemination, which is a positive outcome. Since geographic information can be used in many other unforeseen uses, the positive outcome is potentially multiplied by the quantity and variety of alternate uses of information, an improvement which is at the core of SDI development goals. Sharing part of the computational infrastructure in the establishment of SDIs is now much easier than it was to share spatial data 15 years ago. Therefore, organizations could cooperate in establishing the core of an SDI, and divide among themselves the hosting of Web services for that SDI, thus forming local- or theme-specialized SDIs that could eventually be integrated into INDE. This is reminiscent of cooperation agreements for GIS data sharing, as initiated by some GIS projects in the 1990s, with the difference now that data replication is not necessary. However, achieving broad institutional agreements remains as difficult as they were twenty years ago.

SDIs can be seen as a "publication platform" by data producers. Current data collection and validation methods can go on as usual and periodically a new data release can be copied from maintenance servers to the SDI servers and made available to the users. The coexistence of several data releases, each of which with its own metadata, enables users to perceive temporal evolution. This is similar to the comparison of paper maps from various dates, something that GIS datasets kept us from doing in the last two decades.

Uganda Bureau of Statistics
It should be noted that while the U-Consult Sweden 2004 study aimed at building the internal capacity of UBOS, the last two studies aimed at institutional collaboration and support with UBOS as the custodian of and lead agency for statistical information.

Under the PNSD program, UBOS succeeded in supporting the MDAs with training, software, hardware and awareness creation in order to meet the core objectives of the program. These studies also helped UBOS to set a thematic Data Infrastructure and institutional arrangements for exchange and dissemination of data. It also helped UBOS to develop its own distinctive geo-codes that uniquely identify all administrative areas down to village level for geospatial data acquired in the 2002 census. In addition, due to the central role it played in most of the SDI initiatives, UBOS has been hosting and co-chairing the meetings of the Geo-IM Working Group.

Northern Uganda Data Centre
Originally the KDC was established as a One-Stop Information Centre to support the development of the Karamoja region. It was based on the project 'Master Plan for the Development of Karamoja (KDC)' (October 2000 - September 2005) and was initiated soon after the study conducted by U-Consult Sweden funded by The World Bank for the establishment of Uganda Spatial Data Infrastructure (USDI). It can, therefore, be considered a pilot SDI initiative based on the geographical extent of Uganda.

The Data Centre collaborated with many government agencies as well as other data producers and users. However, it exceeded its mandate by collecting data on 20 thematic areas such as education, water and sanitation, health, conflict, demography, agriculture and others. This was a classic example of duplication of efforts in data collection by different sectors. The Centre also relied heavily on the Italian Government for funding and hence shared the same problem of a lack of sustainability in case of discontinuity of project support. In August 2010, the centre was restructured after the resignation of the core GIS staff. This is likely to affect the program adversely since the new staff ought to start from scratch and renegotiate the funding arrangement.

National Integrated Monitoring and Evaluation Strategy
NIMES was primarily instituted to monitor and evaluate the implementation government programmes funded both by the government of Uganda and development partners. Although the initiative recognized the problems of duplication of data collection effort, lack of documentation of data, variation in data qualities collected and used, poor communication among data producers and users, lack of equipment and human capacity, it fell far short of orientating its core objectives and activities to the principles of Spatial Data Infrastructure. For example, the proposed strategy for implementing NIMES stated 4 elements:


 * specification of data and information needs


 * building an institutional framework for coordination


 * ensuring adequate access to data and information


 * enhancing M&E capacity

It is clear that the neither the four elements stated above and nor the objectives referred to Spatial Data Infrastructure development effort but rather general information required for planning. This lack of a sound objective and implementation strategy directed specifically towards SDI, contributed to the failure to implement the initiative. This was because the SDI was confused [ and fused? ] with the M&E concept and activities. The initiative has (since 2006) closed although approval for the NIMES funding was announced in the Governmental financial budget of 2005/2006.

NIMES objectives and activities were similar to the Karamoja Data Centre, although it had a national mandate. In addition, the two initiatives were situated within the same government agency, that is, the Office of the Prime Minister. This posed the challenge of potential conflict in implementing a similar and parallel program.

Environment Information Network
EIN was a well-planned and funded activity which produced the desired outputs. However, because it depended entirely on donor funds, its operations had to discontinue when the funding was terminated after the second phase of the programme. In addition, alternative sources of funding were neither planned nor secured for sustainability.

