Parameter Sets
The original ECOSYS model ships three regional parameter workbooks. All three
are packaged, each exported to its own SQLite database in ecosys/data/:
| Name | Workbook | Scenario (Szenario!A4, verbatim) |
Original-model comparison |
|---|---|---|---|
valley (default) |
ch1_orig.xls |
Parameter Satz 1 = Talgebiete, normale Vegetation | Three reference runs |
mountain |
ch2_orig.xls |
Parameter Satz 2 = Berggebiete, späte Vegetation | No reference runs |
ticino_valais |
ch3_orig.xls |
Parameter Satz 3 = TI / VS, frühe Vegetation | No reference runs |
In English: valley areas with normal vegetation; mountain areas with late
vegetation; Ticino and Valais (canton codes TI / VS) with early vegetation.
The registry's validated flag records comparison with runs of the original
Excel model; those runs all use set 1. It does not denote validation against
environmental measurements. See validation evidence.
What differs between sets
Elements, nuclides, dose coefficients, transfer factors, soil, skin and shielding data are identical in all three sets. Only the regional scenario data differs:
- plant calendars and phenology (
plant,plant_lai,plant_biomass); - grass growth half-lives (
plant_growth_half_life) and the hay first-interval end (scenario, set 2 only); - seasonal feed rations (
animal_feed_ration); - inhalation and consumption rates (
human_inhalation_rate,consumption_rate).
A test asserts that no other table differs between the packaged databases.
Selecting a set
from ecosys import EcosysEngine
from ecosys.data import PARAMETER_SETS, load_model_data
for parameter_set in PARAMETER_SETS.values():
print(parameter_set.name, parameter_set.title, parameter_set.validated)
engine = EcosysEngine(parameter_set="mountain")
model = load_model_data(parameter_set="mountain")
The registry is ecosys.data.parameter_sets. Each entry records the source
workbook and its SHA-256. An exported database stores its set name in
dataset.parameter_set only when the workbook bytes match a registered
digest, so a modified workbook is never labelled as one of the packaged sets.
Rebuilding the databases
python -m export_excel_data --all --output-dir exported_parameters
python -m export_excel_data --parameter-set mountain --output mountain.sqlite
python -m export_excel_data --source my.xls --output my.sqlite
The first two commands read the bundled workbooks. The third requires an
existing workbook at my.xls. Omitting the output path for a registered-set
export replaces its packaged database; --all without --output-dir rebuilds
all packaged databases. Each replacement occurs only after successful export
and validation, with publication per set.
The importer reports defaults, inferences, skipped rows and omissions with
source coordinates. Each bundled workbook has one diagnostic: turf (Rasen)
is missing from resuspensionmassload, so its resuspension factor and soil
intake default to zero. With --source, --strict rejects these fallbacks;
the unmodified bundled workbooks therefore do not pass strict import.
This import policy is distinct from strict validation of an exported database:
python -m export_excel_data --validate mountain.sqlite