Assessing the Importance of Woodland Landscape Locations for Both Local Communities and Conservation in Gorongosa and Muanza Districts , Sofala Province , Mozambique

In collaboration with two communities living in, and on the edge of, Gorongosa National Park (GNP), Mozambique, we researched the importance of different landscape units to these communities and used the information to develop a management plan for GNP. We conceived the importance of a landscape to local people as a ratio of the benefits they derive from it and the costs of accessing or using those benefits. To test this expectation, we developed Bayesian belief models, for which the parameters were the relative preference weightings derived from community members (the relative preferences for benefits and relative expectations of costs). We then collected field data to confront the models for each of the two sites. In a parallel process, we conducted a vegetation survey to generate a map of the vegetation types, as well as an index of biodiversity importance for each vegetation type of the two 20-km sites. For each site, we simplified and converted the benefit:cost model into a local community importance surface, or map, and then overlaid a conservation importance surface on it in order to identify locations that were of high importance to both conservation groups and the local community. Such areas would require careful management attention. This paper discusses the implications of the research for the planning of GNP, as well as the strengths and weaknesses of the approach. INTRODUCTION The process we used differed from other assessment procedures in two important ways. First, we sought to identify and assign relative importance scores to elements of a landscape using comparable scoring techniques. We did not attempt to identify the value of goods and services either at the margin or as stocks (e.g., Campbell et al. 1995, Lynam et al. 1994), nor did we try to value land. Second, we did not attempt to assign monetary or quasi-monetary values to the landscape units in the manner of Costanza et al. (1997), Lynam et al. (1994), and Campbell et al. (1995). We do not debate value nor do we use the term at all, as it brings with it a great number of preconceptions that are not useful in this analysis (Farber et al. 2002). Our objective was to identify and then compare the relative importance of landscape units to both local communities and conservation scientists and managers, using a simple approach and neutral units. Our approach is closer to the discoursebased valuation processes that incorporate social and equity issues (Wilson and Howarth 2002). We sought During the process of developing a management plan for Gorongosa National Park (GNP) in northern Sofala Province, Mozambique, the presence of people living within the Park and its immediate vicinity was identified as a major management problem. The major objective of the Park was to conserve ecosystems and biodiversity. Local people were recognized as users of natural resources, but Park management had set itself the objective of ensuring that the use of resources did not undermine the achievement of conservation, recreation, and knowledge-generation objectives. Little was known about the spatial patterns of resource use by local communities nor what areas were likely to be heavily impacted by community use of resources. Therefore, our research aimed to develop and test an approach for estimating local importance scores for landscape units, and then relating them to formal biodiversity conservation importance scores. Tropical Resource Ecology Program, University of Zimbabwe Ecology and Society 9(4): 1. http://www.ecologyandsociety.org/vol9/iss4/art1 first to develop spatially explicit answers to the following two questions: 1) How important is each landscape unit to the well-being of the people living in the two communities? 2) How important is each landscape unit to the conservation of vegetation diversity in these same areas? Second, we sought to do this in a way that helped us understand which factors contributed meaningfully to determining the importance of each landscape unit to the communities. Fig. 1. Map of central Mozambique, showing Gorongosa Mountain, Gorongosa National Park (GNP), and the four preliminary sites considered for further investigation; C = Canda, V = Vunduzi, N = Nhanchururu, and M = Muaredzi. Only Nhanchururu and Muaredzi were selected for the analysis. Map insert shows sites in relation to the Mozambican international border.


