Ecology or Economy: Managing the Impact of Infrastructure Projects on Endangered Species

By

  • Share
  • Share URL
  • menu
  • <
  • >
  • 1Short Project Description 
  • 2Summary 
  • 3About MeAbout Our Team 
  • 4Question / Proposal 
  • 5Research 
  • 6Method / Testing and Redesign 
  • 7Results 
  • 8Conclusion / Report 
  • 9Bibliography, References and Acknowledgements 

As the global population grows, it has become increasingly common for infrastructure projects (e.g. dams, solar farms) to adversely impact the habitat quality and overall viability of wildlife in general, and endangered species in particular. The goal of my project is to identify and validate a computer simulation based approach to assess and manage this impact. Using the Panoche Valley Solar Farm project in California and its impact on the endangered giant kangaroo rat population as a real-world case study example, I have been able to validate simulation modeling as a feasible and cost-effective approach to this class of problems.

The goal of my project is to identify a simulation based approach for managing the impact of infrastructure projects on the viability of endangered wildlife. The Panoche Valley Solar Farm project in California and its impact on the giant kangaroo rat (GKR) population is my real-world case study.

Hypotheses: 1. The solar farm as proposed will adversely impact GKR population viability, causing extinction within 100 years; 2. It is possible to identify a modified solar farm footprint that significantly reduces this impact, with minimal reduction in power output.

The control group is the projected GKR population in Panoche if solar farm is not built. Experimental scenarios include the solar farm as currently proposed and a few variants with footprint modifications. I selected Vortex as the computer simulation system for this project. I built a customized stochastic model to accurately capture the habitat and GKR population characteristics. This model was run iteratively to estimate the GKR population over 100 years under each scenario.

Simulation results show that the solar farm as proposed will have a significant adverse impact on the GKR population, with >40% likelihood of extinction. I was able to identify a modified footprint that reduces extinction likelihood to <15% with just an 8% power output reduction.

My project validates computer simulation as an effective approach for managing the impact of infrastructure projects on endangered wildlife. Using the case study, I was able to demonstrate how to effectively balance ecological conservation and economic development.

My Presentation

https://docs.google.com/presentation/d/1YhYqblGkxu-xU_qC5yVjerKSZ-JyIXjoj6ZH7OPmr10/pub?start=false&loop=false&delayms=30000

 

I'm a 14-year old who is passionate about science, specifically environmental science, biology, and medicine. I've always been fascinated by plants/trees and wildlife. I research environment-related topics extensively, and take many nature trips around the US and globally.

I'm very concerned about preserving our open spaces and wildlife for future generations (my blog http://thelaststandblog.wordpress.com/ has some observations). I'm driven to make lasting contributions in promoting the right balance between economic development and ecological conservation. Last year, I worked on an experimental study investigating the use of a native plant (milkvetch) to inhibit the spread of invasive star thistles affecting 15-million acres of California land. This project received a category award at the California State Science Fair. My current project was motivated by my concern for endangered wildlife such as the giant kangaroo rat and San Joaquin kit fox in California's Panoche Valley. I firmly believe that we don't have to choose between ecological conservation and economic development. I want to apply science and technology to help achieve both!

I'm also very interested in the field of medicine and participate in a regenerative medicine project at Stanford University, researching the ability of wnt proteins to support tissue repair and healing. I've won multiple awards for my science research work (several regional/state awards, Broadcom MASTERS national semifinalist). Other interests: Science Bowl (5th nationally), Latin (best-student-in-state 2-years running), astronomy and music (violin, mridangam).

Winning a Google Science Fair prize would be an incredible honor that will validate my passion for environmental conservation research.

The central question of my research is:
Can we devise a computer simulation based approach to accurately assess and effectively manage the impact of an infrastructure project on the population viability of endangered wildlife?

Based on my initial research, it became very apparent that field experimentation is just not feasible for this class of problems -- such experiments will be highly constrained in scope, they will take decades to show results, and the damage done from experimentation may be irreversible. Hence my choice of simulation modeling.

I have used the proposed Panoche Valley Solar Farm project in California and its potential impact on the giant kangaroo rat (GKR) population as the real-world case study example for my research.

My hypotheses are:
1. My first hypothesis is about the impact of the solar farm: The solar farm as proposed will adversely impact GKR population viability in the Panoche Valley, and significantly raise the risk of extinction within the next 100 years;

2. My second hypothesis is about managing the impact: It is possible to identify a modified scope and footprint for the solar farm that (a) significantly reduces the impact on GKR population viability, and (b) keeps the reduction in generated power output low.

