Modelled agroforestry outputs at field and farm scale to support biophysical and environmental assessments

Palma, J.H.N., Oliveira, T., Crous-Duran, J., Graves, A.R., Garcia de Jalon, S., Upson, M., Giannitsopoulos, M., Burgess, P.J., Paulo, J.A., Tomé, M., Ferreiro-Dominguéz, N., Mosquera-Losada, M.R., Gonzalez-Hernández, P., Kay, S., Mirk, J., Kanzler, M., Smith, J., Moreno, G., Pantera, A., Mantovani, D., Rosati, A., Luske, B., Hermansen, J. (2017). Deliverable 6.17 (6.2): Modelled agroforestry outputs at field and farm scale to support biophysical and environmental assessments AGFORWARD project. 18 October 2017. 162 pp.

This report, comprising Deliverable 6.17, in the AGFORWARD project brings together examples of modelled outputs at field and farm scale to support the biophysical, social, and environmental assessment of the innovations selected from work-packages 2 to 5.

After an introduction, the report provides a summary of the main questions, which could be addressed through modelling, that arose from four participative research and development networks in the AGFORWARD project. The networks focused on agroforestry systems of high nature and cultural value, agroforestry with high value tree systems, agroforestry for arable systems, or agroforestry for livestock systems. The report explains how modelling workshops were held in Portugal, Greece and UK to enable data collection and improve the understanding of the management of the systems to be modelled.

The third section of the report describes the additional developments made to the Yield-SAFE model. An important development has been the creation of the CliPick tool to provide daily climate data (needed to run the Yield-SAFE model) for any location in Europe. The Yield-SAFE model has also been developed to include a livestock component based on the utilisable metabolisable energy of the animal feed produced in agricultural, agroforestry or plantation systems (Section 3.2.1). The model also includes an assessment of the possible requirement of livestock for shade (Section 3.2.2) and the production of fruit by trees (Section 3.2.3). The Yield-SAFE model has also been modified to include the effect of trees in modifying the micro-climate and thereby the seasonal production of pasture (Section 3.3). Substantial work has also been undertaken to include the RothC model (describing soil carbon dynamics) within the model (Section 3.4). A full list of the additional variables and outputs included in the updated version of the model is described in Annex VII.


The fourth section briefly describes the steps in calibrating the model to include a new tree or crop species (Section 4). Although the section is short, it describes the approach used to establish a range of calibrations described in the Annexes. These include the default Yield-SAFE soil (Annex I), tree (Annex II), crop (Annex III), livestock (Annex IV) and Roth-C (Annex V) parameters. Annex VII describes new calibrations for ten additional tree species (Table 1). Annex VIII describes updated calibrations for three types of pasture and five crop species (Table 2).

Table 1. The updated Yield-SAFE model includes calibrations for ten additional tree species

Common name Latin name
Blue gum Eucalyptus globulus
Holm oak Quercus rotundifolia
Black walnut Juglans major
Spruce Picea abies
Cherry tree (for fruit production) Prunus avium
Cherry tree (for timber production) Prunus avium
Apple tree Malus domestica
Poplar in short rotation coppice Populus species
Willow in short rotation coppice Salix species
Radiata pine Pinus radiata
Chestnut Castanea sativa


Table 2. The updated Yield-SAFE model includes updated calibrations for nine crop species

Crop Country
Natural Mediterranean pasture/grass Portugal
Pasture (<80% Dactylis glomerata>) Spain
Wheat UK
Barley UK
Barley for RothC calibration UK
Grassland Switzerland
Winter rye Switzerland
Sugar beet Italy
Asparagus Italy


The fifth section of the report (Section 5) describes the use of the updated Yield-SAFE model to describe the provisioning ecosystem services (in terms of food, materials and energy) of agricultural, agroforestry and forestry systems in four situations: montado in Portugal, cherry tree pastures in Switzerland, silvoarable systems in the UK, and short rotation coppice systems in Germany. The results generally show that an increase in tree density led to an increase in the “extractable” energy in the Swiss, British and German systems; in the Portuguese system it was assumed that the bioenergy stored in the trees would not be harvested.

The final section (Section 6) provides a synthesis of the work. It argues that agroforestry system assessment is now supported by an enhanced suite of modelling tools. This includes the improvement of Yield-SAFE and bespoke models such as “Forage-SAFE” and “EcoYield-SAFE”. The process has also enabled a new cohort of European researchers to develop, use, and publish the results of using the models to inform on-farm decisions.