INTRODUCTION
The process we used differed from other assessment procedures in two important ways.First, we sought to identify and assign relative importance scores to elements of a landscape using comparable scoring techniques.We did not attempt to identify the value of goods and services either at the margin or as stocks (e.g., Campbell et al. 1995, Lynam et al. 1994), nor did we try to value land.Second, we did not attempt to assign monetary or quasi-monetary values to the landscape units in the manner of Costanza et al. (1997), Lynam et al. (1994), andCampbell et al. (1995).We do not debate value nor do we use the term at all, as it brings with it a great number of preconceptions that are not useful in this analysis (Farber et al. 2002).Our objective was to identify and then compare the relative importance of landscape units to both local communities and conservation scientists and managers, using a simple approach and neutral units.Our approach is closer to the discoursebased valuation processes that incorporate social and equity issues (Wilson and Howarth 2002).We sought During the process of developing a management plan for Gorongosa National Park (GNP) in northern Sofala Province, Mozambique, the presence of people living within the Park and its immediate vicinity was identified as a major management problem.The major objective of the Park was to conserve ecosystems and biodiversity.Local people were recognized as users of natural resources, but Park management had set itself the objective of ensuring that the use of resources did not undermine the achievement of conservation, recreation, and knowledge-generation objectives.Little was known about the spatial patterns of resource use by local communities nor what areas were likely to be heavily impacted by community use of resources.Therefore, our research aimed to develop and test an approach for estimating local importance scores for landscape units, and then relating them to formal biodiversity conservation importance scores.

PROCESS
We conducted participatory analyses in two villagescale sites (Fig. 1): Muaredzi (Appendix 1), which lies entirely within the boundaries of GNP, and Nhanchururu (Appendix 1) which straddles the boundary of GNP.We used a combination of participatory research methods, Bayesian probability modeling, and spatial data analyses of baseline digital data sets and remotely sensed images to iteratively improve our understanding of the factors determining the importance score that local people assign to specific landscape elements or locations (Appendix 2).
In parallel to this participatory process, we assessed the vegetation diversity of these same areas using standard scientific methods (Appendix 3), interpreting satellite imagery and then field sampling to validate the resultant maps and to fill in the details of species composition in each vegetation type.We scored and ranked vegetation types in order of conservation importance.Conservation importance scores were derived as a function of the relative area of each vegetation type, species diversity of each vegetation type, and the presence of key species of conservation interest.We overlaid the local landscape importance scores with the conservation importance indices to identify areas where conflicts between village use and conservation were likely to be high, i.e., where both conservation and village importance scores were both high.
Community resource use assessment teams (CRUATs) were elected by the people of each village to work with our scientific team.The analysis followed the same pattern in each site.First, for each site.we developed a prior model or hypothesis of the importance of each landscape unit to local villagers.In these models, we defined landscape unit importance as a function of the ratio of benefits derived from the unit to the costs of procuring these benefits.The greater the ratio, the more important the site.
The models were constructed as Bayesian Belief Networks (BBNs).Initial prior models were developed using the weights derived from the CRUAT to define the relative weights of benefits or costs.These models were then updated, using data collected in the field, to yield posterior models.
The CRUAT listed and scored, in terms of relative importance, the basic needs that households require for an adequate quality of life.The CRUAT then mapped the local landscape into locally identified and recognizable units and listed the goods and services that emanated from each unit.Using the scores allocated to basic needs, an index of the gross importance of a landscape unit was estimated as the weighted sum of goods and services derived from the landscape unit or location.The weightings were the local relative importance scores for each good or service.These scores were used as the prior weights in the BBNs.The cost component of the model was estimated as a function of the distance from the village to the location or landscape unit and any institutional or physical barriers that increased the labor costs of procuring or using the resources.Local estimates of the relative contributions of each of these cost components were identified and converted into spatial cost maps using a GIS.Our final estimate of landscape importance was then created as a spatial map of the benefit:cost (B:C) model.
To explore the usefulness of the model, we confronted it with real-world data.Randomly selected locations were visited by members of the CRUAT who scored each location for all model components: benefits, costs, and final importance.We used the resulting data to confront the model and update it.

