Rybicki and Hanski (2013)
Species–area relationships and extinctions caused by habitat loss and fragmentation
· Paper authors:
Joel Rybicki
o PhD from the University of Helsinki and Aalto University in 2016.
o Currently, a Postdoctoral researcher at the Institute of Science and Technology of Austria.
o Computer scientist working on theoretical questions related to distributed and self-organizing systems. In ecology, he is interested in spatial models, metacommunities, habitat loss and fragmentation.
Ilkka Hanski
o One of the most influential ecologists of recent years, developer of the metapopulation theory.
o born in 1953 in Lempäälä, Finland, died in 2016.
o Studied biology at the University of Helsinki and gained his doctorate, on the community ecology of dung beetles, from the University of Oxford in 1979.
o Did research around the globe, from Sweden to Madagascar.
o Recommended book by Hanski: Messages from Islands: A Global Biodiversity Tour, where he presents his ideas on biodiversity at the time that he gives an account of his field work.
· Paper summary
1. Main question:
Background:Since it was first quantified by Arrheius in the 1920s, the species-area relations (SAR) has been of interest to ecologists. Although the mechanisms behind this pattern are not completely understood, it appears that increase in heterogeneity (more habitats) and population sizes (thus reducing extinction risk) with area are the main drivers of it. One application of SAR curves has been the estimation of extinction rates from data on habitat loss. In fact, this has been one of the main tools to quantify human-induced extinction. There is controversy over the best way to apply this equation to assess species extinction, with opposite views on whether it overestimates extinction (given that more area is necessary to detect the first individual than for extirpating the last one) or underestimates it by not considering the extinction debt from unviable populations. As a response to this, researchers have proposed modifications of the curves to assess the effect of habitat loss, mainly the Endemics–area relationship (EAR) which accounts for the species only present in the lost habitat and the remaining species–area relationship (RAR), which gives the fraction of species surviving habitat loss as a power of the fraction of remaining area
Assumptions:to test the behavior of power-law SAR as well as the other relations in predicting species extinction under habitat loss and fragmentation, the authors created a simulated model of a realistic landscape that assumed:
- Several species distributed in a bounded area.
- Spatially variable habitat type.
- No interspecific interactions
- Different ecological traits among species
- Different dispersal capacities among species.
- performance of a local population in a lattice cell is determined by the match between the species phenotype and habitat type.
- Environmental stochasticity.
Main question:how species-area relation and its derivates behave under realistic habitat loss and fragmentation scenarios?
2. Methods
This paper used a simulation approach where a “virtual landscape” was created. The landscape contains different habitats and a set of species with different habitat affinities (simulation environmental heterogeneity). Each site in the landscape will have a colonization and extinction parameter that will be influenced by habitat loss and fragmentation. After running the simulation, it is possible observe the effect of these to processes on the virtual species and determine the behavior of SAR, EAR and RARs under different habitat loss and fragmentation scenarios.
3. Results:
o Each of the relations provided different types of information and behave differently under different fragmentation and habitat loss scenarios. The EAR best describes the species going extinct after habitat loss occur but omits from the calculation unviable populations that would go extinct in the future, on the other hand, the SAR appears to best predict the species loss in the long term.
o The effect of the configuration of the habitat on SAR will depend on the total amount of remaining habitat in the landscape, being only important when there is little remaining habitat.
o Clustering of the remaining fragments results in lower extinction rates after habitat loss.
4. Conclusion:
o This paper demonstrates through simulations that SAR are not overpredicting species loss, as that model is the one that best incorporates species extinction from unviable populations.
o It also demonstrates that the effect of habitat loss above certain threshold will depend on the configuration of the remaining habitat.
o A practical rule of thumb for conservation spatial prioritization is proposed in which the third of the region is selected as conservation landscapes comprising clusters of habitat and within those landscapes, a third is protected.
5. Comments
I find this paper as a unique instance of how highly theoretical work can have very practical applications. The fact that the paper has implications for reserve design and provides advice on how to proceed to achieve the Aichi targets is one of the most interesting aspects of this work to me.
