Repositorio Institucional de Documentos

2012 70 Mattia Trabucchi La evaluación de los servicios de los ecosistemas como herramienta para planificar la restauración ecológica de cuencas hidrográficas Departamento Director/es Geografía y Ordenación del Territorio Comín Sebastián, Francisco A. Director/es Tesis Doctoral Autor Repositorio de la Universidad de Zaragoza – Zaguan http://zaguan.unizar.es UNIVERSIDAD DE ZARAGOZA Departamento Director/es Mattia Trabucchi LA EVALUACIÓN DE LOS SERVICIOS DE LOS ECOSISTEMAS COMO HERRAMIENTA PARA PLANIFICAR LA RESTAURACIÓN ECOLÓGICA DE CUENCAS HIDROGRÁFICAS Director/es Geografía y Ordenación del Territorio Comín Sebastián, Francisco A. Tesis Doctoral Autor 2012 Repositorio de la Universidad de Zaragoza – Zaguan http://zaguan.unizar.es UNIVERSIDAD DE ZARAGOZA Departamento Director/es Director/es Tesis Doctoral Autor Repositorio de la Universidad de Zaragoza – Zaguan http://zaguan.unizar.es UNIVERSIDAD DE ZARAGOZA La evaluación de los servicios de los ecosistemas como herramienta para planificar la restauración ecológica de cuencas hidrográficas Assessment of ecosystem services as a tool for planning ecological restoration of watersheds Ph.D dissertation Mattia Trabucchi II IV La evaluación de los servicios de los ecosistemas como herramienta para planificar la restauración ecológica de cuencas hidrográficas Assessment of ecosystem services as a tool for planning ecological restoration of watersheds “Es necesario que enseñen a sus hijos, lo que nuestros hijos ya saben, que la tierra es nuestra madre. Todo lo que ocurra a la tierra, les ocurrirá también a los hijos de la tierra. Cuando los hombres escupen en el suelo, se están escupiendo así mismos. Esto es lo que sabemos: la tierra no pertenece al hombre, es el hombre el que pertenece a la tierra. Esto es lo que sabemos: todas las cosas están ligadas como la sangre que une a una familia. El sufrimiento de la tierra se convertirá en sufrimiento para los hijos de la tierra. El hombre no ha tejido la red que es la vida, solo es un hilo más de la trama. Lo que hace con la trama se lo está haciendo a sí mismo” Extracto de la carta que envió en 1855 el jefe indio Seattle de la tribu Suwamish al presidente de los Estados Unidos, Franklin Pierce en respuesta a la oferta de compra de las tierras de los Suwamish en el noroeste de los Estados Unidos A todos los becarios, investigadores, trabajadores del Instituto Pirenaico de Ecología que han hecho grande el estar en este centro estos años, por su altruismo y compañerismo, este aire que se respira aquí no se encuentra en otros sitios. VIII Resumen La provisión de servicios por los ecosistemas podría empeorar considerablemente y rápidamente durante la primera mitad del presente siglo si no se restauran eficientemente ecosistemas degradados. Frente a la aproximación clásica de la restauración basada en sistemas de referencia a imitar, existe el reto de obtener metodologías para territorio amplio y complejo y no solo para un sitio con un tipo de ecosistema. Existen muchas opciones para conservar o fortalecer servicios específicos de los ecosistemas de forma que se reduzcan las elecciones negativas que nos veamos obligados a hacer o que se creen sinergias positivas con otros servicios de los ecosistemas. En esta tesis se ha desarrollado una metodología basada en la evaluación de servicios de los ecosistemas, como variables de estado, y del riesgo de erosión, como factor de disturbio, para establecer una jerarquización espacial de actuaciones de restauración a escala de cuenca hidrográfica. Para ello se ha realizado la evaluación de servicios de los ecosistemas, modelización de la erosión y se han utilizado sistemas de información geográfica (SIG) para la elaboración de cartografía jerárquica y análisis espacial. El área de estudio utilizada es la cuenca del Río Martín (Teruel, NE España, 1938 km 2 ) como unidad funcional que, por su susceptibilidad natural a la erosión y con su elevada heterogeneidad paisajística y diferentes usos del suelo (agrícola, minería, ganadera) se presta como un valioso territorio donde aplicar y testar la metodología propuesta. La cartografía elaborada para la estimación de las tasas de erosión ha sido extrapolada con el modelo RUSLE (Ecuación de pérdida de suelo revisada) utilizando un innovador índice de vegetación (GPVI). Este índice fue elaborado mediante una técnica de inteligencia artificial llamada programación genética, la cual fue calibrada con los datos de campo del factor C de RUSLE (muestreo de suelos, transectos de vegetación) del presente estudio. Los datos de campo utilizados para crear el mapa de erosión han sido complementados con imágenes satelitales Landsat 5-TM y mapas disponibles de las características del territorio (litología, uso del suelo, ortofotos aéreas). Las tasas de erosión observadas en la cuenca del Martín tienen una media de 13.8 t ha -1 año -1 siendo notablemente mayores en la parte sur (20 t ha -1 año -1 ) debido a su irregular orografía que en las zonas de llanura del norte (10 t ha -1 año -1 ). Los servicios de los ecosistemas se evaluaron mediante indicadores obtenidos a partir de bases de datos nacionales y regionales complementados con datos de campo. Los datos son expresados para cada servicio en las unidades de medida correspondientes y se basan en el análisis de los mapas de diferentes datos físico-químicos y biológicos. Los datos de los servicios relacionados con el agua han sido proporcionados para la Confederación Hidrográfica del Ebro (CHE), los datos de acumulación de carbono en pies mayores han sido proporcionados por el Departamento de Recursos forestales del Centro de Investigación de tecnología y investigación agraria de Aragón (CITA). Los datos de acumulación de carbono en el suelo son disponibles en el Portal de Suelos Europeo (European Soil Portal). Las rutas de eco-turismo han sido descargadas de la pagina de X rutas wiki-loc y la pagina de senderos de Aragón. La retención de suelo fue modelizada combinando datos del factor C para estimar el porcentual de cobertura vegetal y las tasas de erosión del modelo RUSLE-SIG. Los servicios de los ecosistemas variaron también entre amplios y diferentes rangos. La acumulación de carbono varía entre 0 y 4648 t CO 2 eq en zonas menos densas de vegetación y 40442 y 118073 t CO 2 eq en las zonas forestales densas; la provisión de agua superficial en el norte varía entre 0 y 13 mm y 100 y 210 en el sur de la cuenca, principalmente en fondos de valles; el control de la escorrentía (recarga acuíferos) es más alto en zonas montañosas del sur de la cuenca con valores entre 8 y 81 mm año -1 con valores mínimos entre 8 y 34 mm año -1 en el norte y máximos de 81 mm año -1 en el sur; la retención del suelo se ha expresado en valores relativos que varían de 1 a 5 dependiendo de la relación entre porcentaje de cobertura vegetal y perdida del suelo (estimada por la RUSLESIG en 5 clases de muy baja a muy alta), con valor máximo de retención de suelo a coberturas mayores de 70% y erosión menor de 12 t ha -1 año -1 , y mínimo a zonas de cobertura inferior a 30% y erosión mayor de 17 t ha -1 año -1 . El servicio de eco-turismo se ha evaluado como presencia-ausencia, asignando valor 1 a las áreas de la cuenca que se observan desde los senderos usando la herramienta de visualización de cuenca en SIG (viewshed) y 0 en el resto de la cuenca que no se observa desde los senderos según el modelo digital del terreno utilizado. Tratándose de datos con unidades diferentes, entre ellos se utilizó una agrupación en el rango relativo de 1 a 5 de cada servicio por cortes naturales (Natural Breaks) en SIG, que genera clases cuyos límites se ubican donde hay diferencias relativamente grandes en los valores de los datos por cada servicio. Ecoturismo tenía un valor 0 o 1 según la ausencia o posibilidad de visualización del paisaje en el recorrer los caminos. El valor más elevado de un determinado servicio se considera un área de elevado valor definido como hotspot, que es un área de una importancia máxima para ese servicio. Análisis de solapamiento han sido realizados para entender las relaciones entre servicios. Finalmente a través de la creación de mapas jerárquicos los datos de erosión y servicios ecosistémicos han sido relacionados analizando la congruencia espacial y los patrones espaciales a diferentes escalas anidadas entre ellas, dándonos la posibilidad de analizar el comportamiento de los dos factores, y contrastar el factor de disturbio y las variables de estado a diferentes escalas espaciales. Se ha identificado la zona sur de la cuenca del área de estudio, como el área donde se presentan más servicios y se observan las tasas de erosión más altas debido a factores topográficos, entre otros. En ésta zona, y particularmente en las subcuencas con zonas mineras no restauradas (donde la erosión muestra tasas máximas y los servicios son muchas veces nulos y en subcuencas con altas tasas de erosión y alto número de servicios las acciones de restauración han de ser prioritarias si no se quieren perder servicios que benefician aguas abajo en la cuenca. Claramente según los objetivos del gestor las prioridades pueden modificarse y nuestra metodología fácilmente adaptarse. En la zona norte, llana y mayoritariamente usada para agricultura de cereal de secano, la erosión es relativamente baja y la provisión de servicios de regulación también. Es la zona de menor interés para realizar acciones de restauración dado que la mejora de los servicios no está asegurada y se podría entrar en conflicto con intereses de usos (trade off) de otros servicios (por ej., producción de alimentos) incluidos sociales. También se ha demostrado la utilidad de realizar evaluaciones a diferentes resoluciones espaciales para la mejor identificación de las zonas óptimas de restauración. Se propone un modelo conceptual general de toma de decisiones de restauración a escala de cuenca en función de la provisión de servicios de los ecosistemas y de los factores de alteración ecológica. Finalmente la metodología aquí propuesta, desarrollada con SIG con la creación de mapas jerárquicos, ha resultado fácilmente adaptable a la escala de paisaje. Esto hace que nuestro modelo dependiendo de la disponibilidad de datos, sea una herramienta útil y fácilmente aplicable para la restauración a escala de cuenca hidrográfica o de paisaje, donde los servicios ecosistémicos estén alterados por diferentes factores de disturbio. XII 1. Introduction 14 15 1.1. Ecosystem service trends in basin-scale restoration initiatives Human-induced changes and damage of the Earth’s ecosystems make ecological restoration one of the key strategies of the present and beyond (Hobbs and Harris, 2001). Restoration is vital for stemming both the current loss of biodiversity and the associated decline of ecosystem services (Dobson et al. 1997; Millenium Ecosystem Assessment (MA) 2005). The purpose of restoration is to initiate, or accelerate, the recovery of an ecosystem with respect to its health, integrity and sustainability (SER 2004). Ecological restoration and associated efforts are rapidly increasing and are being implemented throughout the world (Clewell and Aronson 2007). This growth is supported by global and regional policy commitments, such as the Convention on Biological Diversity ([article 8(f)] 2007) and the Commission of the European Community (2008), among others. Restoration can be undertaken at different scales ranging from local and habitat-specific actions to the biome and regional levels. Although small-scale short-term projects can be valuable, these experiments do not resemble real-world ecosystem management. Many authors recognize the urgent need to greatly expand the scale of ecosystem restoration and conservation (Comín 2010; MorenoMateos et al. 2012; Naveh 1994; Palmer 2009; Hobbs and Norton 1996; Wohl et al. 2005). Large-scale ecosystem restoration is required to arrest and reverse the degradation of landscapes around the world, particularly focusing on biodiversity as a positive relationship has been observed between biodiversity and ecosystem services after restoration (Rey-Benayas et al. 2009). Also focus on river systems is encouraged as increasing evidence suggests that the biodiversity of freshwater ecosystems is among the most endangered in the world (Driver et al. 2005; Dudgeon et al. 2006; Jenkins 2003; WWF 2004). The emerging policy focus on ecosystem services represents a significant shift in the objectives of restoration (Bullock et al. 2011). Economic valuation of ecosystem services has accentuated interest in using these services as a basis for restoration and conservation programs (Ehrenfeld 2000). European Environment Agency (EEA) initiated the EURECA project which is intended to contribute to a European Ecosystem Assessment is strong evidence of the institutional interest in integrating ecosystem services in future socio-economic decisions. Recent progress in the assessment and evaluation of ecosystem services is likely to increase the inclusion of ecosystem services in restoration planning and implementation (Fiedler et al. 2008; Martinez et al. 2008; Moberg and Ronnback 2003; Nelson et al. 2009; Reyers et al. 2009). While a single restoration project is unlikely to ameliorate the state of a large degraded basin, ecologists can help to identify combinations of projects that will best restore ecosystem 16 services within watersheds. To obtain a full understanding of the services provided in a study area, research should ideally be conducted at multiple, nested scales, as environmental effects may be uncorrelated across scales (MA 2003), although the large-size, long-term ecological services and functions constrain or control the small-size, periodical ecosystem services and functions (Limburg et al. 2002). Such “strategic” restoration would prioritize the location, size and type of network of restoration projects needed for a watershed that can be compared with the stakeholder needs in order for it to provide optimal levels of ecosystem services (Zedler and Kercher 2005). Biophysical and, increasingly, socio-economic values are currently used to define priority areas for planning conservation and environmental management measures (Raymond et al., 2009) as well as for evaluating the benefits of restoration projects (Aronson et al. 2010; Palmer et al. 2005). However, the degree to which ecosystem services have been incorporated into basin-scale restoration actions to date is unclear. To address this knowledge gap, we conducted a survey of peer-reviewed international scientific literature to reveal global trends. Furthermore, we explored the emerging issues related to ecosystem service classification, mapping approaches, tools and software. We identified opportunities for the increased integration of ecosystem services in basin-scale restoration projects, suggesting a framework based on new hierarchical maps. This is based on congruence among threat maps (e.g., thresholds of impacts) and ecosystem service maps. The resultant new map will facilitate the targeting of threatened service supply at different scales. The inclusion of ecosystem services in restoration projects provides an opportunity for defining clear goals for generating public support and funding sources, which are necessary conditions to enhance the planning and implementation of restoration projects (Choi 2007; Ehrenfeld 2000; Hobbs 2007). 1.1.1. Literature search and data extraction In order to understand how ES have been used in basin-scale restoration we search for peer-reviewed publications in using the ISI Web of Science from 1998-2010 (February) written in the English language, follow the methodology of Egoh et al. (2007). (http://www.newisiwebofknowledge.com). We limited our search to 1998 and beyond because this is when the terminology “ecosystem services” was introduced in the published literature by Daily (1997) and Costanza (1997). This publication, among others, created a clear increase in the number of studies citing ecosystem services (see Fig. 1 in Fisher et al. 2009). We searched for the term “restoration project” in an advanced search on ISI using the Boolean AND associated with a number of terms related to restoration (see Appendix 9.1.). For Data extraction we followed the data extraction methodology of Rey-Benayas et al. (2009) in part (see Appendix 9.1.), examining the titles and abstracts of each reference to determine how closely they aligned with our selection criterion of ecosystem services classification based on MA (2005) within basin areas, thereby determining their inclusion in this review. If the 23 making for a systematic approach that combines the rigor of small-scale studies with the breadth of broad-scale assessments (Tallis et al., 2009). The development and application of these hierarchical maps is a step in this direction, providing the opportunity to obtain an overview of the ecological state of a basin to understand and locate key ecosystem service priority areas for the purpose of maintaining, improving or restoring strategically identified targets. In these cases, the resolution of the available data is key for the downscale approach to be effective. Changing the spatial scale from a basin to prioritized areas requires optimum dataset support, depending on the scale of the target (e.g., at finer scales, a smaller pixel size will be required) to achieve more accurate targeting. Depending on the cell size of our maps, we would be able to downscale gradually from the basin to the subwatershed until we arrive at more defined and specific threatened areas (e.g., slopes, opencast mines, riparian areas, forest patches) In the next chapters we will provide a practical approach to the proposed framework for the creation of hierarchical maps based in erosion and ecosystem services maps in Martín Basin (NE Spain). 