beacons
 
beacons
Home
beacons
About Us
beacons
Contact Us
beacons
beacons

University of Alberta

Laval University

OSU

Memorial University

 

 
Ranker

Ranker is an R-based tool for constructing and ranking candidate reserve networks. Ranker has the ability to:

  • assess representation of existing reserve networks
  • identify networks of benchmarks that best represent a planning region
  • identify networks of benchmarks and existing reserves that best represent a planning region
  • identify catchments that frequently appear within the best network assemblages

Representation is based on statistical dissimilarity metrics for distributions of a set of biophysical attributes.We presently use 4 attributes (described below), but the tool is flexible and any number of attributes can be used.

Construction of Candidate Reserve Networks

Ranker constructs candidate reserve networks based on an area target or a user-specified number of benchmark areas. Networks are constructed from the existing reserve network and/or benchmarks. There are six options available (Fig 1):

 
Fig 1. Ranker options

Options 1 to 4 are parameterized to construct candidate networks until a target area is reached.

Option 1: network is a random sample of benchmarks
Option 2: assessment of existing reserve network
Option 3: network is a random sample of benchmarks and reserves
Option 4: network is comprised of existing reserve network and random sample of benchmarks

Options 5 and 6 construct candidate networks with a user-specified number of benchmark areas. 

Option 5: network is a random sample of benchmarks
Option 6: network is the existing reserve network extended by a user-specificed number of benchmarks

Ranker uses a randomization process to construct candidate reserve networks to a maximum of N=500,000 candidate networks.

Assess Representation of Candidate Reserve Networks

Ranker is presently parameterized to assess representation based on 4 biophysical attributes; however, any number of attributes can be added to the Ranker. The 4 attributes are:

Climate Moisture Index (CMI)

CMI is P minus PET, where P is the yearly average precipitation (rain and snow melt), and PET is yearly potential evapotranspiration, both are expressed in centimetres per year.

Gross Primary Productivity (GPP),
mean 2000-2006

GPP is a measure of the carbon absorption rate of living plants or the amount of carbon absorbed during photosynthesis.

Landcover (NALC 2000, MODIS 2005)

Landcover is plants, waterbodies, wetlands, and other abiotic elements (e.g., rock) that cover the earth.

Lake-edge Density

Lake-edge density is a measure of the density of terrestrial/aquatic edge and represents the abundance of habitat along large waterbodies (lakes and wide rivers).

Representation is quantified by comparing distributions of attributes inside the candidate network against the rest of the study area using statistical dissimilarity metrics derived from two-sample univariate goodness-of-fit statistics (Figs 2 and 3); the closer the two distributions, the more representative the network.

Fig 2: Continuous variables
Fig 3: Binned data

The dissimilarity metrics of the 4 attributes are combined to create a representivity index which can be used to rank networks (Fig 4).

Fig 4: Representativity Index

 

 
 
WHAT'S NEW
 

Jul 24 2017
BEACONs' Website - New content coming soon!

...see all

 

 
beacons
beacons

home | about us | conservation planning | research | tools | datasets | documents| web links| contact us
english | français