AgriLand {landsepi}

Landscape allocation

Description

Generates a landscape composed of fields where croptypes are allocated with controlled proportions and spatio-temporal aggregation.

Usage

AgriLand(
  landscape,
  Nyears,
  rotation_period = 0,
  rotation_sequence = list(c(0, 1, 2)),
  rotation_realloc = FALSE,
  prop = list(c(1/3, 1/3, 1/3)),
  aggreg = list(1),
  algo = "periodic",
  croptype_names = c(),
  graphic = FALSE,
  outputDir = "./"
)

Arguments

landscape

a spatialpolygon object containing field coordinates.

Nyears

an integer giving the number of simulated cropping seasons.

rotation_period

number of years before rotation of the landscape. There is no rotation if rotation_period=0 or rotation_period=Nyears.

rotation_sequence

a list, each element of the list contains indices of croptypes that are cultivated during a period given by "rotation_period". There is no change in cultivated croptypes if the list contains only one element (e.g. only one vector c(0,1,2), indicating cultivation of croptypes 0, 1 and 2).

rotation_realloc

a logical indicating if a new random allocation of croptypes is performed when the landscape is rotated (FALSE=static allocation, TRUE=dynamic allocation). Note that if rotation_realloc=FALSE, all elements of the list "rotation_sequence" must have the same length, and only the first element of the lists "prop" and "aggreg" will be used.

prop

a list of the same size as "rotation_sequence", each element of the list contains a vector of the proportions (in surface) associated with the croptypes in "rotation_sequence". A single vector can be given instead of a list if all elements of "rotation_sequence" are associated with the same proportions.

aggreg

a list of the same size as "rotation_sequence", each element of the list is a single double indicating the degree of aggregation of the landscape. This double must greater or equal 0; the greater its value, the higher the degree of spatial aggregation (roughly, aggreg between 0 and 0.1 for fragmented landscapes, between 0.1 and 0.5 for balanced landscapes, between 0.5 and 3 for aggregated landscapes, and above 3 for highly aggregated landscapes). A single double can be given instead of a list if all elements of "rotation_sequence" are associated with the same level of aggregation.

algo

the algorithm used for the computation of the variance-covariance matrix of the multivariate normal distribution: "exp" for exponential function, "periodic" for periodic function, "random" for random draw (see details of function multiN). If algo="random", the parameter aggreg is not used. Algorithm "exp" is preferable for big landscapes.

croptype_names

a vector of croptype names (for legend in graphic).

graphic

a logical indicating if a graphic of the landscape must be generated (TRUE) or not (FALSE).

outputDir

a directory to save graphic

Details

An algorithm based on latent Gaussian fields is used to allocate two different croptypes across the simulated landscapes (e.g. a susceptible and a resistant cultivar, denoted as SC and RC, respectively). This algorithm allows the control of the proportions of each croptype in terms of surface coverage, and their level of spatial aggregation. A random vector of values is drawn from a multivariate normal distribution with expectation 0 and a variance-covariance matrix which depends on the pairwise distances between the centroids of the fields. Next, the croptypes are allocated to different fields depending on whether each value drawn from the multivariate normal distribution is above or below a threshold. The proportion of each cultivar in the landscape is controlled by the value of this threshold. To allocate more than two croptypes, AgriLand uses sequentially this algorithm. For instance, the allocation of three croptypes (e.g. SC, RC1 and RC2) is performed as follows:

1. the allocation algorithm is run once to segregate the fields where the susceptible cultivar is grown, and

2. the two resistant cultivars (RC1 and RC2) are assigned to the remaining candidate fields by re-running the allocation algorithm.

Value

a gpkg (shapefile) containing the landscape structure (i.e. coordinates of field boundaries), the area and composition (i.e. croptypes) in time (i.e. each year) for each field. A png graphic can be generated if graphic=TRUE.

References

Rimbaud L., Papaïx J., Rey J.-F., Barrett L. G. and Thrall P. H. (2018). Assessing the durability and efficiency of landscape-based strategies to deploy plant resistance to pathogens. PLoS Computational Biology 14(4):e1006067.

See Also

multiN, periodic_cov, allocateLandscapeCroptypes

Examples

## Not run: 
data(landscapeTEST)
landscape <- get("landscapeTEST1")
set.seed(12345)
## Generate a mosaic of three croptypes in balanced proportions
## and high level of spatial aggregation
AgriLand(landscape,
  Nyears = 10,
  rotation_sequence = c(0, 1, 2), prop = rep(1 / 3, 3),
  aggreg = rep(10, 3), algo = "periodic",
  graphic = TRUE, outputDir = getwd()
)

## Generate a dynamic mosaic of two croptypes in unbalanced proportions
## and low level of spatial aggregation,
## the second croptype being replaced every 5 years without changing field allocation
AgriLand(landscape,
  Nyears = 20, rotation_period = 5, rotation_sequence = list(c(0, 1), c(0, 2)),
  prop = c(1 / 3, 2 / 3), aggreg = c(0.07, 0.07), algo = "periodic", graphic = TRUE,
  outputDir = getwd()
)

## Generate a dynamic mosaic of four croptypes in balanced proportions
## and medium level of spatial aggregation,
## with field allocation changing every year
AgriLand(landscape,
  Nyears = 5, rotation_period = 1, rotation_realloc = TRUE,
  rotation_sequence = c(0, 1, 2, 3),
  prop = rep(1 / 4, 4), aggreg = 0.25, algo = "exp", graphic = TRUE, outputDir = getwd()
)

## End(Not run)