1.3. Mapping erosion risk at the basin scale with opencast coal mines to target restoration actions Restoring eroded lands is a major objective to give back value to large parts of the world where erosion is a major environmental problem (Pimentel et al. 1995). However, defining areas for restoration in a vast territory requires establishing the magnitude of the problem and the benefits of the solutions at an adequate spatial scale (Boardman 2003). Soil is often lost through erosion, a natural process that can be fostered by inappropriate land use and intense precipitation, among other factors (Garcia-Ruiz 2010). The European Union considers soil to be a nonrenewable resource, and soil degradation has strong impacts on soil and water resources (Montanarella 2000). The loss of topsoil and changes in its properties will cause the decline of the ecological processes that rely on it. Soil erosion increases the impact on streams through high sediment delivery, which has been identified as a leading cause of river degradation (USEPA 2000). Consequently, soil erosion causes the loss of the services provided by ecosystems (Van Wilgen et al. 1996) and knowing the spatial distribution of erosion rates is a primary step for planning restoration at the watershed scale. In Mediterranean areas, developing efficient tools for decision making regarding land use management is a major objective (Simoncini 2009) because of the multiple environmental problems arising from the intensive use of the land since long ago (Tabara and Ihlan 2008), 24 particularly problems related to erosion (Boardman et al. 2003, Bazzoffi 2009). Opencast mining is one such activity, which contributes mostly to erosion (Wu and Wang 2007). Opencast mines are sources of high sediment yield to rivers if restoration is not properly carried out (Balamurugan 1991; Taylor and Owens 2009). Subsequently, human intervention in failed reclamation areas, especially opencast mines with highly eroded slopes connected with the river network, is necessary to prevent water pollution and to slow irreversible erosion (Pimentel et al. 1995; Palmer et al. 2010). Mapping ecological processes and restoring areas with high sediment delivery would help avoid irreversible degradation that removes nutrients and reduces fertility (DeFries and Eshleman 2004), thus limiting the sedimentation and eutrophication of nearby rivers, which would represent a potential hazard for the long-term sustainability of agriculture and ecosystem services at the basin scale (Krauze and Wagner 2007). For this reason, the number of projects on sediment-related river restoration at the river basin scale is increasing (Kondolf 1998; Ward and Fockner 2001; Pizzuto 2002; Pennisi 2004). Successful restoration projects on river basins require an understanding of sediment transport processes. This understanding is achieved by identifying the suspended sediment sources on the basis of sediment monitoring and modeling (Gao 2008). 1.4. Mapping ecosystem services for management and targeting restoration efforts Human use and manipulation of ecosystems has increased rapidly over the last century. Currently, approximately 60% of worldwide ecosystem services are considered to be either degraded or used in an unsustainable manner (Millennium Ecosystem Assessment 2005). Agriculture and mining are vital human activities that generate essential products for human subsistence and well-being; however, both agriculture and mining have major impacts on the services provided by ecosystems (Power 2010). If we are to retain vital ecological functions, trends in ecosystem degradation need to be either halted or reversed through restoration actions (Global Footprint Network GFN, 2008; Comín 2010). Mapping ecosystem services has become a popular tool for achieving different environmental objectives. Carreño et al. (2011) assessed the tradeoffs between the provisioning of ecosystem and economic services over the course of 50 years of land-use change in Argentina. Egoh et al. (2011) identified spatial priority areas for ecosystem services in grasslands in South Africa and evaluated whether biodiversity priority areas can be aligned with those for ecosystem services. Nelson et al. (2009) used a spatially explicit modeling tool to predict changes in ecosystem services, biodiversity conservation, and commodity production levels in a United States river basin. O’ Farrell et al. (2010) engaged stakeholders and experts in identifying key services for determining the congruence between biodiversity priorities and 25 ecosystem service hotspots. The inclusion of ecosystem services in environmental research will be a major challenge and will bring multiple positive advantages. For example, promoting the variety of ecosystem services that modern agricultural systems can provide would increase the value of agricultural areas in watersheds that require restoration (Swift et al. 2004). The need for aligning restoration objectives and ecosystem services has been recognized, and a growing number of studies are offering examples at appropriate local scales where this alignment has been attempted (Trabucchi et al. Submitted). Planning the management and restoration of a region requires the identification and evaluation of the services provided by different types of land use and the prioritization of areas according to these findings. Two key issues have emerged from such planning. The first relates to the availability of data about ecosystem services (Troy and Wilson 2006). Detailed spatial information is needed to locate and quantify ecosystem services so that ecosystem services can be integrated into plans for management and restoration. The Millenium Ecosystem Assessment (MA 2005) attempted to address the lack of ecosystem service information required for decision making by assessing current knowledge, scientific literature and data. The findings of this study gave rise to the creation of ecosystem service databases at regional and national scales. The second issue pertains to recognizing the need for restoration initiatives that utilize ecosystem service information to reverse ecological degradation, recover habitats and restore biodiversity, ecological functions and services. Such restoration initiatives include erosion control, reforestation, removal of non-native species and weeds, re-vegetation of disturbed areas and the reintroduction of native species (SER 2004). The 2006 Biodiversity Communication and its detailed Action Plan (Commission of the European Community 2006) acknowledged the need for restoration initiatives within the European Union. A recent review by Rey-Benayas et al. (2009) showed that ecological restoration supports biodiversity and ecosystem services by 44 and 25%, respectively, and that increases in both biodiversity and ecosystem services were positively correlated. Ecosystem service identification and evaluation is increasingly used to locate important natural resources and services for conservation, protection, restoration and management (Egoh et al. 2008; Nelson et al. 2009; O’Farrell et al. 2010; Reyers et al. 2009; Viglizzo et al. 2011). Furthermore, this information allows for the prioritization of investments (Johnson 1995). Areas for restoration can be selected in terms of their ability to reduce environmental risks while enhancing ecosystem service delivery. Clearly, we should be developing restoration programs that explicitly state priorities or goals (Forsyth et al. 2012) in the planning stages to guide investment decisions. Spatial congruence between areas targeted for restoration and areas that deliver ecosystem services needs to be examined and possibly aligned beforehand. Multiscale analysis is especially important to Mediterranean ecosystems, which are characterized by high heterogeneity and provide society with a great diversity of ecosystem services at different 26 scales (Martín-Lopez et al. 2012). River basins consist of a mosaic of ecosystems typically classified at subwatershed levels. Planning restoration at such scales requires the prioritization of subwatersheds according to their potential for delivering benefits from the restoration. 1.5. Multi-scale approach for establishing restoration priorities in a degraded Mediterranean landscape through the evaluation of ecosystem services Soil erosion is a major threat to the continued provision of ecosystem services in large parts of the world (Brown 1981), particularly in arid and semi-arid areas (Gisladottir and Stocking 2005; García-Ruiz 2010). The future global change scenario corroborates the negative effects of increasing drought in Mediterranean regions on vegetation (Schroter et al. 2005), with runoff and sediment yields increasing in association with decreasing plant cover (from a certain threshold of cover) (Quinton et al. 1997). These suggested conditions are likely to result in greater amounts of soil being exposed to water and wind erosion (López et al. 1998). Additional factors that determine the predominance of erosion include the spatial scale, topographic thresholds, rainfall magnitude-frequency-duration characteristics, the initial soil moisture content and soil biological activity (Cammeraat 2002). Intensive agriculture and mining are land-use practices that are responsible for increasing erosion rates. These activities cause serious environmental problems across vast areas and result in enforced critical tradeoffs for the associated societies (Zhang et al. 2007; Bernhardt and Palmer 2011; Carreño et al. 2011). A key issue in semi-arid environments is determining how to prioritize areas for restoration to optimize erosion control. However, the challenge is increasingly how to combine this goal with the improved provision of vital ecosystem services, particularly water-related services and reduce the negative consequences for human development (Reynolds et al. 2007). Emerging policies are focused on ecosystem services and their inclusion in measures aimed at the restoration and the control of erosion. This represents a significant shift in the objectives of restoration (Bullock et al. 2011). Different organizations have set targets for ceasing biodiversity losses and the degradation of ecosystem services and restoring them ‘so far as feasible’ (EU Biodiversity Strategy 2020, MA 2003). To meet these policy objectives, there is growing interest in the development of tools and methods for identifying and evaluating ecosystem services and incorporating these measures into policies related to landscape planning, management and the allocation of environmental resources (Ruiz-Navarro et al. 2012; de Groot et al. 2010). This is particularly the case with regard to degraded areas and when attempting to understand trade-offs that arises related to land use and land cover planning (Rodríguez et al. 2006). 27 Mapping of ecosystem services has been identified as a useful aid in decision making during the allocation of efforts aimed at land use planning and management, particularly for the restoration of degraded areas (Reyers et al. 2009; Pert et al. 2010; Carreño et al. 2011). To obtain a complete understanding of the services provided in a study area, research should ideally be conducted at multiple, nested scales, as environmental effects may be uncorrelated across scales (MA 2003). The extent to which ecosystem services can be integrated into basinscale restoration projects that are focused on reversing these trends remains largely untested, despite the recent and growing number studies focused on this broader topic (Fisher et al. 2009). To understand how landscapes affect and are affected by biophysical and socioeconomic activities, we must be able to quantify spatial heterogeneity and its scale dependence (i.e., how patterns change with scale) (Wu 2004). Hierarchy theory is applied to the development and organization of landscape patterns and is best understood if tested across spatial and temporal scales (Bourgeron and Jensen 1993). Disturbance events that maintain landscape patterns and ecosystem sustainability are also spatial-temporal scaledependent phenomena (Turner et al. 1993). Acknowledgment of this situation is critical for the development of management strategies aimed at ecosystem sustainability (McIntosh et al. 1994). Watershed risk analysis procedures can be used to consider the effects of rehabilitation treatments on watershed-level hazards, the consequences of inaction and the resources at stake (Milne and Lewis 2011). The combined analysis of areas that are important for the supply or provision of a suite of services employing erosion maps representing multiple scales should provide useful information for the establishment of priority areas for the restoration of watersheds (Orsi et al. 2011; Su et al. 2012; Trabucchi et al. 2012b). Historic restoration efforts have been primarily focused at a single scale (such as on stands or stream reaches) (Bailey et al. 1993; Milne 1994) and have relied on site-level information to direct restoration actions (Bohn and Kershner 2002). As a result, many restoration programs lack the ability to scale up their findings. This situation has prompted the call for the adoption of a multi-scale approach in planning ecological restoration policies (Ziemer 1997; Hobbs and Harris 2001; Comín 2010). Here, each restoration activity should be evaluated across a hierarchy of scales ranging from a broad region to an individual site, as the success of a local project depends on how well that project contributes to a comprehensive restoration strategy (Ziemer 1999; Palik et al. 2000). Landscape-level empirical studies are required for determining the kinds of scaling relationships that may exist and how variable or consistent they are (Wu 2004). 28 1.6. Objectives The general aim of this study is to check an approach for targeting and prioritizing sites for land management and restoration actions based on the assessment of ecosystem services in a Mediterranean semi-arid watershed with a marked spatial distribution of eroded areas. The specific objectives are: • Modelling the erosion in Martín River Basin. • Evaluating a bundle of ecosystem services (water surface supply and flow regulation, soil retention and accumulation, carbon storage and ecotourism) and creating integrated maps of ecosystem services provision for the bundle of ecosystem services. • Elucidating the spatial patterns of erosion and ecosystem services provision in Martín Basin. • Create a spatial hierarchy of restoration actions against erosion for Martín Basin based on the evaluation of ecosystem services. 29 2. Study area 30 The Martín River watershed the Ebro River basin (NE Spain) ( Fig. 3. Map of t he Martín River part and the lower (North) part of the basin is the Escuriza River. The elevation ranges between 120 “regions” (Fig. 3) : the highlands in the south (mean elevation 1100 m, 765 km lowlands in the north (mean elevation 750 m, 1347 km established. The differences between these two zones are also marked by dams, Escuriza (Fig. 3) and Cueva Foradada (maximum water storage capacities 6 and 22 Hm respectively), which intercept sediments from the entire upstream area and river flow regime, creating a completely human climate is Mediterranean, with continental influence. The dry, and the annual average watershed is a 2112 km 2 territory l ocated in the south the Ebro River basin (NE Spain) ( Fig. 3). he Martín River Basin showing its hydrological network, and the upper (South) part and the lower (North) part of the basin , the last stream on the south east The elevation ranges between 120 and 1620 m. a.s.l. The basin shows two major : the highlands in the south (mean elevation 1100 m, 765 km lowlands in the north (mean elevation 750 m, 1347 km 2 ), where most agricultural established. The differences between these two zones are also marked by the presence of two and Cueva Foradada (maximum water storage capacities 6 and 22 Hm respectively), which intercept sediments from the entire upstream area and disturb the natural a completely human - activity altered environment downstream climate is Mediterranean, with continental influence. The summers and winters are usually and the annual average precipitation of the period 1970-2000 was 31 ocated in the south -central part of showing its hydrological network, and the upper (South) south east of the watershed shows two major : the highlands in the south (mean elevation 1100 m, 765 km 2 ) and the ), where most agricultural lands are the presence of two and Cueva Foradada (maximum water storage capacities 6 and 22 Hm 3 , disturb the natural activity altered environment downstream . The summers and winters are usually was 360 mm and is 32 heterogeneously distributed both in space and time (Fig. 4). A few big storms are recorded every summer, more frequently in the upper (south) part of the watershed. Fig. 4. Spatial rainfall pattern per year of the period 1970-2000 in Martín Basin. The water deficit ranges between 530 mm and 758 mm, extending the dry period from May until October. The mean annual temperature range is 13-16 ºC, with minimum and maximum average temperatures of 5 and 25 ºC, respectively. Dryness, which has increased in recent years (Moreno-de las Heras et al. 2009), is the main limitation for natural plant development in the region and for the development of agriculture, which is the major socioeconomic activity in the lowland part of the basin (Fig. 5), covering 53% of this part of the basin. This land is mostly used for dry cereal farming (Foto 1 p. 39), except in the narrow belts along the river’s sides near the villages, where an old canal network is still in use to irrigate vegetable and fruit tree fields. The meso-Mediterranean garrigue (Quercus ilex), accompanied by sabine (Juniperus sabina) in a few zones in the southern sector, is replaced northward by Kermes oak (Quercus coccifera), rosemary (Rosmarinus officinalis) formations, and steppe with small species (Macrochloa tenacissima, Stipa tenacissima, Ligeum spartum, Tamarix africana, Juniperus phoenicea). The only significant forests are located in the central part of the basin, and they consist mainly of Aleppo pine (Pinus halepensis). Riparian vegetation is extremely degraded because of the extensive cover of agricultural practices, and the intensive effects of some mining derived impacts and regulated river flows. 39 PICTURES OF THE STUDY AREA Foto 1. Tipos de zonas agrícolas frecuentes en la parte norte de la cuenca del Martín: Agricultura mecanizada con métodos tradicionales (en sentido horario desde la izquierda arriba) dos campos utilizados para secano que durante el invierno se quedan completamente expuestos a los agentes erosivos (Escatrón, Hijar). Campo dedicado a secano protegido residuos de cultivo cerca de Albalate del Arzobispo. 40 Foto 2. Diferentes paisajes dominantes en la parte central de la cuenca del Martín. En el sentido horario desde arriba a la izquierda: Olivares, campos de almendros y viñas; paisaje prevalentemente agrícola con una mayor componente natural de matorral; zona más rocosa y abrupta dominada por bajo matorral a la embocadura del embalse de Cueva Foradada conocida por su alto valor recreativo al interior del Parque Cultural del Río Martín; paisaje agrícola en Alloza. 