Diss. (sammanfattning). Uppsala : Sveriges lantbruksuniv., 2003. / Härtill 4 uppsatser.

Thesis (M.A. in agricultural economics)--Washington State University, August 2006. / Includes bibliographical references.

Tanzania, like other developing countries in the tropics is severely affected by the degradation of water resources owing to improper land management practices. Such practices affect water supply through soil erosion which does not only cause sedimentation of rivers and water bodies but also leads to a reduction in the rainwater infiltration capacity of soils. This thesis seeks to demonstrate how the implementation of proper land management measures can reduce soil erosion and increase water supply in the Morogoro River catchment (Uluguru Mountains). The proper practices referred to are the soil and water conservation (SWC) approaches which include contour farming, fanya juu terracing and bench terracing. The thesis combines social science and geoscience methods in a synergetic manner to address this research problem. To understand how and to what degree SWC methods affect water fluxes and sediment yields, the hydrological model SWAT (Soil and Water Assessment Tool) was applied. Before carrying out the modelling procedures, it was necessary to examine the level of SWC adoption among farmers and factors influencing the process so as to establish the baseline. To this end, biophysical and socio-economic factors assumed to affect farmers’ adoption tendency were examined using a household questionnaire. Modelling results indicate that if correctly implemented contour farming, fanya juu terracing and bench terracing would significantly reduce sediment yield at different rates. The reduction would range approximately between 1% - 85% with the highest percentage change achieved by practicing the three SWC methods simultaneously. However, such SWC measures would not increase water flow annually owing to evapotranspiration losses. Nevertheless, according to modelling results groundwater storage would be increased by about 14% and hence contributing to water supply during the dry season. The household questionnaire survey suggests that the adoption of SWC methods in the study area is very low and complex. While age of the head of household, access to extension (professional) services, household annual income and proximity to the farm significantly influenced farmers’ decision to adopt SWC, gender of the head of household, slope characteristics of the farm, number of adults in the household and farmer’s perception on soil erosion effects had no considerable influence on adoption. Therefore, to successfully realize the SWC benefits demonstrated by the modelling results, smallholder farmers upstream of the catchment should be incentivized to implement proper land management practices. Payment for ecosystem services scheme appears to be a suitable strategy. To make this operational, the Tanzanian government should establish a national water fund which will finance watershed management activities. The methodological approach employed in this thesis is transferrable to other sites with problems comparable to the studied catchment.

Tanzania, like other developing countries in the tropics is severely affected by the degradation of water resources owing to improper land management practices. Such practices affect water supply through soil erosion which does not only cause sedimentation of rivers and water bodies but also leads to a reduction in the rainwater infiltration capacity of soils. This thesis seeks to demonstrate how the implementation of proper land management measures can reduce soil erosion and increase water supply in the Morogoro River catchment (Uluguru Mountains). The proper practices referred to are the soil and water conservation (SWC) approaches which include contour farming, fanya juu terracing and bench terracing. The thesis combines social science and geoscience methods in a synergetic manner to address this research problem. To understand how and to what degree SWC methods affect water fluxes and sediment yields, the hydrological model SWAT (Soil and Water Assessment Tool) was applied. Before carrying out the modelling procedures, it was necessary to examine the level of SWC adoption among farmers and factors influencing the process so as to establish the baseline. To this end, biophysical and socio-economic factors assumed to affect farmers’ adoption tendency were examined using a household questionnaire. Modelling results indicate that if correctly implemented contour farming, fanya juu terracing and bench terracing would significantly reduce sediment yield at different rates. The reduction would range approximately between 1% - 85% with the highest percentage change achieved by practicing the three SWC methods simultaneously. However, such SWC measures would not increase water flow annually owing to evapotranspiration losses. Nevertheless, according to modelling results groundwater storage would be increased by about 14% and hence contributing to water supply during the dry season. The household questionnaire survey suggests that the adoption of SWC methods in the study area is very low and complex. While age of the head of household, access to extension (professional) services, household annual income and proximity to the farm significantly influenced farmers’ decision to adopt SWC, gender of the head of household, slope characteristics of the farm, number of adults in the household and farmer’s perception on soil erosion effects had no considerable influence on adoption. Therefore, to successfully realize the SWC benefits demonstrated by the modelling results, smallholder farmers upstream of the catchment should be incentivized to implement proper land management practices. Payment for ecosystem services scheme appears to be a suitable strategy. To make this operational, the Tanzanian government should establish a national water fund which will finance watershed management activities. The methodological approach employed in this thesis is transferrable to other sites with problems comparable to the studied catchment.

No description available.

The purpose of this study was to explore the definitions, benefits, challenges, methods and perceived levels of current collaboration of Virginia’s 47 Soil and Water Conservation Districts, each a political subdivision of state government. The study was guided by the following questions (1) What is collaboration and how is it used by political subdivisions of state government? (2) What collaborative strategies are used specifically by soil and water conservation districts? (3) At what level are districts currently collaborating? (4) At what level do districts prefer to collaborate? A mixed methods research survey was used. The quantitative section measured current perceptions of collaboration based on six indicators of successful collaborations as determined and tested by the Amherst Wilder Foundation—environment, membership, process and structure, communication, purpose, and resources—through use of its Collaborative Factors Inventory. The qualitative portion allowed further exploration into how districts are utilizing collaboration at a grassroots level. Desired levels of collaboration were also captured. The entire district population—district directors, associate directors, and staff—was surveyed and responses analyzed to better understand collaborative efforts. The results indicate that collaborations occur because of both the resource benefits received and the support of a greater cause—or a mix of relational exchange and resource dependency theories. Of the six collaborative indicators, resources proved the greatest area of concern. The process and structure variable was found to be a second needed area of growth. Trust issues with key partners, a component of the membership variable, were also identified as hindering collaboration. Overall, current perceived levels of collaboration occur between coordination and coalition, or a three to four on a five point scale. However, districts identified a desire to operate more often at the coalition level. By focusing on improvement to process and structure needs as well as resource issues, trust will improve and desired levels of collaboration can be reached. This study will enrich the existing literature by expanding on the use of collaboration as it relates to political subdivisions. Findings will be of value to all conservation districts, with greatest value to Virginia. Partner agencies, policymakers, and public administrators will further benefit by gaining insights into the collaborative process.

As estradas rurais não pavimentadas são importantes geradoras de crescimento em países subdesenvolvidos, desempenhando papel essencial na produção e comercialização de produtos primários e no acesso a bens e serviços. Cerca de 78,6% do sistema de transporte brasileiro é composto por estradas não pavimentadas, sem contabilizar as estradas de uso privado de propriedades rurais. No âmbito florestal as estradas não pavimentadas tem sido o principal foco de estudos relacionados à erosão, pois são os agentes que mais contribuem para o escoamento superficial e geração de sedimentos. A aceleração dos processos erosivos proporcionada pelas estradas está vinculada às mudanças dos processos hidrológicos. A interceptação direta da chuva e do escoamento superficial e subsuperficial, a divergência e a concentração do fluxo d\'água nas estruturas das estradas são as principais mudanças ocorridas. Entre os danos ambientais causados pela introdução das estradas se destacam a redução da qualidade da água e o assoreamento dos cursos d\'água. Diante a necessidade de se entender as funções atribuídas às estradas nas relações entre o manejo florestal e a água, as estradas não pavimentadas da Estação Experimental de Ciências Florestais de Itatinga (EECFI) foram submetidas a duas análises. O trabalho foi dividido em duas partes, a primeira identificou os trechos de estradas suscetíveis à erosão (análise qualitativa) e a segunda parte estimou as taxas de produção de sedimentos e volume de enxurrada nas superfícies das estradas (análise quantitativa). Toda a malha viária da EECFI foi mapeada e digitalizada em plataformas de Sistema de Informações Geográficas (SIG). A partir do modelo digital do terreno e dos mapas de tipo e cobertura de solo foram calculados os quatro indicadores (hidrológicos e topográficos) utilizados na identificação dos trechos com maiores riscos de erosão. Por meio de métodos estatísticos multivariados, todos os indicadores de cada trecho