pecan vs seriation

PEcAn is a workflow framework for ecosystem modelling — it standardises meteorological and initial-condition inputs, drives a set of coupled process models such as SIPNET, ED2, BioCro and LDNDC, and runs uncertainty analysis and state data assimilation over the results. Version 1.10.0 adds an events schema with validation and a SIPNET writer, so disturbance and management events can be declared in JSON and turned into model input, and runs the entire SDA workflow in parallel. It also adds CH4 and N2O to the standard variable table and an AmeriFlux meteorological ensemble path with ERA5 fallback.

Two directions are visible. The framework keeps widening what it can represent — more standard variables, more carbon pools in BADM-based initial conditions, more coupled models — and it keeps industrialising the parts that used to be manual, from automated package documentation to qsub job-array expansion to parallel assimilation. The events schema is the first of these that gives users a declarative input format rather than a function to call, which is a different kind of extension than adding another model coupling.

The events schema currently has one model writer, for SIPNET, so the obvious next step is writers for the other coupled models — otherwise the format only benefits a single backend of the several PEcAn drives.