Extinction risk is modeled using the probability of extinction, with values >25% defined as undesirable. Power output reduction <20% is defined as an acceptable trade-off. Panoche GKR population value of 1000 was used as the threshold for extinction.

My background research covered the following key areas:

1. Biology of the giant kangaroo rat
A detailed understanding of the life, reproduction and mortality patterns of GKR was essential to ensure that the simulation model accurately captured the biology of this endangered animal. I had to pay special attention to the modeling of mortality rate, as it was identified to be a critical input parameter based on sensitivity analysis. After obtaining accurate estimates for GKR mortality from the US Dept of Fish and Game sources, I validated them using data for a closely related species, Ord’s kangaroo rat, which has been heavily researched. I also used different mortality rates for juvenile and adult GKR.

2. Characteristics of the Panoche Valley ecosystem
I paid specific attention to modeling habitat quality based on the suitability of different sections of Panoche Valley as GKR habitat. Using a GKR Habitat Suitability Map (also referenced in the Solargen environmental study), I came up with a custom Habitat Quality Index and used it as a population state variable in the Vortex simulation model. Habitat quality was an important guiding factor that helped me identify and test alternative solar farm footprint modifications.

3. Details of the proposed Solargen Solar Farm project in Panoche Valley
I researched the Solargen proposal in detail, to understand the solar farm footprint, proposed phasing, and the potential reduction in power output with footprint changes.

4. Environmental Impact studies performed for Solargen
I studied the Solargen Environmental Impact Report (EIR) thoroughly. I was disappointed to find their study was limited to a conceptual analysis of the environmental impact. I expected more rigor, including simulation analysis. Let me expand on this point since it's at the core of my research project.

The Solargen proposal did consider a few alternatives. Some of them reduced the power output substantially –  to 110 MW in one case and to 180 MW in another. One alternative involved the translocation of giant kangaroo rats, which would be very disruptive with negative consequences on viability. Another called "Alternative A revised" proposed conservation easements for GKR and was similar to my modifications in some respects, but had some serious drawbacks.

In my project, I wanted to find alternatives that had a smaller impact on the power output; and I wanted to retain GKR in their current native habitat. And most importantly, I wanted more rigor in the population viability impact analysis.

5. Stochastic modeling of population viability
I researched available methods for population viability modeling to ensure that I made an appropriate choice given the GKR animal biology and the environmental impact problem at hand.

6. Population Viability Analysis (PVA) software
I looked into several alternatives including Vortex, DARTER as well as custom development, and chose Vortex as the right tool for my project.

7. Vortex PVA software and its use in conservation biology research
I referenced several conservation biology studies to ensure that I was using best practices to build a customized simulation model for my project.

Based on detailed research, I shortlisted the Vortex Population Viability Analysis software as the right modeling system for this problem. Vortex is widely used by conservation biologists and widely cited in the literature. While I briefly considered writing custom software, I realized that (a) anything I write will likely not approach the level of detail built into Vortex over the years, and (b) more importantly, the primary focus of my work is on the environmental science and conservation aspects, not on software development.

Using Vortex, I took a lot of care to build a customized model that captured the specific characteristics of GKR and the Panoche habitat. This was very important in order to avoid the "garbage-in, garbage-out" problem where one gets meaningless results if the simulation model and associated parameters are not rigorously defined.

I defined the following input parameters:

  • Available habitat size
  • Habitat quality: Developed a custom habitat quality index to assign GKR habitat suitability weighting on available area.
  • Ecological carrying capacity
  • Initial GKR population: Assumed 25% reduction with solar farm due to habitat disruptions.
  • Reproductive rates: Modeled in detail using reproduction age, litter size, litters/year, and a customized model to capture density dependent reproduction.
  • Mortality rates: Modeled in detail to account for variation across juveniles and adults, impact from drought, and environmental variability.
  • Inbreeding: Built-in Vortex capabilities were used to capture inbreeding effects
  • Drought conditions: Captured using likelihood of drought (average 1-in-20 years) and its impact on reproduction and mortality through reduction in habitat quality

I had to do a lot of research to estimate these input parameters, several of which were modeled as stochastic variables.

Population size and probability of extinction were defined as the key output parameters.