6. Questions:
o How realistic do you think the model was? Under different assumptions would it provide the same results?
Their main findings about the dangers of habitat fragmentation seemed pretty obvious to me. I'm really surprised fragmentation and heterogeneity were ignored in past models. It seems so intuitive that corridors are more useful for conservation than isolated islands of habitat. And, yes, species have different ecological traits and different needs. Life is complex. I guess this gets at one of my issues with a lot of these all-encompassing models. It's good that we are trying to understand large patterns, but we then need to add realistic details to truly be able to use models effectively.
ReplyDeleteI agree with Maria in that their findings were kind of obvious. But clustering fragmented areas won't always be enough either. Small animals that cannot cross fences or streets will still struggle. I think a large area may maintain fewer species but I feel as though the populations would be more stable (I have no quantitative evidence for this, just a thought) and that would allow the sites to maintain larger animals, as well. Although, it doesn't seem that is a possibility in this situation. They did not mention corridors which surprised me. Putting a corridor between just a few would be beneficial.
ReplyDeleteI found it interesting that different SAR calculation methods can yield different extinction predictions. I'd like to hear more about continental vs island SARs.
ReplyDeleteIn last week's discussion, one of the lingering questions was: how can we apply macroecology to conservation efforts? I thought this paper gave a very specific example and detailed discussion of how they suggest their work be interpreted. Although they may have made a super complicated model to quantify the obvious as suggested above, it is still neat to read their suggested application.
ReplyDeleteThe focus of this paper, looking at how changing models to different situations can impact the effect and accuracy of those models, is a discussion that is needed in any field that relies heavily on predictive modeling. It allows for the field to develop better models and doesn't allow for the over-use of a first initial model that may not be the best for every scenario. With that, research needs to be explored with real sampling of the various habitat scenarios mentioned and tested theoretically. I think it would be interesting to test the models with real data, and that the data would help the models be better understood by more scientists. As Willow mentioned, this has the potential to be useful in conservation efforts, and could be applied to decide when and where corridors need to be built between these fragmented habitats.
ReplyDeleteThe layout of the logical assembly of parameters is highly considered here which is great. Enhanced details and concise reasoning should predict more realistic predictions but also means there will be more chances of errors or miscalculations because life is simply complex. Especially, the author specifically stated that the ignored interspecific interactions which mean that there is a timing issue. I think this simulation model will work if there is no evolution (freeze of time) meaning that in a long-term of habitat loss, organisms still willing to adapt and change to new places and lifestyle. Also, it means that it ignored the new occupants to the remaining areas. I think it needs to be a more dynamic prediction of outcomes that articulates to the next scenarios. This is a great initial step to establish a predictor model that we can use to choose the correct solutions to counteract against habitat fragmentations due to human activities. Better than nothing! I look at patterns of infectious disease patterns which has a great role in extinction patterns which this paper has not been considered. It is a complex paper actually. I am not a mathematician so it takes more time for me to go through the layouts.
ReplyDeleteI'm interested in the conservation aspect of this paper. Because fragmentation would be ineffective in conserving biodiversity, the author suggests using third-of-third rule. Like Altangerel, I also had to go back and forth between the figures and numbers many times to figure out the math.
ReplyDeleteI agree with Maria,some of their conclusions seemed obvious and I was also surprised habitat fragmentation was not analysed in other models. However, I liked that they analysed it in different ways, testing clusters vs. Scattered habitat fragments. As well as Altangel, it took me quite some time to go through the methods, but their main conclusions seemed reasonable in a biological perspective. And I also agree that these models have to be analysed carefully, as they can not include all the real complexity found in each ecosystem. Interspecific (and intraspecific) relations, behavioral and phenotipical platicity are processes that can greatly affect the species survival at short and long term, but I don't know if there is a way to include those processes in the models and if that way would even be accurate at small or large scale modelling.
ReplyDelete