Opportunistic citizen science data of animal species produce reliable estimates of distribution trends if analysed with occupancy models

AbstractSpecies records from volunteers are a vast and valuable source of information on biodiversity for a wide range of taxonomic groups. Although these citizen science data are opportunistic and unstructured, occupancy analysis can be used to quantify trends in distribution. However, occupancy analysis of unstructured data can be resource-intensive and requires substantial expertise. It is valuable to have simple ‘rules of thumb’ to efficiently assess the suitability of a dataset for occupancy analysis prior to analysis.Our analysis was possible due to the production of trends, from our Bayesian occupancy analysis, for 10 967 species from 34 multi-species recording schemes in Great Britain. These schemes had an average of 500 visits to sites per year, and an average of 20% of visited sites received a revisit in a year. Occupancy trend outputs varied in their precision and we used expert elicitation on a subset of outputs to determine a precision threshold above which trends were suitable for further consideration. We then used classification trees with seven metrics to define simple rules explaining when the data would result in outputs that met the precision threshold.We found that the suitability of a species’ data was best described by (i) the number of records of the focal species in the 10% best-recorded years, and (ii) the proportion of recording visits for that taxonomic group with non-detections of the focal species. Surprisingly few data were required to be predicted to meet the precision threshold. Specifically, for 98% confidence that our Bayesian occupancy models would produce outputs meeting the precision threshold, there needed to be ≥29 records of the focal species in the 10% best-recorded years (equivalent to an average of 12.5 records per year in our dataset), although only ≥10 records (equivalent to 4.5 records per year) were required for species recorded in less than 1 in 25 visits.We applied these rules to regional species data for Great Britain. Data from 32% of the species:region combinations met the precision threshold with 80% confidence, and 14% with 98% confidence. There was great variation between taxonomic groups (e.g. butterflies, moths and dragonflies were well recorded) and region (e.g. south-east England was best recorded).These simple criteria provide no indication of the accuracy or representativeness of the trend outputs: this is vital, but needs to be assessed individually. However our criteria do provide a rapid, quantitative assessment of the predicted suitability of existing data for occupancy analysis and could be used to inform the design and implementation of multi-species citizen science recording projects elsewhere in the world.

Download Full-text

Occupancy models for citizen‐science data

Methods in Ecology and Evolution ◽

10.1111/2041-210x.13090 ◽

2019 ◽

Vol 10 (1) ◽

pp. 8-21 ◽

Cited By ~ 16

Author(s):

Res Altwegg ◽

James D. Nichols

Keyword(s):

Citizen Science ◽

Occupancy Models ◽

Science Data

Download Full-text

Modelling the Heterogeneity within Citizen Science Data for Biodiversity Research

Biodiversity Information Science and Standards ◽

10.3897/biss.5.75623 ◽

2021 ◽

Vol 5 ◽

Author(s):

Diana Bowler ◽

Nick Isaac ◽

Aletta Bonn

Keyword(s):

Citizen Science ◽

Heterogeneous Data ◽

Sampling Effort ◽

Occupancy Models ◽

Species Occurrence ◽

Science Data ◽

Global Biodiversity Information Facility ◽

Occurrence Data ◽

Occupancy Detection ◽

Species Population

Large amounts of species occurrence data are compiled by platforms such as the Global Biodiversity Information Facility (GBIF) but these data are collected by a diversity of methods and people. Statistical tools, such as occupancy-detection models, have been developed and tested as a way to analyze these heterogeneous data and extract information on species’ population trends. However, these models make many assumptions that might not always be met. More detailed metadata associated with occurrence records would help better describe the observation/detection submodel within occupancy models and improve the accuracy/precision of species’ trend estimates. Here, we present examples of occupancy-detection models applied to citizen science datasets, including dragonfly data in Germany, and typical approaches to account for variation in sampling effort and species detectability, including visit covariates, such as list length. Using results from a recent questionnaire in Germany asking citizen scientists about why and how they collect species occurrence data, we also characterize the different approaches that citizen scientists take to sample and report species observations. We use our findings to highlight examples of key metadata that are often missing (e.g., length of time spent searching, complete checklist or not) in data sharing platforms but would greatly aid modelling attempts of heterogeneous species occurrence data.

