Growth and Mortality Analysis • CafriplotsR

library(CafriplotsR)

Introduction

This vignette demonstrates how to compute tree growth rates and mortality/recruitment statistics for permanent forest plots with multiple censuses.

Two main functions are available:

compute_growth(): Calculate diameter growth rates between censuses
compute_mortality(): Calculate mortality and recruitment rates

Both functions automatically fetch data from the database and only require plot identifiers as input.

Computing Growth Rates

Basic Usage

# Connect to database
con <- call.mydb()

# Compute growth by plot name
growth <- compute_growth(plot_names = "mbalmayo001", con = con)

# Or by plot ID
growth <- compute_growth(plot_ids = c(1, 2, 3), con = con)

Understanding the Output

The function returns a list with two components:

# Summary statistics per plot and census interval
growth$summary

# Individual-level growth data
growth$individuals

Summary columns:

plot_name: Plot identifier
census_pair: Census interval (e.g., “1-2” for first to second census)
n_individuals: Total individuals measured in both censuses
n_valid: Individuals with valid growth measurements
n_excluded: Individuals excluded due to measurement errors
mean_growth_mm_yr: Mean diameter growth rate (mm/year)
sd_growth_mm_yr: Standard deviation of growth
median_growth_mm_yr: Median growth rate
mean_dbh_1_mm: Mean DBH at first census (mm)
mean_dbh_2_mm: Mean DBH at second census (mm)
mean_interval_years: Mean time between censuses (years)

Growth Calculation Methods

Two methods are available:

# Incremental growth (default): (DBH2 - DBH1) / time
growth_inc <- compute_growth(plot_ids = 1, method = "I")

# Exponential growth: (log(DBH2) - log(DBH1)) / time
growth_exp <- compute_growth(plot_ids = 1, method = "E")

Adjusting Quality Control Parameters

growth <- compute_growth(
  plot_ids = 1,
  mindbh = 100,      # Minimum DBH in mm (default: 100)
  err.limit = 4,     # Error limit for negative growth detection (default: 4)
  maxgrow = 75       # Maximum valid growth in mm/year (default: 75)
)

Quality control:

Trees with DBH < mindbh are excluded
Extreme negative growth (measurement errors) is detected using a statistical model
Growth rates > maxgrow mm/year are flagged as errors

Summary Only

To get only summary statistics without individual data:

growth <- compute_growth(
  plot_ids = 1,
  return_individual = FALSE
)

Computing Mortality and Recruitment

Basic Usage

# Compute mortality by plot name
mort <- compute_mortality(plot_names = "mbalmayo001", con = con)

# Or by plot ID
mort <- compute_mortality(plot_ids = c(1, 2, 3), con = con)

Understanding the Output

The function returns a list with three components:

# Summary statistics
mort$summary

# List of individuals that died
mort$dead_individuals

# List of newly recruited individuals
mort$recruits

Summary columns:

plot_name: Plot identifier
census_pair: Census interval
N_outset: Number of individuals at start of interval
N_dead: Number of deaths
N_survivor: Number of survivors
N_recruits: Number of new recruits
mortality_rate: Exponential mortality rate
mortality_percent_yr: Annual mortality percentage
recruitment_rate: Exponential recruitment rate
recruitment_percent_yr: Annual recruitment percentage
mean_interval_years: Time between censuses
mean_dbh_dead_cm: Mean DBH of dead trees (cm)
mean_dbh_recruit_cm: Mean DBH of recruits (cm)

Mortality Rate Calculation

Mortality rate is calculated using the exponential model:

$\lambda = \frac{\ln(N_0) - \ln(N_s)}{t}$

Where: - $N_0$ = number of individuals at start - $N_s$ = number of survivors - $t$ = time interval in years

Annual mortality percentage is: $(1 - e^{-\lambda}) \times 100$

Examining Dead Trees

# View dead individuals with their characteristics
mort$dead_individuals

# Filter by taxonomy
library(dplyr)
mort$dead_individuals %>%
  filter(tax_fam == "Fabaceae")

Examining Recruits

# View recruited individuals
mort$recruits

# Size distribution of recruits
hist(mort$recruits$dbh_at_recruitment_cm)

Working with Multiple Plots

Both functions can process multiple plots at once:

# Growth for multiple plots
growth <- compute_growth(
  plot_names = c("plot001", "plot002", "plot003")
)

# Results are combined with plot_name column for identification
growth$summary %>%
  group_by(plot_name) %>%
  summarise(
    mean_growth = mean(mean_growth_mm_yr, na.rm = TRUE),
    n_intervals = n()
  )

Complete Workflow Example

library(CafriplotsR)
library(dplyr)

# Connect
con <- call.mydb()

# Define plots to analyze
my_plots <- c("site_A_plot1", "site_A_plot2", "site_B_plot1")

# Compute growth
growth <- compute_growth(plot_names = my_plots, con = con)

# View growth summary
growth$summary %>%
  arrange(plot_name, census_pair)

# Compute mortality
mort <- compute_mortality(plot_names = my_plots, con = con)

# Compare mortality across plots
mort$summary %>%
  select(plot_name, census_pair, mortality_percent_yr, recruitment_percent_yr) %>%
  arrange(plot_name)

# Identify species with high mortality
mort$dead_individuals %>%
  count(tax_sp_level, sort = TRUE)

# Clean up
cleanup_connections()

Troubleshooting

“Need at least 2 censuses”

This error occurs when the selected plots don’t have multiple census records:

# Check if your plot has multiple censuses
query_result <- query_plots(
  plot_names = "your_plot",
  show_multiple_census = TRUE,
  extract_individuals = FALSE
)

# Look at census_features to see available censuses
query_result$census_features

No valid measurements

If growth calculations return few valid measurements, check:

Minimum DBH threshold: Lower mindbh if needed
Date information: Ensure census dates are recorded in the database
Measurement quality: Review excluded individuals in the output

# Check individual results for exclusions
growth <- compute_growth(plot_ids = 1)
table(growth$individuals$accepted_growth)

References

Growth trimming methodology adapted from:

CTFS R Package: http://ctfs.si.edu/Public/CTFSRPackage/

Mortality rate calculations follow standard forest dynamics methodology as described in:

Sheil, D., & May, R. M. (1996). Mortality and recruitment rate evaluations in heterogeneous tropical forests. Journal of Ecology, 84(1), 91-100.