41 Foto 3. Paisajes típicos de la parte sur de la cuenca (en sentido horario desde arriba e izquierda): Penyarroya, una atracción natural del Parque Cultural del Martín, se aprecian los bosques riparios formados por caducifolias; zona forestal en el municipio de Utrillas. “Humanización” del río en Obón, el bosque de ribera ha sido substituido por cultivos y caminos; zonas encañonadas del Río Martín, nótese los fenómenos de desprendimiento y acumulación de roca en las laderas de los montes 43 3. Methods 45 3.1. Estimating erosion with RUSLE-GIS model Erosion rates have been estimated at the regional scale using the RUSLE model (Fu et al. 2005, Onori et al. 2006; Pizzuto 2002; Pennisi 2004). European environmental researchers (Panagos et al. 2011) have recently mapped a soil erodibility dataset at the European scale. The objective was to overcome problems of limited data availability for the application of the USLE (Universal Soil Loss Equation) model and to present a high quality resource for modelers who aim to estimate soil erosion at the local/regional, national or European scale. Following this direction, the location of eroded areas and the estimation of the average annual soil loss from rill and sheet erosion in the Martín Basin (Ebro Basin, Northeast Spain) were determined by using the RUSLE (Renard et al. 1997) and an updated version of USLE (Wischmeier and Smith 1978), coupled with GIS (Geographic Information System). Many authors have used GIS/RUSLE models to estimate sheet wash erosion and non-point source material discharges in watersheds (Fu et al. 2005; Lim et al. 2005; Smith et al. 2007) and for environmental assessment (Boellstorff and Benito, 2005; Erdogan et al. 2007; Ozcan et al. 2008). An increasing number of studies on restoration ecology are using this model to identify potential restoration areas (Güneralp et al. 2003; Vellidis et al. 2003) and to design reclamation plans for degraded areas such as opencast mines (Toy et al. 1999; Martín-Moreno et al. 2008; Moreno-de las Heras et al. 2009). Despite some uncertainties regarding RUSLE, such as the overestimation of soil loss on plots with low erosion rates and the underestimation of soil loss on plots with high erosion rates (Nearing 1998; Risse et al. 1993), we decided to use this model because it requires data that are relatively common and inexpensive to be processed with GIS. One of the highlights is the formulation of results that can be used for comparative or complementary future studies (Millward and Mersey 1999; Wang et al. 2003; Beguería 2006). 3.1.1. The RUSLE model We used GIS commercial software (using a Spatial Analyst tool) to examine spatial variations in erosion using elevation data at a 20-m grid scale within the study area. Digital land cover data are available as shape files at the Aragon Territorial Information Centre (CINTA 2006). The Universal Soil Loss Equation (USLE) was used for this study because it is the most used empirical model that assesses long-term averages of sheet and rill erosion. This model is based on plot data collected in the USA (Wischmeier and Smith 1978). The USLE and its adapted version RUSLE (Renard et al. 1997) have been applied to various spatial scales and region sizes in different environments worldwide (Vrieling 2008). 46 The USLE and RUSLE are statistically based water erosion models related to six erosion factors (for a detailed description of the factors and data collection methods, see the appendix at points 9.3 and 9.4): A = R * K * L * S * C * P Where: A is the average soil loss from sheet and rill erosion, reported here in tons per hectare per year (t ha −1 yr −1 ) (Fig. 17, p. 65). R is the rainfall-runoff factor and represents the erosion energy in MJ mm ha −1 h −1 yr −1 based on the methodology of Renard et al. (1997), and it represents the average annual summation (EI) values in a normal year's rainfall (Fig. 12 B). K is the soil erodibility factor, which represents both the susceptibility of soil to erosion and the rate of runoff, as measured under the standard unit plot condition expressed in (t h MJ −1 mm −1 ) (Renard et al. 1991) (Fig. 12 A). Only R and K have units; those units, multiplied together, give erosion in units of mass per area and time. Each of the other terms scales the erosion relative to specified experimental conditions (>1 is faster than erosion under those experimental conditions, and <1 is slower). The remaining factors are non-dimensional scaling factors. LS is the topographic factor describing the combined effect of slope length and steepness and is calculated with the approach of Moore and Wilson (1992) (Fig. 13 C), Fig. 12. Input data derived from the database of the Martín watershed: A) soil erodibility map (K-factor in RUSLE, Mg h MJ -1 yr -1 ). Input data derived from the database of the Martín watershed: A) soil erodibility map mm -1 ); B) rainfall erosivity map (Rfactor in RUSLE, MJ mm ha 47 Input data derived from the database of the Martín watershed: A) soil erodibility map factor in RUSLE, MJ mm ha -1 h -1 Fig. 13. Input data derived from the database of the Martín watershed: C) length slope factor (LSfactor en RUSLE) and D) crop management. Input data derived from the database of the Martín watershed: C) length slope factor factor en RUSLE) and D) crop management. 48 Input data derived from the database of the Martín watershed: C) length slope factor 55 3.2.4. Carbon storage in woody vegetation The amount of carbon stored and its fixation rate was mapped across a large region, which included the Martín Basin, by the Agrifood Research and Technology Centre of Aragon (CITA unpublished,http://www.aragon.es/estaticos/GobiernoAragon/Departamentos/MedioAmbiente/Areas/03_Cambio_cli matico/06_Proyectos_actuaciones_Emisiones_GEI/estudio.pdf). This report focused on modeling different forest management alternatives for CO 2 sequestration, such as woody vegetation, and understanding the role of forests as CO 2 sinks. The method used estimates of biomass and CO 2 conversion using allometric equations (Montero et al., 2005) and data on tree diameters measured during the National Forest Inventory (IFN3 2005). Allometric equations related the diameter of a single tree species to the dry matter existing in different fractions or parts of the tree, i.e., the trunk, roots, leaves and branches of three different sizes. The information, which was linked to the sampling points of the National Forest Inventory, was extrapolated to surface units using the comprehensive 1:50.000 Spanish Forest Map (developed in coordination with the Third Spanish National Forest Inventory). GIS data layers for storage and sequestration rate, expressed in metric tons of CO 2 equivalent (t CO 2 eq), were available for the Martín River Basin in this cited report. The GIS layers were extracted as a polygon layer and converted to a raster layer to facilitate calculation (Fig. 16 E). 3.2.5. Potential soil retention Soil erosion represents a hazard for the long-term sustainability of agriculture and the delivery of ecosystem services (Hajjar et al. 2008). Reduced soil retention results in increased sediment delivery to freshwater systems and degrades these systems (Gobin et al. 2004). Natural vegetation enhances soil retention and plays a vital role in ameliorating the impact of erosion on freshwater systems (Reyers et al. 2009). Quinton et al. (1997) found that a decrease in soil loss was particularly notable when the percentage of vegetation cover increased from 0 to 30% but there was little difference in the soil loss after vegetation cover values exceeded 70%. Trabucchi et al. (2012a) mapped erosion risk in the Martín Basin (expressed in t ha -1 yr -1 ) using the RUSLE model (Renard et al. 1997). To extrapolate vegetation percentage cover, we used the cover factor of the RUSLE model, called the C factor (see Appendix 9.1), which is the covermanagement term that represents the prior land use, crop canopy and surface cover (Renard et al. 1991) of our study area. Following the methods of Egoh et al. (2008), soil retention was mapped as a function of vegetation cover (%) and soil erosion estimations. Based on these 56 data, vegetation cover densities were distributed in three classes: 0-30%, 30-70% and 70-100% (Fig. 15 D). Areas with vegetation cover greater than 30% and classified as having a very low to low erosion value were defined as having a potential to retain soil. A soil retention hotspot was defined as having a plant cover density greater than 70% with very low to low erosion values. Zones with cover densities of less than 30% and high to very high soil erosion values were extracted and identified as erosion-prone areas. 3.2.6. Soil formation Accumulation of soil organic matter is an important process for soil formation and can be easily altered by habitat degradation and transformation (de Groot et al. 2002; Yuan et al. 2006). Organic carbon content (OCTOP) (%) in the topsoil layer (0-30 cm) was mapped by Jones et al. (2005) for the European Soil Database using a 1 km resolution grid cell (Fig. 16 F). Data were expressed as a percentage weight of organic carbon in the surface horizon by combining refined pedotransfer rules with spatial-thematic data layers of land cover and temperature. We used these data as a surrogate measure for the supporting ecosystem-service soil formation. Areas with a high organic content (>3.45%) were classified as hotspots. 3.2.7. Potential recreation and ecotourism services Landscape as a visual experience holds considerable societal value. For rural tourism, the landscape is often the main attraction and can add significantly to the quality of life of the surrounding residents (Brabyn and Mark 2011). Agriculture and cattle breeding have historically been the most important social and economic activities in the Martín River Basin, with rural society taking shape around the agricultural and livestock cycles. Mining activities during the second half of the 20 th century not only changed the way of life in these rural communities, but it also changed the landscape in many parts of the watershed, particularly in the southern highlands. Since the end of the last century, many efforts have been made to promote tourism in the study area, which is rich in both natural and cultural resources. The basin is popular for its wide open spaces, scenery and the presence of the Martín River Cultural Park (http://www.parqueriomartin.com/en/), which is rich in both cultural heritage, including cave paintings, Iberian settlements and historical monuments, and natural sites, including caves, ravine waterfalls and mountain peaks. All of these cultural and natural sites are on hiking and mountain biking routes. The track locations were downloaded from Wikiloc (2011) and from the official web page of routes in Aragon (Senderos de Aragón 2011). The viewshed tool in ArcGIS (Environmental Systems Research Institute 2008) was used on the selected routes to calculate the potential viewing area (Fig. 14 A), which is important for providing an attractive visible environment for tourists (Reyers et al. 2009). The resultant maps were included as hotspot production areas following the methodology of O’Farrell et al. (2010). 57 While we acknowledge that many other cultural aspects and values exist within this region, these tourism routes and viewsheds capture the potential for attracting visitors and providing socio-economic benefits to the local populations, which are key factors for socio-economic development and could have a major regulating impact on the area. 3.2.8. Mapping spatial distribution of services and hotspots at basin and subwatershed scale Maps of the selected ecosystem services were created following the methods of Egoh et al. (2008) and O’Farrell et al. (2010). In this study, data on surface water supply, flow regulation and soil formation had spatially continuous values that covered the whole basin, while data on the other services had spatially discrete values (e.g., the woody carbon storage layer was limited to forested areas and all other values were considered to be 0). The original values of the ecosystem services in generally had a Poisson distribution, each map were reclassified into five classes that were determined using a Natural Breaks (O’ Farrell et al. 2010) were generated classes are based on natural groupings inherent in the data. Class breaks are identified that best group similar values and that maximize the differences between classes. The features are divided into classes whose boundaries are set where there are relatively big differences in the data values (Environmental System Research Institute 2008). These five classes were renamed as very high, high, medium, low and very low. We assigned the value of 0 to the very low class of surface water supply, flow regulation and soil formation to avoid overlapping these services for the entire area because insignificant values mask potentially interesting results. The rest of the services of our suite have not been modified because they have a lower spatial distribution and include areas with no service flow at all (e.g., carbon storage is limited only in forested areas). Finally, service layers were overlapped one by one, and overlapping percentages were used to describe the spatial relationships between these services. Hotspot maps were created for every single ecosystem service to identify, manage and conserve high service flow areas by extracting high and very high service values. In addition, multiple hotspot zones among services were identified and established by overlapping the hotspot layers of each of the different services following the methods of Egoh et al. (2008). Services were then generalized to the forth order catchments, which attempted to highlight the richness of services in every subwatershed by defining areas of land that are drained by a stretch of river of lower order than the main Martín River system. Sixty seven subwatersheds were distinguished in the Martín Basin. To identify service values for the subwatersheds (Fig. 22 B, p. 74), we utilized basin service maps using the GIS Spatial Analyst-Zonal Statistic tool (Environmental System Research Institute, 2008) and selected the majority statistical option (ArcGis resource center 2012), which determines the value that occurs most often out of all 58 cells in the input in_value_raster that belongs to the same zone as the output cell. In our case, the majority statistical option attributes to every subwatershed the most frequent value of overlapping services for all of the cells in that subwatershed. When equal numbers of cells within a subwatershed received the highest and the second highest value, the lower value was assigned to the subwatershed. Despite this limitation, it is still considered to be the best statistical option for creating a general overview (Wu 2004). Following this overview for the whole Martín Basin (Fig. 22 A, p.74) and hotspot areas (Fig. 22 C, p. 74), the extraction of detailed overlapped-services maps (Fig. 22 C, p. 74) at the subwatershed scale was conducted. The same Zonal tool using the statistical majority option was applied at a subwatershed scale to select hotspot subwatersheds by the number of overlapped hotspot services (Fig. 22 B, D, p.74). This process of downscaling facilitates the selection of areas in the region that are particularly vulnerable to environmental degradation and have a high supply of ecosystem services. We extracted from the erosion map generated by Trabucchi et al. (2012a), the mean erosion value for every subwatershed of the basin using zonal statistics with GIS. We then reclassified the erosion values and generated a new degradation map. Reclassification of this map was based on thresholds for soil formation in the study area defined as lightly (0-12 t ha -1 yr -1 ) (Rojo 1990), medium (12-17 t ha -1 yr -1 ) and highly (>17 t ha -1 yr -1 ) (Moreno-de las Heras et al. 2011) degradation level ( Fig. 23 left, p. 76). This allows us to label subwatersheds according to the provisioning of ecosystem services and degradation status, establish a relative ranking of priorities for restoration actions to recover lost and degraded ecosystem service provisions. Table 1 includes the criteria to prioritize subwatersheds for restoration based on the combination of ecosystem service delivery and environmental risk of erosion on Martín Basin. Our priority is where already service flow and erosion are high because there is an elevated risk of losing these vital services if erosion is not counteracted with restoration/management actions and where restoration can work improving the delivery of ecosystem services. 59 Table 1. Combined ecosystem services delivery and environmental risk criteria for establishing priority areas for restoration in Martín Basin. Environmental risk (erosion) → ------------------------ Ecosystem service delivery ↓ Low High High Very low priority High priority Low Tertiary priority Secondary priority 3.2.9. Soil erosion priority areas Scale-dependent disturbance dynamics have several important implications for land management (Turner et al. 1994). Martín Basin, as many areas in Spain is affected by erosion due to long history of deforestation, cattle grazing and mining (García-Ruiz 2010). Vegetation growth in the region is limited by semi arid condition (García-Fayos and Bochet 2009; Morenode las Heras 2011). Natural ecosystems play a vital role in ameliorating these impacts by retaining soils and preventing soil erosion. Erosion is counteracted mainly by structural aspects of ecosystems, especially vegetation cover and root systems (Gyssels et al. 2005) that can be stimulated with restoration actions, creating synergy among services (Bennett et al. 2009). As example, soil retention can stimulate soil accumulation service that will contribute in the maintenance of water quality in nearby water bodies (de Groot et al. 2002) among many others. Areas requiring these services are those vulnerable to erosion, as determined by the topography, rainfall, soil depth, and texture. Trabucchi et al. (2012a) mapped erosion risk using the RUSLE model in the study area at 20m cell size resolution which is recognized as the most appropriate scale for estimate soil loss in semiarid areas (Ruiz-Navarro et al. 2012). Reclassification of this map was based on thresholds for soil formation in the study area defined as lightly (0-12 t ha -1 yr -1 ) (Rojo 1990), medium (12-17 t ha -1 yr -1 ) and highly (>17 t ha -1 yr -1 ) (Moreno-de las Heras et al. 2011) degradation level ( Fig. 23 right, p.76). Data were extended for every subwatershed as mean using zonal statistics tool. The belonging at one of the three categories above established automatically classified subwatershed of the basin as Low, Medium and High erosion grade. 