Rapid Automated Risk Assessment

Spatial distribution and analysis of aggregated taxonomic and occurrence data.

About this tool

This is a set of open-source tools to dailycollect and serve consolidated taxonomic and occurrence data from various open data sources , perform geospatial analysis that enable rapid risk assessmentof country-level biodiversity , and allow simple and directed visualizationof the data and results.

What made this possible

All this work is possible due to a series of developments in contribution, publishing and technologies related to biodiversity.

Namely, not limited to:

  • DarwinCore standart by TDWG, that set common terms to work upon
  • DarwinCore-Archive design by GBIF, that allowed the exchange in efficient format of information
  • Integrated Publishing Toolkit from GBIF, that made simple and easy access to datasets
  • GBIF push and committiment to help publishers, which allows easy dicovery of source of data
  • IUCN Red Listing Guidelines, that provide a base set of rules for extinction risk assessment

How does this work

Specifically, the daily workflow goes as follows, mostly in parallel:

  • Collect country taxonomic information from choosen checklists published in IPT
  • [TODO]Download and load into local IPT country occurrences from GBIF by families
  • Collect from local and choosen IPTs occurrence datasets
  • For each taxon:
    • Consolidate occurrences from accepted name and synonyms
    • Classify occurrences in groups
      • All
      • Historic
      • Recent
    • [TODO]Rate the quality of occurrences
    • For each group of occurrences perform the following analysis
      • Extent of occurrence (EOO)
      • Area of occupancy with regular grid of 2km side (AOO-2km)
      • Area of occupancy withh grid of variadic size (AOO-variadic)
      • Subpopulations/Locations/Cluster of occurrences based on circular buffer of median distance
      • Rapid Risk Assessment based on EOO, AOO and decline only
  • Display the results and statistics.

Extent of occurrences (EOO)

Utilizes a convex-hull method to calculate the total extent of occurrence for the specie.

Area of Occupancy (AOO)

Using a world grid of 2km of side and of 10% of the maximum distance between occurrences, and matching those which have occurrences on it, distincts them to calculate the area of occupancy of the specie.

Population clusters

Based on 10% of the maximum distance between the species occurrences (to be changed to a minimum-spaning-tree following Rapoport's approach), draws a buffer of this radius around them and groups those that intersects.

Risk Assessment Analysis

Using IUCN category and criteria extinction risk assessment, but based only on geographical distribution and futher simplified.

MetricValueCategory (criteria)
Number of records < 3DD
Area of occupancy < 10km²CR (B2)
Area of occupancy < 500km²EN (B2)
Area of occupancy < 2000km²VU (B2)
Extent of occurrence < 100km²CR (B1)
Extent of occurrence < 5000km²EN (B1)
Extent of occurrence < 20000km²VU (B1)

Decline and number of subpopulation might be taken into account as official guidelines suggest, if available.

Due to lack of a method for some extra information (specially locations and threats), the categories had to be simplified. The methodology can be improved if we have easy access to such data.

Biodiversity Indexer Workflow

Other informations and links

You can read the paper published at CSBC 2016 (WCAMA), entitled "Assessing the risk of extinction of Brazil’s flora: A computational approach based on micro-services and geospatial analysis", that describes the tools and methodolgies.

Follow the progress and participate on the public task manager.

You can also check the source code of the open source tools at github.

Any feedback you can get in contact by email: diogo@diogok.net.

All source code is available at gihub: