copan:LPJmL: A new hybrid DGVM-based modeling framework for dynamic land use and agricultural management

Dynamic global vegetation models have been established as a useful tool in environmental and agricultural sciences for many purposes, e.g. modeling crop growth, fire disturbances, or biosphere-climate interactions. Nevertheless, DGVMs are often very limited in terms of interactions with the anthroposphere, particularly human-Earth interactions. DGVMs such as LPJmL have been successfully connected to integrated assessment models such as Remind-MAgPIE or IMAGE. Still, the model coupling of those approaches often remains loose and static over the simulation period. copan-LPJmL addresses this issue by providing a standardized Python interface, consisting of the LPJmL coupler extension and the pycoupler, that can be used to exchange LPJmL inputs and outputs annually during the simulation period. In addition to LPJmL and the coupling interface, copan:LPJmL also integrates the world-earth modeling framework copan:CORE, which provides useful standardized abstractions of key entities of such models as the world (the simulation space as a whole), the cell, or the individual (an agent in agent-based modeling (ABM)). With copan:LPJmL, any LPJmL output can be retrieved at the world and cell level, and any input to LPJmL can be returned at the same level. This allows for any modeling to be carried out easily using this structure to interact with LPJmL. We here show three examples making use of this: (1) The model of Integrated social-ecological resilient land systems (InSEEDS), which uses a classical ABM approach to model management decisions by farmers, (2) an adaption of an established crop calendar model and (3) a novel LLM (Large Language model) ABM approach. These three examples show the diversity of models that can be implemented using the copan:LPJmL modeling framework to gain new insights into future potential land use and agricultural pathways in the context of global change.