60 3.3. Regional multi-scale spatial analysis 3.3.1. Delineation of subwatersheds among different spatial aggregation levels To perform a multi-scale analysis of erosion and ecosystem services, we distributed the basic information on these variables, available at a 20 m cell size, at three levels, or scales of aggregation, moving gradually towards a finer resolution. We used the ARCGIS watershed tool to perform this analysis. Following this approach, we created three drainage networks for the Martín Basin with different numbers of subwatersheds, which are described here. We use three pixel spatial aggregations suitable for prioritization restoration actions, specifying the limit of pixels for flow accumulation, these being 20000 (level 1), 2000 (level 2) and 1000 (level 3 ). The spatial arrangement of the Martín Basin at subwatershed level 1 contained 67 subwatersheds (Fig. 24 A left, p. 79), which presented a minimum area of 1.27 Km 2 , a maximum of 120.9 Km 2 and an average of 28 Km 2 . The second subwatershed, level 2, included 655 subwatersheds (Fig. 24B left, p.79), with a minimum area of 0.007 Km 2 , a maximum of 12.1 Km 2 and an average of 2.87 Km 2 . Finally, subwatershed level 3 consisted of 2534 subwatersheds (Fig. 24 C left, p. 79), with a minimum area of 0.006 Km 2 , a maximum of 4.15 Km 2 and an average of 0.75 Km 2 . These subwatersheds are the functional ecological units for the delivery of the majority of our selected suite of ecosystem services, determining erosion dynamics and planning of restoration actions. Classifying assessment units directly assists in resource management, including restoration. Ecosystem service bundles and erosion maps were reclassified and summarized for every subwatershed level to create a new prioritization classification consisting of a combination of erosion rate thresholds and a number of services (Fig. 2, p. 22) . 3.4. Comparison of management units To investigate service delivery and erosion at the finest scale, we selected two subwatersheds from the first level presenting contrasting topographic features and land use practices as a case study. Our selection was made to facilitate the assessment and utility of our multi-spatial level approach for prioritizing restoration measures. Subwatershed number 4, located in the northern lowland region and subwatershed number 63, located in the south mountainous area (Fig. 25 A, p.80), were selected for this analysis. They were further investigated at the second and third levels (Fig. 31, p.88) to determine the optimal management area for planning restoration policies and to develop an understanding of how patterns of congruence change 61 with scale. Subwatershed number 4 (Foto 1, p.39) is a fairly homogeneous area that is mostly used for dryland and irrigation agriculture but also contains some remnant patches of shrubland. The erosion rate here was calculated to be 0.2 ± 64 t ha -1 year -1 . In contrast, subwatershed number 63 contains a mix of conifer and hardwood forest, shrubs, grasslandscrublands, abandoned and restored mines (Fig. 9, p.37) and dry agriculture areas. It has a calculated erosion rate of 0.5±165 t ha -1 year -1 . 63 4. Results 71 first generation with lowered bank slopes (15º) (Fig. 10, p. 37). Intermediate erosion rates were estimated in these mine zones that still in exploitation-restoring process (17-25 t ha -1 yr -1 ) recording maximum soil loss of 184 t ha -1 yr -1 with medium value of 174 t ha -1 yr -1 . Microwatersheds with gentle slopes and a drainage network were created for the mines restored under third generation concepts (Fig. 11, p. 38). In these areas, maximum soil loss estimates range between 106 and 98 t ha -1 yr -1 , while the mean values range from 16 to 23 t ha -1 yr -1 . It is clear that applying improved restoration techniques reduces soil loss in mine zones and that non-restored and deficiently restored mines are sites contributing the highest soil loss (Fig. 8, p. 36). 72 4.2. Ecosystem service provision and spatial distribution Water flow regulation, surface water supply and soil formation are all widespread services provided by, approximately, 79.5%, 67% and 61.5% of the study area, respectively (Table 4). Recreation and ecotourism is present in 36%, soil retention in 27% and carbon storage in 21.1%. See, Fig. 22 A, p. 74 for a general watershed view of the spatial distribution of the values of the services in Martín Basin. Table 4. Percentages of the Martín Basin area where the ecosystem services listed are delivered. Between brackets is the percentage of the basin area where these services are delivered as hotspots (with high and very high values for the service). Ecosystem service Area (% of the total watershed area) Water flow regulation 79.5 (42.4) Surface water supply 67 (7.3) Soil accumulation 61.5 (19.4) Recreation/Ecotourism 36 (22) Carbon storage 21.1 (2.4) Soil retention 40.2 (19) Water flow regulation has the largest hotspot area, which is defined as the percentage of an area where a given service is valued as high and very high, with 42.4% and carbon storage had the smallest with 2.4% (Table 4). Water flow regulation is governed by rainfall distribution but is strongly influenced by permeable, underlying geology, which is high in the mostly porous soils of the southern part of Martín Basin and facilitates groundwater recharge. Surface water supply spread throughout the greater part of the basin. The highest values are located in the southern region and coincide with low values of soil formation. Carbon storage and soil retention depend on the density of canopy cover and are mostly distributed according to an altitudinal pattern. Higher values correspond to a range of 600- 73 1100 m above sea level. At higher altitudes, both services decline to intermediate values. Certain riparian areas defy this altitudinal trend, having high values for both of these services and showing no relationship to altitude (Fig. 15 D, Fig. 16 E, p. 53-54). Soil formation is predominantly found in the southern part of the study area, with very low or negligible values identified as one progresses towards the northern lowland areas of the basin. Recreation and ecotourism services are found in some subwatersheds located in the southerncentral and northern-central part of the basin along the river system. Many hiking and mountain biking routes start near the towns of Albalate del Arzobispo, Montalbán and Utrillas and extend outwards. 4.2.1. Relationship between services The greatest overlap of services (3-5 services) was observed in mountainous areas of the south and central parts of the Martín Basin where dense plant cover, woodland and scrubland are located (Fig. 5, p. 33). A relatively small part (14%) of the Martín Basin is not delivering any of the selected suite of services. One and two services are provided in 25% and 25.8% of the basin area, respectively, and three services are provided in 21% of the area (Fig. 22 A). Fig. 22. Ecosystem services richness Martín Basin (A). Services richness as subwatersheds (B) and as hotspot per are just correlative numbers to label them. The spatial overlap among services is parts of the watershed do not show any service ove overlapping (56 services) is observed in small part of the mountain area in the south A). The maximum overlap between services was found between water flow regulation and accounted for 65% of the basin area Ecosystem services richness , number of services from the bundle of ES studied, Martín Basin (A). Services richness as hotspots (C). On the right respective service richness hotspot per subwatershed (D) . The numbers in the subwatersheds are just correlative numbers to label them. The spatial overlap among services is low in general. Most of the north and central parts of the watershed do not show any service ove r lap, while maximum number 6 services) is observed in small part of the mountain area in the south The maximum overlap between services was found between surface water supply and water flow regulation and accounted for 65% of the basin area (Table 5). 74 , number of services from the bundle of ES studied, in the On the right respective service richness per . The numbers in the subwatersheds Most of the north and central lap, while maximum number of services 6 services) is observed in small part of the mountain area in the south (Fig. 22 surface water supply and 75 Table 5. Proportional (%) overlap of ecosystem services in the basin and hotspots (hotspots in brackets). Soil accumulation Carbon storage Soil retention Water flow regulation Surface water Carbon storage 21.1 (1.26) Soil retention 10 (5.1) 18.7 (2) Water flow regulation 61.1 (16.3) 21 (2.23) 38 (13.2) Surface water 59.4 (4.4) 20.7 (0.35) 3.5 (1.95) 65 (6.75) Tourism 13.6 (4.3) 6.8 (0.18) 10 (4.5) 22,1 (11.5) 17.1 (5.6) The percentage area of the basin with overlapped hotspots of these two services was 6.75% and was located in the southern region (Fig. 22 A). The soil retention and water surfacesupply overlap areas accounted for 3.5% and had an overlapped hotspot area of just 1.95% of the basin, which was associated with forest ecosystems. Recreation and ecotourism services have a relatively high overlap with water flow regulation but a small overlap with other services, such as carbon storage and soil retention (Table 5). The map of overlapped hotspot services generated using high and very high values for all of the services shows that a region comprising only 0.12% of the mapped areas incorporated all 6 services. The area is located in the southern part of the basin and corresponds with conifer forest (Fig. 22 C). Conversely, 41% of the basin, mostly in the northern part, is not delivering high or very high values for any service. Most of the areas classified as hotspots delivered one service (25.9%), two services (19%) and three services (9.2%). Only a small portion (0.71%) delivered five (Fig. 22 C). 4.2.2. Subwatershed classification according to ecosystem service provision Applying the GIS Spatial Analyst tool and the majority statistic option within the zonal statistic module used to identify the greatest number of services found within each subwatershed, we did not find a subwatershed that provided all six services. The distribution of the overlapping services by subwatersheds shows the same pattern as for number of services overlapping but let distinguish that subwatersheds 53, 61, 62, 63 and 65 are providing 4-5 services but only subwatersheds 53, 61, 63 and 65 are delivering 4 services as hotspots (Fig. 22 B, D). These subwatersheds its south part. They were also located in areas classified as having low and medium levels of degradation because of erosion Fig. 23). Fig. 23. Left: Spatial s implification of erosion Moderte: 12-17 t ha -1 yr -1 ; high: subwatersheds. Subwatershed number 62 represents a focal point for surrounding subwatersheds that deliver at least 3 services (Fig. 22 C ). In other subwatersheds deliver at least three services (nº25, 22, and account for 7% of the total area. Nineteen subwatersheds deliver two services and accounting for 36.6 % of the basin area. low degraded status only subwatershed 48 and 54 we degradation level ( Fig. 23 right). In contrast, most of the subwatersheds in the northern part of the basin (13 subwatersheds) were delivering just one service, which was most commonly surface water regulation. 4.2.3. Hotspot services at subwatershed scale Only four subwatersheds were classified as their boundaries. They are located in the southern part of the basin (subwatershed 63, 65, 53 D). These subwatersheds occupy 3. 1% of the total area of Martin Basin in They were also located in areas classified as having low and medium levels of because of erosion ( implification of erosion classes in Martín Basin (Light: 012 ; high: >17 t ha -1 yr -1 . On the Right erosion represented per Subwatershed number 62 represents a focal point for surrounding subwatersheds that deliver ). In between the southern and the central part of the basin, nine other subwatersheds deliver at least three services (nº25, 22, 34, 40, 41, 42, and account for 7% of the total area. Nineteen subwatersheds deliver two services and 36.6 % of the basin area. Mostly of these subwatersheds corresponds with a low degraded status only subwatershed 48 and 54 we re classified as having a moderately In contrast, most of the subwatersheds in the northern part of the basin (13 subwatersheds) were delivering just one service, which was most commonly surface water services at subwatershed scale Only four subwatersheds were classified as hotspots and included up to four services within their boundaries. They are located in the southern part of the basin (subwatershed 63, 65, 53 76 1% of the total area of Martin Basin in They were also located in areas classified as having low and medium levels of 12 t ha -1 yr -1 ; erosion represented per Subwatershed number 62 represents a focal point for surrounding subwatersheds that deliver between the southern and the central part of the basin, nine 42, 48, 54 and 12) and account for 7% of the total area. Nineteen subwatersheds deliver two services and Mostly of these subwatersheds corresponds with a re classified as having a moderately In contrast, most of the subwatersheds in the northern part of the basin (13 subwatersheds) were delivering just one service, which was most commonly surface water and included up to four services within their boundaries. They are located in the southern part of the basin (subwatershed 63, 65, 53 77 and 61) (Fig. 22 D). Subwatersheds 63 and 65 incorporate a vast mined area which has been restored (Fig. 3, p. 31), but is still classified as highly degraded, were as subwatershed 61 is mostly covered by conifer and hardwood and has a medium degradation level. All of these subwatersheds are found on steep slopes. In the same part of the basin, there are other subwatersheds (53, 55, 59, 52 and 62) that supply three services and mostly fall with the low and medium degraded level ( Fig. 23 right). 78 4.3. Multi-spatial-scale approach for establishing restoration priorities against erosion through the evaluation of ecosystem services at watershed scale Here we present the methodological approach for establishing a hierarchical spatial classification of restoration zones in Martin watershed based on the spatial analysis of erosion rates and ecosystem services assessments. 4.3.1. Erosion patterns across subwatershed levels The landscape heterogeneity of the Martín Basin is a key determining factor explaining the erosion patterns in the region, with the northern area being predominantly flat and the southern area being mountainous, showing a considerable increase in slope, altitude and rainfall patterns. Contrasting the three spatial levels provides us with insights regarding how changes in spatial detail can facilitate the targeting of degraded areas. For example, in Fig. 24 A, we are able to clearly identify areas with high erosion values grouped in the south and a large portion of the northern area showing a low erosion value. By increasing the scale detail from the first level to the second level, we are able to differentiate three erosion thresholds in the northern region (Fig. 24 B, C). Fig. 24. Erosion map at first (A), second (B) and third (C) level. plotted the relationship between mean erosion and standard deviation Furthermore, some areas identified at level one as showing low erosion were re presenting both medium and high erosion r facilitating more precise identification and location of areas for restoration. The results at Erosion map at first (A), second (B) and third (C) level. On the right of each map are plotted the relationship between mean erosion and standard deviation for each subwatershed Furthermore, some areas identified at level one as showing low erosion were re presenting both medium and high erosion r egions when examined at the finer detail of level 3, facilitating more precise identification and location of areas for restoration. The results at 79 On the right of each map are for each subwatershed . Furthermore, some areas identified at level one as showing low erosion were re -identified as egions when examined at the finer detail of level 3, facilitating more precise identification and location of areas for restoration. The results at different scales mostly highlight a fairly constant pattern across these scales ( mean erosion rates (and the calculated standard deviations) exhibit similar values within single watersheds (Fig. 24 A, B, C , right erosion rates and the calculated standard deviations. Subwatershed erosion rates that exceed the highest erosion threshold, indicating areas subjected identified (Fig. 24 A, B, C ). This pattern is repeated across different scales. However, the data dispersion increases as the d aggregation. This is a fairly typical characteristic of ecological data (Levin 1992; Costanza and Maxwell 1994). At the third level, some subwatersheds with high standard deviations and mean erosion values in the low 4.3.2. Ecosystem service patterns across subwatershed levels There is a clear dist inction in the ecosystem service supply across the study area ( northern, lower, reaches of the watershed showed the lowest values, which increased toward the south of the basin. However, at the third level, the ecosystem service supply was hi differentiated (Fig. 26 C). Fig. 25. Ecosystem services bundle map at first level. Highlighted by the blue circle show subwatershed number 4 (North) and 63 (South) different scales mostly highlight a fairly constant pattern across these scales ( mean erosion rates (and the calculated standard deviations) exhibit similar values within single , right ) and a direct relationship was observed between the mean erosion rates and the calculated standard deviations. Subwatershed erosion rates that exceed the highest erosion threshold, indicating areas subjected to a high erosion risk, can be easily ). This pattern is repeated across different scales. However, the data dispersion increases as the d etail of the analysis increases through the three levels of data aggregation. This is a fairly typical characteristic of ecological data (Levin 1992; Costanza and Maxwell 1994). At the third level, some subwatersheds with high standard deviations and erosion values in the low -tomedium erosion threshold range are identifiable ( Ecosystem service patterns across subwatershed levels inction in the ecosystem service supply across the study area ( northern, lower, reaches of the watershed showed the lowest values, which increased toward the south of the basin. However, at the third level, the ecosystem service supply was hi Ecosystem services bundle map at first level. Highlighted by the blue circle show (North) and 63 (South) . 80 different scales mostly highlight a fairly constant pattern across these scales ( Fig. 24 right). The mean erosion rates (and the calculated standard deviations) exhibit similar values within single ) and a direct relationship was observed between the mean erosion rates and the calculated standard deviations. Subwatershed erosion rates that exceed to a high erosion risk, can be easily ). This pattern is repeated across different scales. However, the data etail of the analysis increases through the three levels of data aggregation. This is a fairly typical characteristic of ecological data (Levin 1992; Costanza and Maxwell 1994). At the third level, some subwatersheds with high standard deviations and medium erosion threshold range are identifiable ( Fig. 24 C) . inction in the ecosystem service supply across the study area ( Fig. 22). The northern, lower, reaches of the watershed showed the lowest values, which increased toward the south of the basin. However, at the third level, the ecosystem service supply was hi ghly Ecosystem services bundle map at first level. Highlighted by the blue circle show Fig. 30. Hierarchy map for subwa tershed 4(A) and 63 (B ecosystem service corresponding at each subwatershed tershed 4(A) and 63 (B ) at second level. On the right of each map are plotted the erosion mean values against numbers of ecosystem service corresponding at each subwatershed . (t ha is a abbreviation of t ha -1 yr -1 ). On the right of each map are plotted the erosion mean values against numbers of Fig. 31. Hierarchy map for subwatershed 4(C) and 63 (D) at third level. On the right of each map are plotted the erosion mean values against numbers of ecosystem service corresponding at each subwatershed subwatershed 4(C) and 63 (D) at third level. On the right of each map are plotted the erosion mean values against numbers of ecosystem service corresponding at each subwatershed . (t ha is a abbreviation of t ha -1 yr -1 ). 