MODEL FOR THE EMPOWERMENT OF A LOCAL COMMUNITY BY POLICY CRITERIA TRADE-OFF ANALYSIS

Fundación Eduquemos / Universidad de Caldas / CONDESAN

Corporación ECOFONDO / COLCIENCIAS

Due to the highly degradable soils and the construction of the hydroelectric complex Miel I, the San Antonio watershed was selected as a site of research with the purpose of developing a model to evaluate critical points and intervention strategies for a conservation and development program. The information generated was integrated in a linear programming model that optimizes the net income of the rural producers, while taking into account soil losses, water supply and employment generation as sustainability and equity criteria. The aim of the optimization exercise was to find the point in which the assessment of water and sediments production would determine changes in present soil uses. Assigning values to water volume increases or sediment decreases does not encourage changes in soil use. Coffee is the system’s central driving force, as well as the crop that causes major soil losses (54 t/ha). A fall in the price for coffee up to 40% does not have an important effect on soil use in the watershed. Under these conditions it is foreseen that poverty levels will increase more than erosion. Community chances of capturing external resources aimed to generate soil use changes are low.

Keywords: models, hillsides, policy criteria, empowerment, hydroelectric plants

Decision-making on soil use priorities is made more difficult by conflicts between competitivity and agroecological sustainability criteria and the different ways whereby the productive process actors (rural producers and water and energy consumers) are benefited. The objective of this research was to develop a pilot model to assess the social benefits that could be generated by introducing soil management practices and what proportion of such benefits could come back to the producers who made it possible to evaluate the critical points of intervention in a conservation and development program in the San Antonio watershed, part of La Miel River watershed, where the energy complex Miel I is being constructed. The San Antonio River is born in the Florencia forest, a region with a special megabiodiversity and very high rainfall levels (6.750 mm).

This type of study is considered pertinent in view of the fact that: (a) this region is characterized by highly degradable soils due to high rainfall levels and slopes; (b) one of the world’s most efficient hydroelectric systems is being constructed, where sediments could have a very high opportunity cost; and (c) the rural producers have few chances of increasing their income through agricultural productivity due to adverse production factors.

To characterize the production systems, information was gathered from secondary and primary sources considering variables such as village (¨veredas¨), climate, topography, hydrography, soil uses, labor distribution, inputs uses, productivity, production values and agricultural techniques. Starting from a multivariate and principal components analysis, a clusters analysis was done in order to typify producers (Escobar & Berdegué, 1990).

Crop water consumption was estimated by means of the CROPWAT model (Smith, 1993), using secondary -source rainfall, temperature, relative humidity, wind speed and radiation information. To verify the water volume supply of the San Antonio River to La Miel watershed, two measurements were taken at different intervals during a year; the river flow speed was calculated twice a month (floater method) and its level was measured daily.

Soil losses for the different farming systems and crop types were estimated by the EPIC model (Wischmeier & Smith, 1978) and were verified through 5 run-off plots in maize, cassava, beans and fallow (Arroyave et al, 1998). The rural producers were trained in plot management and information gathering, so they were an additional strategy for involvement in the crop practices impact on the soil loss. The total solids amount was estimated using information obtained by daily measurements of water turbidity over a year period, with the aim of verifying the total sediment supply.

The ¨dam¨ model, developed by CIAT-CONDESAN (Estrada, R., 1998, pers. com.) was used to simulate sediment opportunity costs, expressed as net present value per ton of sediment, based on the benefits generated for the hydroelectric project, according to the company technical parameters (HIDROMIEL, 1997).

The information generated by the different models was integrated into a linear programming model, which optimizes the watershed producers’ net income (sales minus cash variable costs). The scenario, in which changes in the current soil use occur in order to reduce soil losses and increase the water supply to the dam, was incorporated as a sustainability criterion into the model. The soil use changes impact on employment generation as an equity criterion was also analyzed. The aim of the optimization exercise was to find the point in which assessment of water production (in two different periods) and sediment production would determine changes in present soil uses. The mental model indicates that there is competition for water and sediment production between the agricultural activity and the dam requirements.

The model’s main use was to generate information that could be used by the producers in the process of organizing themselves and negotiating their interests at conciliation meetings with the environmental policymakers and the dam company. With this aim, 8 Village Participatory Committees (VPCs), one from each watershed village was organized adjusting the methodology proposed by Ashby et al. (1996). Two conciliation meetings were set with the participation of community representatives, township authorities such as local politicians, CORPOCALDAS as the state-level environmental authority, and HIDROMIEL, the dam company, with the aim of analyzing the Florencia forest conservation strategies.

Cluster analysis resulted in four producer types—high, medium high, medium low and low areas—who developed different systems as shown in Table 1. The high area is located in the forest at an average altitude of 1352 m. It includes the largest farms and the greatest proportion of brush and forest; the owners are the oldest (55 years) and sell about 20% of the available labor. The medium high area includes the farms with the greatest slopes (205%); they include 4.4 ha of sown coffee from which they get almost all their income. The medium low area has the highest family density (102 of 253); the smaller farms (5.7 ha) and the most intensive use of labor (71 man-days/ha). They are the lowest investment farms, have the lowest income (1.6 minimum monthly wage) and the highest off-farm sale of labor. In the last three areas, coffee is the system’s driving force, representing 45% of the area used, uses 56% of the labor and generates 66% of the total farm income. The low area (located at an average 857 m altitude), is characterized by cattle raising (21 AU/farm), income diversification with sugarcane, and the youngest owners (39 years).

1 USD = $1000

It was expected that there would be intensive land use in the region as a consequence of family size and the few nonagricultural work alternatives. Nevertheless, 49% of the farm area is forest and brush, reaching almost 76% in high area farms. Under the traditional parameters of land quality, farms would not be appropriate for agricultural use due to the very strong slopes (75-300%).

Labor use is relatively intensive: 110,240 man-days in total; that is, 28 man-days/ha. Farms use the total family labor and they also hire 218 man-days. 60% of the labor were used to harvest crops, 33% for weed control and just 7% for tillaging (slash & burn system) and sowing. The traditional sowing system is minimal tillage (called ¨a chuzo¨), which means an additional effort in weed control but allows permanent land coverage.

With regard to the farmers and their families’ investment priorities, it was observed that in the last 5 years, education (US$294/yr), health (US$ 228/year) and housing improvements (US$ 203/year) are the main component of life quality. Additionally, farmers report that loans are assigned with priority to housing and facilities improvement.

A rainfall of 6918 and 7792 mm was registered during the study although the last 20-year average was 6750 mm. A period of maximum rainfall was identified between September and April (689 mm/mo) and another one of lowest rainfall between May and August (457 mm/mo).

Water production in the watershed has impact only on power generation capacity in the hydroelectric power station because all the families have their own sources of water, and there are no water lines beyond the river mouth. The CROPWAT model estimated a consumption of 7000-7500 m3/ha during the maximum rainfall period and almost 3000 m3/ha during the minimum rainfall period, these rates being similar among crops. From the watershed’s total water production of 292 millions m3, crops currently consume 28 million during the growing period. Consequently, 264 million m3/yr would be added to the river volume; that is, 84 m3/sec, without considering domestic and animal consumption. Estimated river volume value according to height and speed measurement was 7.2 m3/sec.

Sediment production related to agricultural activities is 62.254 T, that is, 16 t/ha/yr (Arroyave et al., 1998). The coffee crop causes the highest soil losses (79%) due to its nonshade planting system, low planting density, hand weed control every 3 months and trampling during harvesting and cultural practices. Subsistence crops (maize, beans, cassava, plantains) cause less than 3% of total sediment production despite the fact that they offer low coverage during part of the crop process due to the traditional sowing system. It is necessary to consider that these practices are carried out during the minimum rainfall period, the areas involved are small, and also the land is left in fallow after the harvest. Estimated soil losses in grasslands is 6% of watershed total but does not include cattle trampling losses.

The total solid and river volume measurements indicated a total sediment production at 48 t/ha. January was the lowest supply month (1.5 t/ha) and March, the highest (5.8 t/ha). Agricultural activity would mean only a third of the total solid supply to the water volume. Landslides were reported on 53% of the farms, affecting 3.248 linear m of land. Assuming a 10-m average width, 1-m depth and a 2 t/m3 density, the additional sediment supply due to landslides would be around 65,680 t; that is, nearly 17 t/ha average in the total watershed. The nonestimated third of the sediments would be related to the effects of cattle trampling and the erosion caused by the river on its bed.

