Hunger Mapping - GBIF Data Blog


A hunger map is a map of missing biodiversity data (a biodiversity data gap). The main challenge with hunger mapping is proving that a species does not exist but should exist in a region. Hunger maps are important because they could be used to prioritize funding and digitization efforts. Currently, GBIF has no way of telling what species are missing from where. In this blog post I review some potential ways GBIF could make global biodiversity hunger maps.

This is a companion discussion topic for the original entry at


Found this article on biodiversity hotspots after posting. Shows that the current animal genus count maps we have are likely filled with data gaps.


@jwaller I’ve always wondered why GBIF doesn’t do something like the ES50 plots that OBIS has, e.g.:
This is an attempt to correct for sampling effort by showing the number of species in a random sample of 50 individuals in an area (in the example above these are computed for a 5°⨉5° grid). There’s a description of the method here:

The expected number of marine species in a random sample of 50 individuals (records) is an indicator on marine biodiversity richness.

The ES50 is defined in OBIS as the sum(esi) over all species of the following per species calculation:

  • when n - ni >= 50 (with n as the total number of records in the cell and ni the total number of records for the ith -species)
    • esi = 1 - exp(lngamma(n-ni+1) + lngamma(n-50+1) - lngamma(n-ni-50+1) - lngamma(n+1))
  • when n >= 50
    • esi = 1
  • else
    • esi = NULL

Warning: ES50 assumes that individuals are randomly distributed, the sample size is sufficiently large, the samples are taxonomically similar, and that all of the samples have been taken in the same manner.

Now obviously this makes a bunch of assumptions, but it would be an interesting thing to do with the GBIF data. One could also use bootstrapping to get some idea of how robust the ES50 values are (i.e., by randomly resampling from the set of occurrences within a given cell). This sounds like the sort of thing that could be automated to generate a “live” map of estimated diversity to complement the default GBIF map showing all occurrences. Thoughts?


The ES50 measure is something we are looking into. I think it is probably feasible to do some statistical modelling along those lines. Knowing the current state of checklist datasets I think that some statistical approach is the only way forward.


Cool, look forward to seeing what happens. Personally I think GBIF needs to be thinking about providing people with things beyond summaries of raw data.


I agree Rod. Concidentally I also suggested this to GBIF recently as I’ve found it very useful for marine data (maps and latitudinal gradients). As ES50 randomly samples a geographic cell it generates an average and standard error bars. Also called Hurlberts index and a form of rarefaction. Would be great to do for the whole world, and perhaps sliced by major taxa (e.g., plants, vertebrates, arthropods, other). As part of next IPCC report we need a world “biodiversity hotspots” map. It needs to be based on published literature. At present, we’d be using published maps largely based on plants and vertebrates but comparisons show regions of endemicity and richness can vary between taxa. A truly global ‘best available’ holisitic data from GBIF would be interesting to contrast with these and at least demonstrate the potential of GBIF to provide policy-relevant data.