88 subwatershed 4(C) and 63 (D) at third level. On the right of each map are plotted the erosion mean values against numbers of Our two selected subwatersheds show marked differences in the number of services delivered, with 0-1 ecosystem services being observed for subwatershed 4 at second level associated with an erosion rate of <12 and 12-17 t ha -1 yr -1 (Fig. 30 A) and 3-4-6 services being obtained in subwatershed 63 with an erosion rate > 17 t ha -1 yr -1 (Fig. 30 B). In subwatershed 63 the priority restoration area is represented by 3 and 4 services and an erosion rate > 17 t ha -1 yr -1 , corresponding to the greater part of the subwatershed (Fig. 30 B). Moving from level two to level three, diversification increases (Fig. 31 C, D) and for subwatershed 4, the number of services now ranges from 0 to 3, but they are mostly associated with low erosion thresholds (Fig. 31 C). In subwatershed 63, at level three, the number of services per subwatershed ranges from 3 to 6 and most of the subwatersheds appear to present high erosion thresholds (Fig. 31 D). 90 91 5. Discussion 92 93 5.1. RUSLE for targeting restoration efforts This study demonstrates that the RUSLE model used with appropriate values for each factor is a powerful tool. Using the GP (Genetic Programming) methodology proposed by Puente et al. (2011) was proven as a reliable approach to generating specifically designed indices to estimate the C factor in contrast with traditional indices, such as those of the NDVI and SAVI family (Puente et al. 2011). We identified high-risk areas where soil conservation-restoration practices are needed. In the Martín River Basin, major efforts should be dedicated to retain soil in its southern high relief part and, especially, in the no-restored opencast coal mines to prevent the irreversible degradation of these zones. For this purpose, the results of this study are useful for identifying different zones of erosion risk at the watershed scale and at lower scales (e.g., subwatershed). The average annual soil loss rate estimated using RUSLE and GIS for the Martín River Basin was 13.8 t ha -1 yr -1 . This estimation exceeds the estimated tolerable limits for soil formation of between 2 and 12 t ha -1 yr -1 in Mediterranean environments (Rojo 1990). These results compare well with other studies in similar areas (Renschler et al. 1999; Van Rompaey et al. 2003; Capolongo et al. 2008), confirming that the RUSLE-GIS generated estimates of soil loss in this study appear to be reasonable. The spatial variation of erosion in the Martín Basin appears to be dominated by slope. The higher mean values of potential erosion were associated with zones located in the highlands with steep areas, including opencast coal mines that had the highest erosion rates even though large areas of many coal mine zones have been submitted to a restoration process. Although erosion varies greatly depending on the type of mine restoration, the steepest zones in the opencast mines match the highest erosion rates in the Martín River Basin because of the creation of large (sometimes 1 or more km 2 ) hillslope areas inside and surrounding the mines by means of excavation. The scale of the mined areas (0.14 – 7.2 km 2 ) in comparison with the pixel size of the DEM (400 m 2 ) supports our assumption. Rill and gully networks in these reclaimed systems can markedly limit water availability and modify the spatial distribution of soil moisture at the slope scale by reducing the opportunities for down-slope runoff reinfiltration and by concentrating the water flow along the channeling network (Biemelt et al. 2005; Moreno-de las Heras et al. 2010). During the photographic field survey to evaluate the connectivity and eroded area prediction along the created buffer zone in the stream and river channels areas, appearing in the model analysis as high erosion areas, corresponded to bare rock and rock landslide phenomena ( Foto recognizes riverside degraded areas, as shown in that were degraded in the year of creation of the are now (2012) restored. Fig. 32. Example of highly degraded riversides, founded using RUSLE the bottom left (a), concrete ditch discharging straight in the river Road embankments have not been considered with a special focus in this paper, but during the photog raphic survey, we realized the magnitude of their impact on the river system During the photographic field survey to evaluate the connectivity and eroded area prediction zone in the stream and river channels , we obse rved that some of the areas, appearing in the model analysis as high erosion areas, corresponded to bare rock and Foto 3); however, in the m onitored areas, the model generally recognizes riverside degraded areas, as shown in Fig. 32. We also identified some mining areas that were degraded in the year of creation of the digital elevation model used here and that Example of highly degraded riversides, founded using RUSLE - buffer map. In evidence on the bottom left (a), concrete ditch discharging straight in the river . Road embankments have not been considered with a special focus in this paper, but during the raphic survey, we realized the magnitude of their impact on the river system 94 During the photographic field survey to evaluate the connectivity and eroded area prediction rved that some of the areas, appearing in the model analysis as high erosion areas, corresponded to bare rock and onitored areas, the model generally We also identified some mining areas used here and that buffer map. In evidence on Road embankments have not been considered with a special focus in this paper, but during the raphic survey, we realized the magnitude of their impact on the river system Fig. 33. Road embankments in the north (c) in the south (d , e, f These slopes are often directly connected by channels to the river network bottom left of the picture and through depositional areas. slopes is doomed to failure if the ecological vegetation establishment is not taken into account at the time of road building. This argument is supported by the existence of road cuts with a slope gradient exceeding 45°, where intense erosion occurs, generating very high soil loss and impacts that Road embankments in different parts of the Martín Basin. In the central part , e, f ). often directly connected by channels to the river network and Fig. 33 e) without having any way to intercept the sediment through depositional areas. García-Fayos et al. (2009) argued that the stabilization of road slopes is doomed to failure if the ecological knowledge of the topographic thresholds that limit vegetation establishment is not taken into account at the time of road building. This argument is supported by the existence of road cuts with a slope gradient exceeding 45°, where intense generating very high soil loss and impacts that other studies have 95 the central part (a, b) in often directly connected by channels to the river network (Fig. 32 a on the without having any way to intercept the sediment et al. (2009) argued that the stabilization of road knowledge of the topographic thresholds that limit vegetation establishment is not taken into account at the time of road building. This argument is supported by the existence of road cuts with a slope gradient exceeding 45°, where intense other studies have already 96 highlighted in Mediterranean sites near the study area (Bochet and García-Fayos 2004). We also highlight that some areas are highly degraded but are disconnected from the fluvial channel or are intercepted by depositional areas. These areas are not a direct threat to water bodies because they are not significant contributing areas. Management plans for a watershed should take into account the need to evaluate the importance of these areas with respect to different uses and the potential benefits of restoring these areas, assessing the effective value for the production of ecosystem services and the mechanical and monetary possibility of action (usually steep slopes) to restore it. 5.2. Plant colonization and reclaimed slopes Moreno-de las Heras et al. (2011) suggest that natural plant colonization in Mediterraneancontinental reclaimed environments requires vegetation cover of at least 30% and rill erosion rates below 17 t ha −1 yr −1 . In our case, 59% of the river basin has less than 30% plant cover, and 60% of the watershed has an erosion rate higher than 12 t ha −1 yr −1 and plant cover lower than 30%. This result is due to the very slow rate of plant recolonization and forest expansion, which occupies approximately 21% of the mountainous southern part of the basin. Fifty-six percent of the mine areas are included in the acceptable soil loss range for plant colonization, but the erosion rate is higher than 17 t ha −1 yr −1 in 44% of the mine zones in the Martín Basin. In these latter zones, plant colonization is difficult, enabling the formation of rill networks depending on the degree of disturbance, slope length and available water, among other factors (Moreno de las Heras et al. 2010). The consequence is a high erosion rate that endangers the life span of these newly created habitats and the wetlands created in the pit of the restored mines, which were established by being filled with high loads of mined materials but are filled with eroded sediments. This siltation process also reduces other key ecological processes (e.g., sediment-water column exchanges, organic matter enrichment) and the biological structure of this type of ecosystem (Mitsch and Gosselink 2007; Gell et al. 2009). In most of the (north) lowland and relatively flat part of the basin, which is dominated by agriculture, the estimated erosion rates are much lower (in general, <10 t ha -1 yr -1 ). The high rates in this part of the basin are associated with river dynamics (bank erosion) and land use (Fig. 5, p. 33), prevalent cereal crops and scrublands. In the southern and central zones of the basin, which are covered by conifers and hardwoods, the estimated values of the C factor (the vegetation-related variable in the RUSLE equation; see appendix 9.4.4.) were, as expected, low because of the relatively high cover density. The C factor for vineyards and olive trees had typical intermediate values because of the vegetation-free zones between the rows of plants, which are common in this type of land use. However, for scrubland, the C values obtained 103 livestock production (Delgado 2000). This would require the use of native and adapted species to avoid potential negative impacts on the ecosystems. Special attention must be given to the mining areas because they are the major source of sediment in the basin (Trabucchi et al. 2012a). These mines have been restored using a variety of restoration techniques and strategies at different times (Moreno delas Heras et al. 2008). The opportunity to create new services in restored areas exists and has been demonstrated on several restored mines in the Martín Basin that have been planted with crops and fruit trees. However, in order for these areas to be sustainable, best agricultural practices need to be adopted due to the high susceptibility of their soils to erosion and the very low soil organic carbon content. Furthermore, wetlands created in the old mine pits can provide multiple functions at a smaller scale, including recreation and education, and contribute multiple services at a larger/watershed scale, which could be accomplished in this semi-arid area through re-establishing a network of sites for biodiversity development (Moreno-Mateos et al. 2009). Mapping multiple ecosystem services provides a useful framework for management and restoration planning at the watershed scale. Detailed spatial prioritization of restoration actions will require analysis of ecosystem services and tradeoffs at a finer spatial resolution. Watersheds or basins have fractal characteristics, so fractal methods of analysis can be effective in predicting ecosystem service patterns at multiple scales (Halley et al. 2004). 5.6. Restoration implications from multi-scale analyses Landscapes are complex systems that require multi-scale analyses if they are to be appropriately managed and if the outcomes of interventions are to be anticipated (Hay et al. 2001). Basin-scale analyses (such as that performed in our case study area, the Martín Basin) appears to represent an appropriate extent scale for evaluating our methodology as the basin is considered the optimal functional ecological unit of management or, at least, that where more intensive interactions occur between human use of the resources and ecological processes (Golley 1994), both of which determine ecosystem services. Exploring a variety of spatial scales is a necessary exercise for understanding resource distribution (Lewis et al. 1996; White and Walker 1997). In our case, different spatial scales (levels of analysis) were used to investigate the spatial locations of possible restoration actions and the dynamics of ecosystem services associated with erosion. The type of multiscale spatial analysis performed in the Martín Basin to assess ecosystem services, which has frequently been suggested (Kremen and Ostfeld 2005; Hein et al. 2006; de Groot et al. 2010), proved useful for identifying sites to be 104 targeted for restoration (to ameliorate erosion) that simultaneously increase the provision of a selected bundle of ecosystem services. An initial assessment of the Martín Basin at low spatial detail is able to provide a general understanding of this territory and to identify the general status of broad areas in the basin, which is useful for a preliminary general statement of types of restoration action according to differences in the environmental risk and ecosystem service provision in these areas (Trabucchi et al. 2012a). Chu et al. (2003) described this need for obtaining a broad-scale understanding related to system dynamics so that it will be possible to explain cause-effect relationships in detail. The introduction of additional hierarchies or levels facilitates the integration of more detailed information. Our third level of analysis was found to be key in determining watershed processes and the mechanism of ecosystem degradation (Nakamura et al. 2005). As expected, reducing pixel aggregation increased spatial differentiation and detail and facilitated the location of areas for the prioritization of restoration and management actions. The second and especially, the third level of analysis followed a bottom-up approach. This approach increased the accuracy of the identification of site-scale areas to be targeted for action and provides a defensible basis for hypothesis testing in field experiments. We explored the third level (highest resolution) in detail, as this scale is expected to be the most economically suitable for directing restoration actions. In our case study, this level of analysis corresponded closely to the scale of opencast mine areas, which present a mean average area of 1.5 Km 2 . The fine-scale analysis highlighted subwatersheds or geographical areas in the basin where restoration actions to control erosion should be prioritized hierarchically to maintain or increase the provision of ecosystem services. This would not have been possible if we had only undertaken a single broad level (first level) of analysis. 5.7. Developed approach for including priority restoration areas As a first step in restoration planning, a regional analysis aims at constructing an overview of ecosystem conditions to identify altered areas in need of management action (Nakamura et al. 2005). To manage a river basin efficiently, objectives must be established and restoration priorities identified (Kondolf and Micheli 1995). This understanding is essential to achieve the optimal and efficient allocation of limited resources (Palik et al. 2000; Suding 2011), especially at a broad scale, where costs can grow exponentially. In the Martín Basin, areas presenting few services and low erosion rates were found to be predominant in the flat northern areas, which have historically delivered provisioning services related to food production. In this homogeneous landscape with an oligotrophic environment (low precipitation, low soil organic 105 matter content), restoration actions would be disproportionately expensive compared with the benefits that would be derived from such actions. We have adapted a simple risk decision support matrix previously used in watershed risk analysis (Milne and Lewis 2011) to facilitate the selection of priority areas for restoration (Table 6 p. 86). A hierarchical mapping approach could be used for a variety of purposes, particularly in exercises related to site location (Palik et al. 2000; Palik et al. 2003). Area selection can be further refined by coupling the generation of hierarchy maps for prioritizing subwatersheds with desired biological or physical ecological indicators (e.g., water quality, land use, erosion) (Niemi and McDonald 2004), combinations of which can be chosen to infer cause and effect relationships (such as explanatory environmental variables and responses manifested as changes in ecosystem services) (Nakamura et al. 2005). Furthermore, alternative state models, emphasizing internally reinforced states and recovery thresholds, can help in guiding restoration efforts (Suding et al. 2004). These thresholds could include types of pollution (e.g., nutrients, suspended soil, gas emissions) and general environmental disturbance thresholds (e.g., fires, floods) (Groffman et al. 2006). Ecological problems often require the extrapolation of fine-scale measurements for the analysis of broad-scale phenomena (Turner et al. 1994). The generation of hierarchical maps that allow the evaluation of restoration activity across a hierarchy of scales, ranging from a broad region to an individual site (Ziemer 1999), appears to be a logical and efficient way of locating key potential restoration areas. It is well recognized that restoration and landscape ecology exhibit an unexplored mutualistic relationship (Bell et al. 1997; Li et al. 2003). Our proposed framework integrates multi-scale studies, representing a key interest in landscape ecology (Turner et al. 1994; Hay et al. 2001; Brandt 2003; Burnett and Blaschke 2003; Wu 2004), with the type of hierarchical prioritization used in restoration ecology (Lee and Grant 1995; Palik et al. 2000; Cipollini et al. 2005; Nakamura et al. 2005; Comín et al. 2009) and the growing field of ecosystem service research (Fisher et al. 2009; Reyers et al. 2009; de Groot et al. 2010; Su et al. 2012). Such a multidisciplinary approach has been recommended to make restoration plans more attractive (Benayas et al. 2009; Bullock et al. 2011; Trabucchi, et al. 2012b) and to enhance research and the application of the three disciplines. Here, the focus of ecological restoration shifts from the site-scale studies adopted in the past aimed at the reestablishment of historical abiotic conditions to promote the natural return of the vegetation (Dobson et al. 1997; Bell 1998; Prach et al. 2001) or the reestablishment and improvement of animal habitat (Huxel and Hastings 1999; Bond and Lake 2003) to broad analyses of environmental conditions at regional scales. This vision is supported by modern restoration practices, which acknowledge the importance of ecosystem patterns and processes 106 occurring at landscape scales (Nakamura et al. 2005). During the nested analysis, various spatial and field assessment data can be added as layers to complement and enrich the analyses and improve the precision of prioritization according to the proposed objectives making our methodology extremely adaptable at each single case of research purpose. 