Basic Needs and the Natural Environment
The livelihood systems of both villages that participated in the local valuation of our landscape functions project are dominated by natural resources-based production with very few external inputs (Tables 1 and 2).Food is derived from local agricultural production based on a tree fallow system of nutrient replenishment, from forest products, from wild foods, and from purchased commodities.The latter contribute only about 20% of the total food input, although this increases in drought or flood years.Most household basic needs are also directly derived from natural resources: houses are constructed from cut trees bound with tree fiber and roofs are thatched using grass; water is drawn from shallow ground wells or rivers.The villagers obtain cash through the sale of grain, livestock, and natural products.Nonagricultural food products become very much more important in drought and flood years, eventually supporting the household.

The Importance of Woodland Landscape Units to Local Communities
A very large number of products were used from the landscape of both village sites.We aggregated many of these into classes of product that satisfied specifically identified needs.For example, there were four different types of honey but we classed them all as "honey", in the "wild product" category.Thus, the benefit side of the local valuation was based on the supply of between 13 and 25 categories of goods.
The goods that contributed most to the importance of landscape units were water, land for agriculture and housing, construction materials (these included poles, fiber, thatching grass, and reeds), firewood, general household and craft materials (such as wood for tool handles, reeds for mat construction, or materials for constructing pestles and mortars), and various wild foods.This pattern of importance scores associated with the goods derived from natural resources is similar to those observed elsewhere in southern Africa (Cumming and Lynam 1997).Villagers collect or use resources from areas of about 300 km 2 for a village of 40 to 100 households.Again, this is a similar area to that observed elsewhere in the region (Cumming and Lynam 1997).3 and 4).Distance was not seen as a major constraint at either site as the constraints identified were dominated by the unavailability of inputs such as tools or knowledge.As an attribute of a given location, distance from the village area was the most important cost-determining factor.
Important lessons that emerged from the analysis regarding the factors governing local valuation of landscape functions or locations included the following: • Village landscapes are important for the bundles of ecosystem goods and services that people derive from each location in the landscape (Figs. 2 and 3).• In terms of predicting the importance of a given location, the preference-weighted sum of stocks of resources on a given site was a good predictor of the importance scores local people assigned to that location (Figs. 4 and  5).Costs, distance, and local (traditional) regulations and institutions did not play much of a role in determining the importance assigned to a location by local users.• Strictly enforced regulations, such as are prevalent in some areas of GNP and for some resources, did act to exclude users and thus greatly reduce the importance scores assigned to the given location.Both sites included a range of vegetation types, from open grassland areas through various savanna woodlands to thickets and forests.We identified 13 vegetation types for Muaredzi (Fig. 6, Table 5) and seven for Nhanchururu (Fig. 7, Table 6), although the total number of plant species recorded was similar for both sites (231 for Muaredzi and 246 for Nhanchururu).For both sites, it was the thicket and forest communities that were identified as being of greatest biodiversity conservation importance, both on the basis of their species composition and, particularly, because of their limited occurrence in the overall landscape.

Biodiversity Conservation Importance Scores and Potential Conflicts between Conservation and Livelihood Systems Uses
For both village areas, the thicket and forest ecosystem types had both the highest conservation importance and the highest local livelihood importance scores.These landscape units are likely to be under the greatest threat from village-level consumptive use and, thus, are where the greatest conflict is likely to occur in terms of meeting both conservation and livelihoods needs.In general, the landscape units that had the highest local importance were also those of high conservation importance (Figs. 10 and 11).There were some Ecology and Society 9(4): 1. http://www.ecologyandsociety.org/vol9/iss4/art1landscape elements that were of high importance to the community (e.g., termite mounds in Muaredzi) that we were unable to map because the resolution of the data was insufficient in relation to the size of the units.These fine-scale, localized, high importance areas are not captured in the maps we generated.The expansion of human populations in and adjacent to the Park will inevitably result in greater demands from people for agricultural land and for the resources that the Park seeks to conserve.Thus, it seems inevitable that conflict between the Park and the people whose livelihoods depend on Park's resources will intensify.Further conflict is likely to arise through the build-up of wildlife populations, such as elephants and large predators.

Confronting Conservation Importance Scores with Local Community Importance Scores
Ecology and Society 9(4): 1. http://www.ecologyandsociety.org/vol9/iss4/art1One possible solution for the Park management is to identify key ecosystem units, such as forest communities, and put in place fully enforced regulations governing the clearance of these areas for cultivation.Development of land-use zones, in collaboration with the affected local communities, may be one way of achieving this.Once these areas of both high conservation and high local resource importance have been identified, and their use regulated through zoning, co-management structures and institutions could be developed to provide sustainable multiple-use opportunities to those communities with a high dependency and capacity to manage these resource units.
As well, the Park management needs to develop and maintain functional relationships with these communities (i.e., relationships with low levels of conflict and high levels of cooperation), which will require significant management inputs.Maintaining the communities within the Park will incur additional costs, including both direct costs (e.g., the costs of maintaining rangers' posts in the vincinities of the communities), and indirect costs (e.g., increased fire incidence).For some areas or ecosystem units, these costs may be warranted, but for other areas they may not be.In such instances, GNP management may be better off seeking incentives to persuade communities to relocate voluntarily.

Key Lessons Learned from the Process
The project developed and tested a rich and relatively rapid approach for identifying the current importance of landscape units to rural communities in central Mozambique, as well as the factors underpinning their importance.The approach was shown to be capable of using spatial data where available (i.e., through base maps or aerial photography) or site sampling where spatial data were not available.
Confrontation of the prior model with field data made it clear that the costs side of the model, and hence our prior understanding of the effects of costs on local importance scores, was weak.We expect that just as individual goods and services have different benefit values so do each of them have different costs associated with their collection or use.Therefore, future iterations of the approach should seek to improve the development of the cost side of our understanding.One thing that is not clear is whether the current techniques enabled the CRUAT to separate the costs of procuring or using benefits of a landscape unit from its overall importance assignment.For example, do people mentally calculate a net importance estimate for each location (net of the costs of procurement) or do they develop a gross estimate and then evaluate the costs?Fig. 7. Vegetation map of Nhanchururu.See Table 6 for descriptions of the vegetation types associated with each mapping unit, nmu1 to nmu4.
The development of the conservation importance scores component of the assessment was, if anything, more difficult than the local community valuations.Mostly, this was because it was much more difficult to identify whose perceptions were of consequence.
There was no concentrated community to ask.In contrast, the local community, although diverse, was in one physical location and was able to develop consensus perceptions through the processes used.
Equally difficult from the conservation importance scoring perspective was the identification of importance scores for rarity or endemism.How much more important is an endemic species than a rare one?
Complete biodiversity assessments on the ground were not possible given the time and resources available.In Ecology and Society 9(4): 1. http://www.ecologyandsociety.org/vol9/iss4/art1retrospect, it would perhaps have been more efficient to use local community knowledge to develop the biodiversity estimates, using morpho-species information, rather than trying to go to species identifications.However, the problem of the importance of what to whom would still remain.Our method was weakened because of our failure to develop and use a cross-comparative reference point or importance object.We had no absolute zero or reference point to establish the relative importance scores assigned to goods, services, or landscape units across the sites.Thus, we were limited in our ability to compare the effects of such things as tenure on the importance of landscape across the two sites.
The project would not have been possible without the logistical, technical, and moral support of Roberto Zolho, Administrator of Gorongosa National Park.We are extremely grateful to him.Brit Reichelt has consistently provided us with the most remarkable administrative support in Mozambique.We are equally grateful for the very able field assistance provided by numerous field staff of the Park, particularly those of Muaredzi and Nhanchururu Posts, who facilitated our stays in the field, and played an important role in passing communications to and from Park headquarters and in passing messages to community members.
We thank the District Administrators of Muanza and Gorongosa Districts for giving their permission to implement the project within their respective districts.In Zimbabwe, we thank the administrative staff of TREP for their ongoing support to the project; Ms. Astrid Huelin for assistance with procuring and processing the satellite imagery, and Mr. Isau Bwerinofa for extensive assistance with digital mapping.
We gratefully acknowledge the financial support of CIFOR.We are most grateful to Wil de Jong for his continued encouragement and support, and to Doug Sheil and Miriam Van Heist for sharing their experiences from Indonesia and for contributing toward the initial shaping of the study.
Two anonymous referees provided helpful comments to improve the paper.

