| Summary: | Climate change, a growing global population and soil degradation put significant
stress on food production and threaten food security, both on a global scale and
in individual agricultural communities. This necessitates studies that explore sustainable
agricultural intensification. Traditional farming systems have received
increased attention, as aspects of these systems (such as niche complementarity)
might provide sustainable solutions. This work centers around the three sisters,
a polyculture of maize (Zea mays), bean (Phaseolus vulgaris) and squash (Cucurbita
spp.), and the milpa, a complex Maya polyculture centered around maize
and bean. Building on an existing functional-structural plant (FSP) model for
maize, a novel FSP model for common bean is developed (in the XL language, on
the GroIMP platform), encompassing twining behaviour and physical plant-plant
interactions. This allows us to simulate maize/bean polycultures, where common
bean climbs upwards around the maize stalk. As the model contains many input
parameters, of which some are difficult or costly to parameterise, a global sensitivity
analysis (GSA) is paramount for identifying (un)important parameters in
the model. This decreases dimensionality of the large model parameter space.
Efforts can then be concentrated on accurately estimating the most important
input parameters. GSA is therefore performed on monocultures of maize and
common bean (growing on poles). To this end, the popular Elementary Effects
GSA method is adapted to make it suitable for models with dimensional inputs,
inputs taking values on arbitrary intervals or discrete inputs. Our results show
the benefit of performing GSA on plant models: for both maize and bean, less
than 30% of input parameters where classified as important for most model outputs.
In addition, performing GSA on plant models leads to new insights about
both the model and the plant developmental processes it describes. The hope is
that this work will inspire more plant modellers to routinely incorporate sensitivity
analysis in their research. Subsequently, the model for maize and bean is
used to assess architectural facilitation in light capture in maize/bean polycultures.
Simulation results agree with experimental observations in the literature
of overyielding in polycultures including maize and climbing bean. This indicates
that aboveground processes (also) play an important role in the phenomenon
of overperforming. In addition, it confirms that such agricultural systems may
play a role in sustainable agricultural intensification. The maize/bean model presented
in this work is one of the first examples of an aboveground FSP model of
a polyculture with complex physical plant-plant interaction. Our results suggest
that FSP modelling could be a valuable tool to investigate such agricultural systems.
In this work, we have shown that it is possible to model maize/bean crop
mixtures, making an aboveground model of the three sisters only a small step
away.
|
|---|