Spatial modelling of agro-ecosystem dynamics across scales: A case in the cotton region of West-Burkina Faso

Development of an integrated generic model for multi-scale assessment of the impacts of agro-ecosystems on major ecosystem services in West Africa

“…For instance, in the model the land allocation for crop production depends on crop yields depending on soil fertility and climate. In terms of decision making and rationality, in most existing integrated agro-ecosystem models, farmers decision making is optimization-based (Castella et al, 2007;Sieber et al, 2008;van Ittersum et al, 2008) and the land use and land cover change is represented by probabilistic transition rules (Castella et al, 2007;Jahel et al, 2016). These last authors used statistical approach to simulate the land use and land cover change based on the past and present data as in most existing agro-ecosystem models.…”

“…Each scale is represented as an autonomous component/ entity (or model) encompassing its own dynamics and interacting for information exchange with the other components of the model. This approach provides flexibility and reuse of the model structure and avoids aggregation and disaggregation of information between scales which cause information distortion (Jahel et al, 2016).…”

“…Farm remains an aggregation of plot and the dynamics at the system level is the aggregation of information from the farms. In Jahel et al (2016) although one can add new scale, the existing scales cannot be removed.…”

“…Previous IAM failed to integrate the interactions between different scales without information distortion between scales, through aggregation/disaggregation processes (Castella et al, 2007;van Ittersum et al, 2008) (Jahel et al, 2016). Even more, most of these IAM are restricted to one scale and context (Belem et al, 2011;Castella et al, 2007;Jahel et al, 2016;van Ittersum et al, 2008) making these models site-specific and limiting their role in knowledge integration.…”

“…Previous IAM failed to integrate the interactions between different scales without information distortion between scales, through aggregation/disaggregation processes (Castella et al, 2007;van Ittersum et al, 2008) (Jahel et al, 2016). Even more, most of these IAM are restricted to one scale and context (Belem et al, 2011;Castella et al, 2007;Jahel et al, 2016;van Ittersum et al, 2008) making these models site-specific and limiting their role in knowledge integration. IAM integrate socio-economic, biophysical and policy dimensions but without an explicit formalization of these dimensions and their interactions at different scales (Piorr et al, 2009;Sieber et al, 2008;van Ittersum et al, 2008), reducing thereby their genericity.…”