Two nice articles on using Lichen data from iNaturalist - both are open access:
Munzi S, Isocrono D and Ravera S (2023) Can we trust iNaturalist in lichenology? Evaluating the effectiveness and reliability of artificial intelligence in lichen identification. Lichenologist 55, 193–201. https://doi.org/10.1017/S0024282923000403
Abstract
iNaturalist is a widely-utilized platform for data collection and sharing among non-professional volunteers and is widely employed in citizen science. This platform’s data are also used in scientific studies for a wide range of purposes, including tracking changes in species distribution, monitoring the spread of alien-invasive species, and assessing the impacts of urbanization and land-use change on biodiversity. Lichens, due to their year-round presence on trees, soil and rocks, and their diverse shapes and colours, have captured the attention of iNaturalist users, and lichen records are widely represented on the platform. However, due to the complexity of lichen identification,
the use of data collected by untrained, or poorly trained volunteers in scientific investigation poses concerns among lichenologists. To address these concerns, this study assessed the reliability of lichen identification by iNaturalist users by comparing records on the platform with identifications carried out by experts (experienced lichenologists) in three cities where citizen science projects were developed. Results of this study caution against the use of unchecked data obtained from the platform in lichenology, demonstrating substantial inconsistency between results gathered by iNaturalist users and experts.
It is worth pointing out that no competent researcher would use data (whether from a CS site like iNaturalist, or even a herbarium like our national herbarium PRE) without checking the data first. There are several steps to this, but these involve identification, taxa that can be confused, location errors, outliers, and other checks. In this regard this paper is rather naive.
The recommendations are good though: a higher level for RG (more agreements), additional agreements contributing to surety. (iNat counts agreements and disagreements, so this is useful. They stopped though at weighting experts more. Their ideas on the CV AI identifications ignores the fact that the AI is trained on existing IDs, and if they are unreliable, then the AI is unreliable: they should rather have retrained the AI on good data, and then re-evaluated it. This is the catch 22 with using CS data: you can only expect to get out what you put in: if you just grab the data, then expect it to be useless. But if you encourage observers, train them, provide them with guides and keys and advice, help with IDs and provide feedback, and vet and check the data, then it can be as good as can be. For lichens this means some lichen groups will only be good to family or genus or complex - but that is the way it works.
R. Troy McMullin and Jessica L. Allen (2022) An assessment of data accuracy and best practice recommendations for observations of lichens and other taxonomically difficult taxa on iNaturalist. Botany 100: 491–497 (2022) https://dx.doi.org/10.1139/cjb-2021-0160
Abstract
We assess the identification accuracy of ‘research grade’ observations of lichens posted on the online platform iNaturalist. Our results show that these observations are frequently misidentified or lack the necessary chemical and (or) microscopic information for accurate identification. Lichens are a taxonomically difficult group, but they are ubiquitous and eye-catching and are regularly the subject of observations posted on iNaturalist. Therefore, we provide best practice recommendations for posting lichen observations and commenting on observations. Data from iNaturalist are a valuable tool for understanding and managing biodiversity, particularly at this crucial time when large scale biodiversity decline is occurring globally. However, the data must be accurate for them to effectively support biodiversity conservation efforts. Our recommendations are also applicable to other taxonomically difficult taxa.
The recommendations are very good, but only for a lichen specialist. For your average naturalist or CS, these are way over the top. Basically these are guidelines developed by specialists for specialists, without any real regard for how to make it work for someone who is interested and wants to contribute, but does not want to carry a rucksack or take out a bank loan for some lichen observations. There is a need to develop some lichen-focussed CS, and for these the guidelines would be a most useful holy grail.
They also miss the point that downloading the data from GBIF should be the last step. First step is to go through the data and ID it, comment on it, and check it - this must be done on the source, so that the observations are improved., and a record of issues is provided. So GBIF might point you to the data, but the data needs to be checked and vetted and commented at source: i.e. on iNaturalist.
These recommendations are worth repeating, so I will add them as another journal article.
Thanks @beetledude for encouraging me to do this.
Comentarios
Agregar un comentario