AI models can now continually learn from new data on high-end smart devices such as smartphones and sensors - ScienceDaily

iNaturalist app users play an important role in helping researchers create global maps of plant traits – ScienceDaily

The knowledge missing in the global distribution of plant traits can be filled in with data from species identification applications. Researchers from the University of Leipzig, the German Center for Integrative Biodiversity Research (iDiv) and other institutions were able to prove this based on data from the popular iNaturalist app. Complemented with data on plant traits, the iNaturalist input results in much more accurate maps than previous methods based on extrapolation from limited databases. Among other things, the new maps provide an improved basis for understanding plant-ecology interactions and modeling the Earth system. The study was published in the journal Nature’s environment and evolution.

Nature and climate depend on each other. Plant growth is absolutely dependent on climate, but this in turn is strongly influenced by plants, such as in the forest, which evaporate a lot of water. In order to be able to make accurate predictions about how the living world will develop, a broad knowledge of the characteristics of vegetation at different locations is necessary, for example, leaf surface size, tissue properties and plant height. However, this data usually has to be recorded manually by specialized scientists in a painstaking and time-consuming process. Thus, the plant trait data available worldwide is very little and covers only certain regions.

The TRY database, operated by iDiv and the Max Planck Institute of Biochemistry in Jena, currently provides such data on plant traits for nearly 280,000 plant species. This makes it one of the most comprehensive databases for mapping plant characteristics in the world. To date, global maps of plant traits have been generated using extrapolation (an estimate beyond the original observational range) from the geographically limited database. However, the resulting maps are not particularly reliable.

In order to fill in the large data gaps, the Leipzig researchers have now taken a different approach. Instead of extrapolating geographically located trait data from the TRY database, they linked it to the extensive dataset from the iNaturalist Citizen Science Project.

With iNaturalist, users of the linked smartphone app share their observations of nature, providing species names, images and geolocation. In this way, more than 19 million data points have been recorded, worldwide, for terrestrial plants alone. The data also feeds into the world’s largest biodiversity database, the Global Biodiversity Information Facility (GBIF). This is publicly available and also serves as an important database for biodiversity research.

In order to test the accuracy of the maps based on the iNaturalist observation set and TRY plant traits, they were compared to the plant trait assessments based on sPlotOpen; The iDiv sPlot platform is the world’s largest archive of plant community data. It contains nearly two million datasets with complete lists of plant species occurring at the sites (plots) studied by professional researchers. The database has also been improved with plant trait data from the TRY database.

Conclusion: The new map based on iNaturalist matches the sPlot data map more closely than previous map products based on extrapolation. “The new maps, based on citizen science data, appearing to be more accurate than extrapolation, were surprising and impressive,” says first author Sophie Wolf, a doctoral researcher at the University of Leipzig. “Especially because iNaturalist and our sPlotOpen reference are completely different in structure.”

“Our study convincingly demonstrates the potential for research in voluntary data,” says recent author Dr. Tija Kattenborn from the University of Leipzig and iDiv. “It is encouraging to further benefit from the synergies between data collected from thousands of citizens and professional scientists.”

Says co-author Professor Miguel Mahesha, Head of the Working Group Department for Modeling Approaches in Remote Sensing at the University of Leipzig and iDiv. “Data for free provides an absolute prerequisite for a better understanding of our planet.”

#iNaturalist #app #users #play #important #role #helping #researchers #create #global #maps #plant #traits #ScienceDaily

Leave a Comment

Your email address will not be published. Required fields are marked *