1 Topic by Darwyn Sumner

  • Darwyn Sumner
  • Darwyn Sumner
  • Veteran
  • Offline
  • From: Dipterists Forum
  • Registered: 10-05-2006
  • Posts: 250

Topic: Produce an IUCN category from occurrence data

Empirical methods of producing these categories (e.g. A Provisional Assessment of the Status of Acalyptratae flies in the UK are sure to give way to more observational methods over time as occurrences build up in the GBGs and/or our Recorder 6 databases. The reports and reassessments are repeated at 10 year intervals and I'm guessing the next ones will use digitised datasets extensively.

What I'm looking for is a method of producing a report (in R6 or using Gateway/Atlas) which performs the calculations and presents the data in such a way as to be usable to propose a Category (as detailed in IUCN Red List Categories and Criteria )
Some of it is clearly not feasible for taxa where abundance cannot me measured but site occupancy should be calculable.

This is not an entirely theoretical exercise even for LERC workers - presumably you use these reports for site management advice.

2 Reply by HarryClarke

  • HarryClarke
  • Veteran
  • Offline
  • From: Surrey
  • Registered: 11-10-2011
  • Posts: 99

Re: Produce an IUCN category from occurrence data

Darwyn, if I understand your question correctly. IUCN Red Lists are based on changes in distribution and/or abundance, as well as number of squares occupied. There is no report that R6 or any other database can produce will provide you with an answer to how the trend is changing over time. Rather the database can provide the raw data which is used for some complex statistical analysis on, to provide a trend in abundance or distribution. For most taxonomic groups, trends in abundance is not possible, and hence the only analysis that can be performed is on distribution. The latest distribution analysis uses Occupancy modelling.
I have performed some occupancy modelling of butterfly data for Surrey butterflies using the R package Sparta that was developed by CEH, which took about 15 hours to run (after extracting the data from the database). I understand for the national dataset, the model takes about a week to run.

For butterfly data, site managers can use abundance trends. However, this is a long term exercise (and a lot of records), as you need at least 5 years worth of data before you can start to calculate a meaningful abundance trend. Anything shorter and you are just measuring seasonal effects.

Harry Clarke
Surrey County Butterfly Recorder

3 Reply by Darwyn Sumner

  • Darwyn Sumner
  • Darwyn Sumner
  • Veteran
  • Offline
  • From: Dipterists Forum
  • Registered: 10-05-2006
  • Posts: 250

Re: Produce an IUCN category from occurrence data

Thanks for that Harry. Very thought-provoking.
As I understand it, biogeographical analyses can be carried out at three levels: presence/absence, abundance, species distribution modelling.
For the very small numbers of occurrences in these provisional assessments, there's currently a considerable dependence upon personal knowledge and expertise. The most populous of the taxa in my scheme have just 122 occurrences - ever. No prospect of surveys, just chance records from interested recorders (thank goodness for iRecord and photo sites but even the latter only get the larger photogenic species) and abundance data rarely gets recorded as the vast majority are chance single swept records unless one is lucky enough to stumble across a mass emergence.
The formal IUCN Species Status Review reports are founded on a data table demonstrating hectad or tetrad count changes between defined time periods. So a report which coughs up the number of occupied hectads over specified time periods (e.g. decades) for each of the taxa stored in a rucksack seems to be feasible, even if one is obliged to download missing records from the Gateway into R6.
A Grid map from the Gateway will produce such figures simply by specifying appropriate date ranges, but that's not a quick method.
To give an idea of numbers, my national scheme only has 4083 occurrences in total, far fewer than in many taxon groups at the local level.

(Most interested in SDM but I suspect that's not going to help with this particular problem. Even if I factor in European records meaningful associations are tricky)