de estrada, simultaneamente, foram divididos em grupos com características similares. A identificação dos grupos de trechos suscetíveis à erosão foi realizada pela interpretação da análise de variância (One Way ANOVA). Para a estimativa proposta na segunda parte do trabalho, foram instaladas 16 parcelas com diferentes declividades, com e sem realização de manutenção nas superfícies das estradas. As taxas de produção de sedimentos e de volume de enxurrada foram analisadas pela análise de covariância, sendo que os efeitos das variáveis e covariáveis foram analisados pelo modelo linear generalizado misto. Os resultados mostraram que tanto os indicadores quanto o método utilizado para a separação foram capazes de identificar dois grupos de trechos com características distintas quanto à suscetibilidade à erosão. Essas características foram proporcionadas pelo posicionamento das estradas em relação às curvas de nível do terreno. A declividade e a manutenção das superfícies das estradas aumentaram significativamente a produção de sedimentos, com maior destaque para as declividades. No que diz respeito ao aumento do volume de enxurrada, somente a realização da manutenção apresentou efeito significativo. A adequação das técnicas de manutenção, assim como a adoção de técnicas complementares em trechos de estradas considerados mais suscetíveis, reduzirão a perda de solo e água e os danos ambientais decorrentes. / Unpaved rural roads generate important growth in developing countries, they play an essential role in the production and marketing of primary products and access to goods and services. Approximately 78.6% of Brazil\'s transport system consists of unpaved roads, not counting the roads of private use of rural properties. In the forest sector the unpaved roads have been the main focus of studies related to erosion, as are the agents that contribute most to the runoff and sediment production. The erosion acceleration provided by roads are linked to changes in hydrological processes. The main changes are rainfall and surface and subsurface flow interception, and the divergence and the concentration of water flow on the roads\' structures. Among the environmental damage caused by the introduction of roads it\'s possible to emphasize the reduction of water quality and stream aggradation. In view of the need to understand the functions of roads in the relations between forest management and water, the unpaved roads of the Experimental Station of Forest Sciences Itatinga (EECFI) were subjected to two analyzes. The study was divided into two parts, the first identified road segments susceptible to erosion (qualitative analysis) and the second part estimated the sediments production rates and volume runoff on road surfaces (quantitative analysis). The entire road network of EECFI was mapped and digitalized using System Geographic Information (GIS) platforms. From the digital elevation model (DEM) and the soil type and land cover maps were calculated four indicators (hydrological and topographical) used in the identification of segments with greater erosion risks. Through multivariate statistical methods, all indicators of each road segment, simultaneously, were divided into groups with similar characteristics. The identification of segments groups susceptible to erosion was performed by the interpretation of the analysis of variance (one way ANOVA). For the estimative proposed in the second part, 16 plots were installed with different slopes, with and without carrying out maintenance on the surfaces of the roads. The sediment production rates and runoff volume were analyzed by analysis of covariance, and the effects of the variables and covariates were analyzed using generalized linear mixed model. The results showed that both the indicators and the method used for the separation were able to identify two segments groups with different characteristics for susceptibility to erosion. These characteristics were provided by the positioning of roads in relation to the terrain contour lines. The slope and the maintenance of road surfaces significantly increased the production of sediments, most notably the slopes. Regarding the increase of the runoff volume, only maintenance activities showed significant effect. The adjustment of maintenance techniques, as well as the adoption of complementary techniques in segments roads considered more susceptible, will reduce the soil and water losses and resulting environmental damages.

