Edge effects and amphibian population dynamics
Introduction
Habitat loss and modification are leading drivers of amphibian declines and biodiversity loss across the globe (Cushman, 2006). Land use change can lead to habitat fragmentation and increase the amount of edge habitat in the landscape. An increase in edge habitat can lead to edge effects, which can cause changes to the biotic and abiotic environment at edges and lead to changes in spatial distribution of populations (Andren, 1994; Murcia, 1995; Cushman, 2006). Edge effects can also influence fragment connectivity and function, thereby further influencing population structure (Santos-Barrera & Urbina-Cardona, 2011).
Habitat fragmentation resulting from land use and land cover change (e.g. deforestation or urbanization) leads to patches of habitat in the landscape. These patches are surrounded by matrix, which is the habitat between fragments or patches (Jules & Shahani, 2003). The theory of island biogeography suggests that a larger island will have a greater number of species than a smaller island (MacArthur & Wilson, 1967). This theory has long been applied to terrestrial landscapes where patches are likened to islands and the surrounding matrix has been viewed as similar to the inhospitable ocean.
However, the theory of island biogeography may not be accurate to use in terrestrial ecosystems because it has been shown that matrix quality, the type of matrix habitat, can profoundly influence within-patch dynamics (Jules & Shahani, 2003; Vandermeer & Lin, 2008; Umetsu et al., 2008; Santos-Barrera & Urbina-Cardona, 2011). Matrix types with more similar structure to the patch had higher quality for organisms due to a higher functional connectivity between patches (Prevedello & Vieira, 2010). For example, an old growth forest patch with a secondary forest matrix likely provides higher quality habitat for organisms than an old growth forest patch with a pasture matrix. Lieberman et al. (1986) found significant differences in amphibian population dynamics between disturbed and undisturbed habitats, where disturbed habitats had fewer species, higher overall abundance, and lower diversity than undisturbed habitats. Vandermeer & Carvajal (2001) determined that relative high matrix quality may not always lead to positive impacts on population dynamics, however. Therefore, more research needs to be done on the effect of patch and matrix habitat types on population dynamics.
This figure adapted from Fahrig et al. (2019) shows the difference between patch scale and landscape scale. Studies examining edge effects on biota have often used a patch scale analysis where several populations in various sized patches are compared to each other. However, a landscape scale analysis of multiple patches with interacting edges may give better insight into how edges are actually affecting species in fragmented landscapes.
La Selva Biological Station
La Selva Biological Station is located 3 km south of Puerto Viejo de Sarapiquí, Puerto Viejo de Sarapiqui, 41001, Costa Rica (10.4306° N, 84.0070° W). It is located in Braulio Carrillo National Park and was established in 1968 as a research station by the Organization for Tropical Studies. Since that time, the station has been widely used by biologists for a wide variety of tropical field work on local flora and fauna.
Click the map to zoom in on La Selva and see its location.
The Landscape
This image shows the area of Heredia, Costa Rica where La Selva Biological Station is located. It is clear from the satellite imagery that this landscape is fragmented with lots of forest patches and various matrix types.
La Selva Biological Station Property
The area of La Selva Biological Station can be seen in white over the satellite image of the landscape. The forest at the station is considered to be lowland tropical rainforest. There is slight temperature variation during the year. Daily temperatures can fluctuate from 19-31°C. The months with the highest amount of rainfall are July, November, and December. A short dry season occurs from February to April.
Land Use and Land Cover
While, the land area of the station itself does not appear as fragmented as the surrounding landscape, there are likely still edges in this landscape due to the variety of land use and land cover types throughout the property.
Simplified Land Use
Because of the high amount of land use and land cover types in the original layer, I used a crosswalk to simplify the land use and land cover data to include only five land cover types rather than fourteen.
Using the simplified land cover and land use data, I calculated the area of each land use type. I was then able to create a pie chart of land use and land cover types in La Selva. Primary Forest (50.09%) covers the highest amount of land area at La Selva while Buildings (0.65%) cover the least amount.
Forest cover
Pictured are raster layers of forest cover from 2000 with a resolution of 30m downloaded from the Hansen et al. (2013) Global Forest Change project. The map on the left is the raw forest cover data with values ranging from 0 (black) to 100 (white) where 0 is equal to no forest cover and 100 is equal to total forest cover. The map on the right is the reclassified forest cover raster where forest cover above 70% was classified as forest (dark green) and forest cover below 70% was classified as matrix (light green). Move the slider to see how the landscape changes with the reclassification of the raster to a simpler forest cover metric.
I used Fragstats, a program which can calculate patch and landscape metrics from a raster image, to determine patch level and landscape level metrics in the La Selva landscape. Using this program, I was able to calculate patch area, total edge, and largest patch index.
Fragstats Output
This raster was created with the output from the Fragstats program. In ArcMap, I joined the output table from Fragstats to the attribute table for this raster layer. I was then able to change the properties of the raster to reflect the size of patches in the landscape. The yellow color represents the smallest patches in the landscape while dark green represents the largest patch in the landscape. The largest patch in the landscape accounts for 77.5% of the total landscape area. The total amount of edge in this landscape is 162.28km of edge.