Background to Muaredzi
The Muaredzi community is situated on the north and south sides of the Muaredzi River where it joins the Urema http://www.ecologyandsociety.org/vol9/iss4/art1River, downstream of Lake Urema (Fig. A1.1).Maunza, the nearest town, is approximately 35 km to the northeast and Chitengo, the GNP headquarters, is about the same distance to the west.There is no regular transport from Muaredzi to Maunza and, other than the occasional visit by national parks staff, very few vehicles come to the village.he village area, comprising all households and fields, is relatively compact, being contained within an area of about 2 km 2 .Although we do not have a full count of people living in Muaredzi, 40 households were identified in T November 2001.These were split roughly equally north and south of the Muaredzi River.The community falls under the jurisdiction of two different Regulos.Regulo Nguinha controls the area to the north of the Muaredzi River and Regulo Nhantaze controls the area to the south.Within Muaredzi, there were four Fumos 1 .
Residents are forbidden by park regulations to venture to the west of the Urema River.The village area does not appear to have any clear boundaries to the east, south or north.eading away from Muaredzi.One leads north for some 18 km along the edge of the Urema flood plain to Goinha (also known as Muanza Baixo).The other is a path that runs for some 5 km to the south of the village, to a crossing point on the Urema River known as Jangada.
In addition to the road to Muanza, there are two other tracks l Ecology and Society 9(4): 1. http://www.ecologyandsociety.org/vol9/iss4/art1 Across the river, this connects to the road to GNP headquarters at Chitengo.Before the civil war, there was a pontoon here (hence the name Jangada), but now the only means of crossing is by a dugout canoe.

Tin a
the valley floor area.All thicket types (riverine, allu l uaredzi area, but the forest types appear to be absent.ds and clearly do cause some destruction to crops.A number of smaller animals are also commonly seen close to the village, including nyala, impala, bushbuck, oribi, warthog, and wild pig.Lake Urema is reported The Nhanchururu site is situated astride the western boundary of GNP, some 15 km southeast of Gorongosa ast of Villa Gorongosa (Fig. A1.2).It is part of the Barue Plateau, the altitude of which varies between about 200 and 340 m above sea level.The terrain is deeply dissected, with rivers ia boehmii, B. spiciformis, Erythrophloeum africanum, Julbernardia globiflora, and Pterocarpus angolensis.There are some rea is roughly rectangular in shape, about 10 km south to north and 8 km east to west.Nhanchururu is bounded to the east by the national park, to the west by Nhangeia village, to the south by Nhanthemba village, and Cynodon dactylon and Digitaria swazilandensis lawns.The latter form the bulk of the flood plains on the south and northwest sides of Lake Urema.The medium grassland largely comprises two communities, one dominated by Setaria eylesii and the other by Echinachloa stagina.The tall grasslands are characterized by a Vetiveria nigritana community, which grows to 225 cm in height.These different grassland communities occur as a mosaic that grades into the savanna areas above the flood plain.Historically, there would have been a large biomass and diversity of herbivores associated with these grasslands but, during and after the war of independence, these populations were decimated.Only small populations of mostly smaller herbivores, such as impala, now occur in the Muaredzi area.There are, however, infrequent visits to the area by hippopotami and elephants.Tinley also noted an aquatic community based on seasonally flooded pans in the flood plain.
Tinley identified six savanna woodland types growing on the rift valley floor: • Marginal flood plain woodland (Acacia albida, A. xanthophloea); • Sand savanna (Burkea africana, Terminalia sericea); • Mopane savanna (Colophospermum mopane); • Palm savanna (Hyphaene benguellensis, B ley lso identified four thicket types and two forest types from via fan, tree-base, and termitaria thickets) appear to occur in the M As it is situated within the national park, the village is exposed to wildlife.Elephant move within the village area and surroun to harbor a healthy population of crocodiles, and hippos are also present.