I defined 4 major scenarios: the control group is the projected Panoche GKR population with no solar farm over 100 years. The Panoche solar farm as proposed is the second scenario; I defined 2 additional scenarios with modified footprints for the solar farm.

In the first modified footprint, I eliminate the Southeast extension that is a high suitability GKR habitat. This will cut down about a third of phase 5, reducing about 35 MW of power. In the second modification, I remove all of phase 5, dropping a total of 100 MW of power.

I ran several hundred simulations for each scenario. The simulations were run on a Windows PC. Test runs with 25-50 simulations took around 30 minutes per scenario. I ran 250-500 simulations for the actual experimental runs, which typically took several hours per run for each scenario. The simulation runs were repeated several times for each scenario in order to obtain additional data points and improve the accuracy of results.

I performed sensitivity analysis to identify input parameters with a big influence on the outcome. Mortality rate was identified as a critical parameter which led to additional data validation and model refinements; other parameters like initial population size and carrying capacity didn't materially affect the outcome, and therefore more tolerant to estimation errors.

Simulation results show that the Panoche Valley Solar Farm as proposed will likely have a significant adverse impact on the GKR population, with an estimated >40% likelihood of GKR extinction over 100 years.

I was able to identify an alternative footprint for the solar farm with the potential to significantly mitigate the impact and reduce the likelihood of GKR extinction to <15% over the 100 year horizon. This alternative only reduced power output by 8% -- from 420 MW to 385 MW -- and therefore presented a very reasonable trade-off.

My second alternative footprint further reduced the extinction probability to 11%; however, this modification had a 24% reduction in power output -- from 420 MW to 320 MW -- and hence much less desirable.

Panoche GKR Probability of Extinction

The summary chart below shows the likelihood of GKR extinction under each of the four scenarios. Probability of extinction is calculated in Vortex as the percentage of iterations where the estimated total population in year 100 falls below the 1000 threshold. The control group presents no risk of extinction. The solar farm footprint as proposed poses the highest risk. Modification 2 is expected to involve the least risk of GKR extinction. However, when evaluated in conjunction with the reduction in power output, Modification 1 is the best overall option that optimizes both criteria.

Panoche GKR Mean Total Population

The chart below shows the mean total GKR population in Panoche Valley over the 100 year horizon for each scenario. Mean total GKR population is calculated in Vortex as the estimated total GKR population in year 100, averaged across the iterations of a simulation run. The population stays fairly stable under the control scenario, affected only by reproduction and mortality, and events such as droughts. The solar farm in its proposed footprint is likely to push the mean total population below 2,000 over the next 100 years, a level that is a little too close to the extinction threshold of 1,000. The two modifications investigated appear very promising and will likely provide a significant positive uplift in GKR population.

Sensitivity Analysis

Sensitivity analysis indicated that Mortality Rate is a key input parameter with a significant influence on GKR population viability outcomes. The chart below shows that +2% or +5% changes in mortality rate have a measurable and significant impact on the estimated total mean GKR population. Based on this finding, I conducted additional research to validate mortality rate estimates (e.g. using data for a closely related species, Ord's kangaroo rat) to ensure accuracy, and introduced model refinements (i.e. different estimates for juveniles and adults). Other parameters such as carrying capacity and initial population size did not have a major influence on the outcome, and therefore not as critical.

I've also provided below some detailed graphs from an example simulation run, mostly for illustration purposes.

Scenario: Control Group (No Solar Farm)



Scenario: Solar Farm as proposed

Scenario: Solar Farm Modification 1

Scenario: Solar Farm Modification 2

Overall, my research project validates computer based simulation as a feasible, accurate and cost-effective approach to the assessment and management of the impact of infrastructure projects on the viability of endangered wildlife species.

In the real-world Panoche Solar Farm case study, my results from simulation confirm both my hypotheses:

  1. The Panoche Solar Farm project as proposed will have a significant adverse impact on the population viability of the giant kangaroo rat population, with an estimated >40% risk of extinction over the 100 year horizon.
  2. I was able to identify an alternative proposal for the solar farm footprint that reduces extinction risk to <15%, with only a modest 8% reduction in power output, representing a good balance across ecological conservation and economic development.

In terms of future research, I would like to extend my study in the following ways:

  • Extend the model to include the impact of associated infrastructure and traffic e.g. habitat fragmentation from roads
  • Broaden the study to cover the impact on both the giant kangaroo rat and the San Joaquin kit fox using a predator-prey model.

