Addressing the underrepresentation of African ecosystems in plant traits, adaptation, and biogeochemical cycles: Mapping traits for model parameterisation

Plant trait observations provide a critical bridge between organismal strategies and ecosystem-scale biogeochemical cycling, yet trait coverage and the evidence base used for land-surface model (LSM) parameterisation remain geographically uneven. Current LSM parameterisations are biased because African floras and trait data are underrepresented in global syntheses. This limits how well LSMs represent the diversity of plant strategies across African landscapes, contributing to uncertainty and regional bias in simulations of carbon and water cycling under rapid environmental change. A key need is therefore a transparent measurements-to-models pathway that converts heterogeneous trait and botanical information into model-ready functional groupings that can directly support parameterisation and evaluation of LSMs.

Here we operationalise a transparent trait-to-PFT translation layer for LSM parameterisation. We used reproducible, rule-based workflow that could be applied across floristic regions to map African plant species represented in the TRY plant trait database to the Joint UK Land Environment Simulator (JULES) plant functional types (PFT) taxonomy. We assigned classification attributes including growth form, leaf type, leaf phenology, photosynthetic pathway, and climatic zone, and implemented decision rules to generate consistent PFT assignments. This process mapped 1603 plant species from 137 families to JULES PFTs. Our output provides JULES-ready PFT labels alongside decision metadata that document classification rules, provenance, and the inputs that supported each assignment, enabling reuse across different PFT taxonomies, as well as sensitivity testing of alternative classification choices.

Building on the capabilities of the existing TRY categorical lookup table, this exercise has yielded a five-fold increase in the number of plant trait observations linked with JULES PFTs across the continent of Africa. Using these records, we derive PFT-level trait distributions for Africa and translate them into trait-informed ranges for critical JULES vegetation parameters. We quantify how these Africa-specific ranges differ from current global default PFT values. These constraints provide a practical route to more defensible parameter choices and more targeted sensitivity analyses. Together, the dataset and parameter summaries support improved integration of existing and future plant trait data into PFT parameterisations in land surface models and similar large scale modelling exercises, to enhance the representation of African ecosystems and better constrain the uncertainty when modelling global systems.