Made available in DSpace on 2016-12-23T13:51:43Z (GMT). No. of bitstreams: 1 Dissertacao - Danilo Costa Fukunaga.pdf: 8515561 bytes, checksum: ba38cee72ad050285f5ea5c0e36c53d6 (MD5) Previous issue date: 2012-08-24 / O aumento da atividade econômica e do número de habitantes no mundo, juntamente com o uso não sustentável dos recursos hídricos, os despejos de efluentes sem tratamento no ambiente e o manejo inadequado das bacias hidrográficas, têm contribuído para diminuição da disponibilidade de água no planeta. Apesar de ser um recurso natural renovável, a água que retorna da atmosfera não ocorre de maneira uniformemente distribuída, nem no espaço geográfico, nem no tempo, tornando a sua boa gestão essencial para diminuição das consequências desastrosas dos eventos extremos de cheia e de escassez. Mas, uma gestão exitosa requer, além de profissionais qualificados, o uso de ferramentas para suporte à tomada de decisão. Neste contexto, a utilização de modelos hidrológicos pode auxiliar no planejamento e avaliação de práticas conservacionistas de solo e água e de manejo de bacias hidrográficas. Do exposto, o presente estudo teve o objetivo de avaliar a aplicabilidade do SWAT para a estimação de vazões em bacias hidrográficas no sul do Estado do Espírito Santo. Foram estudadas: a microbacia do córrego Jaqueira (MBJ), com aproximadamente 22,6 ha, e a sub-bacia do rio Itapemirim, à montante da estação fluviométrica de Rive (BIMR), com aproximadamente 2.237 km². Devido ao alto número de falhas e ao curto período de monitoramento hidrológico, não foi possível calibrar e verificar a aplicabilidade do modelo para a MBJ. Para a BIMR, na análise das vazões médias, estimou-se 45,5 m³ s-1 antes da calibração, e 32,6 m³ s-1 após a calibração, muito próxima da vazão média monitorada de 36,9 m³ s-1. Para o período de validação, a vazão média monitorada foi 22% maior (45,0 m³ s-1), enquanto a estimada foi de 33,9 m³ s-1, permanecendo próxima da vazão média estimada calibrada. Em análise gráfica, observou-se que, antes da calibração, as vazões mínimas estimadas ficaram abaixo das observadas, porém as vazões de pico eram superestimadas. Após a calibração, as vazões mínimas tiveram melhores correlações, enquanto as vazões de pico foram subestimadas. Na análise dos índices estatísticos, o R² passou de 0,74 para 0,76 após a calibração, mas foi reduzido para 0,63 no período de validação. Para o coeficiente de Nash Sufcliffe, houve uma melhora significativa do valor após a calibração, passando de -0,38 para 0,75, permanecendo com um valor aceitável de 0,57 no período de validação. O SWAT foi capaz de simular vazões em bacias hidrográficas do Sul do Espírito Santo, desde que não houvesse muitas falhas nos dados monitorados, sendo capaz inclusive de simular dados de outros períodos para os quais os parâmetros não foram ajustados / The increasing of world economic activity and the number of inhabitants, along with the unsustainable use of water resources, untreated wastewater release into the environment and inadequate watersheds management, has contributed to water availability decrease on the planet. Despite being a renewable natural resource, the water that returns from the atmosphere does not occur evenly distributed, either in geographic space or in time, making its good management essential to reduce disastrous consequences of extreme events of floods and droughts. But a successful management requires, as well as qualified professionals, the use of support decision tools. In this context, the use of hydrological models can assist in planning and assessment of conservation practices for soil and water and watershed management. In this light, the present study aimed to evaluate the SWAT capability to estimate flow in river basins in the southern of Espírito Santo state, Brazil. The Jaqueira stream watershed (MBJ), with approximately 22.6 ha, and Itapemirim river basin, upstream of the Rive fluviometric station (BIMR), with approximately 2,237 km², were studied. Due to the high number of gaps and short-term hydrological monitoring, was not possible to calibrate and verify model applicability for the MBJ. For BIMR, the analysis of the average flow rate was estimated by 45.5 m³ s-1, before calibration, and 32.6 m³ s-1 after calibration, closely to the