Vincent de Leijster

90 Chapter 5 (C. alliodora) between 2001 and 2003 (Collazos Quintana, 2004; FNC, 2011). We asked the participating owners or managers of the farms in which year they transitioned their farm from a coffee monoculture to an agroforestry system, which we used as the ‘time since agroforestry’ variable. For more information about the validation of the chronosequence and for a representation of the number of farms per ages classes since transition, we refer the reader to the appendix. 5.2.3 Field data collection The data of this study was collected through field measurements and farmers surveys in April and May 2018, and again in April and May 2019. In each farm, we selected a site with productive coffee plants that the owner or manager considered representative of the farm because of its cropping density, production capacity and applied inputs, while avoiding the edge of the farm. In this site, a square plot of 20 by 20 m was randomly selected, henceforth referred to as ‘plot’. Through a survey with the owner or manager of the farm, we obtained general farm and management characteristics. 5.2.3.1 General farm and management characteristics We used a structured questionnaire to survey 74 farmers (60 agroforestry coffee, 14 monoculture coffee) who owned or managed the farms on which the biophysical data was collected. Farmers provided information about the general characteristics of their farm (farm size and coffee planting density), coffee production (coffee yields of ’18 and when this was registered of ’17, ’16, and ‘15, in units of dried green coffee) and since when they planted shading trees. Further, we collected information about the management of pest control, weeding, and fertilization and pruning of coffee plants. For each management activity we collected information about how it was done (manually, mechanically, organically, or chemically), labor hours, quantity of agrochemicals, and costs of agrochemicals (Appendix Table A5-1). These variables were then standardized by difference to the minimum and dividing by the difference between the maximum and minimum. Finally, we visually recorded the spatial arrangements of the shade trees using the following categories: ‘Dispersed’ where the majority of the trees were dispersed and randomly located throughout the farm, ‘Alleys’, where the majority of the trees were placed in alley formation on the farm, or ‘Living fences’, where the majority of the trees were bordering the farm.