NAME

r.fidimo performs analysis of fish dispersal based on leptokurtic dispersal kernels. It calculates fish dispersal along river corridors based on a users stream network input, fish source populations and species-specific dispersal parameters.

OPTIONS

Stream parameters

river
Name of input stream network map (thinned raster map) on which the calculations of dispersal is performed. In cases of very maeandering rivers (with bends narrower than the spatial resolution of the analyis (cell size) it is recommended to run v.generalize and r.thin and/or to use r.fidimo.river in advance. The stream network should only consist of tree like structures, as braiding rivers and networks with two-side connected side arms will not work. Check the raster beforehand carefully!
outflow_point
Coordinates of outflow point of the basin outlet (see r.stream.basins) for the calculationof the flow direction within the provided river network. The coordinate pair must be comma separated e.g. X,Y and in the same metric projected coordinate system as the river network raster. An example outflow point looks like: 
510975,5331275
barriers
Name of input barrier text-file indicating the geographical position and passability of barriers e.g. weirs. The file should contain the X and Y coordinate and a value for permeability (0-1 where 0 is non-passable and 1 is 100% passable). The values must be |-separated: 
3543350.8001|6064831.9001|1
3535061.5179|6064457.5321|0.3

Source populations

The source populations can be provided either as random points, or as fixed source population raster:
n_source
For source populations set randomly; number or percentage of cells containing source populations. The model selects randomly cells within the river network (number specified by the user) and assigns a starting density of 1 to each occupied cell.
source_populations
Input raster map indicating the starting density per cell. All cells with densities > 0 will act as source populations for the model. The raster map must have the resolution as the river raster and all source population cells must also be part of the river raster. Source points at the end points of the network (most upstream cell etc.) might cause problems with the calculation.

Dispersal parameters

species
Selected species with predefined L and AR (see R-package 'fishmove').
L
Length of fish which should be modelled. Increasing L is positively correlated with larger dispersal distances. Setting L will overwrite any species-settings (see R-package 'fishmove').
AR
Aspect ratio of the caudal fin of a fish which should be modelled. Increasing AR is positively correlated with larger dispersal distances. etting AR will overwrite any species-settings (see R-package 'fishmove').
T
Time interval for one modelling step. The dispersal kernel is time dependent and increasing T is positively correlated with larger dispersal istances (see R-package 'fishmove').
p
Share of the stationary component of the population. The value is set to 0.67 by default (Radinger and Wolter, 2014, Fish and Fisheries).

Output

output
The base name of the output file(s). The output raster files will be %output%_fit respectively %output%_lwr and %output%_upr if a statistical interval is set.
statistical_interval
Statistical interval (confidence or prediction) derived from the regression analyis (see R-package 'fishmove'). If a statistical interval is set, three output maps will be created (fit, lwr and upr).

Examples

A GRASS GIS sample location with an extracted river sample river network can be found here: https://github.com/jradinger/fidimo_example_mapset. Inside the GRASS-FIDIMO sample mapset it is possible to run following examples:

Example 1

r.fidimo.py river=example_stream@PERMANENT coors="510975,5331275" n_source="3%" species="Custom species" L=250 AR=1.2 output=fidimo_out_test statistical_interval="Confidence Interval"
This example calculates fish dispersal for a 250 mm sized (L) species with and aspect ratio (AR) of 1.2. As starting points (source populations) for dispersal 3% of the river network cells are selected randomly. In addition also the confidence interval (CI based of the multiple regression model, see Radinger and Wolter 2014) is calculated and three output maps are produced with the suffix: _fit, _upr, and _lwr. Note: the calculation of the CI tripples the calculation time.

Dependencies

SEE ALSO

r.stream.order, r.stream.basins, r.watershed

REFERENCES

Radinger, J. and Wolter, C. (2014) Patterns and predictors of fish dispersal in rivers. Fish and Fisheries 1-18. DOI: 10.1111/faf.12028

AUTHOR

Johannes Radinger, Leibniz-Institute of Freshwater Ecology and Inland Fisheries, Berlin (Germany)

Last changed: $Date$