Ecological Indicators 48 (2015) 396–407 Contents lists available at ScienceDirect Ecological Indicators jo ur nal ho me page: www.elsevier.com/locate/ ecolind Assessing the impact of agricultural pressures on N and P loads and eutrophication risk Rémi Dupas a,b,∗ , Magalie Delmas c , Jean-Marcel Dorioz d , Josette Garnier e , Florentina Moatar f , Chantal Gascuel-Odoux a,b a INRA, UMR1069, Sol Agro and hydroSystem, F-35000 Rennes, France b Agrocampus Ouest, Sol Agro and hydroSystem, F-35000 Rennes, France c INRA, Infosol, F-45075 Orléans, France d INRA, UMR Carrtel, F74203 BP 11 Thonon les bains, France e CNRS, Univ Paris 06, UMR Sisyphe 7619, F-75252 Paris, France f Univ Tours, Fac Sci & Tech, EA 6293, F-37200 Tours, France a r t i c l e i n f o Article history: Received 26 May 2014 Received in revised form 11 July 2014 Accepted 6 August 2014 Keywords: Nitrogen Phosphorus Mass-balance model Eutrophication Indicator Agriculture a b s t r a c t Excessive nutrient delivery into freshwater bodies results in increased eutrophication risk worldwide. Because high-frequency monitoring cannot be generalised to all rivers, methods are needed to assess eutrophication risk in contexts with scarce data. To this end, we present an assessment framework which includes: (i) a mass-balance model to estimate diffuse N and P transfer/retention in unmonitored catch- ments and (ii) a set of indicators based on N:P:Si molar ratios to assess the risk of eutrophication in freshwaters. The model, called Nutting, integrates variables that describe both agricultural pressures and physical attributes of catchments (climate, topography, soil). Nutting refines previous mass-balance mod- els by describing nutrient pressures with soil N surplus and soil P content instead of N and P inputs, and by considering physical attributes not only as lumped variables over the entire area but also within river corridors. The model was calibrated on a set of 160 independent catchments across France and applied to all headwater catchments. We found that apparent N and P retention represented 53 ± 24% and 95 ± 29% of soil N and P surplus, respectively, and was mainly controlled by the climate and a hydrology-related connectivity index. The spatial organisation of the landscape was of secondary importance compared to the refined description of agricultural pressures. Estimated eutrophication risk was highly sensitive to assumptions about P bioavailability, hence the potential range of headwaters at risk of eutrophica- tion spanned 26–63% of the catchments, depending on assumptions. This framework provides a generic method to assess the relative contribution of agriculture to nutrient loads and the subsequent risk of eutrophication. © 2014 Elsevier Ltd. All rights reserved. 1. Introduction Degradation of surface water resulting from excessive nitrogen (N) and phosphorus (P) inputs is a major concern for drinking water quality and ecosystem health (Carpenter et al., 1998; Vitousek et al., 1997). Due to improvement of point-source control in recent decades, research and management efforts to decrease nutrient pollution have been redirected towards diffuse sources (Van Drecht et al., 2009). In industrialised countries, most source apportion- ment studies have shown that agriculture is currently a major ∗ Corresponding author at: INRA, UMR1069, Sol Agro and hydroSystem, F-35000 Rennes, France. Tel.: +33 223487047. E-mail address: rdupas@agrocampus-ouest.fr (R. Dupas). source of nutrients in surface waters (e.g. Bouraoui and Grizzetti, 2011; Grizzetti et al., 2012; Windolf et al., 2012). Nutrient transfers from agricultural landscapes result from combined hydrological and biogeochemical processes controlling their mobilisation and delivery in the terrestrial and aquatic compartments of catchments (Bouwman et al., 2013; Haygarth et al., 2005; Seitzinger et al., 2006). These processes are difficult to understand and to model due to the complexity of agricultural landscapes (Burt and Pinay, 2005; Strayer et al., 2003). Low-order catchments (i.e. below Strahler order 5) are known to contribute large amounts of nutrient loads to downstream water bodies (Alexander et al., 2007; Lassaletta et al., 2010; Peterson et al., 2001). In terms of scientifically understanding the processes controlling N and P diffuse transfer/retention in agricultural land- scapes, low-order catchments are relevant spatial units because http://dx.doi.org/10.1016/j.ecolind.2014.08.007 1470-160X/© 2014 Elsevier Ltd. All rights reserved.