Skip to contents

Derive one species' fitting window from where its references have support

growth_auto_window.Rd

Nobody should have to nominate an age range by hand. Both bounds are dictated by the data and by a species attribute that is not being fitted, so both can be read off directly.

Usage

growth_auto_window(
  reference,
  longevity,
  age_floor = 20,
  age_quantile = 0.95,
  senescence_frac = 0.45,
  mort_shp = NULL,
  sources = c("SORTIE", "TIPSY", "VDYP")
)

Arguments

reference

A data frame of reference observations, with columns source ("SORTIE", "TIPSY", "VDYP" or "Ground plots"), age and aboveground_c_mg_ha.

longevity

Numeric. The species' longevity.

age_floor

Numeric. Youngest age to score.

age_quantile

Numeric. Quantile of observed plot ages to close at.

senescence_frac

Numeric. Fraction of longevity at which to close, used only when mort_shp is NULL.

mort_shp

Numeric MortalityCurve for this species, or NULL. When supplied it sets the cap and senescence_frac is ignored.

sources

Character. Modelled reference series to consider.

Value

Numeric length-2.

Details

The window OPENS at age_floor. Stands younger than that are essentially unmeasured – the ground-plot programs do not sample them – and it is also the range where LANDIS-II is known to overestimate biomass for reasons that have nothing to do with these four parameters, so scoring there would import a bias the sweep cannot fix.

The window CLOSES at the earliest of three limits: the age_quantile of the observed plot ages, beyond which the cloud thins to a handful of stands; the end of the modelled reference curve; and senescence_frac x longevity. LANDIS-II ramps mortality up as a cohort approaches longevity and the curve then falls to exactly zero and stays there, so an open-ended window scores the modelled die-off rather than the level the stand holds.

WHERE THAT CAP BELONGS IS A PROPERTY OF MortalityCurve. The extension defines it as a position in the lifespan (2.12.4: 5 puts onset at 10% of life span, 25 at 85%), so the age at which a species leaves its plateau varies by nearly twofold across the documented range. Supply mort_shp and the cap is that species' own onset, via growth_mortality_onset_frac(). Measured on one calibration, the departure from 95% of peak biomass ran 0.43-0.48 x longevity at MortalityCurve 10, 0.63-0.70 at 15 and 0.82-0.84 at 25 – so a single fraction cannot separate a species that breaks up early from one that holds its stand almost to the end, which is the distinction the parameter exists to make.

The cap is at the ONSET of mortality, not at peak biomass, and is therefore conservative: biomass keeps rising for a period after onset while growth still exceeds mortality. That is deliberate. Peak location depends on GrowthCurve as well, and GrowthCurve may still be swept – a peak-based cap would then score different candidates over different ranges and could not rank them fairly. Onset depends on MortalityCurve alone, so as long as that is assigned rather than swept, every candidate for a species sees one window.

senescence_frac remains as the fallback when mort_shp is not supplied. Its default of 0.45 was calibrated against a parameterisation that gave every species a MortalityCurve near 23; it does not generalise, and on a set carrying 10s the earliest 95%-of-peak departure falls to 0.433, below the cap itself.

See also

Other growth calibration helpers: extract_landscape_cohort_structures(), growth_add_objective(), growth_best_candidates(), growth_bin_observations(), growth_calibration_design(), growth_calibration_partition(), growth_calibration_work_root(), growth_climatic_distance(), growth_climatic_weight(), growth_expand_over_pseudo_species(), growth_factorial_ratio_grid(), growth_fitting_windows(), growth_identifiability(), growth_inflation_factor(), growth_mortality_onset_frac(), growth_plot_palette(), growth_pseudo_species_name(), growth_reference_curves(), growth_reference_inflection(), growth_score_fit(), growth_scoring_for(), growth_smooth_observations(), growth_structure_cell_curves(), growth_structure_cohort_table(), growth_structure_design(), growth_structure_summary(), growth_window_for(), plot_growth_calibration(), plot_growth_candidate(), plot_growth_factorial_sensitivity(), plot_growth_structures(), read_growth_scoring(), read_landscape_cohort_structures(), scale_linetype_growth_reference(), write_growth_review_bundle()