Abstract: • Evidence use in EU Impact Assessment reports is limited. • Many scientific studies used farm models for assessment of policies in the EU. • Scientific challenges include understanding farmer decision-making and interactions. • Model codes and data should be published, including evaluation. • Stronger science-policy interaction is required. Farm models are potentially relevant tools for policy impact assessment. Governments and international organizations use impact assessment (IA) as an ex-ante policy process and procedure to evaluate impacts of policy options as part of the introduction of new policies. IA is increasingly used. This paper reviews both the use of farm models in such policy IAs in the European Commission, and the development and use of farm models for policy IA by the scientific community over the past decade. A systematic review was performed, based on 202 studies from the period 2007–2015 and results were discussed in a science-policy workshop. Based on the literature review and the workshop, this paper describes progress in the development of farm models, challenges in their use in policy processes and a research and cooperation agenda. We conclude that main issues for a research agenda include: 1) better understanding of farmer decision-making and effects of the social milieu, with increased focus on the interactions between farmers and other actors, the link to the value chain, and farm structural change; 2) thorough and consistent model evaluation and model comparison, with increased attention for model sensitivity and uncertainty, and 3) the organization of a network of farm modellers. In addition, the agenda for science-policy cooperation emphasizes the need for: 4) synthesizing research evidence into systematic reviews as an institutional element in the existing science-policy-interfaces for agricultural systems, 5) improved and timely data collection, allowing to assess heterogeneity in farm objectives, management and indicators, and 6) stronger science-policy interaction, moving from a research-driven to a user-driven approach.

Abstract: This deliverable lays out the work as done as part of MACSUR CropM on data, with the  focus on improving data management and have shared data curation for future use. The  issue was tackled with help from the MACSUR central hub coordination in the form of Jason  Jargenson from University of Reading. The data management as proposed and  implemented in this deliverable is very much a bottom up process, in which partners in a  meeting in Spring 2013 in Aarhus investigated the best way forward for data management  across activities in CropM.As a follow up to this, the work was mainly divided in three  parts:  1. The  Open  Data  Journal  for  Agricultural  Research,  mainly  focused  on  long  term  data  archival  and  citation  of  data  sets,  as  input  and  outputs  to  the  modelling  work,  as  part  of  MACSUR,  lead  by  Wageningen  UR  2. The  Geonetwork  data  catalog  hosted  at  Aarhus  Universitet,  that  allows  for  operational  access  and  storage  of  data  sets  as  part  of  the  ongoing  work,  also  for  restricted  access  of  the  consortium,  and  as  a  first  step  to  visualization,  lead  by  Aarhus  Universitet.  3. The  work  on  rating  data  sets,  that  provides  a  tool  for  improving  data  set  access  in  an  early  phase  for  connecting  them  to  models,  lead  by  Reading  University.  At the end of the deliverable some next steps are giving for data activities in the context  of AgMIP and beyond. No Label