monitored average flow, that was 36.9 m³ s-1. For the validation period, the average flow monitored was 22% higher (45.0 m³ s-1), while the estimated flow was 33.9 m³ s-1, remaining closely to the average calibrated estimated flow. In graphical analysis, it was observed that before calibration, the estimated minimum flow rates were below those observed, but the peak flows were overestimated. After calibration, the minimum flows had better correlations, while the peak flows were underestimated. In the analysis of statistical indexes, the R² increased from 0.74 to 0.76, after calibration, but was reduced to 0.63 in the validation period. For the Nash Sufcliffe coefficient, was verified a significant improvement of its value after calibration, from -0.38 to 0.75, remaining with an acceptable value in the validation period, 0.57. The SWAT was able to simulate flows in river basins in the southern of Espírito Santo state, since there were not many gaps in the monitored data, and also being able to simulate data from other periods for which the parameters were not adjusted

Made available in DSpace on 2016-12-23T13:51:54Z (GMT). No. of bitstreams: 1 Introducao_Revisao.pdf: 957704 bytes, checksum: e17ebeae5294a2c7986614add6f88f96 (MD5) Previous issue date: 2012-08-24 / O aumento da atividade econômica e do número de habitantes no mundo, juntamente com o uso não sustentável dos recursos hídricos, os despejos de efluentes sem tratamento no ambiente e o manejo inadequado das bacias hidrográficas, têm contribuído para diminuição da disponibilidade de água no planeta. Apesar de ser um recurso natural renovável, a água que retorna da atmosfera não ocorre de maneira uniformemente distribuída, nem no espaço geográfico, nem no tempo, tornando a sua boa gestão essencial para diminuição das consequências desastrosas dos eventos extremos de cheia e de escassez. Mas, uma gestão exitosa requer, além de profissionais qualificados, o uso de ferramentas para suporte à tomada de decisão. Neste contexto, a utilização de modelos hidrológicos pode auxiliar no planejamento e avaliação de práticas conservacionistas de solo e água e de manejo de bacias hidrográficas. Do exposto, o presente estudo teve o objetivo de avaliar a aplicabilidade do SWAT para a estimação de vazões em bacias hidrográficas no sul do Estado do Espírito Santo. Foram estudadas: a microbacia do córrego Jaqueira (MBJ), com aproximadamente 22,6 ha, e a sub-bacia do rio Itapemirim, à montante da estação fluviométrica de Rive (BIMR), com aproximadamente 2.237 km². Devido ao alto número de falhas e ao curto período de monitoramento hidrológico, não foi possível calibrar e verificar a aplicabilidade do modelo para a MBJ. Para a BIMR, na análise das vazões médias, estimou-se 45,5 m³ s-1 antes da calibração, e 32,6 m³ s-1 após a calibração, muito próxima da vazão média monitorada de 36,9 m³ s-1. Para o período de validação, a vazão média monitorada foi 22% maior (45,0 m³ s-1), enquanto a estimada foi de 33,9 m³ s-1, permanecendo próxima da vazão média estimada calibrada. Em análise gráfica, observou-se que, antes da calibração, as vazões mínimas estimadas ficaram abaixo das observadas, porém as vazões de pico eram superestimadas. Após a calibração, as vazões mínimas tiveram melhores correlações, enquanto as vazões de pico foram subestimadas. Na análise dos índices estatísticos, o R² passou de 0,74 para 0,76 após a calibração, mas foi reduzido para 0,63 no período de validação. Para o coeficiente de Nash Sufcliffe, houve uma melhora significativa do valor após a calibração, passando de -0,38 para 0,75, permanecendo com um valor aceitável de 0,57 no período de validação. O SWAT foi capaz de simular vazões em bacias hidrográficas do Sul do Espírito Santo, desde que não houvesse muitas falhas nos dados monitorados, sendo capaz inclusive de simular dados de outros períodos para os quais os parâmetros não foram ajustados / The increasing of world economic activity and the number of inhabitants, along with the unsustainable use of water resources, untreated wastewater release into the environment and inadequate watersheds management, has contributed to water availability decrease on the planet. Despite being a renewable natural resource, the water that returns from the atmosphere does not occur evenly