Geo-Information Society of Uganda
The evidence available indicates that the Uganda Spatial Data Infrastructure (USDI) formed by the Geo-Information Society of Uganda (GISU), organized several start-up meetings with prospective members and government officials, especially the Assistant Commissioner [ of the? ] M&E Office of the Prime Minister. It organized the National Workshop on Geo-spatial Information and Development which carried the theme: NSDI for enhancement of information exchange held on 7th June 2006. It also developed a draft constitution to guide and bind its members. However, apart from activities listed above, this initiative which lasted for about a year closed in 2006 without much impact in steering the development of SDI in Uganda. The choice of Makerere University as the centre for leading such effort meant that its resolutions could not engage Government in implementation. Instead it tried to co-opt the Assistant Commissioner [ of the? ] M&E Office of the Prime Minister as the government focal point at the same time as the NIMES initiative was being set up. A University can and should contribute to SDI development through education and research.

Lessons Learned and Guiding Principles

 * From the experience of the examples listed above and others below is a list of general lessons and insights to be learnt as well as a set of guiding principles to be considered while formulating and implementing SDI programs.


 * Remain sensitive to momentum and ensure SDI implementation sustainability through successive, smaller, "quick win" deliverables.


 * While the initial implementation of SDI components in Africa is often assisted through donor and other NGO funding, it is sometimes the handover and lack of continuity where SDIs fail when donor projects come to a close.


 * Be sensitive to and have a mitigation plan for technological challenges during the early phases of SDI implementation activities.


 * While the longer-term architecture and application areas of an SDI will be generic and cross-disciplinary, focus initially on specific application areas that address well-articulated needs. Absent concrete application areas, SDI initiatives tend to lose traction.


 * Leverage and build upon existing efforts, structures, and institutional arrangements to the extent possible, especially in cases with a relative absence of political support or a legislative mandate.


 * SDI institutions and the top stakeholders should have a business case or reasons for operating the SDI and working for its success. As may be noted in the case studies in this report, often the leadership or secretariat has been placed in a land administration agency or environmental agency that is both an information provider as well as an information user.  Such lead SDI agencies must deal with others to do their functions and therefore they need SDI.  They will not let it fail.


 * Legislation is a powerful SDI enabler.


 * All stakeholders need to remain engaged and encouraged to remain at the table.

Selected World Bank-Supported SDI-related Projects
The Bank is a vital source of financial and technical assistance to developing countries around the world. In its mission to fight poverty and encourage economic development the Bank has supported various sectors in many developing countries worldwide. As spatial data is applicable to most development activities, some of them have involved the use of GIS and the development of spatial data infrastructure related products.

This section discusses some selected Bank supported GIS/SDI related projects, their successes and failures and lessons that can be drawn for future Bank supported SDI related projects. Projects reviewed include:

The Environmental Support Program, Zambia The program aimed to mainstream environmental and natural resources management in Zambia's development processes at the national and local levels;

The Natural Resource Management Project, Phase I, Ghana. The project was meant to assist the Government of Ghana in implementing its policy of protecting, rehabilitating and sustainably managing national land, forest and wildlife resources;

The Regional Environmental Information Management Program (REIMP), Central Africa. This was an inter-country project to improve planning and management of natural resources in the Congo Basin by providing the various stakeholders with appropriate environmental data;

The Integrated Transport (ITP) Project, Lesotho. The main goal of the project is to contribute to the Government's poverty reduction and economic growth strategy by improving access to basic services, markets and employment opportunities for rural and urban Basotho, as well as to support the Government in improving its transport sector management.

Project Sample
The selected projects were reviewed against a set of questions about objectives, project design, successes and failures and lessons learned with particular focus on GIS/SDI aspects. Implementation Completion Reports with other related documents were used as reference documents. This section uses one of the reviewed projects, the Natural Resource Management Project Phase I in Ghana, to highlight the SDI issues involved in most projects. Complete reviews of other projects can be found in Appendix [ ? ].

The Natural Resources Management Project (NRMP-1), which ran between June 1998 and June 2003, was meant to assist the Government of Ghana (GoG) in the implemention of its policy of protecting, rehabilitating and sustainably managing national land, forest and wildlife resources by: institutionalizing viable sustainable land, forest and wildlife management systems nationwide; Establishing effective national policy and institutional frameworks and developing collaborative natural resource management systems.