Download Full-text

Endemic and Threatened Amazona Parrots of the Atlantic Forest: An Overview of Their Geographic Range and Population Size

Diversity ◽

10.3390/d13090416 ◽

2021 ◽

Vol 13 (9) ◽

pp. 416

Author(s):

Viviane Zulian ◽

David A. W. Miller ◽

Gonçalo Ferraz

Keyword(s):

Population Size ◽

Citizen Science ◽

Atlantic Forest ◽

Site Occupancy ◽

Geographic Range ◽

Occupancy Models ◽

Science Data ◽

Abundance Estimates ◽

Conservation Actions ◽

One Year

Amazona is the largest genus of the Psittacidae, one of the most threatened bird families. Here, we study four species of Amazona (Amazona brasiliensis, A. pretrei, A. vinacea, and A. rhodocorytha) that are dependent on a highly vulnerable biome: the Brazilian Atlantic Forest. To examine their distribution and abundance, we compile abundance estimates and counts, and develop site-occupancy models of their geographic range. These models integrate data from formal research and citizen science platforms to estimate probabilistic maps of the species’ occurrence throughout their range. Estimated range areas varied from 15,000 km2 for A. brasiliensis to more than 400,000 km2 for A. vinacea. While A. vinacea is the only species with a statistical estimate of abundance (~8000 individuals), A. pretrei has the longest time series of roost counts, and A. rhodocorytha has the least information about population size. The highest number of individuals counted in one year was for A. pretrei (~20,000), followed by A. brasiliensis (~9000). Continued modeling of research and citizen science data, matched with collaborative designed surveys that count parrots at their non-breeding roosts, are essential for an appropriate assessment of the species’ status, as well as for examining the outcome of conservation actions.

Download Full-text

The Use of Citizen Science to Achieve Multivariate Management Goals on Public Lands

Diversity ◽

10.3390/d13070293 ◽

2021 ◽

Vol 13 (7) ◽

pp. 293

Author(s):

Sara Souther ◽

Vincent Randall ◽

Nanebah Lyndon

Keyword(s):

Citizen Science ◽

Land Management ◽

Ecological Integrity ◽

Public Land ◽

Suitable Habitat ◽

Science Data ◽

Public Land Management ◽

Federal Land ◽

Botanical Research ◽

Federal Land Management

Federal land management agencies in the US are tasked with maintaining the ecological integrity of over 2 million km2 of land for myriad public uses. Citizen science, operating at the nexus of science, education, and outreach, offers unique benefits to address socio-ecological questions and problems, and thus may offer novel opportunities to support the complex mission of public land managers. Here, we use a case study of an iNaturalist program, the Tribal Nations Botanical Research Collaborative (TNBRC), to examine the use of citizen science programs in public land management. The TNBRC collected 2030 observations of 34 plant species across the project area, while offering learning opportunities for participants. Using occurrence data, we examined observational trends through time and identified five species with 50 or fewer digital observations to investigate as species of possible conservation concern. We compared predictive outcomes of habitat suitability models built using citizen science data and Forest Inventory and Analysis (FIA) data. Models exhibited high agreement, identifying the same underlying predictors of species occurrence and, 95% of the time, identifying the same pixels as suitable habitat. Actions such as staff training on data use and interpretation could enhance integration of citizen science in Federal land management.

Download Full-text

Abundant Citizen Science Data Reveal That the Peacock Butterfly Aglais io Recently Became Bivoltine in Belgium

Insects ◽

10.3390/insects12080683 ◽

2021 ◽

Vol 12 (8) ◽

pp. 683

Author(s):

Marc Herremans ◽

Karin Gielen ◽

Jos Van Kerckhoven ◽

Pieter Vanormelingen ◽

Wim Veraghtert ◽

...