Background to Nhanchururu
Mountain, and some 25 km northe draining south to the Mucodza River and north or northeast to the Vunduzi River.The community lies on the upper portion of the rift escarpment, on the watershed between the Mucodza and Vunduzi Rivers.
The vegetation of the Nhanchururu area is largely miombo savanna woodland, but with some evergreen thickets on the deeper sands of the interfluve crests.The dominant woodland species are Brachysteg narrow patches of thick riverine forest along the Vunduzi and Mucodza Rivers but these are very limited in extent.
Sketch maps drawn by community members provided more specific background data for Nhanchururu.The village a to the north by Safumira village.The boundaries with adjacent villages appear to be reasonably clear.These comprise the Mucodza River to the south, the Vunduzi River to the north, and a minor drainage called the Rio Nhachituzui to the west.
To the east, the boundary between the village and the park is less clear.The community members were adamant at the entire village was outside of the park, and that the park started immediately to the east of the village, with the boundary being marked by a line of low hills and the small Rio Nhachiru.However, at the approach to the and then continues east into the park (and in former times apparently all the way through to Chitengo).There th village along the main access road from the west, shortly after entering the village area, there is an official sign stating that one is now entering Gorongosa National Park.According to this, the bulk of the village falls within the national park.Regardless of this situation, the community members seemed to feel much more secure than the Muaredzi residents, and there was never any suggestion of fears that the park may in future attempt to move them.
In terms of roads and major paths, the main access road follows the watershed between the Vunduzi and Mucodza Rivers, bisecting the village into southern and northern portions.It leads through the village to the Rangers' post, were no other significant tracks to the east.To the south, there are two routes that cross the Mucodza River, both Ecology and Society 9(4): 1. http://www.ecologyandsociety.org/vol9/iss4/art1 of which are located towards the western end of the village.One of these is a shortcut to Villa Gorongosa, if traveling by foot or bicycle.As far as vehicles are concerned, this route appears not to have been used for some time, is in a very poor state of repair, and the crossing over the Mucodza would not be passable until late into the dry season.To the west, in addition to the main access road, there is one other footpath that crosses the Rio Nhachituzui and continues to the neighboring village.To the north, there are a number of routes that lead off the main access road towards the Vunduzi River.Two of these reach to the Vunduzi, but neither appears to cross the river.
A total of 107 households were identified within the village, split roughly equally to either side of the main access road.Households tend to be scattered individually rather than clumped.Nhanchururu has four fumos.Of these, Fumo Almeida appears to be the most influential, and the other three of lesser significance.The responsible ovements occurred during the war for independence and the subsequent period of continued fighting.
Regulo lives outside of the village to the south of the Mucodza River.

