Indicator: pathways associated with alien species introductions

[timadriaens]

Description

The indicator shows the number of non-native plant and animal species introduced in Belgium via a certain pathway. It is based on a checklist of alien species, composed of various existing sources and databases. The information will be updated and refined as the checklist is further supplemented. The available information on introduction pathways was organized following the Convention on Biological Diversity standard (CBD 2014).

This indicator uses the same data output as issue #17. It is a specific group by (per pathway) on the same dataframe.

[stijnvanhoey]

To start the discussion on figure/table: the example of the previous version, made as indocator:

[stijnvanhoey]

As within the scope of the Trias project, pathways naming was adapted, I created a rough version based on the updated checklist information:

Not polished, but can be interesting for the discussion on the new looks...

[stijnvanhoey]

And a simplified version based on the main categories only:

[ToonVanDaele]

We (Tim and I) suggest tables with numbers with the following characteristics:

1. lumping categories 'to be determined by experts' and 'unknown' to one catergory -> 'unkown'

2. a single category 'unkown' at pathway level 1 (no unknowns at pathway level 2)

3. three seperate tables for plantae, non-vertebrates and vertebrates

4. three columns:

   • pathway level 1 (vocabulary partof = NA),
   • pathway level 2 (vocabulary),
   • number of species

5. The background color of the rows indicating the intentionality of introduction (intentional = white, unintentional = light grey). See grouping in vocabulary.

6. a nice to have feature: a list with the names of the species by clicking (or hovering) over the numbers) (up to 5 species or so).

Here an example for the table:

Table Pathways Plants

pathway category	patwhay subcategory	Number of species
Release	fishery in the wild	20
	biological control	50
Escape	horticulture	790
	agriculture	30
	aquaculure / mariculture	10
...	...	...

rows with 'unintentional' introduction -> background color 'light grey'

[peterdesmet]

Regarding 1. to be determined by experts won't occur in the new checklist data (was just something from an old example)

Regarding 6. Since this will likely be a static output, the example species could be added as an extra column. Should its selection be random?

[damianooldoni]

@timadriaens:

About 3. So, you would like to have maximum(!) 6 tables: one per kingdom, except kingdom Animalia which will be split in two: vertebrates and non-vertebrates, isn't?

For the rest, I think I can deal with it the next weeks.

[damianooldoni]

First version available in branch pathways. @peterdesmet : I see that kableExtra package allows popover messages on static pages: see kableExtra popover section So, this could be a way of accomplishing 6 of @timadriaens. I use already that package in order of colouring rows based on intentionality of introduction. What do you think about it?

[damianooldoni]

Although possible, adding popovers is not so straightforward as thought and it would not appear in any printable (to paper or print to file = PDF) version of the report. The version I implemented last week (https://github.com/trias-project/pipeline/commit/05742ac69a8eb5c45853e051cdb367b238195d9e) has an extra column with a random sample of the species per each category. The number of species to put in this column can be changed as it is an input parameter of a function.

[damianooldoni]

As discussed with @timadriaens, we all agree on this approach.

[damianooldoni]

While working on updating checklist indicators based on unified checklist, I see in pathway indicator that all unknowns at pathway level 2 are dumped together in category unknown at pathway level 1. This means the following: species where pathway level 1 is present, but no information on pathway level 2, are dumped together in big category unknown at level 1. Based on taxa in unified checklist, this means that 187 animal species and 119 plants where pathway level 1 is present, but not pathway level 2, are dumped in category unknown at level 1. Doing so, these taxa are added to 129 animals, 164 plants, and 18 Fungi where no pathway information at level 1 neither level 2 is present.

I implemented this strategy months ago, but I find it strange, and @timadriaens as well. Still, it seems that @timadriaens and @ToonVanDaele decided to follow this strategy in this issue: see comment above, point 2: https://github.com/trias-project/indicators/issues/19#issuecomment-381630412.

@timadriaens : could you please confirm this strategy? Or do you prefer to change it and have unknowns at level 2 too?

As example, table would look like this:

pathwa_level1	pathwa_level2	n	examples
unknown		10	a, b, c, h, w
contaminant	food	2	d, f
contaminant	unknown	3	g, k, x

while now it looks like this:

pathwa_level1	pathwa_level2	n	examples
unknown		13	a, b, g, h, x
contaminant	food	2	d, f

[timadriaens]

I would keep the information we dispose of. In other words, if level 2 is not documented, but we do have level 1, then we should keep that level 1 information. So the first table is what we want I think.

[damianooldoni]

Thanks @timadriaens : I will update the TrIAS function and, next, the indicator results. I will close issue only after it.

[timadriaens]

after discussion with @damianooldoni : do we choose a cut-off date for this one (e.g. pathways of introduction since 1990 or since 2000) ? We know a number of pathways are not active anymore, and pathways change over time (this would be possible to explore when dashboarding the checklist). But for now, the indicator mostly wants to add a layer onto the current pathway analysis for the Regulation so I would say since 1990 (this leaves about 900 species but not all of them have pathway information of course)? Checking how the date cut-off influences the amount of information.

[damianooldoni]

Table pathway without cutoff:

pathway_level1	pathway_level2	n
contaminant	animal_parasite	8
contaminant	food	1
contaminant	habitat_material	95
contaminant	nursery	20
contaminant	on_animals	499
contaminant	on_plants	28
contaminant	seed	537
contaminant	timber	11
contaminant	NA	40
corridor	water	22
corridor	NA	56
escape	agriculture	80
escape	aquaculture	19
escape	food_bait	27
escape	fur_farm	1
escape	horticulture	968
escape	ornamental	7
escape	pet	27
escape	research	1
escape	zoo	2
escape	NA	51
natural	dispersal	56
release	biological_control	8
release	conservation	1
release	hunting	2
release	landscape_improvement	3
release	other	9
release	NA	80
stowaway	ballast_water	35
stowaway	container	14
stowaway	hull_fouling	20
stowaway	organic_packing	1
stowaway	other	17
stowaway	people_luggage	16
stowaway	ship	1
stowaway	vehicles	9
stowaway	NA	50
unaided	NA	10
unknown	NA	1
NA	NA	320

Number of taxa - pathways combinations: 3153.

Cut-off: first_observed >= 1990 Number of taxa - pathways combinations: 1055.

Cut-off: first_observed >= 2000 Number of taxa - pathways combinations: 840.

[timadriaens]

OK that still leaves about 1000 pathway x species combo's since 2000, so would go for this cut-off since policy on pathways is mostly relevant for pathways that are currently active.

[damianooldoni]

Thanks. I will put this in pathway indicator pipeline. We close as it is done.

[damianooldoni]

This issue is solved by https://github.com/trias-project/indicators/pull/56. Soon webpage will be updated too. Then I will close this issue.