Completeness Metric

Completeness Metric in Meyers paper:

Completeness is defined as the ratio of expert-predicted species richness documented by GBIF records to total expert-predicted species.

Records : Taken from GBIF and cleansed for taxonomic validation ( must be an actual bird) and geographical validation ( must actually belong to that area; no zoos etc.). Point data is plotted onto world map and divided into regions of varying granularity, the smallest being 110 sq. km.

Expert-predicted richness: From a discussion with Robert Guralnick and Jorrit on Nov 5 2015 it appears that the export richness predictions are derived from International Union for Conservation of Nature (IUCN, http://iucn.org).

Issues in using this metric:

Issue 1 : Cant validate the accuracy of their data since the experts are unknown. Resolution: Assume that since it’s published in a large journal, it’s accurate enough, at least for a first iteration. Alternate Solution: Can we mine GBIF's dataset ourselves and determine at least that part of the ratio?

Issue 2: The data provided is in squares of 1100 sq. km; we need island specific data. Resolution:

• For the Galapagos, since it’s the only land mass on the area, it’s reasonable to assume that all species in that square belong to that island.
• For Kauai and Moorea, we can model them as a system of interlinked islands with two interaction patterns. For "sea-birds", the number of species in each island is equal to the number of species on all islands, since long-distance flight negates the insularity effect of water bodies. For "terrestrial-birds", the number of birds on an island is a function of 1) the distance between the various islands and 2) the relative size of that island, based off the central tenet of the McArthur-Wilson equilibrium model. Assumption: The ratio of expected number of sea-birds to land-birds is assumed to be the same as discovered number of sea-birds to land-birds (to be found on effechecka).
• For the Friday Harbor, there are too many variables to isolate, so I'm not sure how to proceed.

####Possible list of data points for each region:

• Completeness
• Number of/ list of discovered species
• Number of undiscovered species
• Common discovered species between 2 or more regions
• Links to existing work and names of prominent researches