distributed, either in geographic space or in time, making its good management essential to reduce disastrous consequences of extreme events of floods and droughts. But a successful management requires, as well as qualified professionals, the use of support decision tools. In this context, the use of hydrological models can assist in planning and assessment of conservation practices for soil and water and watershed management. In this light, the present study aimed to evaluate the SWAT capability to estimate flow in river basins in the southern of Espírito Santo state, Brazil. The Jaqueira stream watershed (MBJ), with approximately 22.6 ha, and Itapemirim river basin, upstream of the Rive fluviometric station (BIMR), with approximately 2,237 km², were studied. Due to the high number of gaps and short-term hydrological monitoring, was not possible to calibrate and verify model applicability for the MBJ. For BIMR, the analysis of the average flow rate was estimated by 45.5 m³ s-1, before calibration, and 32.6 m³ s-1 after calibration, closely to the monitored average flow, that was 36.9 m³ s-1. For the validation period, the average flow monitored was 22% higher (45.0 m³ s-1), while the estimated flow was 33.9 m³ s-1, remaining closely to the average calibrated estimated flow. In graphical analysis, it was observed that before calibration, the estimated minimum flow rates were below those observed, but the peak flows were overestimated. After calibration, the minimum flows had better correlations, while the peak flows were underestimated. In the analysis of statistical indexes, the R² increased from 0.74 to 0.76, after calibration, but was reduced to 0.63 in the validation period. For the Nash Sufcliffe coefficient, was verified a significant improvement of its value after calibration, from -0.38 to 0.75, remaining with an acceptable value in the validation period, 0.57. The SWAT was able to simulate flows in river basins in the southern of Espírito Santo state, since there were not many gaps in the monitored data, and also being able to simulate data from other periods for which the parameters were not adjusted

As estradas rurais não pavimentadas são importantes geradoras de crescimento em países subdesenvolvidos, desempenhando papel essencial na produção e comercialização de produtos primários e no acesso a bens e serviços. Cerca de 78,6% do sistema de transporte brasileiro é composto por estradas não pavimentadas, sem contabilizar as estradas de uso privado de propriedades rurais. No âmbito florestal as estradas não pavimentadas tem sido o principal foco de estudos relacionados à erosão, pois são os agentes que mais contribuem para o escoamento superficial e geração de sedimentos. A aceleração dos processos erosivos proporcionada pelas estradas está vinculada às mudanças dos processos hidrológicos. A interceptação direta da chuva e do escoamento superficial e subsuperficial, a divergência e a concentração do fluxo d\'água nas estruturas das estradas são as principais mudanças ocorridas. Entre os danos ambientais causados pela introdução das estradas se destacam a redução da qualidade da água e o assoreamento dos cursos d\'água. Diante a necessidade de se entender as funções atribuídas às estradas nas relações entre o manejo florestal e a água, as estradas não pavimentadas da Estação Experimental de Ciências Florestais de Itatinga (EECFI) foram submetidas a duas análises. O trabalho foi dividido em duas partes, a primeira identificou os trechos de estradas suscetíveis à erosão (análise qualitativa) e a segunda parte estimou as taxas de produção de sedimentos e volume de enxurrada nas superfícies das estradas (análise quantitativa). Toda a malha viária da EECFI foi mapeada e digitalizada em plataformas de Sistema de Informações Geográficas (SIG). A partir do modelo digital do terreno e dos mapas de tipo e cobertura de solo foram calculados os quatro indicadores (hidrológicos e topográficos) utilizados na identificação dos trechos com maiores riscos de erosão. Por meio de métodos estatísticos multivariados, todos os indicadores de cada trecho de estrada, simultaneamente, foram divididos em grupos com características similares. A identificação dos grupos de trechos suscetíveis à erosão foi realizada pela interpretação da