Skip to main content

​
Overview

The .demo_landscape() function is an internal utility that creates synthetic raster time series with random pixel values. This is useful for testing, demonstrations, and learning how to use OpenLand without real data.
This is primarily an internal function for package development and testing. For production analysis, use real raster data with contingencyTable().

​
Function Signature

.demo_landscape(
  year,
  nrows = 100,
  ncols = 100,
  res = 1,
  xmn = 0,
  xmx = 100,
  ymn = 0,
  ymx = 100,
  crs = NA,
  category = 1:5,
  prob = NULL
)

​
Parameters

​
year
numeric vector
required
A vector of years for the time series (e.g., 2000:2005). First and last years are included.
​
nrows
numeric
default:"100"
Number of rows in the generated raster.
​
ncols
numeric
default:"100"
Number of columns in the generated raster.
​
res
numeric
default:"1"
Spatial resolution of the raster cells.
​
xmn
numeric
default:"0"
Minimum x coordinate (left edge).
​
xmx
numeric
default:"100"
Maximum x coordinate (right edge).
​
ymn
numeric
default:"0"
Minimum y coordinate (bottom edge).
​
ymx
numeric
default:"100"
Maximum y coordinate (top edge).
​
crs
character
default:"NA"
Coordinate reference system. Use proj4 string format or NA for no projection.
​
category
numeric vector
default:"1:5"
Vector of land use category values to randomly assign to pixels.
​
prob
numeric vector
default:"NULL"
Probability weights for each category. If NULL, all categories have equal probability.

​
Return Value

Returns a named list of RasterLayer objects, one for each year. Each raster:
  • Has dimensions nrows × ncols
  • Contains pixel values randomly sampled from category
  • Is named following the convention landscape_{year}
  • Can be used directly with contingencyTable()

​
Usage Examples

​
Basic Demo Landscape

Create a simple 5-year landscape with default settings: 
library(OpenLand)

# Create demo landscape for 2000-2004
demo <- .demo_landscape(year = 2000:2004)

# View the structure
names(demo)
# [1] "landscape_2000" "landscape_2001" "landscape_2002" 
# [4] "landscape_2003" "landscape_2004"

# Check the first raster
demo[[1]]

​
Custom Extent and Projection

Create a landscape with UTM projection: 
# Create demo with UTM Zone 21 South
demo_utm <- .demo_landscape(
  year = 2010:2015,
  nrows = 200,
  ncols = 200,
  res = 30,  # 30-meter resolution
  xmn = 500000,
  xmx = 506000,
  ymn = 7500000,
  ymx = 7506000,
  crs = "+proj=utm +zone=21 +south +ellps=GRS80 +units=m +no_defs"
)

​
Custom Categories and Probabilities

Create a landscape dominated by forest with some agriculture: 
# Define categories
# 1 = Forest, 2 = Agriculture, 3 = Urban, 4 = Water
demo_custom <- .demo_landscape(
  year = 2015:2020,
  nrows = 150,
  ncols = 150,
  category = 1:4,
  prob = c(0.6, 0.25, 0.10, 0.05)  # 60% forest, 25% agriculture, etc.
)

​
Using with OpenLand Workflow

Create demo data and run a complete analysis: 
library(OpenLand)

# 1. Create demo landscape
demo_rasters <- .demo_landscape(
  year = 2000:2005,
  nrows = 100,
  ncols = 100,
  category = 1:3,
  prob = c(0.5, 0.3, 0.2)
)

# 2. Convert to RasterStack
library(raster)
demo_stack <- stack(demo_rasters)

# 3. Create contingency tables
ct <- contingencyTable(
  input_raster = demo_stack,
  pixelresolution = 1
)

# 4. Customize legend
ct$tb_legend$categoryName <- factor(
  c("Forest", "Agriculture", "Urban"),
  levels = c("Forest", "Agriculture", "Urban")
)

ct$tb_legend$color <- c("#228B22", "#FFD700", "#FF0000")

# 5. Run intensity analysis
results <- intensityAnalysis(
  dataset = ct,
  category_n = "Urban",
  category_m = "Forest"
)

# 6. Plot results
plot(results$interval_lvl)

​
Use Cases

​
Testing and Development

# Quick test of contingency table creation
test_data <- .demo_landscape(year = 2020:2022, nrows = 50, ncols = 50)
test_stack <- stack(test_data)
ct_test <- contingencyTable(test_data, pixelresolution = 1)

​
Learning and Tutorials

# Create simple example for teaching
simple_demo <- .demo_landscape(
  year = 2010:2012,
  nrows = 20,
  ncols = 20,
  category = 1:2,
  prob = c(0.7, 0.3)
)

# Easy to visualize and understand
plot(stack(simple_demo))

​
Benchmark Testing

# Create large landscape for performance testing
large_demo <- .demo_landscape(
  year = 2000:2020,
  nrows = 1000,
  ncols = 1000,
  category = 1:10
)

# Time the contingency table creation
system.time({
  ct_large <- contingencyTable(large_demo, pixelresolution = 30)
})

​
Important Notes

The generated landscapes have random pixel values. They do not represent realistic spatial patterns or actual land use change processes. Use real data for production analysis.
Since this is an internal function (prefixed with .), its interface may change in future package versions. For stable workflows, use contingencyTable() with real raster data.

​
See Also

Build docs developers (and LLMs) love

Get started for freeTalk to us