Mapping Land Rights in Mozambique

Abstract

Mapping of land rights in Third World areas can be done by modestly trained and assisted local residents using highly enlarged (1:5000 or larger) aerial photography. Three technical challenges are discussed: (1) cartographic quality attained from non-standard methods, (2) image enlargement (8X to 40X) accomplished with photo-digital techniques, and (3) implementation methods using pre-literacy and distance education techniques. Examples of images and GIS-cartography results of two pilot studies are shown and discussed in the context of Mozambique's recent and highly progressive land-rights laws. Implications for improving local education and for collection of additional geography-based data (soils, biodiversity, demographics, etc.) are discussed.

Introduction to the Challenges

The fundamental scientific challenge (i.e., not socio-political challenge) of mapping land rights in Third World areas is: How to obtain the land boundary information and use it to create meaningful maps? That scientific challenge can be resolved by combining the two greatest information resources on the topic: the local residents' knowledge of their locale and the image detail found in aerial photography. To solve this fundamental challenge requires solving three subsequent technology challenges concerning (1) high cartographic standards, (2) adequate image quality at low costs, and (3) technology transfer for implementation by local residents. When the three technology challenges are solved, the land-rights boundaries can be entered by standard methods into a geographic information system (GIS) for modern data storage, analysis, and map production to serve the individuals, their communities, and their government entities.

Background about Mozambique and Its Land-Rights Regulations

With 800,000 sq km and 18 million people, Mozambique has a size and population slightly larger than Texas, but with an extremely low level of national development. The current population is approximately 85 percent rural subsistence farmers with little or no education. The United Nations (1999) ranked Mozambique as the sixth poorest country in the World. The causes include a war of independence (1964-1974), near-total loss of technical skills when the Portuguese left in 1975, major guerrilla warfare that ended in 1992, and a socialist political and economic regime from 1975 to 1994 (Ramsay, 1999).

Virtually no land ownership records from the earlier periods are usable. For example, the existing cadastral map of the area in the pilot study (Bairro Mali) is a tattered, brittle sheet of drafting velum with penciled property boundaries used as an overlay on a 1:50,000-scale topographic sheet that dated to the colonial era (Source: DINAGECA, Maputo, Mozambique). Most of the marked properties measure approximately 0.6 by 2 km (120 hectares). No evidence of these property units was visible on the ground during the fieldwork nor in the aerial photographs of 1989 and 1996. The area is only 25 kilometers from the heart of the national capital, has had road access for many decades, and has therefore received more favorable mapping coverage than vast areas in remote provinces.

On the positive side, a high level of progressive thinking is evident in recent legal arrangements about land rights. Although now rejecting socialism, the national government of Mozambique continues to own all the land, but grants land-use rights to individuals and entities after consultation with communities in the area. Any person who has occupied land for ten years or more, without any needed paperwork, automatically has rights to continue to use that land (Lei, 1997; Companha Terra, 1998). However, recent political stability and international economic investment have spurred development to the point that land-rights conflicts are a major national issue, especially regarding areas with desirable resources. At the "National Seminar on Community Land Delimitation and Management" in August 1998, the impossibility of establishing new cadastral records fast enough using standard "demarcation" mapping techniques was officially recognized (Comissão de Terras, 1999).

Mapping for "Delimitation"

To overcome the demarcation impass, an intermediate legal arrangement, called "delimitation," was established for recording community land rights (Decrees and Regulations, 1998; 1999). These legal arrangements allow rural communities to organize themselves, "delimit" (i.e., describe) their collective area (and also delimit individual lands within the community area), confirm such limits with surrounding neighboring communities, and present specific documents (including a sketch-map esboço) to the national agency for cadastral records, DINAGECA (Direcção Nacional de Geografia e Cadastro). The distinction between delimitation (legally recognized written agreement by the local community and its neighbors) and demarcation (official recognition with field verification and cement monuments by the national agency) is a crucial difference that permits the community of local residents to become the organized and legally recognized unit for land development, without the wait and expense of official demarcation. Delimitation paves the way for resolving land-rights conflicts and encouraging investment from outside the local community.

