ED-MORE is a size- and age-structured terrestrial biosphere model with fully coupled carbon, energy and water dynamics, written in Fortran 2018. A forest is cohorts of plants on patches of land, with fusion and fission keeping both bounded. At every sub-daily step, ED-MORE solves the carbon, energy and water dynamics of the canopy air, the leaves, the wood, the snow and the soil together, so leaf temperature, canopy humidity and soil moisture are computed rather than read from a driver file. The demography then grows, kills and recruits plants on the carbon this produces. A run can be a single site, a polygon divided into sites with their own soil, elevation and slope, or a region in which every ERA5-Land grid cell within a box is its own polygon.
Status: v0.4.1, the first ED-MORE release, not yet benchmarked. The examples fit some parameters to observations (a flux tower, a forest census, leaf area and litter) and set a regional run against satellite leaf area, but no regression suite against reference runs exists yet, and no systematic evaluation against observations has been made. What is verified is internal: the test suite on two compilers, the per-step conservation budgets, and thread invariance. Treat the numbers a run produces as a working model's numbers. The changelog records which numbers each release moved.
A forest regrowing from near-bare ground over 300 years (example 03): the leaf-area profile beside a stand cross-section (left), and a landscape whose trees keep their place as they grow (right).
mamba env create -f environment.yml # once: netCDF, CMake and the Python tools
conda activate edmore
FC=ifx cmake --preset dev # the dev build, in build-dev/; netCDF from $CONDA_PREFIX
cmake --build build-dev -j
ctest --test-dir build-dev --output-on-failure
# A forest regrowing from near-bare ground for 300 years (example 03; about a minute)
PYTHONPATH=python python examples/example03_demography/run_demography.py --start bare
python examples/example03_demography/plot_site_timeseries.py examples/example03_demography/output/bare.nc -o timeseries.pngnetCDF is required: environment.yml brings it, with CMake and the Python
packages the examples use, and without conda scripts/install_netcdf.sh installs it. gfortran, ifx
and nvfortran are supported. docs/building.md covers the four build kinds
(debug, dev, profile, and production for long runs), the parallel builds and the compiler traps. The
dev build also makes the shared library the Python front end loads
(python/README.md). Each example's README says which data it needs and how to
run it.
| If you want to | Read |
|---|---|
| learn what is in the model: each process and its equations | docs/science/, one page per process, indexed in docs/README.md |
| run ED-MORE, coming from ED2 | docs/ed2_comparison.md: the main differences first, then process by process, keyed to ED2IN |
| set up a run | docs/configuration.md: the two TOML files and the three ways to start |
| see worked examples | examples/: leaf gas exchange, phenology, demography, a flux tower, forest regrowth, a region |
| prepare forcing, census and trait inputs | scripts/prepare_input/: ERA5-Land, flux-tower, census, CO₂ and PFT-trait tools |
| calibrate the fast processes at a flux tower | scripts/calibrate_fast/ |
| drive it from Python | python/README.md: the edmore package |
| plot the output | example 03 (its stand file) and example 05 (edmore's diagnostic files) |
| change the code | Developing ED-MORE below, then CONTRIBUTING.md |
| know what changed | CHANGELOG-v0.md |
The name states four principles. The rules that follow from them, each with the check that enforces
it where one exists, are in AGENTS.md: written so that AI coding agents can follow
them, and binding on everyone who changes the code. How to propose, send and merge a change is in
CONTRIBUTING.md.
- Modular: one job per piece, and one place for each piece. The source splits by timescale, a
fast sub-daily loop and a slow daily one, over a state layer both share. A process kernel sees only
the stores it works on, dependencies point one way, and there is no hidden global state: config and
state are passed as arguments, so every routine can be tested alone.
src/README.mdsays where a new file goes. - Open: anyone can read the model, check it and reuse it. The code is under Apache 2.0. Every
equation is published in
docs/science/, versioned with the code that implements it, and the code cites it. Comments say what the code does and why, in plain words. Every parameter value carries its source, and decisions are made in the open, on issues and pull requests. - Reproducible: the same inputs give the same answer, and every change to the answer is declared. Every parameter comes from the config files, and a run refuses what it cannot do right rather than fall back in silence. Output is identical at any thread count, a state file restarts a run bit for bit, and every budget is checked every step. Every pull request states its effect on results.
- Extensible: new science goes in without rewiring what is there. A process is a module behind the interface its driver already calls, a choice of scheme is a selector, and adding a config key, a per-cohort field or an output variable follows a fixed recipe. Outside code drives the model through its C API, with the same operators the model uses.
The code is modern Fortran 2018: modules, derived types, explicit interfaces, allocatable
ownership, parameterized kinds, pure and elemental kernels, implicit none everywhere, and at
most 132 columns. docs/building.md lists the compiler traps: code that one compiler builds
correctly and another miscompiles.
ED-MORE carries forward, from version 0.4.1, the development of its prototype, MEDS. The MEDS repository, with its history, issues and pull requests up to v0.4.0, is archived at https://github.com/xiangtaoxu/MEDS.
Cite the version you used, as CITATION.cff gives it; GitHub's "Cite this
repository" button formats it. Zenodo archives every release with its own DOI; the DOI
10.5281/zenodo.23674060 stands for all of them and
resolves to the latest. The science ED-MORE builds on:
- Moorcroft, P. R., Hurtt, G. C. and Pacala, S. W. (2001). A method for scaling vegetation dynamics: the Ecosystem Demography model (ED). Ecological Monographs 71(4), 557–586. https://doi.org/10.1890/0012-9615(2001)071[0557:AMFSVD]2.0.CO;2
- Medvigy, D., Wofsy, S. C., Munger, J. W., Hollinger, D. Y. and Moorcroft, P. R. (2009). Mechanistic scaling of ecosystem function and dynamics in space and time: Ecosystem Demography model version 2. Journal of Geophysical Research: Biogeosciences 114. https://doi.org/10.1029/2008JG000812
- Xu, X., Medvigy, D., Powers, J. S., Becknell, J. M. and Guan, K. (2016). Diversity in plant hydraulic traits explains seasonal and inter-annual variations of vegetation dynamics in seasonally dry tropical forests. New Phytologist 212(1), 80–95. https://doi.org/10.1111/nph.14009
- Longo, M., et al. (2019). The biophysics, ecology, and biogeochemistry of functionally diverse, vertically and horizontally heterogeneous ecosystems: the Ecosystem Demography model, version 2.2 – Part 1: Model description. Geoscientific Model Development 12, 4309–4346. https://doi.org/10.5194/gmd-12-4309-2019
These papers describe ED and its successor ED2, which is developed at https://github.com/EDmodel/ED2.
ED-MORE is released under the Apache License 2.0. NOTICE carries the
copyright notice and the attributions to MEDS and ED2, and AUTHORS lists the copyright
holders. Contributions are welcome under the same licence; see CONTRIBUTING.md.