Participatory Research Methods
The same approach was followed for both sites: a traditional ceremony was held first, followed by an open tive group of community informants (community resource use assessment team, or CRUAT) was selected, a modus operandi was established with the informant group and, thereafter, the unity members were told that the project team sought an improved understanding of household and community livelihoods.We explained that we wished to work with a limited group of informants, and that lsewhere (Lynam 1999(Lynam , 2001)), they are not described here.
In the initial model, the importance of a landscape unit to the community was expressed as a simple ratio of nefits divided by costs-B:C).Thus, the larger the B:C ratio, the more important the landscape unit or location was expected to be.The benefit side of the model was defined as a function of three APPENDIX 2. community meeting; then, a representa process of data collection was begun.For Muaredzi, this was achieved over a series of three field trips (September 2001, November 2001, and April 2002).For Nhanchururu, the traditional ceremony was held in April 2002 and the remainder of the activities and collection of community livelihood data were carried out during a single field trip in May 2002.
The initial community meetings provided an opportunity to explain the aims and needs of the project to those present.The comm these informants should be representative of the major socio-economic groups within the community.These representatives would form the CRUAT.In Muaredzi, the CRUAT comprised 14 men and 8 women; in Nhanchururu, 10 men and 8 women.
Three basic tools were used to conduct the analysis: spidergrams, sketch mapping, and open discussion.As each of these have been described in detail e Model Development benefits to costs (i.e., be inputs: i) the relative importance or preference for each of the goods and services (GS) derived from a given landscape unit or location; ii) the number of such GS; and iii) the density of GS per unit area in the landscape unit.Thus, the gross benefit derived from a unit of the landscape was a simple weighted sum of the importance score and density across all GS (Eq.1)., where density ranges between 0 (none) and 1 (maximum).ce was physical barriers, such as rivers, wetlands or steep terrain.The third cost-contributing source was the institutional barriers goods and services of high importance that were not governed by limiting institutions, which were close to the household or community, and which had no barriers impeding access.Landscape units or locations of D i = the d The cost component of the model was deemed to be a function of three major cost sources.First, the distance traveled to obtain the good or service, where this distance was the weighted sum of distances along major routes and distances off routes.The off-route distances would be more costly.The second cost sour or rules and regulations governing access to a given resource or landscape unit.This latter group was complicated by the elements associated with institutional costs-in the context of this project, the probability of transgressions being discovered and then the associated fine or punishment for deviations.This was simplified in the model to reflect only an opportunity cost associated with regulations-the resource use opportunities forgone due to the regulations.
The conceptual model defined our expectations of the determinants of landscape importance.Explicitly, the expectations derived from the model were that the importance of landscape units would be highest where there were multiple low importance would occur under the reverse conditions.
A computer implementation of the model was developed as a Bayesian Belief Network (BBN) using Netica (www.norsys.com).

Refinement of the Model
from the CRUATs was subsequently used to shape and update the model for each study site.In particular, this enabled the detailing of goods, services, and cost functions for each site, and the assignment of se factors.The result was the development of specific prior models for either site.These models were at a stage where, when information regarding the status of each of the peripheral nodes (goods arry out a sampling process, in order to generate field data with which to confront the model, and to provide the basis for further refinement and updating of the model.cations within each village, together with CRUAT members and, for each location, to record their scores for each of the goods and services present at the site, for all cost factors, Nhanchururu, two men and two women. Information obtained relative weights to each of the and services and cost functions) for a particular point location was input, the model would provide an estimate of the most probable landscape importance for that location.

Field Sampling for Model Confrontation
The final step in terms of collection of field data was to c The general approach was to visit a number of lo and then an overall landscape unit importance score.The scores for goods and services and for cost factors were subsequently fed into the model, and the model then generated an estimated value for each sample.These estimates were then compared with the CRUAT importance scores for each sample.
In order to increase the number of samples possible within the available time, the CRUAT group at each site was split into three or four subgroups.Each subgroup comprised several community members, plus a data recorder (facilitator).For Muaredzi, each subgroup comprised two men and four women; for Sampling was done along line transects.Each subgroup covered a single transect per day.Transects were selected

Fig. 1 .
Fig. 1.Map of central Mozambique, showing Gorongosa Mountain, Gorongosa National Park (GNP), and the four preliminary sites considered for further investigation; C = Canda, V = Vunduzi, N = Nhanchururu, and M = Muaredzi.Only Nhanchururu and Muaredzi were selected for the analysis.Map insert shows sites in relation to the Mozambican international border.

Fig. 2 .
Fig. 2. Correlation between the natural log of the benefit:cost importance calculated by the BBN model at a sample location (LNBCVALUE) and the natural logarithm of the local importance score (LNSCORE) given to that location for Muaredzi (Pearson correlation coefficient r = 0.617, n = 75).Histograms show the distributions of values for each variable.

Fig. 3 .
Fig. 3. Correlation between the natural log of the benefit:cost importance calculated by the BBN model at a sample location (LNBCVALUE) and the natural logarithm of the local importance score (LNSCORE) given to that location for Nhanchururu (Pearson correlation coefficient r = 0.727, n = 79).Histograms show the distributions of values for each variable.

