APPLICATION OF PROBA-1/CHRIS DATA TO DETERMINE DEGRADATION INDICATORS WITHIN THE SEMI-ARID WETLAND AREA (NATIONAL PARK LAS TABLAS DE DAIMIEL, SPAIN) Thomas Schmid a, *, José Antonio Domínguez b , Jesús Solana c , José Gumuzzio d , Magaly Koch e a CIEMAT, Avda. Complutense 22, 28040 Madrid, Spain - thomas.schmid@ciemat.es b CEDEX, Paseo Bajo de la Virgen del Puerto, 3, 28005 Madrid, Spain - jose.a.dominguez@cedex.es. c Universidad Complutense de Madrid, Faculty of Geography and History, 28040 Madrid, Spain – jesus.solana@ghis.ucm.es. d Universidad Autónoma de Madrid, Science Faculty, Madrid, Spain – jose.gumuzzio@uam.es. e Center for Remote Sensing, Boston University, Boston, MA 02215, USA – mkoch@bu.edu. KEY WORDS: Semiarid wetlands, PROBA-1/CHRIS, degradation indicators, endmembers, MTMF ABSTRACT: The National Park of Las Tablas de Daimiel in Central Spain is a wetland area of great value regarding biodiversity and wildlife habitat. This type of ecosystem is unique within a semi-arid climate and especially sensitive to degradation processes caused by growing pressure on natural resources (agricultural practices, groundwater depletion and land use change) as well as climatic changes. The objective of this work is to determine semi-arid wetland characteristics that are related to indicators of quality or degradation that represent the wetland state as a result of human-induced activities applying multi-angle Proba-1/CHRIS data. Data acquisitions were taken when the wetland area was under the influence of dry or wet conditions as a result of management practices intended to regulate the ecosystem habitat. The methodology is focused on obtaining wetland information related to wetland indicators (vegetation, water mass, soil and sediments). Endmembers are identified and a partial spectral unmixing is carried out applying a Mixture Tuned Matched Filtering. Results show that image-derived endmembers from the multi-angular Proba-1/CHRIS data identified a series of surface components, which are associated to the conditions and quality of the wetland area. These include important wetland indicators such as palustrine vegetation species (Pragmites australis and Cladium mariscus) and shallow water bodies. Fractional abundance maps were obtained for endmembers that were well represented within the wetland area. In this case, the identification and spatial distribution of Phragmites australis can be used as an indicator that is affecting the emblematic Cladium mariscus. * Corresponding author. 1. INTRODUCTION Wetland areas in semi-arid regions are very important and vulnerable ecosystems. These complex and fragile ecosystems are especially important in maintaining and controlling the environmental quality and biodiversity and often undergo extreme changes from the wet to the dry season (Millennium Ecosystem Assessment, 2005; Carrasco, 2006). In Central Spain, small wetland areas are spread throughout the region and are mainly threatened by human-induced activities causing rapid processes of degradation with loss of wetland surface and, eventually leading to their disappearance (Cirujano and Medina, 2002). Activities such as intensive irrigation, overexploitation of groundwater, drying and artificial drainage of wetlands as well as altering and channelling of rivers have severely affected the wetlands and their natural functions. This is the case of the semi-arid wetland “Las Tablas de Daimiel”, National Park included in the UNESCO Man and Biosphere Reserve, Special Protection Areas for Birds (SPA) and Ramsar Convention. This wetland is a refuge for endemic and threatened species of flora and fauna, but the human-induced changes lead to a loss of characteristic vegetation and large colonies of anatid invernates. Nowadays, degradation processes such as desiccation, salinization, euthrophication and contamination are seriously affecting the wetland area (Sánchez-Carrillo and Álvarez-Cobelas, 2001; Alvarez-Cobelas et al., 2007). The monitoring of the conditions of wetlands is important in order to manage and preserve these natural ecosystems. The challenge is to detect, with adequate precision and at different temporal and spatial scales, any rapid changes in the composition of the surface land cover (soil, vegetation and water). These changes can be detected with Earth Observation (EO) data, especially because wetlands are dynamic systems requiring frequent observations to capture seasonal and inter annual dynamics related to hydrological processes and human impact. Wetland areas under semi-arid conditions have been carried out in previous studies applying different types of optical sensors with hyperspectral and multispectral data acquisition capabilities (Koch et al., 2001; Schmid, et al., 2005; Andrew and Ustin, 2008). An integrated approach has been carried out where EO information from different sensors has been applied to these types of wetlands (Schmid et al., 2008a; 2008b). The objective of this work is to determine semi-arid wetland characteristics related to indicators of quality or degradation that represent the wetland state of the National Park of Daimiel as a result of human-induced activities applying multi-angle Proba- 1/CHRIS data. The work is further carried out with field and imaging spectroscopy data, soil and water analyses, field survey observations, topographic information and ancillary data. This includes selecting wetland indicators that are representative for emerging and aquatic vegetation, shallow water bodies, soils and sediments, developing a spectral library containing wetland