análise de variância (One Way ANOVA). Para a estimativa proposta na segunda parte do trabalho, foram instaladas 16 parcelas com diferentes declividades, com e sem realização de manutenção nas superfícies das estradas. As taxas de produção de sedimentos e de volume de enxurrada foram analisadas pela análise de covariância, sendo que os efeitos das variáveis e covariáveis foram analisados pelo modelo linear generalizado misto. Os resultados mostraram que tanto os indicadores quanto o método utilizado para a separação foram capazes de identificar dois grupos de trechos com características distintas quanto à suscetibilidade à erosão. Essas características foram proporcionadas pelo posicionamento das estradas em relação às curvas de nível do terreno. A declividade e a manutenção das superfícies das estradas aumentaram significativamente a produção de sedimentos, com maior destaque para as declividades. No que diz respeito ao aumento do volume de enxurrada, somente a realização da manutenção apresentou efeito significativo. A adequação das técnicas de manutenção, assim como a adoção de técnicas complementares em trechos de estradas considerados mais suscetíveis, reduzirão a perda de solo e água e os danos ambientais decorrentes. / Unpaved rural roads generate important growth in developing countries, they play an essential role in the production and marketing of primary products and access to goods and services. Approximately 78.6% of Brazil\'s transport system consists of unpaved roads, not counting the roads of private use of rural properties. In the forest sector the unpaved roads have been the main focus of studies related to erosion, as are the agents that contribute most to the runoff and sediment production. The erosion acceleration provided by roads are linked to changes in hydrological processes. The main changes are rainfall and surface and subsurface flow interception, and the divergence and the concentration of water flow on the roads\' structures. Among the environmental damage caused by the introduction of roads it\'s possible to emphasize the reduction of water quality and stream aggradation. In view of the need to understand the functions of roads in the relations between forest management and water, the unpaved roads of the Experimental Station of Forest Sciences Itatinga (EECFI) were subjected to two analyzes. The study was divided into two parts, the first identified road segments susceptible to erosion (qualitative analysis) and the second part estimated the sediments production rates and volume runoff on road surfaces (quantitative analysis). The entire road network of EECFI was mapped and digitalized using System Geographic Information (GIS) platforms. From the digital elevation model (DEM) and the soil type and land cover maps were calculated four indicators (hydrological and topographical) used in the identification of segments with greater erosion risks. Through multivariate statistical methods, all indicators of each road segment, simultaneously, were divided into groups with similar characteristics. The identification of segments groups susceptible to erosion was performed by the interpretation of the analysis of variance (one way ANOVA). For the estimative proposed in the second part, 16 plots were installed with different slopes, with and without carrying out maintenance on the surfaces of the roads. The sediment production rates and runoff volume were analyzed by analysis of covariance, and the effects of the variables and covariates were analyzed using generalized linear mixed model. The results showed that both the indicators and the method used for the separation were able to identify two segments groups with different characteristics for susceptibility to erosion. These characteristics were provided by the positioning of roads in relation to the terrain contour lines. The slope and the maintenance of road surfaces significantly increased the production of sediments, most notably the slopes. Regarding the increase of the runoff volume, only maintenance activities showed significant effect. The adjustment of maintenance techniques, as well as the adoption of complementary techniques in segments roads considered more susceptible, will reduce the soil and water losses and resulting environmental damages.