The delimitation procedures include at least a minimal form of territorial mapping called an esboço ("sketch-map"), defined in the Decrees and Regulations (1999) as "a drawing, at a conventional scale, of the configuration of the piece of land, containing drawn or written references sufficient (tendentes a) to locate it on the Cadastral Atlas, including, when necessary, the geo-referencing of points and/or boundary lines not visible on existing [l:50,000-scale] topographic maps." The sketch map is not required to have a known projection, standard cartographic symbols, ground surveying, nor geo-referencing for metes and bounds descriptions of features such as road junctions and stream courses that are on topographic maps. The sketch-map and a required descriptive narrative (memoria descritiva), plus other documents signed by community leaders, are considered to be sufficient to comply with the delimitation requirements.

To obtain an adequate "sketch-map," any combination of mapping techniques can be used. When testing the "delimitation" process in 20 pilot studies in early 1999, as authorized by the interministerial Commission of Lands, some obtained any needed geo-referencing at 1:50,000 scale by using a common GPS unit. Others used surveying techniques and/or traditional aerial photography. All involved substantial fieldwork, teams of mapping specialists, and substantial costs and equipment. At the time of this writing, the only alternative method proposed for recording the community (and individual property) delimitations is what is described in this article.

Enlarged Aerial Photographs to Produce Delimitation Sketch-Maps

The proposed alternative method is based on aerial photography (available from DINAGECA) that is highly enlarged photographically and digitally so that the local residents can easily see their lands at scales appropriate for distinctive landmarks (streams and roads, at 1:10,000 scale, trails and fields at 1:5000 scale, and houses and individual trees at 1:2000 or 1:1000 scale). The local residents participate in a circular flow of information and resources between national and local entities using regional support groups (see the outer circle in Figure 1).

Local individuals who are knowledgeable of the community and individual property boundaries are given brief instruction (from the regional support groups) on scaled measurements and how to understand the aerial ("bird's eye") view. They practice on visible features of their actual local area before they (and/or the semi-skilled assistants) mark the community boundary lines on images (or on overlays over the aerial photographs). The people can literally walk and mark their community boundary line, producing the sketch-map and documentation needed to complete the community land registration process.

Those marked images or overlays are checked for completeness by the regional support workers and prepared for entry into a geographic information system (GIS). The resultant maps become part of the recorded national database, with copies given back to the local community for its records and for recording changes.

Pilot Field Testing of the Mapping Method

Bairro Mali, a rural area 25-km north of Maputo, Mozambique, was selected for the field testing because of accessibility (seven kilometers on a good dirt road from a main hard-surfaced highway) and "typical" rural characteristics. Mali has no electricity, no industry except one cooperative chicken coop, no commerce except a single mini-store with fewer than 50 items on hand, and near-total dependence on subsistence agriculture. After abandoning the area during the war years, the former residents of Mali started returning in 1992. The distances between houses (of cane and thatch or of low-cost cinder-blocks) are from 50 to 300 meters. A small primary school has reopened; the education level of most residents is three years of schooling or less. The main language is Ronga, but Portuguese is understood and spoken by many residents as a second language.

From November to December 1998 and from August to October 1999, the author conducted pilot studies. On the first day, the author and two technicians from DINAGECA tested the methodology for five hours, marking the boundaries of individual properties on an overlay sheet on an enlarged (1:10.000-scale) 1989 aerial photograph (1:40.000 original). This first pilot study dealt with the more difficult issue of individual property boundaries because it was conducted prior to the governmental decision to focus on community boundaries. The ground truth was provided by four local residents led by the Bairro Mali "Secretario" (appointed official), Mr. Machel David, a local resident especially knowledgeable of the property boundaries.

On the second visit a week later, the researchers used a l:5000-scale enlargement of a 1996 aerial photo (1:10,000 original), with noticeably improved results. The local residents were actively and willingly involved in the mapping activities, but the focus was primarily on the tangible map product, not the instruction of the local people. The basic results of this first pilot effort involving seven individual properties were entered into a GIS (ArcView 3.1).

