Repository contains all the code needed to reproduce the analysis presented in the manuscript Recent climate change reduced Spanish forests' carbon sink capacity. The analysis pipeline can be used to infer woody biomass measurements from satellite imagery, analyse spatio-temporal patterns and relationship with recent climatic anomalies.
The repository is organized around multiple components, each representing a modelling step. Within each component, core algorithmic logic modules are located within the core/ directories, and wrapper scripts to run different pipeline stages are found below the scripts/ directories. Each component has an associated configuration file that centralizes parametrization.
Conda environment files needed to execute different parts of the pipeline are located within the environments directory at the repository root.
The most convenient way of running the pipeline and make modifications is to install the codebase as a package in editable mode by running: pip install -e . in repository root. The code can be used by running the dedicated scripts in each component or by executing pre-made recipes (in repository root) that orchestrate and chain a series of these scripts to execute multiple steps at once.
Most of the computing is too demanding for a single-machine and cannot be done in a personal laptop. We executed the pipeline in a local cluster of 5 workstations with 24 cores and 192 Gb of RAM each. The canopy height modelling component trains Neural Networks and requires access to GPUs.
ALS height data can be downloaded from: https://pnoa.ign.es/web/portal/pnoa-lidar/presentacion National can be downloaded forest inventory data from: https://www.miteco.gob.es/es/biodiversidad/temas/inventarios-nacionales/inventario-forestal-nacional/cuarto_inventario.html
For any questions please contact me at: diego.bengochea@mncn.csic.es