Keyword(s):

Citizen Science ◽

Western Europe ◽

Environmental Changes ◽

Early Summer ◽

Science Data ◽

The Public ◽

Online Portal ◽

Spring Generation ◽

Reproductive Cycles ◽

New Generation

The peacock butterfly is abundant and widespread in Europe. It is generally believed to be univoltine (one generation per year): adults born in summer overwinter and reappear again in spring to reproduce. However, recent flight patterns in western Europe mostly show three peaks during the year: a first one in spring (overwintering butterflies), a second one in early summer (offspring of the spring generation), and a third one in autumn. It was thus far unclear whether this autumn flight peak was a second new generation or consisted of butterflies flying again in autumn after a summer rest (aestivation). The life cycle of one of Europe’s most common butterflies is therefore still surprisingly inadequately understood. We used hundreds of thousands of observations and thousands of pictures submitted by naturalists from the public to the online portal observation.orgin Belgium and analyzed relations between flight patterns, condition (wear), reproductive cycles, peak abundances, and phenology to clarify the current life history. We demonstrate that peacocks have shifted towards two new generations per year in recent decades. Mass citizen science data in online portals has become increasingly important in tracking the response of biodiversity to rapid environmental changes such as climate change.

Download Full-text

Explaining Spatial Variation in the Recording Effort of Citizen Science Data across Multiple Taxa

PLoS ONE ◽

10.1371/journal.pone.0147796 ◽

2016 ◽

Vol 11 (1) ◽

pp. e0147796 ◽

Cited By ~ 36

Author(s):

Louise Mair ◽

Alejandro Ruete

Keyword(s):

Spatial Variation ◽

Citizen Science ◽

Science Data

Download Full-text

A Conceptual Probabilistic Framework for Annotation Aggregation of Citizen Science Data

Mathematics ◽

10.3390/math9080875 ◽

2021 ◽

Vol 9 (8) ◽

pp. 875

Author(s):

Jesus Cerquides ◽

Mehmet Oğuz Mülâyim ◽

Jerónimo Hernández-González ◽

Amudha Ravi Shankar ◽

Jose Luis Fernandez-Marquez

Keyword(s):

Data Quality ◽

Citizen Science ◽

Graphical Model ◽

Real Life ◽

Scientific Journals ◽

Probabilistic Framework ◽

Science Data ◽

Label Aggregation ◽

Evaluation Of Data ◽

Model Formalism

Over the last decade, hundreds of thousands of volunteers have contributed to science by collecting or analyzing data. This public participation in science, also known as citizen science, has contributed to significant discoveries and led to publications in major scientific journals. However, little attention has been paid to data quality issues. In this work we argue that being able to determine the accuracy of data obtained by crowdsourcing is a fundamental question and we point out that, for many real-life scenarios, mathematical tools and processes for the evaluation of data quality are missing. We propose a probabilistic methodology for the evaluation of the accuracy of labeling data obtained by crowdsourcing in citizen science. The methodology builds on an abstract probabilistic graphical model formalism, which is shown to generalize some already existing label aggregation models. We show how to make practical use of the methodology through a comparison of data obtained from different citizen science communities analyzing the earthquake that took place in Albania in 2019.

Download Full-text

Using citizen science data to monitor the Sustainable Development Goals: a bottom-up analysis

Sustainability Science ◽

10.1007/s11625-021-01001-1 ◽

2021 ◽

Author(s):

Laura Ballerini ◽

Sylvia I. Bergh

Keyword(s):

Sustainable Development ◽

Citizen Science ◽

Sustainable Development Goals ◽

Comparative Case Study ◽

Science Data ◽

The Sustainable Development ◽

Case Study Analysis ◽

Wide Range ◽

Development Goals

AbstractOfficial data are not sufficient for monitoring the United Nations Sustainable Development Goals (SDGs): they do not reach remote locations or marginalized populations and can be manipulated by governments. Citizen science data (CSD), defined as data that citizens voluntarily gather by employing a wide range of technologies and methodologies, could help to tackle these problems and ultimately improve SDG monitoring. However, the link between CSD and the SDGs is still understudied. This article aims to develop an empirical understanding of the CSD-SDG link by focusing on the perspective of projects which employ CSD. Specifically, the article presents primary and secondary qualitative data collected on 30 of these projects and an explorative comparative case study analysis. It finds that projects which use CSD recognize that the SDGs can provide a valuable framework and legitimacy, as well as attract funding, visibility, and partnerships. But, at the same time, the article reveals that these projects also encounter several barriers with respect to the SDGs: a widespread lack of knowledge of the goals, combined with frustration and political resistance towards the UN, may deter these projects from contributing their data to the SDG monitoring apparatus.

Download Full-text