From August to October 1999 a second, expanded pilot study focused on the delimitation of community boundaries. The research included experimentation with numerous options for digital scanning and enlargements. The physical delimitation of the Bairro Mali community was completed using a combination of easy delimitations along roads (using 1:40,000-scale aerial photographs from 1989), and medium to difficult delimitations through

undifferentiated fields and woodlands (using enlargements at scales up to 1:1000, as described below). Therefore, maps produced at several scales gave far more detail than what was needed for a "delimitation" on a l:50,000-scale topographic map. Also, the nodes and lines were briefly described in a table to approximate a "descriptive narrative." An example of the Bairro Mali landscape is seen in .

The process involved approximately 30 hours of the author's time, but much of that was spent in training efforts and demonstrations of methods (compass orienteering, measured paces, and examining the photographs) to the over 20 people who participated on different days in the 1999 pilot study. At the conclusion of the mapping, the Secretario Machel David commented that he had never before seen a map that showed the extent and limits of Bairro Mali.

The pilot experiences are considered to have been highly successful. Technical details are given below with a discussion of how the three technical challenges (as stated in the Introduction) can be solved.

Challenge of Cartographic Quality

The proposed mapping method with highly enlarged aerial photographs goes against the mainstream of professional mapping practices that emphasize absolute (latitude-longitude) location precision. Because the enlarged photographs are not rectified, they have inconsistent and non-standard scales, plus the impact of radial displacement. Although those sources of error cannot be eliminated within the Mozambican cost constraints of the mapping method, five factors provide major compensation:

• Half of the challenge of obtaining acceptable cartographic quality was answered by the legal decisions in Mozambique to accept "sketch-maps" instead of requiring prime quality cadastral surveying, as discussed previously.
   
• Although the delimitation mapping is conducted at scales of 1:5000 and larger, the resultant mapping goes into a Cadastral Atlas at 1:50,000 scale, thereby reducing by cartographic generalization the error from the non-rectified image.
   
• The use of photogrammetic-quality aerial photography provides a tremendous amount of relative (metes and bounds) accuracy of the marked locations.
   
• Global positioning system (GPS) data collection can be added to the mapping database at any time before, during, or after the boundary delimitation by the local residents. However, GPS with differential correction is expensive, is not readily available, and is more in keeping with demarcation than with delimitation mapping. On the other hand, low cost non-differential GPS units have less accuracy (even if accurate to 10 meters) than the photographic accuracy of highly specific metes and bounds documentation, such as "two meters straight south of the road junction with the E-W trail that is about 35 meters north of the building in the center of Figures 2 and 3," which happens to be the location of an old cadastral cement marker in Bairro Mali.
   
• The aerial photography remains available for rectification at any later date if full photogrammetric precision is ever needed.

Challenge of Image Acquisition

Scales and Materials

Of the procedural elements in the outer circle in Figure 1, the first, second, and fourth components (national leadership and reasonably trained technicians) are either already operational or reasonably easy to develop, except for one technical requirement. That missing requirement is the provision of "highly enlarged" aerial photography, upon which the mapping method is totally dependent. "Highly enlarged" is defined here as scales of 1:5000 and larger.

Aerial photography is the base material. Black-and-white (B/W) images at a scale of 1:40,000 for both positives and negatives are available at DINAGECA in Maputo. The entire country has been photographed, but many areas only have images from before 1975.

Only occasionally are large-scale (such as 1:10,000) images of rural areas obtained, as in the case of urban fringe zones (eg., Bairro Mali) or in high-priority development areas. Therefore, for the proposed mapping method, the process of making 8X (1:5000) to 40X (1:1000) enlargements of l:40,000-scale aerial photographs is a critical issue in terms of image quality.

A crucial criterion that must be met is a pixel resolution of 1-meter or less. As shown in , a resolution of 1 meter is adequate to make individual trees, houses, trails, etc., readily identifiable at a scale of 1:2500 without the distracting pixelization effect. also identifies the focal test site that shows the Bairro Mali store (measured as eight meters on its eastern side) and the surrounding trees, road, and trails.

Image Costs

Logistically, the Mozambique archive of aerial photograph negatives is a national resource for which there are only reproduction expenses to be incurred by the community-boundaries mapping projects. A national expenditure for new photography (even at 1:80,000 scale from high altitude photography) would be desirable.

