Effects of density on spacing patterns and habitat associations of a Neotropical Glassfrog
- 1. Prime real estate matters: effects of population density on spatial distributions of a Neotropical glassfrog Nicole F. Angeli1,2, Grace V. DiRenzo2, Alexander Cunha3, and Karen R. Lips2 1Applied Biodiversity Sciences, Texas A&M University 2Department of Biology, University of Maryland, College Park 3Organismal and Evolutionary Ecology Program, Harvard University
- 2. Random Different patterns predict processes Many Neotropical stream frogs are aggregated (Atelopus varius, Pounds and Crump, 1987; Craugastor punctariolus, Ryan, Lips and Eichholz, 2008) Uniform Clustered (Clark, 1946; Diggle, 1985; Ripley, 1991; Bivand et al., 2008)
- 3. Clustered animals may fight for resources ‘Limiting wars’ Maynard Smith and Price, 1973 Espadarana prosoblepon; Jacobson, 1985 Espadarana prosoblepon are aggregated (CD=3.215, Witters and Lips, unpub) Why do E. prosoblepon cluster, and why do we care?
- 4. Declines in amphibian community at Omar Torrijos H.D. NP, El Copé, Coclé, Panama t =-24.44, df=486, P < 0.0001 Lips et al. 2006 PNAS Quantify lethal and non-lethal effects of declines
- 5. Males: night, on stream banks; Fighting, calling, mating; lifespan ~ 5 years (Jacobson, 1985; Guayasamin et al., 2009) Females: cryptic (Savage, 2002) Juveniles: hatch and fall into stream leaf litter to develop (Cisneros-Heredia et al., 2006) Espadarana (Centrolene) prosoblepon Espadarana prosoblepon
- 6. Can the spatial organization of E. prosoblepon change over time and space? 1. Does clustering occur on the streams? Map male E. prosoblepon along streams 2. Will clustering change before and after Bd? Detect changes in dispersion at varying densities 3. Are clusters based on resources or interactions? Quantify animal arrangement and habitat
- 7. 4-200 m permanent stream transects Loop Cascada Silenciosa Guabal Visual Encounter Surveys extending 2 m onto stream banks Adults sexed by calling and presence of humeral spines Individual Toe Clips (Heyer et al., 1994)
- 8. Microhabitat variables Canopy Veg at 0.5 m Veg at 1.0 m Veg at 1.5 m Sand Gravel Cobble Boulder CWD Stream Width Stream Depth
- 9. Complete Data Set: X=13 years 225 surveys 480 meters of habitat data N= 1,678 male E. prosoblepon 881 unique males - X-Y locality data - Individual marks
- 10. Density of captures during 225 surveys annualized over 13 years Loop Guabal Cascada Silenciosa Percent change in abundance after Bd -46.3% -5.9% -68.3% -31.8% Years with frogs: 11 11 8 7
- 11. Low intensity, no frogs a. Visualize clusters using density maps smoothed with Gaussian kernel (Baddeley and Turner, 2005) b. Kuldorff (2006) SatScan Statistic to identify clusters independent of space, time Distance from stream (meters) 1. Identify and compare hotspots over time TransectLength(meters)
- 12. 2. Detect changes in dispersion at varying densities a. Detect patterns using linearized Ripley’s K b. Use derived inflection points to calculate nearest neighbor and inter-cluster distance (Bikhofer et al. ,2006; Cowling, 1998; O’Driscoll ,1998; Stoyan ,1992; Ripley, 1981)
- 13. 3. Predict frog locality by microhabitat traits a. Transform variables with PCA b. Use mixed effects generalized linear models to assess predictability
- 14. 1. Frog hotspots before and after Bd Twenty-one hotspots, 9 persist over all years (42.8%) Post-decline: 8 disappear and 4 appear …..independent of local density.
- 15. 2000 2001 2002 2003 2004 2005 2007 2010 2011 Results: Loop * * ** * * ** * * * * * ** * * * * * Clusters p<0.01* TransectLength(meters)
- 16. Results: Cascada * 2000 2001 2002 2003 2004 2005 2006 2007 2010 2011 2012 * * * * * * ** * * * * * * * * ** * * * ** Clusters p<0.01* TransectLength(meters)
- 17. Results: Guabal 2000 2001 2002 2003 2004 2005 2011 2012 * ** * ** * * ** * * ** * Clusters p<0.01* TransectLength(meters)
- 18. Results: Silenciosa 2000 2001 2002 2003 2004 2006 2012 * * * * * * * * * * ** * *** * Clusters p<0.01* TransectLength(meters)
- 19. P=0.45 P=0.45 P=0.29 P=0.03 2. Results: Cluster arrangement differs among transects but not PRE and POST
- 20. P=0.36 P=0.17 P=0.49 P=0.53 Results: Nearest Neighbors does not change PRE and POST, or among transects Nearest-neighbor 0.692±0.74 m
- 21. 3. Can we predict frog abundance by PC-transformed habitat? Pre-decline Post-decline
- 22. Summary: No changes in dispersion occur before and after Bd arrives in the system Individuals interact similarly, and cluster arrangement varies by transect Pre Post Pre Post Pre Post Pre Post
- 23. Are clusters hotspots for resources? E. prosoblepon clusters of new individuals persist in the same locations with similar habitat-associations
- 24. New and unique contributions E. prosoblepon clusters persist in the same locations, individuals interact similarly, …..independent of local density.
- 25. Community analyses of disease transmission Microhabitat attributes of disease See changes tadpole community at by DiRenzo, Graziella et al., at 4:00 in L100B (COS 71). Future Work
- 26. Acknowledgments Lips lab at University of Maryland, College Park Appalachian Lab: Dr. Robert Gardner and Dr. Matt Fitzpatrick National Science Foundation provided support to KRL TAMU herpetology for support and comments Tweet me for R code and conservation news @auratus_nicole
- 27. Can we predict frog abundance by PC-transformed habitat? 2001 GLMM Model Coefficient T-value P value Intercept 1.534 5.788 >0.001 PC1 -0.413 -2.294 0.034 PC2 0.098 0.383 0.705 PC3 0.062 -0.363 0.720 2012 GLMM Intercept 0.802 3.264 >0.001 PC1 0.129 0.930 0.365 PC2 0.201 1.334 0.199 PC3 0.278 1.586 0.131 PRE POST
- 28. Cluster arrangement varies by transect, but individuals interact similarly Nearest-neighbor 0.692±0.74 m
- 29. Summary: Vegetation is always important 1. Vegetation is important PRE and POST Bd on streams 2. Changes in substrate (small to large) and canopy (open) in the Post years PRE POST Microhabitat Canopy Veg at 0.5 m Veg at 1.0 m Veg at 1.5 m Sand Gravel Cobble Boulder CWD Stream Width Stream Depth Microhabitat variables Canopy Veg at 0.5 m Veg at 1.0 m Veg at 1.5 m Sand Gravel Cobble Boulder CWD Stream Width
- 30. El Copé, Panama Espadarana prosoblepon distribution (BerkeleyMapper 2.0) 700 m elevation 8°40’ N, 80°37’17” W
- 31. Costa Rica Colombia Caribbean Sea Pacific Ocean m
- 33. Loop Cascada Guabal Silenciosa
- 34. Vegetation is always important