Abstract
Identifying and understanding limiting conditions is at the centre of ecology and biogeography. Traditionally, associations between climate and occurrences of organisms are inferred from observational data using regression analysis, correlation analysis or clustering. Those methods extract patterns and relationships that hold throughout a dataset. We present a computational methodology called redescription mining, that emphasizes local patterns and associations that hold strongly on subsets of the dataset, instead. We aim to showcase the potential of this methodology for ecological and biogeographical studies, and encourage researchers to try it.Redescription mining can be used to identify associations between different descriptive views of the same system. It produces an ensemble of local models, that provide different perspectives over the system. Each model (redescription) consists of two sets of limiting conditions, over two different views, that hold locally. Limiting conditions, as well as the corresponding subregions, are identified automatically using data analysis algorithms.We explain how this methodology applies to a biogeographic case study focused on China and southern Asia. We consider dental traits of the large herbivorous mammals that occur there and climatic conditions as two aspects of this ecological system, and look for associations between them.Redescription mining can offer more refined inferences on the potential relation between variables describing different aspects of a system than classical methods. Thus, it permits different questions to be posed of the data, and can usefully complement classical methods in ecology and biogeography to uncover novel biogeographic patterns.A python package for carrying out redescription mining analysis is publicly available.
Highlights
Among the central perspectives in ecology and biogeography is uncovering patterns in the organization of ecological systems and as semblages, and the processes that underlie them (Cox et al, 2020; Dansereau, 1957; MacArthur and Wilson, 1967; Ovaskainen and Abrego, 2020)
Our study focuses on China and southern Asia, which is a pivotal region for biogeographic analyses, due to the complex Asian monsoon climate system and biogeography, affecting the living conditions of approximately one-third of the global human population
For illustration of clustering we use all and variables except HOD and loph count (LOPT), which we found to be mostly constant within the focus area
Summary
Among the central perspectives in ecology and biogeography is uncovering patterns in the organization of ecological systems and as semblages, and the processes that underlie them (Cox et al, 2020; Dansereau, 1957; MacArthur and Wilson, 1967; Ovaskainen and Abrego, 2020). Contemporary biogeographical studies are data inten sive, span increasingly large spatial and temporal scales and require rigorous computational approaches (Pearse and Peres-Neto, 2017). Such analyses typically aim at extracting generic patterns and relations from large observational datasets and highlighting contrasts between different subsets of the data. Kreft and Jetz (2010) and Vavrek (2016) compared clus tering methods to identify biogeographic patterns from species distribu tion data and fossil datasets, respectively. Kreft and Jetz (2010) and Vavrek (2016) compared clus tering methods to identify biogeographic patterns from species distribu tion data and fossil datasets, respectively. Kreft and Jetz (2010) found that the clusters identified this way were overall similar to the classic primary geographical divisions of the world’s biota, and exhibit notable dif ferences in the assignment of some subregions, such as, in particular, Madagascar, the Sahara, northern Africa and the Arabian Peninsula
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