Limitations
When the area of La Selva is overlayed onto this raster layer, it is clear that there is not much edge habitat within La Selva according to the Fragstats output. Most of the edges in this landscape occur outside of the La Selva property.
The amount of edge could also change depending on different forest cover parameters. I reclassified forest cover where forest cover above 70% was classified as forest and forest cover below 70% was classified as matrix or not forest. Had I used more than two classifications for forest cover (e.g. high, medium, and low) I might see different results for the amount of edge in La Selva.
Amphibian Population Dynamics
Costa Rica hosts a unique and diverse amphibian fauna which has been studied widely in conservation and landscape ecology research. Many amphibians are particularly susceptible to environmental changes because they have strict physiological, life history, and niche requirements (Rothermel & Semlitsch 2002; Becker et al. 2007; Wells 2007). Edges can influence processes that affect survival and fitness of many taxa, but amphibians are particularly susceptible (Gardner et al. 2011). Therefore, land use and land cover and associated edges in the landscape may be influencing the spatial distribution of amphibians at La Selva Biological Station.
Pictured here is Oophaga pumilio, the strawberry poison frog. Oophaga pumilio is a species that resides in lowland habitats, like La Selva. As adults, this species can survive in both continuous forest and matrix habitats (Rivera-Ordonez et al., 2019). The adults of this species prefer anthropogenically influenced areas, which tend to be associated with a large amount of edges (personal observation). I used GIS to examine O. pumilio records at La Selva Biological Research Station in Costa Rica and determine how this species is spatially influenced by land use/land cover type and edges in the landscape.
Oophaga pumilio at La Selva
The pink circles on the map indicate 366 individual records of O. pumilio sightings downloaded from VertNet and GBIF (iNaturalist). I used these data to determine the impact of edges and land use/land cover type on O. pumilio populations at La Selva.
How does land use and land cover type affect O. pumilio presence?
Because the output from Fragstats showed few edges within the La Selva property, I decided to focus on how land use and land cover influence O. pumilio. This map shows the crosswalked land use and land cover data with the individual records of Oophaga pumilio sightings at La Selva.
This pie chart shows the percent of O. pumilio in each land use type at La Selva. Most frogs can be found in primary forest habitat with the second highest amount of sightings in anthropogenically influenced habitat (buildings).
There are several reasons we could be seeing these trends.
- Buildings and human modified areas are likely better sampled areas
- O. pumilio tends to prefer human modified habitats (buildings)
- Older records of O. pumilio sightings from the 1970s and early 1980s often have the same latitude and longitude coordinates for different individuals. These coordinates are at the very center of the La Selva property which lies in Primary Forest. Therefore, these sightings may have occurred at different locations and land use/land cover types at La Selva but were given the central coordinates due to lack of accurate technology at the time. A huge bias towards primary forest land use and land cover type is likely the result of this limitation.
This is a comparison of the percentage of land use and land cover type within La Selva to the percentage of O. pumilio in each land cover type. The most interesting finding is the high percentage of frogs found in human modified land cover (27.32%) when this land use/land cover type only accounts for 0.65% of the land area at La Selva. Again, this could be due to an uneven sampling effort, but it also supports the prediction that O. pumilio prefer human modified landscapes which tend to have many edges.
How do edges influence the presence of O. pumilio?
I used the Near function in ArcMap which calculates distance between features to calculate the distance between the individual records of O. pumilio and the edges in the land use and land cover layer. This allows me to determine where individual frogs are located in relation to edges.
I created a histogram to show the distribution of O. pumilio in relation to the edges in this environment. I grouped all individuals above 1000m away from the edge due to the fact that most of these frogs had the same latitude and longitude measurements and may be skewing the data. It also biologically makes sense to group these individuals together since they are likely not experiencing strong edge effects at a distance of 1000m from the nearest edge.
When we examine distances 0 to 1000m from the edge, most frogs were found between 200 and 300m from the closest edge. Edge effects can penetrate far into a patch of forest (Laurence et al. 2002). We see very few individuals between 0 and 100m from the edge. Due to the land use and land cover and resulting edges at La Selva, there is an edge effect on Oophaga pumilio at least 200m into the forest.
Takeaway points
Land Use and Land Cover
Edge effects due to land use and land cover are influencing the spatial distribution of Oophaga pumilio at La Selva. Most frogs were found in primary forest and human modified land cover. Few O. pumilio are located between 0 and 200m from the edge, indicating an edge effect. Evidence of edge effects on this species was surprising because of their prevalence in anthropogenically influenced habitat (personal observation).
These results could be influenced by sampling bias (e.g. higher sightings found near buildings due to higher presence of observers there) and inaccuracies in the data (e.g. inaccurate latitude and longitude for older sightings).
It would be interesting to continue this research and examine the landscape level effects of edges more closely. I examined edge effects at the patch level (using distance to the nearest edge), which may not be the best way to examine edge effects because it does not take into account multiple edges in the landscape. It would be possible to study landscape effects in ArcMap using buffers around individual O. pumilio points to determine how many edges are present within the landscape around individuals. Perhaps O. pumilio prefer edges more than the distance to the nearest edge measurements suggest.