5.8. Investigation of possible trade-offs in restoration prioritization Ecosystem service trade-offs are defined as situations in which one service is increased or improved at the expense of another (Bennett et al. 2009) and can arise from the differing interests of social agents (Martín-López et al. 2012). Analyzing the spatial patterns of ecosystem service bundles allows us to understand how services are distributed across a landscape, how the distributions of different services compare and where trade-offs and synergies among ecosystem services might occur (Raudsepp-Hearne et al. 2010). The presented approach highlights where potential ecosystem service improvement can be achieved through restoration and consequently, which trade-offs can be established between the services evaluated here (carbon storage, soil formation and retention, water flow regulation, surface water provisioning, eco-tourism), which contribute positively to natural resource enhancement and those that contribute negatively to natural resource conservation, which are typically provisioning services based on human extractive activities as intensive agriculture and mining. Conventional agricultural practices degrade the soil structure and soil microbial communities due to mechanical activities such as plowing, but management practices can also protect the soil and reduce erosion and runoff (Lupwayi et al. 1998; Holland 2004). The Martín Basin, especially its northern region, displays clear evidence of trade-offs between regulatory and provisioning services, which is an issue that has been noted in many other regions of the world (Rodríguez et al. 2006; Power 2010). Management decisions often focus on the immediate provisioning of a commodity or service at the expense of this service or another ecosystem service at a distant location or in the future (Power 2010). However, win-win scenarios are possible when appropriate land-use practices, such as conservation tillage, crop diversification and legume intensification, are applied (López et al. 1998; Prosperi et al. 2006; Trabucchi et al. 2012a). The potential success of integrating these approaches depends on the maintenance of ecological integrity and cohesion (Gómez-Sal and GonzálezGarcía 2007). Therefore, it may be possible to manage agro-ecosystems to support a diversity of ecosystem services while still maintaining or even enhancing certain provisioning services (Power 2010; Nainggolan et al. 2011). Understanding the benefits and costs of different types of management practices is necessary to allow the establishment and maintenance of sustainable agro-ecosystems (Dale and Polasky 2007). 107 Due to the predominant natural land cover in the southern part of the Martín Basin, the tradeoffs among ecosystem services in this region are of another type and are more difficult to identify because they also exhibit many synergies and dependent ecological processes (section 5.4. p.84). For example, most of the ecosystem services produced in perennial vegetation areas, such as under forest cover, are related to water (e.g., purification, regulation) and these, in turn, are linked to soil (e.g., accumulation, retention) (Klijn et al. 1996; Milne and Lewis 2011; Powlson et al. 2011). While there are clear synergies, there are also potential trade-offs. For example, increasing carbon storage through the planting of fast-growing trees for CO 2 accumulation (a carbon storage service linked to climate regulation) or cellulose production (a provisioning service) may reduce the surface water supply and could also result in the salinization and/or acidification of soils, with consequent decreases in ecosystem services associated with grasslands and reduced resilience of such systems (Bot and Benites 2005; Cespedes-Payret et al. 2009). Identifying trade-offs is an important step that allows policy makers to understand the longterm effects of preferring one ecosystem service over another and the consequences of focusing only on the present provision of a service, rather than the future (Rodríguez et al. 2006). 5.9. Possible methodological limitations and future research needs 5.9.1. Data management Spatial analysis typically involves GIS overlay analysis and geoprocessing to combine diverse sources of input layers in deriving a desired map. This analysis is often complicated by differences in parent scales, years of creation, accuracy levels, modeled data and minimum mapping units for each input layer (Troy and Wilson 2006). There is no single “correct” or “optimal” scale for characterizing spatial heterogeneity, but comparisons between landscapes using pattern indices must be based on the same spatial resolution and extent. Indeed, a comprehensive empirical database containing pattern metric “scalograms” and other forms of multiple-scale information on diverse landscapes is crucial for achieving a general understanding of landscape patterns and developing spatial scaling rules (Wu 2004). The relationship between ecosystem service delivery and the regulation of environmental factors, such as erosion, may also change according to the spatial scale of analysis (Jackway and Deriche 1996). An analyst's job will often include assembling many layers with different resolutions to obtain a final map that is suitable for management purposes. Ecosystem services, such as the ecological functions and processes from which they are derived, may 108 change in relation to the spatial pattern of observation (Hein et al. 2006; Hurteau et al. 2009), posing a major challenge for mapping these services. It is difficult to define the most appropriate scale of a study, as the resolution at which the phenomena of interest operate and are operated upon may not be immediately apparent (Rutchey and Godin 2009). Thus, in most cases, the best practice may be to adopt the highest resolution affordable (Haines-Young and Chopping 1996) but there must be a threshold for increasing the resolution (decreasing the grain size) of the analysis which once surpassed provides not so useful information as it is not related to functional aspects of the ecosystem (basin in our approach) or could result on excess of resources used in the analysis versus value of the information obtained. Furthermore, high-quality databases and new sampling approaches that support research at broader spatial and temporal scales are critical for enhancing ecological understanding and supporting further development of restoration ecology as a scientific discipline (Michener 1997). 5.9.1. Statistical analysis Selecting appropriate statistical procedures and asking the right questions is vital for meeting targets (Marcot 1998). This study employed one of several available methods for aggregating spatial data to analyze ecosystem service bundles. We used the majority rule method because of our interest in identifying the major number of services present at each spatial level (Trabucchi, submitted). Although this is probably the most commonly used rule in ecological and remote sensing applications (Wu 2004), it would be interesting to compare how different aggregation methods affect the characteristics of ecosystem service bundles. The use of rules, such as maximum, minimum and average rules and others available in GIS zonal statistical tools can have a marked effect on the obtained results (Smith et al. 2007). 5.9.2. Validation of the framework In the Martín Basin, some subwatersheds at the third level, classified as being of high priority for restoration (presenting erosion of > 17 t ha -1 yr -1 and >3 ecosystem services), coincide with closed mines. This finding confirmed both the appropriateness of the size of the subwatersheds generated at this level as well as the erosion and ecosystem service categorization applied for prioritizing restoration. However, future studies are needed to investigate the application of hierarchical maps at a mine scale, where these data are available, to further validate the approach presented here. 5.10. Flexibility of the Identifying goals for restoration and prioritizing restoration efforts are subjective processes to some extent (Palik et al. 2000) and achieve different restoration targets. Table 7 Schematic approach based on ecosystem services prioritize restoration action. A more generalized scheme to prioritize restoration actions at watershed scale is proposed with references to the r elative value of the disturbance factor and the ecosystem services (Table 7 ). In general, restoration priority is given to zones with relatively high risk of degradation (high values of the disturbance factor) as decreasing this risk should contribute to maintain the existing high values of ecosystem services and/or to increase l lower priority is given to zones with low environmental improvement of ecosystem services is not at risk in these zones. Here zones with high environmental risk and high number of services (or value) are priorit th e risk of degradation and loss with high environmental risk and low ecosystem services as it is expected t service provision will increase after amelio restoration is assigned to zones in a watershed with low environmental risk and low ecosystem ser vices, as it is expected that ecosystem services disturbance. And final position Flexibility of the framework Identifying goals for restoration and prioritizing restoration efforts are subjective processes to al. 2000) and the framework showed here can easily be modified to achieve different restoration targets. Schematic approach based on ecosystem services (E.S.) and environmental risk for A more generalized scheme to prioritize restoration actions at watershed scale is proposed elative value of the disturbance factor and the ecosystem services ). In general, restoration priority is given to zones with relatively high risk of degradation (high values of the disturbance factor) as decreasing this risk should contribute to maintain the existing high values of ecosystem services and/or to increase l lower priority is given to zones with low environmental disturbance as maintenance or improvement of ecosystem services is not at risk in these zones. Here zones with high environmental risk and high number of services (or value) are priorit ized for restoration due to e risk of degradation and loss of highly valued services. Second priority is given to zones with high environmental risk and low ecosystem services as it is expected t provision will increase after amelio rating the disturbance factor. Thir restoration is assigned to zones in a watershed with low environmental risk and low ecosystem vices, as it is expected that ecosystem services provision will increase after decreasing the position in the range of restoration priority is assigned to zones with low 109 Identifying goals for restoration and prioritizing restoration efforts are subjective processes to can easily be modified to and environmental risk for A more generalized scheme to prioritize restoration actions at watershed scale is proposed elative value of the disturbance factor and the ecosystem services ). In general, restoration priority is given to zones with relatively high risk of degradation (high values of the disturbance factor) as decreasing this risk should contribute to maintain the existing high values of ecosystem services and/or to increase l ow values. And as maintenance or improvement of ecosystem services is not at risk in these zones. Here zones with high for restoration due to of highly valued services. Second priority is given to zones with high environmental risk and low ecosystem services as it is expected t hat ecosystem rating the disturbance factor. Thir d priority for restoration is assigned to zones in a watershed with low environmental risk and low ecosystem provision will increase after decreasing the in the range of restoration priority is assigned to zones with low environmental risk and high ecosystem s provision. This generalization can be re (1983; 1997) to express the potential relationships among parts of an ecosystem under different environmental conditions which work both in the same direction to favor restoration (synergy: erosion × ecosystem services) or which both set some ecosystem characteristics th inking the restoration priority framework for a potential synergistic effect among restored (with the above criteria) zones in a watershed. Restoration to decrease environmental risks of zon will ensure the provision of ecosystem services with lower ecosystem services ecosystem services provision in zones with low environmental risk. Further develo the approach presented will be necessary to show if this dynamic aspect of zones of a watershed restored with the priority criteria presented here cases of restoration at watershed scale performed with this way. Fig. 35 Mandala for criteria priorization environmental risk and high ecosystem s ervices as no risk exists for ecosystem services This generalization can be re -ordered in similar way to those plankton ma ndalas by Margalef to express the potential relationships among parts of an ecosystem under different environmental conditions which work both in the same direction to favor restoration (synergy: erosion × ecosystem services) or which both set against (erosion/ecosystem services) some ecosystem characteristics (Fig. 35.). This type of mandala can be interpreted inking the restoration priority crite ria for different zones of a watershed, but also as a framework for a potential synergistic effect among restored (with the above criteria) zones in a watershed. Restoration to decrease environmental risks of zon e s with high ecosystem services the provision of ecosystem services with effects in other zone s of the watershed with lower ecosystem services and high environmental risk, which in turn may favor increasing provision in zones with low environmental risk. Further develo the approach presented will be necessary to show if this dynamic aspect of influences between zones of a watershed restored with the priority criteria presented here takes place cases of restoration at watershed scale performed with this approach will provide advances in priorization for different zones of a watershed 110 ervices as no risk exists for ecosystem services ndalas by Margalef to express the potential relationships among parts of an ecosystem under different environmental conditions which work both in the same direction to favor restoration against (erosion/ecosystem services) can be interpreted for reria for different zones of a watershed, but also as a framework for a potential synergistic effect among restored (with the above criteria) zones in a s with high ecosystem services s of the watershed in turn may favor increasing provision in zones with low environmental risk. Further develo pment of influences between takes place . Practical this approach will provide advances in 111 6. Conclusions 112 119 • La inclusión de los servicios de los ecosistemas en los estudios de restauración a escala de cuenca se ha incrementado desde el año 2006 bajo el impulso del Millenium Ecosystem Assessment, pero los enfoques adoptados para este fin hasta ahora son diversos. Esto se debe a la herencia de la utilización de enfoques ad hoc en los planes de restauración del pasado (ej. restauración de hábitat de una única especie) y a la inexistencia en la actualidad de un marco general a seguir para la inclusión de los servicios de los ecosistemas en los planes de restauración (basado en una metodología y definiciones generales) que podrían hacer la localización y evaluación y localización de los servicios más directa y sencilla. La complementación cartográfica de los servicios con factores que amenazan la continua provisión de los mismos parece una solución para priorizar jerárquicamente las necesarias acciones de restauración. • Se ha usado el modelo RUSLE (Ecuación de pérdida de suelo Revisada), un modelo de tipo empírico para evaluar el riesgo de erosión en la cuenca del Río Martín utilizando un innovador índice de vegetación para la evaluación de un factor clave (factor C cobertura) del modelo elegido que ha demostrado obtener mejor resultados que los disponibles en la actualidad. La limitación de este modelo es que no incluye fenómenos de deposición y retransporte que ocurren en un perfil ladera-abajo ya que resulta en un valor de erosión laminar. La implementación de RUSLE SIG (Sistema de Información Geográfica) ha permitido la estimación de la tasa media de erosión, obteniendo valores muy similares a los deducidos en otras áreas con características similares. • El aporte medio anual de sedimentos en la cuenca del Río Martín fue de 13.8 t ha -1 año -1 . La parte sur de la cuenca resulta ser la más afectada por erosión influenciada por la orografía acentuada de la zona (24 t ha -1 año -1 ). La parte baja, norte, de la cuenca (principalmente campos agrícolas), es donde se registran las menores tasas de erosión (10 t ha -1 año -1 ). • El mapa de erosión generado por el modelo permitió analizar las principales áreas fuentes de sedimento. El pastizal resulta ser el área con más altas tasas medias de erosión (25 t ha -1 año -1 ) aunque solo representa el 1% del área de la cuenca, los matorrales; las zonas de suelo desnudo como las minas y