The project had five main components:

High forest resource management – to establish forest resource management policy and legal framework; design and test integrated forest reserve management systems and mechanism to encourage private sector involvement in both on- and off-reserve settings.

Savanna resource management – to develop new multidisciplinary institutional arrangements for dry lands management; and pilot test and implement improved systems for community-based management of savanna woodland and other resources.

Wildlife resource management – to retrain personnel in the Wildlife Department and prepare and implement park and reserve management plans.

Environmental management coordination – to support district level environmental management planning and monitoring capacity; and improve information flow among sectoral agencies.

Biodiversity conservation - to develop and implement integrated reserve management plans.

The other project information is summarized in Table 5.1.

Most of the project components had some GIS/SDI related aspects but the Environmental Management Coordination component had a direct connection to SDI development and included: Development of information framework for the management and coordination of geospatial information; formation of workgroups on data policy and standards; strengthening of partnerships between participating ministries; regional and district capacity building; land use/land cover map production and update; and contribution of Geodata metadata by participating ministries.



Table 5.1: The Natural Resources Management Project (NRMP-1) Project Identification Summary

Successes and Failures
Successes for the borrowers: The Environmental management coordination component scored a number of successes and achieved most of its targets:

Major targets achieved: One of the major targets - the development of regional environmental information systems to improve information flow among sectoral agencies - was met and regional EIS systems were installed in 10 regions and 6 pilot districts. The project also carried out capacity building for staff with some being sent abroad to study courses relevant to the sector. The production and update of land use/landcover maps was another target that was successfully met.

Successful development of SDI for Ghana: A key outcome from this component was the National Framework for Geospatial Information Management (NAFGIM) which was successfully set up as the Spatial Data Infrastructure for Ghana. The development of this SDI continued with participating institutions creating and contributing metadata for their respective spatial data. A national SDI structure including national workgroups for data policy and data standards development was also constituted.

Failures for the borrowers: The project also had its own shortcomings. The main factors that affected the whole project were a complex project design, inadequate institutional capacities and ambitious objectives for such a short implementation period. Other failures were:

Poor coordination of funding from multiple donors: Delays in implementation of some parts of the project were partly caused by the Ministry of Land and Forestry (MLF) having to work through 4 other ministries (inefficient communication or administrative channels).

Inadequate Government funding: The GoG consistently failed to provide counterpart funding on time and in the amounts estimated by the project.

Successes for the Bank

Bank Supervision: Supervision by the Bank team was intensive and provided a number of useful recommendations to the implementing agency. For example the supervision missions were responsible for identifying human capacity and morale problems within the MLF and for recommending that MLF begin to own some of the project administrative responsibilities.

Failures for the Bank:

Overly Ambitious Project Designs as a stumbling block to implementation: The project design included too many components, sub-components and activities to be carried out in a two year time-frame. This imbalance should have been detected during the project appraisal stage and the project redesigned to ensure good project implementation.

Proper Assessment of Government Commitment to Reform: This was not determined early enough. It turned out that GoG's level of financial commitment to the project was low and the project suffered from inadequate and untimely release of funds.

Monitoring and Evaluation: Monitoring and evaluation needs and capacities for the borrower were not adequately assessed.

Lessons learned
Project Design

The project was prepared by the Government with the active participation of implementing agencies. There was adequate representation from Ministry of Lands and Forestry, the Forestry Commission and Ministry of Finance. This part of the project was rated satisfactory.

However, the overly ambitious project designs and moderate institutional capacities became stumbling blocks to implementation. Project design should therefore avoid being unnecessarily complex and must be tailored to the skills and institutional short and medium term implementation capacities of the implementing agency. For an agency with lower capacity, the design should focus on fewer, core project activities.

Project Implementation

Several lessons arise from the implementation of the project:

Co-financing

to achieve the benefits of a partnership with donors, it is essential that joint financing be sought and other mechanisms of funding be explored that ensure that funds come in simultaneously and well before implementation e.g., basket funding.

Proper Assessment of Government Commitment to Reform

In this project, GoG's level of financial commitment to the project was low and therefore the project suffered from inadequate and untimely release of funds. The key lesson here is that there is a need to make a determination, during preparation and the early stages of appraisal, about the level of Government commitment to the project. If it is perceived to be lacking, there is a need to engage in discussions with the government, at a higher level, to determine where the problem lies. Early diagnosis is important to ensure that this does not adversely affect the project.

