Skip to contents

Ponderosa Pine Forest with Custom Configurations

EmpericalPatternR Package

2026-02-01

Source: vignettes/ponderosa-pine.Rmd
ponderosa-pine.Rmd

Overview

This vignette demonstrates how to create custom configurations for different forest types using create_config(). We’ll simulate a ponderosa pine forest with:

  • Custom target values (different from pinyon-juniper)
  • Multiple species (PIPO, PSME, ABCO)
  • Larger trees (montane forest)
  • Different allometric relationships
  • No nurse tree effect

Why Custom Configurations?

The pre-built pj_huffman_2009() configuration is great for pinyon-juniper woodlands, but other forest types have different characteristics:

Characteristic Pinyon-Juniper Ponderosa Pine
Density ~900 trees/ha ~450 trees/ha
Mean DBH 15 cm 35 cm
Mean Height 8 m 18 m
Species 2 (PIED, JUSO) 3+ (PIPO, PSME, ABCO)
Spacing Clumped (R~1.0) Regular (R~1.4)
Nurse Effect Yes No

Step 1: Load Package

Step 2: Get Ponderosa Allometric Parameters 

# Get species-specific allometric equations
ponderosa_params <- get_ponderosa_allometric_params()

# View crown radius parameters
ponderosa_params$crown_radius$PIPO
# $a: -0.204  (intercept)
# $b:  0.649  (DBH coefficient)  
# $c:  0.421  (height coefficient)

# View height parameters (Chapman-Richards)
ponderosa_params$height$PIPO
# $a: 27.0    (asymptotic height - 1.3)
# $b: 0.025   (growth rate)

# View foliage mass parameters (Miller)
ponderosa_params$foliage$PIPO
# $a: -2.287  (intercept)
# $b:  1.924  (DBH coefficient)

These parameters come from published equations: - Crown: Reese et al. (log-log with height) - Height: Chapman-Richards growth model - Foliage: Miller et al. (log-log)

Step 3: Create Custom Configuration 

config <- create_config(
  # Target stand structure
  density_ha = 450,
  species_props = c(PIPO = 0.70, PSME = 0.20, ABCO = 0.10),
  mean_dbh = 35.0,
  sd_dbh = 12.0,
  mean_height = 18.0,
  sd_height = 6.0,
  canopy_cover = 0.45,
  cfl = 0.85,
  clark_evans_r = 1.4,
  
  # Simulation parameters
  plot_size = 100,
  max_iterations = 5000,
  temp_initial = 0.01,
  cooling_rate = 0.9999,
  
  # Optimization weights
  weight_ce = 30,
  weight_dbh_mean = 15,
  weight_dbh_sd = 10,
  weight_height_mean = 10,
  weight_height_sd = 5,
  weight_species = 60,
  weight_canopy_cover = 60,
  weight_cfl = 50,
  weight_density = 80,
  weight_nurse = 0,  # No nurse effect
  
  # No nurse trees
  nurse_distance = 0,
  use_nurse_effect = FALSE,
  mortality_prop = 0.10
)

# View configuration
print(config)

Configuration Explanation

Target Structure: - density_ha = 450: Lower density than P-J (managed stand) - species_props: 70% ponderosa, 20% Douglas-fir, 10% white fir - mean_dbh = 35: Much larger trees (mature forest) - clark_evans_r = 1.4: Regular spacing (managed/natural fire regime)

Optimization Weights: - weight_density = 80: Critical to match tree count - weight_species = 60: Important for composition - weight_canopy_cover = 60: Important for structure - weight_nurse = 0: Disabled (no nurse effect in ponderosa)

Simulation Parameters: - max_iterations = 5000: Fewer needed (lower density) - cooling_rate = 0.9999: Slower cooling (more exploration)

Step 4: Run Simulation 

set.seed(456)

result <- simulate_stand(
  targets = config$targets,
  weights = config$weights,
  plot_size = config$simulation$plot_size,
  max_iterations = config$simulation$max_iterations,
  initial_temp = config$simulation$temp_initial,
  cooling_rate = config$simulation$cooling_rate,
  verbose = TRUE,
  plot_interval = 500,
  nurse_distance = config$simulation$nurse_distance,
  use_nurse_effect = config$simulation$use_nurse_effect,
  mortality_prop = config$simulation$mortality_prop
)

Step 5: Analyze Results 

analyze_simulation_results(
  result = result,
  targets = config$targets,
  prefix = "ponderosa_forest",
  save_plots = TRUE,
  nurse_distance_target = 0,
  target_mortality = config$simulation$mortality_prop * 100
)

Expected Output 

================================================================================
SIMULATION RESULTS SUMMARY
Stand: ponderosa_forest
================================================================================

STAND METRICS:
                          Achieved    Target      Diff     Status
  Density (trees/ha)      448         450         -2       ✓ GOOD
  Mean DBH (cm)           34.8        35.0        -0.2     ✓ GOOD
  Mean Height (m)         17.9        18.0        -0.1     ✓ GOOD
  Canopy Cover (%)        44.7        45.0        -0.3     ✓ GOOD
  CFL (kg/m²)             0.84        0.85        -0.01    ✓ GOOD
  Clark-Evans R           1.38        1.40        -0.02    ✓ GOOD

SPECIES COMPOSITION:
  PIPO (Ponderosa Pine)   69.8%       70.0%       -0.2%    ✓ GOOD
  PSME (Douglas-fir)      20.1%       20.0%       +0.1%    ✓ GOOD
  ABCO (White Fir)        10.1%       10.0%       +0.1%    ✓ GOOD

