B072-08
STOAT - A cloud-based toolbox for the intersection and analysis of remote sensing and spatiotemporal biodiversity data

Friday, 11 December 2020: 19:28
Virtual
Walter Jetz1, Richard Li1, Adam Wilson2, Robert P Guralnick3, Ajay Ranipeta4, John Wilshire5, Jeremy Malczyk6 and Michelle Duong1, (1)Yale University, Ecology and Evolutionary Biology, New Haven, CT, United States, (2)SUNY Buffalo, Geography, Buffalo, NY, United States, (3)University of Florida, Boulder, FL, United States, (4)Yale University, New Haven, United States, (5)Yale University, New Haven, CT, United States, (6)Descartes Labs, Inc., Los Alamos, NM, United States
Recorded Presentation
Abstract:
We present a new platform for the precise and flexible fusion of spatiotemporal biodiversity data and environmental data for ecological analysis: STOAT, Spatiotemporal Observation Annotation Tool (https://mol.org/stoat).

Spatiotemporal biodiversity data of various forms are accumulating rapidly in near real-time. They include: incidental museum or citizen science point observations, survey data of varying protocols, sensor-based inventories, animal movement/tracking data, among others. Notably, these data greatly differ in their spatial grains (e.g. from meters to many kilometers), temporal grains (from minutes to months), and their associated spatiotemporal uncertainties. A vast and growing set of global remotely-sensed and other environmental data offer tremendous opportunities for linking these biodiversity data to environmental data flows (‘environmental annotation’). However, ad hoc fusion of such heterogeneous biodiversity data with environmental data can lead to unreliable results, and even mischaracterizations of the environmental signatures of the biodiversity data.

Assessment and monitoring of environmental biodiversity dynamics thus requires tools that are appropriate for the available data, questions, and applications. Our platform aims to support biodiversity and remote sensing scientists by offering with customizable spatial and temporal environmental characterization of biodiversity records. Buffers in space and time can be used to adjust the effective grain size and/or uncertainty of heterogeneous biodiversity data, giving scientists the power to integrate data from heterogeneous biodiversity data sources in a scientifically rigorous way. The platform supports a growing range of remotely sensed and modeled datasets, including Landsat, daily MODIS, and EarthEnv and Chelsa products. The platform provides both an R package and a range of web-based tools as well as fully integration with Map of Life (mol.org) where it is joining a suite of other tools and reports addressing global biodiversity data.

<< Previous Abstract | Next Abstract