Conclusion
Forests play multiple important roles for the Earth system. Sound,
quantitative knowledge of forest functioning, structure and diversity is
therefore essential, especially in times of global change. However, many
scientific questions regarding forest properties and dynamics remain
partly unresolved, ranging from understanding tree community assembly
and projecting forest responses to environmental changes, to assessing
the management of forest ecosystems. We illustrated how different forest
modelling approaches, due to their continuous development, their
complementarity and mutual enrichment, represent an invaluable toolkit
to address multiple ecological questions that require a renewed research
effort.
The development of forest models crucially benefits from the
interactions among scientists from various fields, within and across
modelling communities, but also with field ecologists, physiologists,
data scientists, computer engineers, remote-sensing researchers, and a
variety of stakeholders. Owing to their long and successful history in
integrating data and knowledge from these various sources, the models
used to simulate forests have progressively reached maturity to tackle a
broader array of ecological problems. For instance, forest models prove
essential to understand the multiple drivers of forest productivity and
biomass by combining field and remote-sensing data across space and
time, and, as a result, provide informed quantification of carbon stocks
and fluxes. Forest models also provide tractable platforms to perform
virtual experiments still out of reach of empirical approaches on forest
systems that are characterized by slow dynamics and large spatial
extents. This notably allows to shed light on the complex links between
forest biodiversity, functioning and resilience in the long term.
Furthermore, forest models can disentangle the drivers of community
assembly in forest communities, thus complementing theoretical
approaches that typically remain limited to simplified systems. Last but
not least, ongoing global change and the resulting biodiversity crisis,
changing climate and disturbance regimes crucially increase the demand
of informed projections on forest socio-ecosystems, for which forest
models have a proven long history, while new developments allow for the
integration of an increasing number of interacting factors.
We further demonstrated that the converging trajectories of the
different modelling approaches used to simulate forests have provided
new opportunities for comparisons among their outputs. This allows for
the quantification of simulation uncertainties and the identification of
their sources, and hence informs and fosters new model developments as
well as empirical investigations. Overall, iterative model-data fusion
approaches and the resulting cycles of simulation-assessment-improvement
are continuously increasing the scope of model applications. Forest
models will thus keep on contributing to a deeper understanding of
forest structure and functioning, and they offer promising routes to
fill remaining knowledge gaps and to take on future challenges of forest
ecology.