Fig. 4 .
Fig. 4. Correlation between the natural log of the total benefit score at a sample location (LNGOODTOT) and the natural logarithm of the local valuation score (LNSCORE) given to that location for Muaredzi (Pearson correlation coefficient r = 0.628, n = 75).Histograms show the distributions of values for each variable.

Fig. 5 .
Fig. 5. Correlation between the natural log of the total benefit score at a sample location (LNTOTALBENEF) and the natural logarithm of the local valuation score (LNSCORE) given to that location for Nhanchururu (Pearson correlation coefficient r = 0.355, n = 82).Histograms show the distributions of values for each variable.

Fig. 6 .
Fig. 6.Vegetation map of Muaredzi.See Table5for descriptions of the vegetation types associated with each mapping unit, mmu1 to mmu7.

Fig. 8 .
Fig. 8. Three-dimensional view of the benefit:cost (B:C) surface of the Muaredzi village area taken from the southwest.The z-axis is magnified 10 times to highlight the spatial variation in predicted landscape importance.The landscape coloring represents the predicted B:C (i.e., importance) of the landscape to local community members.Highest importance units in the landscape are those in white and gold (the peak in the center of the image).Thereafter, areas in light to darker blue and then red to dark red reflect decreasing landscape importance.The major routes and tracks are marked by thin red lines, with the households of the village marked in light blue.The blue swath of the Urema River is evident in the bottom left corner and the Muaredzi River crosses from right (east) to left (west) just to the foreground side of the village area.The two light blue patches to the east of the village area (along the main road to Muanza) are patches of dry forest that are of very high importance to the community.

Fig. 9 .
Fig. 9. Three-dimensional view of the benefit:cost (B:C) surface of the Nhanchururu village area taken from the southeast.The z-axis is magnified 10 times to highlight the spatial variation in predicted landscape importance.The landscape coloring represents the predicted B:C (i.e., importance) of the landscape to local community members.Highest B:C scores are shown in white and gold, with decreasing importance scores shown by light to dark blue and then light to dark red.The black line running over the surface is the GNP boundary.

Fig. 10 .Fig. 11 .
Fig. 10.Muaredzi site, with shading showing the range in scores of the joint conservation and community use data.Major tracks, roads, and rivers are shown for reference purposes.The two highest importance patches to the east of the village area are two small dry-forest patches (Nsitu or MMU6).
People were moved from the rift valley areas of Gorongosa National Park in the 1950s to the Barue plateau area, including what is now Nhanchururu.Further disruptions and m 1 Fumos are the next level of traditional leadership down from the Regulo.

Table 1 .
Final set of goods and services that were identified by the Muaredzi CRUAT.Standardized relative importance weight (RIW) used in the Bayesian belief network (BBN).

Table 2 .
Overall list of natural resources used within Nhanchururu.Importance scores reflect the relative importance of each resource to an average household within Nhanchururu achieving an adequate standard of living.All scores are relative to the least important resources (wildlife, aquatic plants and two types of honey).

Table 3 .
Overall factors limiting access to natural resources in Muaredzi.Importance scores reflect the relative importance of each factor as regards its contribution toward limiting access to natural resources by an average household within Muaredzi village.All scores are relative to the least important factor (weakness).

Table 4 .
Overall listing of factors limiting access to natural resources in Nhanchururu.Importance scores reflect the relative importance of each factor as regards its contribution toward limiting access to natural resources by an average household within Nhanchururu.All scores are relative to the least important factor (the need to water vegetable gardens).

Table 5 .
Descriptions of the mapping units used in the vegetation map of Muaredzi with the vegetation types associated with each unit.MMU = Muaredzi Mapping Unit.

Table 6 .
Descriptions of the mapping units used in the vegetation map of Nhanchururu, with the vegetation types associated with each unit.NMU=Nhanchururu Mapping Unit. http://www.ecologyandsociety.org/vol9/iss4/art1