In late 1999, the cost of an aerial photograph from DINAGECA was approximately US$6.00. Enlargements of 2X and 4X cost approximately US$16 and US$64 each. Therefore, a pair of aerial photographs at 1:40,000 scale and an enlargement of one image to 1:10,000 scale costs approximately US$76.00, or slightly more than one dollar per square kilometer. This is a bearable cost, but the images are far from the needed scales, and photographic costs escalate rapidly.

An 8X enlargement of an aerial photograph from 1:40,000 to 1:5000 scale can be accomplished photographically, but at an expense of 64 times the cost of copying the single photo at its original scale. This becomes quite costly, especially when at least three identical copies are needed (for national records, technical assistance, and local mapping). Furthermore, photographic enlargements require technicians, enlargers, and darkroom facilities, all available at DINAGECA, but not sufficient in number to handle voluminous requests for highly enlarged photographs.

Alternatively, digital images provide numerous cost advantages: printing on regular paper, inexpensive multiple copies, computer storage of the images at minimal cost, scale control of printouts, and ease of selecting only the specific areas that require enlargements. For example, the digital images for the Bairro Mali pilot studies cost less than US$10.00.

Steps for Enlargement, with Options

Several steps and various options provide numerous alternatives for obtaining appropriate enlarged images. The following experiments and options have produced satisfactory results during the pilot studies:

Step A. Obtain Source Images

(1) Aerial photography (discussed above).

(2) Digital imagery: The 1-meter resolution satellite images from the IKONOS satellite will probably be appropriate, but images of Mozambique are not yet available for testing (Space Imaging, 1999). When available, these images or those from aircraft-borne digital cameras might be cost-effective substitutes.

Step B. Enlarge the Image

(1) Photographic 4X enlargement: A one-step R/W enlargement to a scale of 1:10,000 produces a single photo positive measuring almost one meter square for each of alternating (10 percent endlap) aerial photographs. Note: Smaller photographic sheets can be used if enlarging only the central areas of all of the aerial photographs in the flight line: this results in essentially the same costs for photographic materials. With economies of scale, the enlarged-photograph cost-per-square-kilometer of land surface will be very reasonable, and considerably less than re-flying the area at the desired 1:10.000 scale.

(2) Greater than 4X photographic enlargements: Not attempted because of costs.

Step C. Convert to Digital Imagery

(1) Industrial/professional digital scanning: One-step B/W commercial scanning can be done with resolutions available from 160 ìm (159 dpi) down to 5 ìm (5080 dpi). If starting with a 1:40,000-scale aerial photographic negative, scanning at 25 ìm (1016 dpi) produces a 1-meter ground resolution, appropriate for quality enlargements to the final scales of 1:5000 and 1:2500. Use of 10-ìm (2540 dpi) scanning provides a 0.4-meter ground resolution for 1:2000 or 1:1000 enlargements (IDOT, undated). An experiment using a l:10,000-scale original image scanned at 50 ìm was very successful and was used to produce the images in

One advantage of this high-resolution industrial-quality scanning is that only a single operation is used to go from the l:40,000-scale original into digital format, thereby avoiding the need for the photographic enlargement in Step B. This advantage must be weighed against the need to purchase the scanner or contract the scanner services, plus any resultant time delays for processing.

(2) Digital scanning of the l:10.000-scale image (i.e., of the 4X enlargement of the l:40.000-scale original); Good quality, standard "personal/consumer" flat-bed scanners, with resolutions of 600 dpi and even 1200 dpi, produce images that correspond to the quality of the industrial/professional scanning in Step C(l) above (see and compare directly with , noting that the photographs are different).

(3) Digital photography: Using the macro-lens feature of a Sony Mavica 88 digital camera, the author obtained high-resolution (approx. 1-meter pixels) digital imagery of small areas of the 1:10,000-scale photographic enlargements of the original l:40.000-scale aerial photographs. Although initially of interest, this method was more time consuming than the use of the flatbed scanner and caused some image distortion.

Step D. Print at Least Three Copies of Each Image (for National, Technical, and Local Uses)

(1) Personal printers: Standard ink-jet and laser printers produce analog B/W images that are quite adequate for the needed fieldwork. Also, they can be inexpensively photocopied (with some loss of image quality), thereby overcoming the problem of expensive, single-copy photographic reproduction.

(2) Large-format printers/plotters: Images on larger sheet sizes show more extensive areas, facilitate the interpretation process, and reduce the work to make large mosaics. At least sheet size A3 (11 by 17 inches) is desirable.