MORTALITY SIMULATION:
  Trees killed:           45 (10.0%)
  Surviving trees:        403 (90.0%)

Step 6: Compare Allometric Equations

One key difference between forest types is allometric relationships:

# Compare crown radius for 40cm DBH tree
test_dbh <- 40
test_species <- "PIPO"

# Default (P-J) parameters
default_params <- get_default_allometric_params()
default_height <- calc_height(test_dbh, test_species, default_params)
default_radius <- calc_crown_radius(test_dbh, default_height, test_species, default_params)

# Ponderosa parameters
ponderosa_height <- calc_height(test_dbh, test_species, ponderosa_params)
ponderosa_radius <- calc_crown_radius(test_dbh, ponderosa_height, test_species, ponderosa_params)

cat(sprintf("40cm DBH Ponderosa Pine:\n"))
cat(sprintf("  Default equations:  Radius = %.2fm, Height = %.1fm\n", 
            default_radius, default_height))
cat(sprintf("  Ponderosa equations: Radius = %.2fm, Height = %.1fm\n",
            ponderosa_radius, ponderosa_height))
cat(sprintf("  Difference:         +%.2fm (+%.1f%%), +%.1fm (+%.1f%%)\n",
            ponderosa_radius - default_radius,
            (ponderosa_radius/default_radius - 1) * 100,
            ponderosa_height - default_height,
            (ponderosa_height/default_height - 1) * 100))

Ponderosa pines typically have: - Larger crowns for given DBH (more open grown) - Taller heights (montane environment) - More foliage mass (higher photosynthetic capacity)

Custom Configuration Scenarios

Scenario 1: Dense Young Stand 

config_young <- create_config(
  density_ha = 800,
  species_props = c(PIPO = 1.0),
  mean_dbh = 15.0,
  sd_dbh = 5.0,
  mean_height = 8.0,
  sd_height = 3.0,
  canopy_cover = 0.50,
  cfl = 0.60,
  clark_evans_r = 1.0,  # Clumped
  plot_size = 100,
  max_iterations = 8000
)

Scenario 2: Old-Growth Stand 

config_oldgrowth <- create_config(
  density_ha = 200,
  species_props = c(PIPO = 0.60, PSME = 0.30, ABCO = 0.10),
  mean_dbh = 60.0,
  sd_dbh = 25.0,
  mean_height = 28.0,
  sd_height = 10.0,
  canopy_cover = 0.40,
  cfl = 1.20,
  clark_evans_r = 1.6,  # Very regular
  plot_size = 100,
  max_iterations = 3000
)

Scenario 3: Post-Fire Stand 

config_postfire <- create_config(
  density_ha = 150,
  species_props = c(PIPO = 0.80, PSME = 0.15, ABCO = 0.05),
  mean_dbh = 45.0,
  sd_dbh = 20.0,
  mean_height = 20.0,
  sd_height = 8.0,
  canopy_cover = 0.30,
  cfl = 0.50,
  clark_evans_r = 1.5,
  plot_size = 100,
  max_iterations = 3000,
  mortality_prop = 0.05  # Low mortality (survivors)
)

Troubleshooting Custom Configs

Issue: Energy Not Converging

Symptoms: Final energy > 5000, metrics don’t match targets

Solutions: 1. Increase max_iterations (try 10000, 20000) 2. Decrease cooling_rate (try 0.99995 for slower cooling) 3. Check for conflicting targets (e.g., high density + high canopy) 4. Adjust weights to prioritize most important metrics

Issue: Species Proportions Off

Symptoms: One species dominates or is missing

Solutions: 1. Increase weight_species (try 80-100) 2. Check that species_props sums to 1.0 3. Ensure all species have allometric parameters defined

Issue: Canopy Cover Mismatch

Symptoms: Achieved cover much higher/lower than target

Solutions: 1. Adjust both canopy_cover target AND weight_canopy_cover 2. Check allometric equations (larger crowns = more cover) 3. Consider interaction with density and mean DBH

Issue: Spatial Pattern Wrong

Symptoms: Trees clumped when should be regular (or vice versa)

Solutions: 1. Increase weight_ce to prioritize spatial pattern 2. Adjust clark_evans_r target (1.0 = random, >1.0 = regular, <1.0 = clumped) 3. Increase iterations for spatial optimization

Saving and Sharing Configurations 

# Save configuration for later use
save_config(config, "my_ponderosa_config.rds")

# Load saved configuration
config_loaded <- readRDS("my_ponderosa_config.rds")

# Share as R script (editable template)
generate_config_template(
  file = "ponderosa_template.R",
  config_name = "my_ponderosa_config",
  base_config = "custom"
)

Advanced: Integrating Custom Allometry

Note: Current version uses built-in allometric equations during simulation. To fully integrate custom equations, you would:

  1. Modify calc_tree_attributes() to accept allometric_params argument
  2. Pass custom parameters through simulate_stand()
  3. Apply custom equations to all attribute calculations

This is a planned enhancement. For now, custom allometry can be: - Applied post-hoc to results - Used in custom analysis scripts - Incorporated in derived metrics

Next Steps

  • Template System: Use generate_config_template() for easier customization
  • Multiple Runs: Compare different configurations/scenarios
  • Sensitivity Analysis: Test how results change with parameter variations
  • Field Validation: Compare simulated stands to field measurements

See Also