According to the results of the ¨Dam¨ model, the consumer’s savings in power due to the decreased levels of sediment production become important only 83 years after the dam construction, benefits that represent US$2.56/t sediment (expressed as NPV). The model was very sensitive to the discount rate changes; the rate used was 5%, which corresponds to the normal interest rate for this type of investment.

Farmers’ net income -considering the inputs and external labor costs as outcomes and agricultural products sales and the wages earned for their own labor as incomes- is equivalent to two minimal legal wages (1 MLW = US$172). In general a positive correlation between farm size and income was observed, except for the low area which is marginal for coffee growing, where regardless the relatively greater farm size, net income is lower because of the benefits being shared with cattle owners.

To validate the linear programming model, the options were restricted only to the current farm activities and parameter solidity was verified. Results show an objective function of $3,655,000; that is, $13,000/day for family labor and $8000 for paid labor. After discounting the use for agricultural activities, the water supply to the river volume was 264 million m3/yr. The estimated sediment supply was 60,121 t/yr.

The increase of the objective function after the optimizing exercises, was only 8%, showing that soil use is well adjusted to the agroecological, technical and economical regional conditions. Main changes in soil uses refer to the reduction in grass area and increase in coffee area. Farmer logic, non-incorporated in the model and which explains such differences is based on the need of grass availability for cattle feeding intended as capital saving, risk prevention and on-farm consumption, independent of the net income economic criteria. With the proposed model structure, water production remains relatively constant, but sediment deposits increase to 63.950 t.

Model sensitivity to the assigned marginal water production values (from $5-40/m3) was minimal. The low response capacity was due to the high rainfall levels and the similarity of the estimated water consumption among different soil covers. Model sensitivity to reduced sediment deposits assigned values was greater than the water valuation for encouraging soil use changes. Table 2 shows the sediment deposition reductions when a value is assigned to it; the starting response value is about $11,000/t non-produced sediments. Consequently, important changes in soil use in the watershed should not be expected by assigning a value of $2560/t to the reduced sediment deposits, which the dam company should be willing to pay for its benefits. It is also important to consider that the valuation of non-produced sediment has a negative impact on employment generation (Figure 1).

A reduction in coffee prices of up to 40% does not affect current soil use, but it has a remarkable impact on employment and income generation (Figure 2). A way of reducing the non-produced sediment value is developing crop practices in subsistence crops and coffee area, which have a potential for reducing soil losses and at the same time capability for increasing crop yield. Assuming investments of $60,000/ha/yr for conservation practices on 751 ha along the watershed, farmers’ income could be sustained while reducing sediment production levels; it would represent $6.000/t of non-produced sediment.

The VPC's were the support for socializing the project results and making the community aware about its reality. The VPC's development reached made them decide independently to create the Association of the villages of the Florencia forest and the San Antonio Watershed (ASVESELVA) as a legal entity that would congregate and represent them. Thanks to its good will and influence, the work done by ASVESELVA facilitated the diagnosis and education process. The watershed’s productive potential, employment-generation capacity, future opportunities for developing the Florencia forest biodiversity-based enterprises (e.g., its intrinsic value, ecotourism, environmental services, sustainable agriculture, extractivity and bioprospection; Vogel, 1996) were recognized. The issues discussed during the education events led to a better positioning of the farmers in the land sales process, forcing the dam company to revise its initial appraisal and purchasing strategy. The optimization model results indicate that there are few chances for capturing external resources aimed at generating changes in soil use that would reduce the current levels of sediment production and increase the water supply for the hydroelectric project. Nevertheless, Law 99 of 1993 (environmental legislation) makes it mandatory for the power project managing company to transfer 6% of its global power sales to conservation and development activities. If yearly global power sales is estimated at US$70 million, the company should reinvest US$4.2 million, of which one million should be reinvested in the townships within the watershed. The San Antonio watershed represents 5.2% of the La Miel watershed total area (3972 out of 77,000 ha); but, according to these research results, it supplies 8.5% of the total project water (7.2 of 84.3 m3/sec). It means that the community now has a pressing mechanism for priorities defining and resources assignment. If resources are distributed according to the area, San Antonio watershed should get US$54,600/yr. However, if distribution is made based on water-volume supply, the watershed should get US$89,250/yr; that is, 63% more than the assignment based on area.