Forest Cover
I wasn't able to use Fragstats the way I intended, though I did learn more about the amount of edges in the landscape. With the parameters I chose to reclassify the raster, it was difficult to find evidence of edges in the La Selva property. It would be interesting to continue examining how more than two categories of forest cover might influence the results.
References
Andrén, H. (1994). Effects of habitat fragmentation on birds and mammals in landscapes with different proportions of suitable habitat: A Review. Oikos, 71, 355-366.
Becker, C. G., Fonseca, C. R., Haddad, C. F. B., Batista, R. F., & Prado, P. I. (2007). Habitat Split and the Global Declines of Amphibians. Science, 318, 1775-1777.
Cushman, S. A. (2006). Effects of habitat loss and fragmentation on amphibians: A review and prospectus. Biological Conservation, 128(2), 231–240. doi:10.1016/j.biocon.2005.09.031.
Fahrig, L., Arroyo-Rodriguez, V., Bennett, J. R., Boucher-Lalonde, V., Cazetta, E., Currie, D. J., Eigenbrod, F., Ford, A. T., Harrison, S. P., Jaeger, J. A. G., Koper, N., Martin, A. E., Martin, J., Metzger, J. P., Morrison, P., Rhodes, J. R., Saunders, D. A., Simberloff, D., Smith, A. C., Tischendorf, L., Vellend, M., & Watling, J. I. (2019). Is habitat fragmentation bad for biodiversity? Biological Conservation, 230, 179-186.
Gardner, J., Peters, A., Kearney, M., Joseph, L., & Heinsohn, R. (2011). Declining body size: a third universal response to warming? Trends in Ecology and Evolution, 26, 285-291.
Hansen, M.C., Potapov, P.V., Moore, R., Hancher, M., Turubanova, S.A., Tyukavina, A., Thau, D., Stehman, S.V., Goetz, S.J., Loveland, T.R., Kommareddy, A., Egorov, A., Chini, L., Justice, C. O., & Townshend, J. R. G. (2013). High-resolution global maps of 21st-century forest cover change. Science, 342, 850–853.
Jules, E. S., & Shahani, P. (2003). A broader ecological context to habitat fragmentation: why matrix habitat is more important than we thought. Journal of Vegetation Science, 14(3), 459-464.
Laurence, W .F., Lovejoy, T. E., Vasconcelos, H L., Bruna, E. M., Didham, R. K., Stouffer, P. C., Gascon, C., Bierregaard, R. O., Laurance, S. G., and Sampaio, E. 2002. Ecosystem decay of Amazonian forest fragments: a 22-year investigation. Conservation Biology, 16:605-618.
Leiberman, S. S. (1986). Ecology of the leaf litter herpetofauna of a neotropical rain forest: La Selva, Costa Rica. Acta Zoológica Mexicana (N.S.), 15, 1-72. doi:10.21829/azm.1986.13152050.
MacArthur, R. H., & Wilson, E.O. 1967. The theory of island biogeography. Princeton Univ. Press, Princeton, NJ.
Murcia, C. (1995). Edge effects in fragmented forests: implications for conservation. TREE review, 10(2), 58-62.
Prevedello, J. A., & Vieira, M. V. (2010). Does the type of matrix matter? A qualitative review of the evidence. Biodiversity Conservation, 19(5), 1205-1223.
Rivera‐Ordonez, J. M., Justin Nowakowski, A., Manansala, A., Thompson, M. E., & Todd, B. D. (2019). Thermal niche variation among individuals of the poison frog, Oophaga pumilio, in forest and converted habitats. Biotropica, 2019, 1-10.
Rothermel, B. B., & Semlitsch, R. D. (2002). An experimental investigation of landscape resistance of forest versus old-field habitats to emigrating juvenile amphibians. Conservation Biology, 16, 1324–1332. https://doi.org/10.1046/j.1523-1739.2002.01085
Santos-Barrera, G., & Urbina-Cardona, N. (2011). The role of the matrix-edge dynamics of amphibian conservation in tropical montane fragmented landscapes. Revista Mexicana de Biodiversidad, 82(2), 679-687.
Savage, J. 2002. The Amphibians and Reptiles of Costa Rica. Chicago: University of Chicago Press.
Umetsu, F., Metzger, J. P., & Pardini, R. (2008). Importance of estimating matrix quality for modeling species distribution in complex tropical landscapes: a test with Atlantic forest small mammals. Ecography, 31(3), 359-370.
Vandermeer, J., & Carvajal, R. (2001). Metapopulation dynamics and the quality of the matrix. American Naturalist, 158(3), 211-220.
Vandermeer, J., & Lin, B. B. (2008). The importance of matrix quality in fragmented landscapes: understanding ecosystem collapse through a combination of deterministic and stochastic forces. Ecological Complexity, 5(3), 222-227.
Wells, K. D. (2007). The ecology and behavior of amphibians. The University of Chicago Press, Chicago.
