Vignette: VCA parameter reference with citations

[NewGraphEnvironment]

Problem

fl_scenarios() and fl_params() ship pre-defined flood factor scenarios and parameter metadata (#28), but there's no worked example showing how to use them with fl_valley_confine().

Proposed Solution

Add a section to the valley-confinement vignette (or a new vignette) demonstrating:

1. Load scenarios with fl_scenarios()
2. Run fl_valley_confine() at ff=2, ff=4, ff=6 on the bundled test data
3. Compare outputs side-by-side — active channel vs functional floodplain vs valley bottom
4. Show fl_params() for parameter documentation lookup
5. Demonstrate custom scenario CSV for site-level work (e.g., 1m lidar with smaller thresholds)

This proves the functions work end-to-end and gives users a template for multi-scenario analysis.

Relates to #28

[NewGraphEnvironment]

Updated scope (2026-03-19)

The multi-scenario comparison is landing in the main valley-confinement vignette as part of #28. What remains is a dedicated parameters vignette explaining the model components and their literature sources.

Already shipping with #28

• Scenario comparison section in valley-confinement vignette (ff=2, ff=4, ff=6)
• fl_params() and fl_scenarios() tables rendered in vignette
• Hardcoded summary table replaced with fl_params() output

What this issue now covers

New vignette (vca-parameters.Rmd) as the reference for "what are VCA parameters and where do they come from":

1. Bankfull regression — the equation, R², limitations, PNW→BC applicability
2. Two independent width models — bcfishpass (Thorley et al. 2021) vs VCA/Hall 2007, where each is used (Document channel width sources and trust model #25)
3. flood_factor interpretation — valley bottom vs functional floodplain, DEM resolution effects
4. Parameter-by-parameter documentation — rendered from fl_params() CSV
5. DEM resolution guidance — how to adjust params for 1m lidar vs 25m TRIM
6. Proper citations — references.bib in vignettes/ with Nagel 2014, Hall 2007, Thorley et al. 2021

Content graduates from inst/research/vca_parameter_rationale.md (internal notes → published vignette).

Relates to #25
Relates to #28

[NewGraphEnvironment]

Updated scope (2026-03-19)

The multi-scenario comparison landed in the main valley-confinement vignette as part of #28. What remains is a dedicated parameters vignette explaining the model components and their literature sources.

Already shipping with #28

• Scenario comparison section in valley-confinement vignette (ff=2, ff=4, ff=6)
• fl_params() and fl_scenarios() tables rendered in vignette
• Hardcoded summary table replaced with fl_params() output
• effect column added to flood_params.csv

What this issue now covers

New vignette (vca-parameters.Rmd) as the reference for "what are VCA parameters and where do they come from":

1. Bankfull regression — the equation, R², limitations, PNW→BC applicability
2. Two independent width models — bcfishpass (Thorley et al. 2021) vs VCA/Hall 2007, where each is used (Document channel width sources and trust model #25)
3. flood_factor interpretation — valley bottom vs functional floodplain, DEM resolution effects
4. Parameter-by-parameter documentation — rendered from fl_params() CSV
5. DEM resolution guidance — how to adjust params for 1m lidar vs 25m TRIM
6. Proper citations — references.bib in vignettes/ with Nagel 2014, Hall 2007, Thorley et al. 2021

Content graduates from inst/research/vca_parameter_rationale.md (internal notes → published vignette).

Relates to #25
Relates to #28