(3) Virtual images: Images on computer screens will have increased usefulness when computer resources can be brought to local rural areas. The digital image is especially useful when the mapping work is integrated in the GIS. The example in uses the digital photographic image as the background for the individual property boundaries. The GIS work is viewed as a task, not as an additional challenge, because even larger projects are routinely conducted by powerful geographic information systems.

Challenge of Technology Transfer to Local Residents

For over a hundred years, aerial photography has been increasingly recognized as a valuable resource for mapping (Anderson, 1982, pp. 4-6), but the photographs have primarily been used as a high-technology resource for professionals. On the other hand, educators have long known that school children and general citizens can sufficiently understand an aerial photograph to identify their local surroundings. Such local usage is photointerpretation at the preliminary and most basic levels, that of "object detection" and " recognition" (Vink. 1964: Anderson, 1982, pp. 19-26). The Mozambique mapping method proposes that use of "low-level" interpretation of photogrammetric-quality images can play a major role in a large mapping effort.

The implementation of this proposed method of mapping depends upon the ability to educate/train the local residents and their advisors to be able to conduct the mapping activities. The three levels of implementation (Figure 1) involve project leadership, technical support, and local residents. Each must be supported by tailored instruction and training. The national leadership instruction focuses on high level management and methods (including GIS applications and cost issues). Instruction for the institutional implementation is like college-level classes or technical training with both theoretical and practical learning. Both the national and regional levels are straightforward variations of standard educational practices. The third one, instructing the local residents, is the major challenge. Factors that assist in meeting this challenge of technology transfer to local residents are as follows:

• Intimate familiarity of the people with their local surroundings. Rural Mozambicans live and work outdoors, generally do not travel far (estimate five kilometers) from home, and almost invariably walk to their destinations.
• Presence of Non-Governmental Organizations (NGOs) and local primary schools (few but increasing in number) to assist with instruction and data collection.
• The relative ease of mapping the community boundaries instead of all of the individual properties (as was originally planned and tested in 1998).
• The already established use of picture-books for adult literacy education and to inform local people about the land laws (six booklets were issued by the Campanha Terra during 1998 and 1999).
• The author's substantial experience teaching photointerpretation skills in Brazil (1978-1982), including the use of distance education techniques to reach students who cannot attend regular classes.
• The favorable preliminary experiences during the pilot studies.
  

The development of the instructional methods and materials is part of the next stage of the project in early 2000.

The technology transfer for mapping can be seen also as a major opportunity to enhance the educational level of the local residents, both through the community delimitation activity and through the local school where children can have access to the mapping materials and GIS results. Geography, mathematics, and environmental biology classes can benefit from having the mapping materials available to show to children in their local communities.

In support of the previous paragraph, three map-based surveys of land-use in Great Britain can be cited (Walford, 1997). They were conducted mainly by school children, with the first done in the 1930s by Sir Dudley Stamp. The Third World circumstances in Mozambique make the intended mapping and education tasks more challenging and potentially even more beneficial than the UK experiences.

Although it is important to note that the community-boundary mapping activity does not involve the local residents with advanced issues of photointerpretation (such as classification systems, height measurement, grid systems, data analyses, and area planning), there is nothing to prevent the adults or the school children from learning about those issues.

Even if the local people only continue with "delimitations," they will gradually improve their local knowledge and local database. For example, with the local image in hand and key reference points clearly located on the ground and image, the task of collecting additional data on almost any topic with a geographic (spatial) component is as easy as marking points, lines, and areas on additional overlays on the same highly enlarged aerial photographs. Such topics include biodiversity, soils, agricultural production, infrastructure, health, community services, hazards (such as landmines in Bairro Mali), demographic data, taxation base, and all forms of planning. Considering the local, regional, and national interests in gathering these data, it is reasonable to hope that a wide variety of governmental and non-governmental entities will be supportive of the locally conducted mapping efforts.

Discussion and Conclusion

Third World development is a major international concern, and Mozambique is somewhat a model of strife-torn impoverished countries attempting to make great strides in short periods. When such countries lack the basics of development, including mapping, land-rights records, and established educational systems, the opportunities for innovative solutions are abundant, and the results could be non-standard, as in the delimitation instead of demarcation mapping.

