PREDICTS_symposium

View Github Repository Open presentation in a new window

Martin-Jung

See all presentation from Martin-Jung

PREDICTS_symposium

0 0

PREDICTS_symposium

My talk for the PREDICTS symposium

On Github Martin-Jung / PREDICTS_symposium

Recommendations for broad-scale diversity models from an independent study
Martin Jung Center for Macroecology, Climate and Evolution

Biodiversity models

Predicted change in species richnes from Newbold et al. (2015)

Potential Problems

  • Many broad-scale models generalize over a wide range of variables in a single factor (land use )
  • It is tempting to apply those models in a new spatial context
  • Based on available ( biased ? ) data

Question :

How well does a broad-scale model reflect species diversity and abundance on a local land-use gradient where local conditions are known?

Methods

  • Timed point counts
  • Total 147 sites: 172 species, 2700 individual counts

Auxiliary data for the comparison

  • Remote-sensing and census data (INDVI and meanNDVI, Human population density, Forest-Cover year 2000 )
  • Functional traits (Range size, Threat status, Forest specialization )

Results

Summary

  • Cropland higher number of species than the average more intense cropland site (Agroforestry)
  • More forest-specialists in cropland
  • Community differences in plantation forest maybe driven my type (Eucalyptus)
  • However: Primary forest likely lower diversity due to size and fragmentation

Conclusion

  • Large-scale models succeed at detecting overall impacts of land use change.
  • However they might lose accuracy if they are used to predict local impacts on biodiversity, if local conditions do not conform
  • Species traits and environmental co-variates could provide a window of opportunity to improve predictions in regions, where coarse categories don´t capture all of the variabilty.

Acknowledgements

  • Tim Newbold and Neil Burgess for scientific advice
  • PREDICTS contributors for making data available (http://predicts.org.uk)
  • The CHIESA project partners for support and coordination
  • Everyone at CMEC and UNEP-WCMC
  • Danida Fellowship Centre for financial support

Thanks for listening!

Recommendations for broad-scale diversity models from an independent study Martin Jung Center for Macroecology, Climate and Evolution