streamnet

Morphology analysis of stream networks

Installation

You can install streamnet from github with:

# install.packages("devtools")
devtools::install_github("jsta/streamnet")

In addition, many functions require a system installation of GRASS 7 along with the v.stream.order extension.

There is a helper function to install v.stream.order at streamnet:::install_grass_extensions.

Usage

Calculate morphology metrics

library(sf)
library(nhdR)
library(streamnet)
library(ggplot2)
library(raster)

data(nhd_sub_lines)
data(nhd_sub_lakes)

outlet_reach   <- terminal_reaches(network = nhd_sub_lines, 
                                   approve_all_dl = TRUE, quiet = TRUE)
outlet_point   <- st_cast(st_line_sample(outlet_reach, sample = 1), "POINT")

ggplot() +
  geom_sf(data = nhd_sub_lines) +
  geom_sf(data = outlet_point, aes(color = "red")) +
  scale_color_manual(labels = "outlet", values = "red") +
  labs(colour = "") + theme_minimal()

calc_metrics(nhd_sub_lines, nhd_sub_lakes)

Simplify stream networks

data(nhd_sub_lines)

# Combine(dissolve) adjacent reaches with no junctions
nhd_sub_simple <- simplify_network(nhd_sub_lines)
avg_link_length(nhd_sub_simple)
#> 2444.693 [m]
avg_link_length(nhd_sub_lines)
#> 1312.988 [m]

Round-trip igraph and sf lines

tree <- create_reversed_tree(15)
class(tree)
#> [1] "igraph"
plot(tree)

tree_sf <- igraph2sf(tree)
plot(tree_sf)

Create synthetic stream networks

# Diffusion limited aggregation
dt <- sim_dla()
viz_dla(dt, which.max(dt))

# Generate from a binary raster
foo <- matrix(0, ncol = 9, nrow = 9)
foo[1:4,3] <- 1
foo[5,4]   <- 1
foo[6:9,5] <- 1
foo <- raster(foo, xmn = 1, xmx = 9, ymn = 1, ymx = 9)
origin <- which.min(apply(
              which(as.matrix(flip(foo, "y")) == 1, arr.ind = TRUE), 1, sum))
res <- raster2network(foo, origin)

par(mfrow = c(1, 2))
plot(foo)
plot(foo); plot(res, add = TRUE)