HYDROLOGICAL PROCESSES Hydrol. Process. 26, 255–271 (2012) Published online 5 July 2011 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/hyp.8199 Impact of the spatial arrangement of land management practices on surface runoff for small catchments F. Colin, 1 * R. Moussa 2 and X. Louchart 2 1 Montpellier SupAgro, UMR LISAH, 34060 Montpellier Cedex 1, France 2 INRA, UMR LISAH, 34060 Montpellier Cedex 1, France Abstract: Predicting the impact of land use changes on the hydrological response is crucial for water resource management. In the particular case of small catchments (1–10 km 2 ), distributed models could provide useful answers regarding the effects of cultivation practices and man-made works on water fluxes. However, the impacts of specific land use spatial arrangements are difficult to predict because of the prohibitive number of possible cases to consider. Focusing on surface runoff, this article describes a strategy based on a water particle tracking routine to be plugged-in a distributed model that is designed to determine the spatial arrangements of land management practices that have the greatest impact on volume, peak discharge and lag time at the catchment outlet. A case study is described; the hydrological response of the Roujan catchment (Herault, France) is simulated with the MHYDAS model. The Roujan catchment contains a vineyard in a Mediterranean climate in a landscape in which weeding practices highly influence the partition between soil infiltration and runoff. Results showed that the proposed strategy is much more efficient than a random approach to design the spatial arrangements of the vineyard weeding practices with the greatest impact. Therefore, the proposed strategy may lead to innovative policies for the spatial planning of land management practices. Copyright  2011 John Wiley & Sons, Ltd. KEY WORDS distributed model; MHYDAS; water particle routine; plug-in; Mediterranean climate; vineyard; weeding practices Received 6 December 2010; Accepted 2 June 2011 INTRODUCTION Modelling the hydrological impacts of landuse changes in small agricultural catchments Changes in land use may have a large impact on the hydrological fluxes at local, hillslope, small catchment, mesoscale or large catchment scales, as shown in recent reviews (Brown et al., 2005; Fohrer et al., 2005). Under- standing and predicting these impacts are major chal- lenges for hydrological research (De Fries and Eshleman, 2004). Physically based hydrological models are able to provide a response to this issue. Use of these models has changed over the past few decades. Because of their development and availability, such models now allow catchments to be managed to anticipate flood events, the pollutant fate and transport or even soil erosion. Recent studies have shown that many models yield positive feed- back in various contexts. Lukey et al. (2000) reported that the SHETRAN model could be used to simulate the impact of reforestation on both runoff and erosion within a badlands environment in south-eastern France. In addition, the SWAT model (Arnold et al., 1998) has produced satisfactory results in the simulation of pro- cesses that affect water fluxes, sediment and nutrient loads inside a catchment (Abbaspour et al., 2007; Bosch, 2008). Martinez-Carreras et al. (2007) reported that the * Correspondence to: F. Colin, Montpellier SupAgro, UMR LISAH, 2, place Viala, 34060 Montpellier Cedex 1, France. E-mail: colinf@supagro.inra.fr KINEROS2 model could be successfully used to simulate badland erosion in a small Mediterranean mountain catch- ment. These studies show that physically based models are relevant to quantify and compare situations that are either derived from actual conditions or constructed from scenarios. From a land management perspective, this type of model allows for the testing of different scenarios in order to develop a scenario hierarchy. Environmental policies involve the implementation of land management practices that are compatible with the preservation of water resources. A hydrological diagno- sis is necessary in order to choose the most relevant action zones and to make recommendations for alterna- tive land uses, cultivation practices and/or their spatial arrangements (Wu et al., 2001; Montero and Brasa, 2005; Pandey et al., 2009). Such an evaluation must be exam- ined at different scales. For example, water resources are managed at the mesoscale or larger, while land man- agement, in particular agricultural production, is gener- ally planned at the field or the hillslope scale accord- ing to farm strategy and coordination (Heathwaite et al., 2005; Heckrath et al., 2008). Between these two lim- its, small catchments (1–10 km 2 ) are at the appropriate scale for both physically based hydrological modelling (Quinn, 2004) and decision making related to field man- agement. Small catchment scales allow both to produce a fine description of the catchment and to character- ize the hydrological impacts of land use, cultivation practices and their spatial arrangements (Bormann et al., 1999). Copyright  2011 John Wiley & Sons, Ltd.