algunas zonas de matorralpastizal, fueron las responsables de las más altas tasas de erosión generadas en la cuenca (23, 24 t ha -1 año -1 respectivamente), todas ellas en su parte sur. • La principal ventaja de esta metodología es la sencillez de su implementación a partir de fuentes de información relativamente fáciles de adquirir hoy en día. • La metodología desarrollada en este estudio ha demostrado la utilidad de las técnicas de teledetección para realizar estudios básicos y aplicados, tanto a escala de cuenca como a escala regional (10-10,000 km 2 ). Sobre la base de nuestros resultados 120 experimentales, creemos que RUSLE-SIG, no obstante sus limitaciones y el uso del índice de vegetación como el GPVI, podrían mejorar la predicción de las tasas de erosión del suelo y la consecuente planificación, gestión, conservación y restauración del suelo a escala de cuenca. • Mediante la evaluación y el cartografiado de un conjunto de servicios ecosistémicos en la cuenca del Río Martín se ha observado que la regulación hídrica, la producción de agua dulce superficial (escorrentía) y la fertilidad del suelo son los servicios más extendidos con respectivamente el 80%, 67% y 62% del área de la cuenca, mientras que los servicios de ecoturismo (36%), control de la erosión (27%) y regulación climática (captura y almacenamiento de carbono en pies mayores) (21.1%) se proveen en extensiones menores de la cuenca. Las áreas hotspot están concentradas en las partes sur, más alta, de la cuenca. La regulación hídrica tiene el área más vasta de hotspot cubriendo el 42.4% de la cuenca y la regulación climática la más reducida con un 2.4%, ambas en el sur de la cuenca. • Los servicios relacionados con el agua siguen el patrón espacial de la lluvia. La producción de agua dulce superficial está presente en toda la cuenca y coincide con bajos valores de acumulación de suelo en el sur de la cuenca. La regulación climática y el control de la erosión dependen de la densidad de la vegetación y están distribuidos mayoritariamente siguiendo el patrón espacial de aumento de altitud. La acumulación de suelo predomina en la parte sur reduciéndose progresivamente hacia la parte norte, baja, de la cuenca. El servicio de ecoturismo se distribuye en distintas áreas del sur y centro-norte de la cuenca y, en muchos casos, se centran cerca de los pueblos de Albalate del Arzobispo, Montalbán, y Utrillas, ya que las rutas senderistas empiezan en estos núcleos urbanos o sus cercanías. • El más alto valor de solapamiento, de 3 a 5 servicios se ha observado en áreas montañosas del sur y del centro de la cuenca correspondiendo con una alta densidad de cobertura vegetal forestal y de matorral. Uno y dos servicios son provisionados en el 25% y 25.8% de la cuenca respectivamente y tres servicios en el 21%. Cuatro y cinco servicios representan el 10% y 2,6%, respectivamente, de la cuenca. Seis servicios juntos se encuentran solamente en un área pequeña, 0,67%, en el entorno natural alrededor del pueblo de Montalbán, mientras un 14% de la cuenca no proporciona ninguno de los servicios seleccionados en ese estudio. El solapamiento espacial entre servicios es grande en general. Los servicios relacionados con el agua son los que tienen el más alto porcentaje de solapamiento con el 65% de la cuenca y 6.75% de los correspondientes hotspot en la parte sur de la cuenca, lo cual indica sinergia en la presencia de estos servicios. • Como se ha observado por otros autores, las partes montañosas con su relieve heterogéneo y mayores precipitaciones son capaces de generar un mayor número de servicios y de mayor valor que las zonas de llanura altamente antropizadas y dedicadas a producción agrícola con clima semiárido cual es el caso de la cuenca del Martín. El 121 mismo relieve aportador de heterogeneidad bio-geofísica y riqueza paisajística puede acentuar la disminución de estos servicios en estas zonas si su gestión y uso no son los apropiados poniendo en peligro la continua producción de servicios capaces de influir en el buen estado ecológico de toda la cuenca si existen factores alteradores importantes, como la erosión. Este es el caso de algunas subcuencas o zonas de la parte sur de la cuenca del Martín con escasa cobertura vegetal y otras en donde se ubican zonas mineras en las que se han estimado tasas de erosión mayores de 17 t ha -1 año -1 y 1-2 servicios de los ecosistemas. • Se ha elaborado un marco para la definición y jerarquización de zonas de restauración a escala de cuenca hidrográfica basado en mapas jerárquicos de la erosión, como factor de alteración, y de los servicios de los ecosistemas, como variables de estado, para identificar zonas prioritarias de restauración basadas en umbrales de erosión y números de servicios producidos utilizando datos básicos en mapas con pixel de 20 m × 20 m. Esta metodología incluye tres niveles espaciales de análisis, en este caso de la cuenca del Martín, definidos por tres niveles diferentes de agregación espacial de los pixeles que forman los mapas de erosión y de provisión de servicios, que son agregados para evaluar su congruencia espacial. La elaboración de mapas y patrones multi-escala (con diferente agregación de pixeles) ha permitido identificar la resolución ideal de análisis espacial para seleccionar áreas prioritarias de restauración de la erosión para mejorar la provisión de servicios ecosistémicos en la cuenca del Martín. Para la clasificación de zonas se establecieron tres umbrales de erosión, que coinciden con límites ecológicos para el establecimiento de la vegetación en el área de estudio, contrapuestos con el número de servicios. Graficando los resultados notamos como reduciendo la agrupación espacial de los pixeles, creando subcuencas más pequeñas, el grado de precisión y la diversificación en la definición de zonas de la cuenca con diferentes valores combinados de erosión y de servicios ecosistémicos aumenta. Entre las subcuencas generadas al tercer nivel de agregación de pixeles, destacan zonas mineras que tienen un área media de 1,5 Km 2 , al sur de la cuenca, donde se observan tasas de erosión mayores de 17 t ha -1 año -1 y alto número de servicios. Una escala de análisis a escala de cuenca fluvial sirve para tener una visión amplia de condiciones diferenciadas entre partes amplias de la cuenca. • Esta aproximación, combinando en una escala el factor de alteración, la erosión, y el factor de estado, la provisión de servicios de los ecosistemas, constituye un enfoque lógico y práctico para la selección y establecimiento de una jerarquía de áreas de restauración; y su representación gráfica mediante SIG, una herramienta útil para la selección y establecimiento de una jerarquía espacial de áreas de restauración en la cuenca hidrográfica. La disponibilidad de los datos con una resolución óptima y el análisis de necesidades de los interesados se requieren para que este enfoque pueda mostrar todo su potencial. 122 • De este estudio se deriva un marco conceptual adaptable y fácilmente aplicable a la definición de zonas de actuación de mejora del funcionamiento ecológico natural en otras cuencas o territorios que tienen perturbaciones ambientales que amenazan la provisión de servicios del ecosistema (exceptuando los de producción). A la escala de cuenca hidrográfica, se recomienda ordenar las actuaciones de restauración priorizando las zonas identificadas a la escala espacial adecuada como de alto factor de disturbio y de provisión de servicios (por existir riesgo de alteraciones y de pérdida de servicios); seguidas de zonas con alto factor de disturbio y baja provisión de servicios (por haber una potencial ganancia de servicios derivada de la restauración que obraría disminuyendo el factor de alteración), y con menor interés de realizar actuaciones de restauración ambiental las zonas con bajo riesgo de alteración ambiental y donde puede existir una mejora en la provisión de servicios; dejando sin actuaciones las zonas con bajo impacto y alta provisión de servicios, ya que no existe riesgo para la provisión de estos servicios. • Como conclusión general se ha comprobado con este trabajo que la evaluación de los servicios de los ecosistemas es una base útil para la planificación de la restauración ecológica y la gestión de un territorio formado por un mosaico de ecosistemas. 123 8. References cited 124 Akaike, H., 1974. A new look at the statistical model identification. Automatic Control, IEEE Transactions on, 19(6), pp.716–723. Angulo-Martínez, M. & Beguería, S., 2009. Estimating rainfall erosivity from daily precipitation records: A comparison among methods using data from the Ebro Basin (NE Spain). Journal of Hydrology, 379(1-2), pp.111–121. Aragon Territorial Information Centre (CINTA) (2006) Aragon DEM 20920 m pixel and others layers. Available on: http://sitar.aragon.es Aronson, J. et al., 2006. Nature conservation as if people mattered. Journal for nature conservation, 14(3-4), pp.260–263. Aronson, James et al., 2010. Are Socioeconomic Benefits of Restoration Adequately Quantified? A Meta-analysis of Recent Papers (2000-2008). Restoration Ecology, 18(2), pp.143–154. Bailey, P.A. et al., 1993. A fish community analysis of the Minnesota River basin. Minnesota Pollution Control Agency, St. Paul, Minn. Balamurugan, G., 1991. Tin mining and sediment supply in Peninsular Malaysia with special reference to the Kelang River basin. The Environmentalist, 11(4), pp.281–291. Barbier, E.B. et al., 2008. Coastal Ecosystem-Based Management with Nonlinear Ecological Functions and Values. Science, 319(5861), pp.321 –323. Battin, J. et al., 2007. Projected impacts of climate change on salmon habitat restoration. Proceedings of the national academy of sciences, 104(16), p.6720. Bazzoffi, P., 2008. Soil erosion tolerance and water runoff control: minimum environmental standards. Regional Environmental Change, 9, pp.169–179. Beguería, S., 2006. Identifying erosion areas at basin scale using remote sensing data and GIS: a case study in a geologically complex mountain basin in the Spanish Pyrenees. International journal of remote sensing, 27(20), pp.4585–4598. Bell, G.P., 1998. Ecology and Management of Arundo donax, and Approaches To Riparian Habitat Restoration In Southern California. Bell, S.S., Fonseca, M.S. & Motten, L.B., 1997. Linking Restoration and Landscape Ecology. Restoration Ecology, 5(4), pp.318–323. Bellot, J. et al., 1999. Effect of different vegetation type cover effects on the soil water balance in a semi-arid area of south eastern Spain. Phys. Chem. Earth (B), 24, pp.24 353–357 Bennett, E.M., Peterson, G.D., & Gordon, L.J., 2009. Understanding relationships among multiple ecosystem services. Ecology Letters, 12(12), pp.1394–1404. Biemelt, D. et al., 2005. Overland flow, erosion, and related phosphorus and iron fluxes at plot scale: a case study from a non-vegetated lignite mining dump in Lusatia. Geoderma, 129(1), pp.4–18. 126 Boardman, J., Poesen, J. & Evans, R., 2003. Socio-economic factors in soil erosion and conservation. Environmental science & policy, 6(1), pp.1–6. Bochet, E. & García-Fayos, P., 2004. Factors controlling vegetation establishment and water erosion on motorway slopes in Valencia, Spain. Restoration Ecology, 12(2), pp.166–174. Boellstorff, D. & Benito, G., 2005. Impacts of set-aside policy on the risk of soil erosion in central Spain. Agriculture, Ecosystems & Environment, 107(2-3), pp.231–243. Bohn, B.A. & Kershner, J.L., 2002. Establishing aquatic restoration priorities using a watershed approach. Journal of Environmental Management, 64(4), pp.355–363. Bond, N.R. & Lake, P.S., 2003. Local habitat restoration in streams: Constraints on the effectiveness of restoration for stream biota. Ecological Management & Restoration, 4(3), pp.193–198. Bot, A. & Benites, J., 2005. The Importance of Soil Organic Matter: Key to Drought-resistant Soil and Sustained Food Production, Food & Agriculture Org. Bourgeron, P.S. & Jensen, M.E., 1994. An overview of ecological principles for ecosystem management. Ecosystem Management: Principles and Applications. US Department of Agriculture, Forest Service, Pacific Northwest Research Station, Portland, Oregon, pp.45–57. Box Jr., J.E., 1981. The effects of surface salty fragments on soil erosion by water. Soil Science Society of America Journal 45, pp.111–116. Brabyn, L. & Mark, D.M., 2011. Using viewsheds, GIS, and a landscape classification to tag landscape photographs. Applied Geography, 31, pp.1115–1122. Brandt, J., 2003. Multifunctional landscapes--perspectives for the future. Journal Of Environmental Sciences China, 15(2), pp.187–192. Brown, L.R., 1981. World population growth, soil erosion, and food security.[Adapted from book]. Science (Washington, DC);(United States), 214. Bullock, J.M. et al., 2011. Restoration of ecosystem services and biodiversity: conflicts and opportunities. Trends in Ecology & Evolution, 26(10), pp.541–549. Burnett, C. & Blaschke, T., 2003. A multi-scale segmentation/object relationship modelling methodology for landscape analysis. Ecological Modelling, 168(3), pp.233–249. Burrough, P.A., 2001. GIS and geostatistics: Essential partners for spatial analysis. Environmental and Ecological Statistics, 8(4), pp.361–377. Cammeraat, L.H., 2002. A review of two strongly contrasting geomorphological systems within the context of scale. Earth Surface Processes and Landforms, 27(11), pp.1201–1222. Capolongo, D. et al., 2008. Spatial and temporal variations in soil erosion and deposition due to land-levelling in a semi-arid area of Basilicata (Southern Italy). Earth Surface Processes and Landforms, 33(3), pp.364–379. 127 Carpenter, S.R. et al., 2009. Science for managing ecosystem services: Beyond the Millennium Ecosystem Assessment. Proceedings of the National Academy of Sciences, 106(5), pp.1305 – 1312. Carreño, L., Frank, F.C. & Viglizzo, E.F., 2011. Tradeoffs between economic and ecosystem services in Argentina during 50 years of land-use change. Agriculture, Ecosystems & Environment, On line first. Cespedes-Payret, C. et al., 2009. The irruption of new agro-industrial technologies in Uruguay and their environmental impacts on soil, water supply and biodiversity: a review. International Journal of Environment and Health, 3(2), pp.175–197. Choi, Y.D., 2007. Restoration Ecology to the Future: A Call for New Paradigm. Restoration Ecology, 15(2), pp.351–353. Chu C., Minns C.K. & Mandrak N.E., 2003. Comparative regional assessment of factors impacting freshwater fish biodiversity in Canada. Canadian Journal of Fisheries and Aquatic Sciences, 60(5), pp.624–634. Cipollini, K.A., Maruyama, A.L. & Zimmerman, C.L., 2005. Planning for restoration: A decision analysis approach to prioritization. Restor. Ecol., 13(3), pp.460–470. Clewell, A.F. & Aronson, J., 2007. Ecological restoration: principles, values, and structure of an emerging profession, Island Pr. Cobourn, J., 1999. Integrated watershed management on the Truckee River in Nevada. J. Am. Water Resour. Assoc., 35(3), pp.623–632. Comín, F. A. et al., 2009. Establishing priorities for the management and restoration of river basins with opencast coal mines. River Basin Management V, pp.315–326. Comín, F. A., 2010. Ecological restoration: a global challenge, Cambridge Univ Pr. Commission of the European communities, 2006. Halting the loss of biodiversity by 2008 – and beyond; Sustaining ecosystem services for human well–being, Annex 1 action plan to 2010 and beyond. Ref: SEC (2006) 621. Commission of the European Communities, 2008. Providing Restoration Guidelines for Natura 2000 Habitats and Species. http://biodiversitychm.eea.europa.eu/events/providingrestoration-guidelines-natura-2000/network/other/Global/society-ecological-restorationinternational/UK-SER_2008_rev.doc. Convention on Biological Diversity (CBD), 2007. The Ecosystem Approach (UNEP/CBD/COP/5/23 Decision V/6. Nairobi, Kenya. Costanza, R. et al., 1997. The value of the world’s ecosystem services and natural capital. Nature, 387(6630), pp.253–260. Cuevas, J.G. & van Leersum, G., 2001. Project “conservation, restoration, and development of the Juan Fernandez islands, Chile.” Rev. Chil. Hist. Nat., 74(4), pp.899–910. 128 Daily, G. C., 1997. Nature’s services: societal dependence on natural ecosystems, Island Pr. Daily, G.C., 2000. Management objectives for the protection of ecosystem services. Environmental Science & Policy, 3(6), pp.333–339. Dale, V.H. & Polasky, S., 2007. Measures of the effects of agricultural practices on ecosystem services. Ecological Economics, 64(2), pp.286–296. DeFries, R. & Eshleman, K.N., 2004. Land-use change and hydrologic processes: a major focus for the future. Hydrological processes, 18(11), pp.2183–2186. Delgado Enguita, I., 2000. Base forrajera para el establecimiento de ganaderías de ovino en el secano cerealista, Institución Fernando el Católico. Dobson, Andy P., Bradshaw, A.D. & Baker, A.J.M., 1997. Hopes for the Future: Restoration Ecology and Conservation Biology. Science, 277(5325), pp.515–522. Dodds, W.K. et al., 2008. Comparing ecosystem goods and services provided by restored and native lands. BioScience, 58(9), pp.837–845. Driver, A.M., Rouget, K., Lombard, M., Nel, A.T., Turpie, J.L., Cowling, J.K., Desmet, R.M., Goodman, P., Harris, P., Jonas, J., Reyers, Z., Sink, B., Strauss, K., 2005. National Spatial Biodiversity Assessment 2004: Priorities for Biodiversity Conservation in South Africa Strelitzia 17. Dudgeon, D. et al., 2006. Freshwater biodiversity: importance, threats, status and conservation challenges. Biological Reviews, 81(2), pp.163–182. Dunjó, G., Pardini, G. & Gispert, M., 2003. Land use change effects on abandoned terraced soils in a Mediterranean catchment, NE Spain. CATENA, 52(1), pp.23–37. Durán Zuazo, V.H. & Rodríguez Pleguezuelo, C.R., 2008. Soil-erosion and runoff prevention by plant covers. A review. Agronomy for Sustainable Development, 28(1), pp.65–86. Egoh, Benis et al., 2007. Integrating ecosystem services into conservation assessments: A review. Ecological Economics, 63(4), pp.714–721. Egoh, Benis et al., 2008. Mapping ecosystem services for planning and management. Agriculture, Ecosystems & Environment, 127(1-2), pp.135–140. Egoh, Benis N. et al., 2011. Identifying priority areas for ecosystem service management in South African grasslands. Journal of Environmental Management, 92(6), pp.1642–1650. Ehrenfeld, J.G., 2000. Defining the limits of restoration: the need for realistic goals. Restoration Ecology, 8(1), pp.2–9. Environmental Systems Research Institute (ESRI). 