Acknowledging the mean chances of capturing resources for encouraging agricultural activities of lower impact on the dam has taken the community to prioritize other development-stimulating strategies such as education. Thanks to this approach, there is a local Environmental Educational Program, implemented by Caldas University. Additionally, the national apprenticeship service (SENA) is carrying out a training program for young peasants within the Natural Resources Qualify Workers Program.

As a further step in the local community empowerment, along with the technical information and education, farmers have been working on a proposal for civil society natural reserve formation as independent strategy for guaranteeing a sustainable preservation, restoration and development of the Florencia forest according to Panayotou (1996): natural resources as state assets become no ones property and soon become everybody’s non-resources. The proposal is supported by the same Law 99 of 1993, which gives priority to the civil society’s role in the planning and managing of natural resources. The most important results of the two conciliation meetings up to date are to have gained the acknowledgment of the community as the main actors in the natural resources and conservation and development process and to have opened a negotiation and decision-making space with community participation.

The parameters for water consumption and sediment production do not indicate a high environmental impact from the agricultural activities taking place in the San Antonio watershed. The results show an important rationale for soil management and adjustment of the production system in order to make them sustainable in the long run by reducing erosion losses: the proportion of brush and forest areas are relatively high; only the subsistence crops present poor soil cover at certain times, but the sown areas are small; the remaining crops offer a high soil coverage and last over 5 years in the rotation system. The planting system is minimal tillage, and the main tool for weed control is the ¨machete¨, which favors a permanent vegetative coverage, thereby reducing erosion problems. For both economic reasons and the present coffee crop crisis, farmers are bound to grow grasses to expand areas and increase productivity (this crop offers permanent soil coverage and a protective root system, which helps reduce erosion losses.

Assessing impact on biodiversity was not intended; however, the fact that a very high proportion of the watershed is kept as forest and brush indicate that farmers are maintaining some level of biodiversity. The former arguments should be enough to make render the decision regarding the State’s purchasing Florencia’s forestlands and the removal of the present watershed inhabitants, untenable. History shows that declaring forest resources existing prior to the II World War as state property has resulted in over 50% extinction, with very low social profit (Panayotou, 1996). It is therefore urgent to take advantage of the open spaces offered by the conciliation meetings to identify participative mechanisms for planning, execution and management of the purchased lands in order to avoid replication of negative experiences.

Chances in the crop practices suggests such a small impact on available water production that is not an attractive scenery for the dam company and, therefore, it is not an important negotiation point for farmers, even more considering their inability to control the supply to the dam project. A similar situation occurs with sediment production for the dam when soil losses are low. Besides, in order to obtain substantial changes in soil use, an investment greater than the one the company would be willing to pay (to internalize the externalizes for sediment reduction) would be required.

Current situation seems to be very stable since appears to be almost impossible to modify present crop practices. According to the model, 40% coffee price reduction would not be enough to encourage changes in soil use, due to the fact that if family labor were available, coffee growing would still be the best option for small farmers. It should be also considered that there are not production alternatives different to grasses, which means that a very impact on erosion increases could be expected from modifying crop practices. Under these conditions, is easier to foresee a poverty increase, due to the coffee price reduction, rather than erosion increase.

The landslide seems to have a greater effect on dam performance. Research carried out did not identify technical and infrastructural factor (road construction, e.g.) which determine this landslide presentation and magnitude.

The information generated for the model was a starting point for the valuation of farmer resources, the appropriation of such values, and the way as they could put negotiating terms with those who design environmental policies and who are directly benefited from the conservation process (power consumers and the project managing company). As Panayotou (1996) said, civil society organization plays a critic rol in the economy and society transition from its current inefficient and unsustainable situation to another one more efficient and sustainable. An empowered local community is already being started to consolidate in the Florencia forest thanks to education, formal organization, generated and socialized information, progressive appropriate of processes and the opening of negotiating spaces. According to Panayotou (1996), such process is useful for: a) to generate and disseminate information on causes and effects of environmental problems and negative impact of some policies, b) to increase the choice and opinion opportunities, and c) to congregate the will of social groups with low politic and economic power to press governments and to propitiate politic reforms.