The R package Species Presence/Absence - R Trends Analyses The Details The Task Estimating trends in species occurrence from unstructured data • Wildlife recording has a long history in the UK • Data collected details what species were recorded, where and when • Results are traditionally presented in atlases • Technological advances have made recording simpler The Problems Bias in time and space • The number of records made each year has increased over time • Recording is more intense in areas inhabited by prolific recorders • Sparta brings together six methods into one R package. • The package is hosted on GitHub where you can find tutorials and bug and issue reporting For code, tutorials and installation instructions visit: https://github.com/BiologicalRecordsCentre/sparta You can install sparta directly from your R console too... > library(‘devtools’) > install_github(‘sparta’, username = ‘BiologicalRecordsCentre’) Examples of sparta in use • Trends for over 2000 species, calculated using sparta, were used in the State of Nature Report • Red-listing of British bees, wasps, ants and plants • Species risk assessments using sparta trends and species distribution models The State of Nature Report, led by the RSPB, used trend estimates generated using sparta An R package for estimating trends in species’ status from unstructured, presence-only data Sparta Tom August, Colin Harrower and Nick Isaac – Biological Records Centre, Centre for Ecology & Hydrology, Wallingford [email protected] @tomaugust85 AugustT Records from volunteer biological recording are traditionally used to create atlases detailing the distribution of species over a set time period. There are now a range of websites and smart phone applications that allow members of the public to record wildlife 10 100 1000 10000 100000 1000000 1970 1980 1990 2000 2010 log(Number of records) Butterflies Bryophyte Bees Wasps Ants For many groups the number of records made increases dramatically over time The Solutions Bring together available methods • Some methods remove biased data prior to analyses • Others attempt to quantify bias and account for this in their analyses A map of sampling effort as estimated by Frescalo. This is used to account for recorder effort. White indicate areas of high recording intensity • Statistical methods have been developed to account for temporal and spatial bias. How do I get Sparta? The distribution of records made by the top Orthoptera recorder • Simpler recording has led to more data collection • This data lends itself to finer temporal and spatial analysis than has previously been possible • However, data collected in this unstructured fashion can contain biases
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
The R package Species Presence/Absence - R Trends
Analyses
The Details
The Task
Estimating trends in species occurrence from unstructured data
• Wildlife recording has a long history in the UK
• Data collected details what species were recorded, where and when
• Results are traditionally presented in atlases
• Technological advances have made recording simpler
The Problems
Bias in time and space
• The number of records made each year has increased over time
• Recording is more intense in areas inhabited by prolific recorders
•Sparta brings together six methods into one R package.
•The package is hosted on GitHub where you can find tutorials and bug and issue reporting
For code, tutorials and installation instructions visit: https://github.com/BiologicalRecordsCentre/sparta
You can install sparta directly from your R console too...
> library(‘devtools’)
> install_github(‘sparta’, username = ‘BiologicalRecordsCentre’)
Examples of sparta in use
•Trends for over 2000 species, calculated using sparta, were used in the State of Nature Report
•Red-listing of British bees, wasps, ants and plants
•Species risk assessments using sparta trends and species distribution models
The State of Nature Report, led by the RSPB, used trend estimates generated using sparta
An R package for estimating trends in species’ status from unstructured, presence-only data
Sparta Tom August, Colin Harrower and Nick Isaac – Biological Records Centre, Centre for Ecology & Hydrology, Wallingford
@tomaugust85
AugustT
Records from volunteer biological recording are traditionally used to create atlases detailing the distribution of species over a set time period.
There are now a range of websites and smart phone applications that allow members of the public to record wildlife
10
100
1000
10000
100000
1000000
1970 1980 1990 2000 2010
log(
Nu
mb
er o
f re
cord
s)
Butterflies
Bryophyte
Bees
Wasps
Ants
For many groups the number of records made increases dramatically over time
The Solutions Bring together available methods
•Some methods remove biased data prior to analyses
•Others attempt to quantify bias and account for this in their analyses A map of sampling effort as estimated by Frescalo.
This is used to account for recorder effort. White indicate areas of high recording intensity
•Statistical methods have been developed to account for temporal and spatial bias.
How do I get Sparta?
The distribution of records made by the top Orthoptera recorder
• Simpler recording has led to more data collection
• This data lends itself to finer temporal and spatial analysis than has previously been possible
• However, data collected in this unstructured fashion can contain biases
Related Documents