I would also like to apply this approach to model the environmental impact of other ongoing and future infrastructure projects in California and globally.

I’m very happy to see that my research confirms the following:

  1. Simulation based population viability modeling can be used as a rigorous and cost-effective approach in gaining a detailed understanding of the ecological impact of infrastructure projects
  2. More importantly, it can be effectively used for managing and mitigating the impact of these projects by identifying win-win alternatives.

While Population Viability Analysis software such as Vortex is often used in conservation biology research, their use for environmental impact assessment in the context of infrastructure projects represents a new and creative application. I'm happy that my research work will make an original and lasting contribution towards this class of problems with significant practical value.

With global growth, there will be more and more such projects that will need to be assessed and carefully managed in order to minimize the environmental impact. Blocking these projects is not the right approach -- we need to build roads; and, we need to develop alternative energy sources such as solar farms and reduce our dependence on fossil fuels, for example. The right answer, in my view, is to find ways to build these infrastructure projects, while containing their impact on the ecosystem and the endangered animals. I'm excited to see that science and technology can provide us the right tools to help find these answers based on detailed analysis.

In summary, I firmly believe that we should not have to choose between economic development and ecological conservation – we can and must do both!

Bibliography

  1. San Benito County, Panoche Valley Solar Farm Project  web site: http://www.cosb.us/Solargen/
  2. U.S. Fish and Wildlife Service, Sacramento Fish and Wildlife Office, Sacramento, CA: 5 Year Review, Giant Kangaroo Rat (Dipodomys ingens), February 2010.
  3. The IUCN Red List of Threatened Species, page on Giant Kangaroo Rat: http://www.iucnredlist.org/details/6678/0
  4. VORTEX Population Viability Analysis software. http://www.vortex9.org/vortex.html
  5. Endangered Species Recovery Program Giant Kangaroo Rat page: http://esrp.csustan.edu/publications/pubhtml.php?doc=sjvrp&file=chapter02H00.html
  6. A Course in Mathematical Modeling: By Douglas D. Mooney, Randall J. Swift; March 1999
  7. Population viability analysis in conservation planning: an overview. Akçakaya H.R. and P. Sjögren-Gulve. 2000.  Ecological Bulletins 48:9-21.
  8. Orangutan population biology, life history, and conservation. Perspectives from population viability analysis models. Andrew J. Marshall, Robert Lacy, Marc Ancrenaz, Onnie Byers, Simon J. Husson, Mark Leighton, Erik Meijaard, Norm Rosen, Ian Singleton, Suzette Stephens, Kathy Traylor-Holzer, S. Suci Utami Atmoko, Carel P. van Schaik and Serge A. Wich. Chapter from the book "Orangutans: Geographic Variation in Behavioral Ecology and Conservation"; 2008.
  9. Population viability analysis as a tool in wildlife conservation policy: With reference to Australia. David B. Lindenmayer, Tim W. Clark, Robert C. Lacy, and Virginia C. Thomas. Environmental Management. Nov/Dec 1993, Volume 17, Issue 6, pp 745-758.
  10. Miller, P.S., and R.C. Lacy. 2005. VORTEX: A Stochastic Simulation of the Extinction Process. Version 9.50 User’s Manual. Apple Valley, MN: Conservation Breeding Specialist Group (SSC/IUCN).

Note: I've used the Environmental Impact Report (EIR) on the Panoche Valley Solar Farm project prepared for the County of San Benito as one of my references. This document and associated details were made available for public review online, and references to content from these documents and their use do not involve copyright violation.

Acknowledgements

  1. Dr. Robert Lacy, Conservation Scientist, Chicago Zoological Society and primary author of Vortex Population Viability Analysis software
    • Made Vortex software available free of cost
    • Dr. Lacy also helped me troubleshoot and resolve a software bug in Vortex that I encountered during my project.
  2. Mr. Daniel Sommer, The Harker School (mentor)
    • My science teacher Mr. Sommer acted as a sounding board through the whole project.
  3. Mr. William Spangler, Field Biologist, H. T. Harvey & Associates
    • Mr. Spangler, who is doing work on an environmental study in the Carrizo Plain, CA  answered a few specific questions related to model parameters.
  4. My parents, Hari and Meera Sankar
    • They encouraged my interest in environmental science and took me on numerous road trips and hikes, including several trips to Panoche Valley
Back to Dashboard