Monitoring and Evaluation

Monitoring and evaluation needs and capacities should be assessed in the first two or three supervision missions to determine whether the team is correctly setting up the M&E system. Where M&E experience is lacking corrective measures should be taken.

Legislation as a Requisite for Implementation

Legislative procedures and creation of by-laws is a slow process that could confound the project if these are a necessary first step for core implementation. As was the case, it took a long time to get Cabinet and Parliament to ratify and approve policy and institutional reforms. In some cases, legislation should be attained prior to effectiveness to ensure that this process is completed quickly.

Technical issues versus Institutional/Governance Issues

Most of the difficulties in this project were attributed to governance issues such as communication amongst the agencies involved, monitoring and evaluation and coordination. On the other hand, technical issues seemed to be more easily achievable considering that most of the successes scored in the Environmental Management coordination component were technical in nature. Going forward, it is vital to look at institutional and governance issues more closely before the implementation of the project.

Concluding Comments
Although the overall borrower performance was rated unsatisfactory, the project achieved most of its targets related to GIS/SDI. This success can be attributed to the level of agreement between the different stakeholders involved and the capacity building arrangements set up during the project implementation period.

The development of SDI always involves considerable interaction between agencies. This main thematic focus of the project was Forestry, but it is clear from the number of stakeholders that the developments in this sector are interlinked with other sectors. For SDI development to be successful, effective institutional arrangements and partnerships must be set among all stakeholders. The all-encompassing nature of SDI can be seen also from the other projects.

This project suffered from a complex project design and its objectives were seen to be over ambitious for the project's implementation period. This is the area the bank should scrutinize to ensure planned projects are in-line with clients' level of capacity, commitment and the planned project duration. The Environmental Support Program also suffered from such problems, as the to effectively run the program was missing in the implementing agency.

Lastly, poor coordination of funding and the failure by the GoG to meet its agreed proportion of counterpart funding affected the implementation of the project. This is another factor that affected most of the programs.

Challenging Conditions in a Developing Country Context
In all countries studied there are growing pressures to improve geospatial information and to make it available to public sector authorities and civil society. However there are differing demands and practices in most countries which make SDI extremely challenging, indeed complex. These bottlenecks and barriers impede development and introduce obstacles in such areas as GIS and SDI spatial and environmental research, public and commercial innovative services, and strategy creation. To consider some of the unique challenges of SDI development and implementation that may be faced in developing countries, the experiences of Uganda and Jordan are presented. Problems span the range of SDI components to include specific technical, operational and institutional challenges.

Technical
fragmented and diffused reference spatial data across cities, regions, the country and neighbouring countries;


 * Different formats of reference data, i.e. coordinate system, grid system, map projection, files and data base formats;


 * Different classification systems and definitions of objects, e.g. roads, cities, rivers, administration borders;


 * Different cartographic presentations and symbols representing the same objects;


 * Lack of common rules for edge matching between data bases in different cities, rural areas, regions, especially between countries;


 * Lack of one common data model format;


 * Different formats of data bases, different object classifications which prevent data aggregation;


 * Lack of common or well-known access points and names for information about data sources (metadata), or to compile datasets from many sources and aggregate them to become consistent data describing the data bases;


 * Lack of data access points available to allow above mentioned services;


 * Lack of policy overall and other factors including national security and legal restrictions which can also stand in the way of progress; and


 * Lack of national laws applying regulations for data access and disclosure including pricing policies.

Unclear Mandates
There are often inconsistent mandates or multiple ministries doing similar geo-data collection tasks in different areas with no overall plan, no common specification or common scales, though all are allocated a merge portion of the national mapping budget leaving little for the mandated national mapping agency.

Lack of Annual Budget for Key Data Sets
Often ministries with key data set mandates are allocated little or no budget by government with which to undertake the geospatial data collection or key mapping work. Such is the case with the Ministry of Lands, Housing and Urban Development in Uganda, where for 30 to 40 years little or no topographic mapping has been possible because of the lack of funding (save for a small percentage of the country which is funded by bi-and multi-lateral development agencies).

Lack of Incentives for Cooperation
Most often in many countries there are no incentives and rewards for managers who cooperate with others in other agencies on joint projects that lead toward SDI. Indeed, often the reverse is true with managers rewarded for spending their annual budgets regardless of the objectives and outcomes.