Participatory land use modelling, pathways to an integrated approach Richard Hewitt a, b, * , Hedwig van Delden c, d , Francisco Escobar a, e a University of Alcalá, Departamento de Geografía, Facultad de Filosofía y Letras (Colegio de Málaga), C/Colegios, 2, 28801 Alcalá de Henares, Madrid, Spain b Observatorio para una Cultura del Territorio, C/Duque Fernán Nuñez, 2 e 1  , 28012 Madrid, Spain c Research Institute for Knowledge Systems (RIKS), Maastrichter Pastoorstraat 14, 6211 BV Maastricht, Netherlands d School of Civil, Environmental and Mining Engineering, University of Adelaide, Engineering North N136, North Terrace Campus, SA 5005, Australia e CEPS/INSTEAD, Département de Géographie et Développement, 3, avenue de la Fonte, L-4364 Esch-sur-Alzette, Luxembourg article info Article history: Received 21 March 2013 Received in revised form 14 October 2013 Accepted 15 October 2013 Available online 18 November 2013 Keywords: Cellular-automata Integration Calibration Land use modelling Participation Stakeholders abstract The increasing adoption of land use models in planning and policy development highlights the need for an integrated approach that combines analytical modelling techniques with discursive ‘soft-science’ methodologies. Recent scientific contributions to the discipline have tended to focus on analytical problems such as statistical assessment of model goodness of fit through map comparison techniques, while the problem of integrating stakeholder information into land use models has received little attention. Using the example of a land use model developed for the Guadiamar basin in South West Spain, location of the emblematic Doñana natural area, an integrated methodology for participatory calibration and evaluation of model results is presented which combines information from key stake- holders across a range of sectors with analytical model calibration techniques. Both discursive and analytical techniques are presented side by side to demonstrate that including participatory approaches is likely to improve both calibration results and model applicability. Integration of participatory methods into land use models is more likely to be successful if stakeholders are selected carefully so as to make best possible use of their time and knowledge, and are involved in the modelling process from the beginning of the project cycle. Ó 2013 Elsevier Ltd. All rights reserved. 1. Introduction 1.1. Research background Over the past decades the adoption of land use models in planning and policy making has increased dramatically (Seaton, 2001; Oxley et al., 2004; Encinas et al., 2006; Engelen et al., 2007). This has required the deployment of methods that cross disciplines and research communities, linking “soft” (humanistic, discursive) and “hard” (analytical, natural) science approaches. Soft-science approaches try to take into account the inherent unpredictability of human behaviour and the capacity of human agents to change the system from within. Hard-science approaches assume the collection of beliefs and perceptions which make up our view of the world to be static for the purpose of investigating a particular theory or problem (Winder, 2004). Soft-science methods are useful in cases where human behaviour or interaction is important (e.g. land use policy), and may often involve participa- tory or social enquiry techniques which provide qualitative or approximate information (Lemon et al., 1994). Hard-science ap- proaches are relevant to the study of natural phenomena (e.g. degradation of a natural resource), and involve mathematical and quantitative methods which provide precise, numerical data. In cases of humaneenvironment interaction, as in a land use change model, both kinds of information are necessary and integrative approaches that combine hard and soft-science methodologies are therefore important. As land use models have become more widely used, spatial modelling frameworks such as Metronamica (RIKS, 2011; Van Delden and Hurkens, 2011) and CLUE (Veldkamp and Fresco, 1996; Verburg et al., 2008) have been developed, obviating the need to design a new system every time. Apart from the clear advantage of time-saving, the principal benefit of applying existing modelling frameworks to new regions rather than developing models from scratch for each new research project is that the model concepts and mechanisms tend to become better tested over time. Thus, the emphasis has come to rest on calibration, that is, the adaptation of these existing frameworks to a particular case study * Corresponding author. University of Alcalá, Departamento de Geografía, Fac- ultad de Filosofía y Letras (Colegio de Málaga), C/Colegios, 2, 28801 Alcalá de Henares, Madrid, Spain. Tel.: þ34 673603561, þ34 918854482. E-mail addresses: richardjhewitt@hushmail.com (R. Hewitt), hvdelden@riks.nl (H. van Delden), francisco.escobar@uah.es (F. Escobar). Contents lists available at ScienceDirect Environmental Modelling & Software journal homepage: www.elsevier.com/locate/envsoft 1364-8152/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.envsoft.2013.10.019 Environmental Modelling & Software 52 (2014) 149e165