The piloted innovative method of mapping community boundaries requires no new technologies or theories. Instead, the method utilizes many existing technologies in a combination specifically suited to the Mozambique situation. The pilot studies have shown ways to resolve the three technical challenges of cartographic precision, image enlargement, and technology transfer to local people. But in the process, the proposed method violates established practices of cartography and image utilization.

First, aerial photography was not intended to be enlarged eight to forty times. Conventional practice in the industrialized countries is to re-photograph the area at the desired scales. Third World economics prohibit such expenses, unless a wealthy donor can be convinced to pay the bill and ignore the advantages of "appropriate technologies."

Second, the international leadership of cartography has never encouraged (nor hardly tolerated) national mapping efforts outside the norms established for mapping in the industrialized and highly educated countries. Throughout the Third World, the burden of such unrealistic standards has been almost impossible to avoid. The proposed methods represent a shift away from those standards.

Third, large-coverage (even national) mapping efforts are seldom conducted by local people who lack both experience and training. In normal practice, the vast majority of mapping using aerial photography has been highly dominated by educated adults and mapping professionals who have advanced skills. But Third World realities include a scarcity of trained personnel, and even less money to pay for the numerous personnel needed to do the task. On the other hand, unemployment and under-employment of the general population mean a vast supply of available person-hours, especially when the work is in the local area and for local benefits. Furthermore, low-level technical training is a useful addition in communities trying to establish literacy and basic education to people in many age cohorts. Efforts for the proposed technical transfer can have favorable results beyond the mapping skills taught.

From conceptualization to the end of the pilot studies, the work reported here was conducted during 16 months with three brief multi-purpose stays (a total of 55 days) in Mozambique, and very few resources other than air travel expenses. The author has returned to Mozambique (August 1999 through July 2000) as a Fulbright Professor for refinement, evaluation, and implementation efforts for community delimitation mapping. The method described in this article has been presented to DINAGECA; a response is pending. Progress reports are to be available via links from the author's homepage at http://www.ilstu.edu/~psanders/.

Acknowledgments

The author performed the studies reported herein as a Fulbright Professor of Geography to Mozambique 1999-2000.

References

Anderson, Paul S. (editor), 1982, Fundamentos para Fotointerpretação, [Fundamentals for Phototinterpretation], Sociedade Brasileira de Cartografia, Rio de Janeiro, Brazil, 136 p.

Comissão de Terras, 1999, Relatorio de Actividades (1995-1999), [Report of Activities 1995-1999], Ministry of Agriculture, Maputo, Mozambique, 114 p.

Campanha Terra, 1998, Manual para Melhor Comprender a Nova Lei de Terras, [Manual for Better Understanding of the New Land Law]. Campanha Terra, Maputo, Mozambique, 20 p.

IDOT, undated, Digital Scanning Conversion Chart--Metric, Aerial Survey Section, Illinois Department of Transportation, Springfield, Illinois.

Kloeck-Jenson, Scott, 1999, Consulting Local Communities in Mozambique: Issues of Representation, Process, and Documentation, Land Tenure Center-Mozambique, Universidade Eduardo Mondlane, Maputo, Mozambique, 10 p.

Ramsay, E. Jeffress, 1999, Mozambique Country Report, Global Studies: Africa, Eighth Edition, Dushkin/McGraw Hill, Sluice Dock, Connecticut, pp. 152-154.

Space Imaging, 1999, First Image from the IKONOS Satellite Shows Washington, D.C., press release, 12 October 1999. http://www.spaceimaging.com/newsroom/releases/1999/firstimage.htm

United Nations, 1999, Human Development Report-1999, United Nations Development Program, New York, 262 p.

Vink, A.P.A., (1964), Some Thoughts on Photointerpretation, International Institute for Aerial Survey and Earth Science (ITC), Publication B 25, Enschede, The Netherlands, 20 p.

Walford. Rex (editor) 1997, Land Use-UK: A Survey for the 21st Century, The Geographical Association, UK, 160 p.

Laws, Decrees, and Regulations

Lei, 1997. Law No. 19/97, "Lei de Terras" [major sections are in Kloeck-Jenson (1999)].