2008. ArcGIS Desktop (ArcInfo) Software. ESRI, Redlands, California, USA. 135 Moreno-de las Heras, M., Merino-Martín, L. & Nicolau, J.M., 2009. Effect of vegetation cover on the hydrology of reclaimed mining soils under Mediterranean-Continental climate. Catena, 77(1), pp.39–47. Moreno-de las Heras, M., Nicolau, J.M. & Espigares, T., 2008. Vegetation succession in reclaimed coal-mining slopes in a Mediterranean-dry environment. Ecological Engineering, 34(2), pp.168–178. Moreno-Mateos, D. et al., 2012. Structural and functional loss in restored wetland ecosystems. PLoS biology, 10(1), p.e1001247. Moreno-Mateos, D., Pedrocchi, C. & Comín, F. A., 2009. Avian communities’ preferences in recently created agricultural wetlands in irrigated landscapes of semi-arid areas. Biodiversity and Conservation, 18(4), pp.811–828. Myers, N., 1996. Environmental services of biodiversity. Proceedings of the National Academy of Sciences, 93(7), pp.2764 –2769. Naidoo, R. et al., 2008. Global mapping of ecosystem services and conservation priorities. Proceedings of the National Academy of Sciences, 105(28), p.9495. Naidoo, R. & Ricketts, Taylor H, 2006. Mapping the Economic Costs and Benefits of Conservation. PLoS Biol, 4(11), p.e360. Nainggolan, D. et al., 2011. Farmer typology, future scenarios and the implications for ecosystem service provision: a case study from south-eastern Spain. Regional Environmental Change. Nakamura, F., Inahara, S. & Kaneko, M., 2005. A hierarchical approach to ecosystem assessment of restoration planning at regional, catchment and local scales in Japan. Landscape and Ecological Engineering, 1(1), pp.43–52. Naveh, Z., 1994. From Biodiversity to Ecodiversity: A Landscape-Ecology Approach to Conservation and Restoration. Restoration Ecology, 2(3), pp.180–189. Nearing, M.A., 1998. Why soil erosion models over-predict small soil losses and under-predict large soil losses. Catena, 32(1), pp.15–22. Nelson, E. et al., 2009. Modeling multiple ecosystem services, biodiversity conservation, commodity production, and tradeoffs at landscape scales. Frontiers in Ecology and the Environment, 7(1), pp.4–11. Nicolau, J.-M., 2003. Trends in relief design and construction in opencast mining reclamation. Land Degradation & Development, 14(2), pp.215–226. Niemi, G.J. & McDonald, M.E., 2004. Application of ecological indicators. Annual Review of Ecology, Evolution, and Systematics, pp.89–111. Nienhuis, P.H. et al., 2002. Ecological rehabilitation of the lowland basin of the river Rhine (NW Europe). Hydrobiologia, 478(1), pp.53–72. 136 Novotny, V., 1999. Integrating diffuse/nonpoint pollution control and water body restoration into watershed management. J. Am. Water Resour. Assoc., 35(4), pp.717–727. O’Farrell, P.J. et al., 2010. Multi-functional landscapes in semi arid environments: implications for biodiversity and ecosystem services. Landscape Ecology, 25, pp.1231–1246. Onori, F. et al., 2006. Soil erosion prediction at the basin scale using the revised universal soil loss equation (RUSLE) in a catchment of Sicily (southern Italy). Environmental Geology, 50(8), pp.1129–1140. Ormerod, S.J., 2003. Restoration in applied ecology: editor’s introduction. Journal of Applied Ecology, 40(1), pp.44–50. Orsi, F., Church, R.L. & Geneletti, D., 2011. Restoring forest landscapes for biodiversity conservation and rural livelihoods: A spatial optimisation model. Environmental Modelling & Software. Ozcan, A.U. et al., 2008. Use of USLE/GIS technology integrated with geostatistics to assess soil erosion risk in different land uses of Indagi Mountain Pass—Çankırı, Turkey. Environmental Geology, 53(8), pp.1731–1741. Palik, B. J et al., 2000. Using landscape hierarchies to guide restoration of disturbed ecosystems. Ecological Applications, 10(1), pp.189–202. Palik, Brian J., Buech, R. & Egeland, L., 2003. Using an ecological land hierarchy to predict seasonal-wetland abundance in upland forests. Ecological Applications, 13(4), pp.1153–1163. Palmer, M. a et al., 2005. Standards for ecologically successful river restoration. Journal of Applied Ecology, 42(2), pp.208–217. Palmer, M.A. et al., 2010. Mountaintop mining consequences. Science, 327(5962), pp.148–149. Palmer, Margaret A. et al., 2009. Climate Change and River Ecosystems: Protection and Adaptation Options. Environmental Management, 44(6), pp.1053–1068. Panagos, P. et al., 2011. Soil erodibility estimation using LUCAS point survey data of Europe. Environmental Modelling & Software, (0). Pardini, G., Gispert, M. & Dunjó, G., 2003. Runoff erosion and nutrient depletion in five Mediterranean soils of NE Spain under different land use. The Science of The Total Environment, 309(1-3), pp.213–224. Pausas, J.G. et al., 2004. Pines and oaks in the restoration of Mediterranean landscapes of Spain: New perspectives for an old practice – a review. Plant Ecology (formerly Vegetatio), 171(1/2), pp.209–220. Gao, P., 2008. Understanding watershed suspended sediment transport. Progress in Physical Geography, 32(3), pp.243–263. Pennisi, E., 2004. The grand (Canyon) experiment. Science, 306(5703), pp.1884–1886. 137 Peña Monné, J.L., Sánchez Fabre, M. & Cuadrat, J.M., 2002. El clima de la provincia de Teruel. Pert, P.L. et al., 2010. A catchment-based approach to mapping hydrological ecosystem services using riparian habitat: A case study from the Wet Tropics, Australia. Ecological Complexity, 7(3), pp.378–388. Peterson, C.H. & Lipcius, R.N., 2003. Conceptual progress towards predicting quantitative ecosystem benefits of ecological restorations. Marine Ecology Progress Series, 264(11). Pimentel, D. et al., 1995. Environmental and economic costs of soil erosion and conservation benefits. Science, 267(5201), p.1117. Pizzuto, J., 2002. Effects of dam removal on river form and process. BioScience, 52(8), pp.683– 691. Poli, R., Langdon, W.B. & McPhee, N.F., 2008. A field guide to genetic programming, Lulu Enterprises Uk Ltd. Power, A.G., 2010. Ecosystem services and agriculture: tradeoffs and synergies. Philosophical Transactions of the Royal Society B: Biological Sciences, 365(1554), pp.2959 –2971. Powlson, D.S. et al., 2011. Soil management in relation to sustainable agriculture and ecosystem services. Food Policy, 36, Supplement 1(0), pp.S72–S87. Prach, K. et al., 2001. The role of spontaneous vegetation succession in ecosystem restoration: A perspective. Applied Vegetation Science, 4(1), pp.111–114. Prosperi, J.-M. et al., 2006. Morphologic and Agronomic Diversity of Wild Genetic Resources of Medicago sativa L. Collected in Spain. Genetic Resources and Crop Evolution, 53(4), pp.843– 856. Puente, C. et al., 2011. A genetic programming approach to estimate vegetation cover in the context of soil erosion assessment. Photogrammetric Engineering and Remote Sensing, 77(4), pp.363–376. Quinton, J.N., Edwards, G.M. & Morgan, R.P.C., 1997. The influence of vegetation species and plant properties on runoff and soil erosion: results from a rainfall simulation study in south east Spain. Soil Use and Management, 13(3), pp.143–148. Raudsepp-Hearne, C., Peterson, G. D. & Bennett, E. M., 2010. Ecosystem service bundles for analyzing tradeoffs in diverse landscapes. Proceedings of the National Academy of Sciences, 107(11), pp.5242 –5247. Rayburn, A.P. & Schulte, L.A., 2009. Landscape change in an agricultural watershed in the US Midwest. Landscape and Urban Planning, 93(2), pp.132–141. Raymond, C.M. et al., 2009. Mapping community values for natural capital and ecosystem services. Ecological Economics, 68(5), pp.1301–1315. 138 Renard, K.G. et al., 1997. Predicting soil erosion by water: a guide to conservation planning with the revised universal soil loss equation (RUSLE). Agriculture Handbook (Washington), (703). Renard, Kenneth G et al., 1991. RUSLE: Revised Universal Soil Loss Equation. Journal of Soil and Water Conservation, 46(1), pp.30–33. Renschler, C.S., Mannaerts, C. & Diekkrüger, B., 1999. Evaluating spatial and temporal variability in soil erosion risk–rainfall erosivity and soil loss ratios in Andalusia, Spain. Catena, 34(3-4), pp.209–225. Rey Benayas, J.M. et al., 2009. Enhancement of biodiversity and ecosystem services by ecological restoration: a meta-analysis. Science (New York, N.Y.), 325(5944), pp.1121–1124. Reyers, B. et al., 2009. Ecosystem Services, Land-Cover Change, and Stakeholders: Finding a Sustainable Foothold for a Semiarid Biodiversity Hotspot. Ecology and Society, 14(1), p.38. Reynolds, J.F. et al., 2007. Global desertification: building a science for dryland development. science, 316(5826), p.847. Ries, J.B., Langer, M. & Rehberg, C., 2000. Experimental investigations on water and wind erosion on abandoned fields and arable land in the central Ebro Basin, Aragon, Spain (with 6 figures and 3 tables). Zeitschrift fur geomorphologie supplementband, pp.91–108. Risse, L.M. et al., 1993. Error assessment in the Universal Soil Loss Equation. Rodríguez, J.P. et al., 2006. Trade-offs across space, time, and ecosystem services. Ecology and Society, 11(1), p.28. Rojo, L., 1990. Plan nacional de restauración hidrológico-forestal y control de la erosión, Memoria. Romkens, M.J.M. & Wang, J.Y., 1987. SOIL ROUGHNESS CHANGES FROM RAINFALL. Transactions of the American Society of Agricultural Engineers, 30(1), pp.101–107. Van Rompaey, A.J.J. et al., 2003. Validation of soil erosion risk assessments in Italy, Joint Research Centre, Commission of the European Communities. Ruiz-Navarro, A. et al., 2012. Effect of the spatial resolution on landscape control of soil fertility in a semiarid area. Journal of Soils and Sediments, 12(4), pp.471–485. Rutchey, K. & Godin, J., 2009. Determining an appropriate minimum mapping unit in vegetation mapping for ecosystem restoration: a case study from the Everglades, USA. Landscape ecology, 24(10), pp.1351–1362. Samraj, P. et al., 1988. Hydrological behaviour of the Nilgiri sub-watersheds as affected by bluegum plantations, part I. The annual water balance. Journal of Hydrology, 103(3–4), pp.335–345. 139 Sánchez-Andrés, R. et al., 2010. Litterfall dynamics and nutrient decomposition of arid mangroves in the Gulf of California: Their role sustaining ecosystem heterotrophy. Estuarine, Coastal and Shelf Science, 89(3), pp.191–199. Sauer, T. & Ries, J.B., 2008. Vegetation cover and geomorphodynamics on abandoned fields in the Central Ebro Basin (Spain). Geomorphology, 102(2), pp.267–277. Senderos de Aragon, 2011. http://senderos.turismodearagon.com/ Schaich, H., 2009. Local residents’ perceptions of floodplain restoration measures in Luxembourg’s Syr Valley. Landscape and Urban Planning, 93(1), pp.20–30. Schroter, D., 2005. Ecosystem Service Supply and Vulnerability to Global Change in Europe. Science, 310(5752), pp.1333–1337. Schulte, L.A., Pidgeon, A.M., et al. (2005). "One hundred fifty years of change in forest bird breeding habitat: Estimates of species distributions." Conservation Biology 19(6), pp. 19441956. SER, 2004. The SER International Primer on Ecological Restoration, Science & Policy Working Group. http://www.ser.org/content/ecological_restoration_primer.asp. Simoncini, R., 2009. Developing an integrated approach to enhance the delivering of environmental goods and services by agro-ecosystems. Regional Environmental Change, 9(3), pp.153–167. Smith, M.J.D., Goodchild, M.F. & Longley, P., 2007. Geospatial Analysis: A Comprehensive Guide to Principles, Techniques and Software Tools, Troubador Publishing Ltd. Smith, S. V. et al., 2007. Soil erosion and significance for carbon fluxes in a mountainous Mediterranean-climate watershed. Ecological Applications, 17(5), pp.1379–1387. Sophocleous, M., 2002. Interactions between groundwater and surface water: the state of the science. Hydrogeology Journal, 10(1), pp.52–67. Di Stefano, C., Ferro, V. & Porto, P., 2000. Length slope factors for applying the revised universal soil loss equation at basin scale in southern Italy. Journal of agricultural engineering research, 75(4), pp.349–364. Su, C. et al., 2012. Ecosystem management based on ecosystem services and human activities: a case study in the Yanhe watershed RID B-1493-2009. Sustainability Science, 7(1), pp.17–32. Suding, K.N., 2011. Toward an Era of Restoration in Ecology: Successes, Failures, and Opportunities Ahead. Annual Review of Ecology, Evolution, and Systematics, 42(1), pp.465– 487. Suding, K.N., Gross, K.L. & Houseman, G.R., 2004. Alternative states and positive feedbacks in restoration ecology. Trends in Ecology & Evolution, 19(1), pp.46–53. 140 Swift, M.J., Izac, A.-M.N. & van Noordwijk, M., 2004. Biodiversity and ecosystem services in agricultural landscapes”are we asking the right questions? Agriculture, Ecosystems & Environment, 104(1), pp.113–134. Tabara, J.D. & Ilhan, A., 2008. Culture as trigger for sustainability transition in the water domain: the case of the Spanish water policy and the Ebro river basin. Regional Environmental Change, 8(2), pp.59–71. Tallis, H. et al., 2008. An ecosystem services framework to support both practical conservation and economic development. Proceedings of the National Academy of Sciences, 105(28), p.9457. Tallis, Heather & Polasky, S., 2009. Mapping and Valuing Ecosystem Services as an Approach for Conservation and Natural-Resource Management. Annals of the New York Academy of Sciences, 1162(1), pp.265–283. Taylor, K.G. & Owens, P.N., 2009. Sediments in urban river basins: a review of sediment– contaminant dynamics in an environmental system conditioned by human activities. Journal of Soils and Sediments, 9(4), pp.281–303. R Development Core Team & others, 2010. R: A language and environment for statistical computing. R Foundation for Statistical Computing Vienna Austria, (01/19). Tong, C. et al., 2007. Ecosystem service values and restoration in the urban Sanyang wetland of Wenzhou, China. ecological engineering, 29(3), pp.249–258. Toy, T.J., Foster, G. R & Renard, K. G, 1999. RUSLE for mining, construction and reclamation lands. Journal of Soil and Water Conservation, 54(2), pp.462–467. Trabucchi, M., Ntshotsho, P., et al., 2012b. Ecosystem service trends in basin-scale restoration initiatives: A review. Journal of Environmental Management, 111, pp.18–23. Trabucchi, M.,Puente, Cesar, et al., 2012a. Mapping erosion risk at the basin scale in a Mediterranean environment with opencast coal mines to target restoration actions. Regional Environmental Change. Online first Troy, A. & Wilson, M.A., 2006. Mapping ecosystem services: Practical challenges and opportunities in linking GIS and value transfer. Ecological Economics, 60(2), pp.435–449. Turner, M. G. et al., 1994. Multiscale organization of landscape heterogeneity. Ecosystem management: principles and applications. US Department of Agriculture, Forest Service, Pacific Northwest Research Station, General Technical Report PNW-GTR-319. Portland, Oregon USA, pp.73–79. Turner, Monica G. et al., 1993. A revised concept of landscape equilibrium: Disturbance and stability on scaled landscapes. Landscape Ecology, 8(3), pp.213–227. USDA-SCS (U.S. Department of Agriculture, Soil Conservation Service), 1972. National Engineering Handbook, Hydrology Section 4, chapter 4-10. 141 USEPA (2000) National management measures to control nonpoint source pollution from agriculture. USEPA office of water, Washington, DC. Vellidis, G. et al., 2003. Nutrient Transport in a Restored Riparian Wetland. J. Environ. Qual., 32(2), pp.711–726. Vesk, P.A. et al., 2008. Time lags in provision of habitat resources through revegetation. Biol. Conserv., 141(1), pp.174–186. Viglizzo, E.F. et al., 2011. Ecosystem service evaluation to support land-use policy. Agriculture, Ecosystems & Environment, Online first . Vrieling, A. et al., 2008. Timing of erosion and satellite data: A multi-resolution approach to soil erosion risk mapping. International Journal of Applied Earth Observation and Geoinformation, 10(3), pp.267–281. Wallace, K.J., 2007. Classification of ecosystem services: Problems and solutions. Biological Conservation, 139(3–4), pp.235–246. Wang, G. et al., 2003. Mapping multiple variables for predicting soil loss by geostatistical methods with TM images and a slope map. Photogrammetric engineering and remote sensing, 69(8), pp.889–898. Ward, J.V. & Tockner, K., 2001. Biodiversity: towards a unifying theme for river ecology. Freshwater Biology, 46(6), pp.807–820. Welsch, D.J., 1991. Riparian forest buffers: function and design for protection and enhancement of water resources, US Dept. of Agriculture, Forest Service, Northeastern Area, State & Private Forestry, Forest Resources Management. Weltz, M.A., Renard, K.G. & Simanton, J.R., 1987. Revised universal soil loss equation for western rangelands. General technical report RM-Rocky Mountain Forest and Range Experiment Station, US Department of Agriculture, Forest Service. Wendland, K.J. et al., 2010. Targeting and implementing payments for ecosystem services: Opportunities for bundling biodiversity conservation with carbon and water services in Madagascar. Ecological Economics, 69(11), pp.2093–2107. West, T.O. & Post, W.M., 2002. Soil Organic Carbon Sequestration Rates by Tillage and Crop Rotation. Soil Science Society of America Journal, 66(6), p.1930. Wikiloc, 2010. http://es.wikiloc.com/wikiloc/home.do WWF, 2004. Living Planet Report. World Wide Fund for Nature, Gland, Switzerland. Van Wilgen, B.W., Cowling, R.M. & Burgers, C.J., 1996. Valuation of ecosystem services. BioScience, 46(3), pp.184–189. Winjum, J.K. & Schroeder, P.E., 1997. Forest plantations of the world: their extent, ecological attributes, and carbon storage. Agricultural and Forest Meteorology, 84(1–2), pp.153–167. 142 Wischmeier, W.H. & Smith, D.D., 1978. Predicting rainfall-erosion losses from cropland east of the Rocky Mountains: Guide for selection of practices for soil and water conservation, Agricultural Research Service, US Dept of Agriculture in cooperation with Purdue Agricultural Experiment Station. Wohl, E. et al., 2005. River restoration. Water Resources Research, 41(10), p.W10301. Wu, J., 2004. Effects of changing scale on landscape pattern analysis: scaling relations. Landscape Ecology, 19(2), pp.125–138. Wu, J. et al., 2000. Multiscale analysis of landscape heterogeneity: scale variance and pattern metrics. Annals of GIS, 6(1), pp.6–19. Wu, Q. & Wang, M., 2007. A framework for risk assessment on soil erosion by water using an integrated and systematic approach. Journal of Hydrology, 337(1), pp.11–21. Yuan, J. et al., 2006. Soil Formation and Vegetation Establishment on the Cliff Face of Abandoned Quarries in the Early Stages of Natural Colonization. Restoration Ecology, 14(3), pp.349–356. Zedler, J.B. & Kercher, S., 2005. Wetland resources: status, trends, ecosystem services, and restorability. Annu. Rev. Environ. Resour., 30, pp.39–74. Zhang, W. et al., 2007. Ecosystem services and dis-services to agriculture. Ecological Economics, 64(2), pp.253–260. Ziemer, R.R., 1999. Some scale considerations for watercourse restoration and rehabilitation. Notes: We recommend that you also print this page and attach it to the printout of the article, to retain the full citation information. This article was written and prepared by US Government employees on official time, and is therefore in the public domain. Ziemer, R.R., 1997. Temporal and spatial scales. Watershed restoration: principles and practices. American Fisheries Society, Bethesda, Maryland, pp.80–95. 143 9. Appendix 144