aammd
commented
7 years ago Alas! there is a problem. One important violation to the assumption that taxonomy reproduces traits is ontogeny. When adults and larvae have different traits (but the same species name), there are then two rows in the trait table -- one for the adult, and one for the larvae. There are also probably two rows in any bromeliad survey -- again, one, for adult and one for larvae. however, if we join by taxonomy alone we will have 2 x 2 = 4 rows, all for the same taxonomic species -- and worse, we do not know (without inspecting the data) which is the adult and which the larvae.
What is a good solution? One choice is possibly to add a new column: life stage. This column would have three levels (larval, adult or both). Note that this is different from the AS trait, which characterizes when a taxon is aquatic. Instead, this new column would say what life stages are included in this morphospecies. So it would normally be larva for most flies etc, but could be both for things with simple life history (e.g. leeches, ostracods) and either larvae or adult for those rare taxa which are present in both stages.
@cereghino @dsrivast any thoughts before I go through with this?
In order to get the new traits in, we must,
[x] download only the trait which are needed -- species id, "canonical" tratis and taxonomic information for each morphospecies
[x] download the trait data from a safe place (presently google Drive, later Zenodo)
[x] integrate the two
[x] delete everythng which cleans and processes those earlier traits, which we no longer need.
[x] means to identify taxa which are not matched by anything in the traits table