Source: UNIVERSITY OF CALIFORNIA, BERKELEY submitted to
FACILITATING DATA-DRIVEN DECISIONS FOR CONSERVATION & NATURAL RESOURCE MANAGEMENT
Sponsoring Institution
National Institute of Food and Agriculture
Project Status
COMPLETE
Funding Source
HATCH
Reporting Frequency
Annual
Accession No.
1010171
Grant No.
(N/A)
Cumulative Award Amt.
(N/A)
Proposal No.
(N/A)
Multistate No.
(N/A)
Project Start Date
Oct 1, 2016
Project End Date
Sep 30, 2021
Grant Year
(N/A)
Program Code
[(N/A)]- (N/A)
Project Director
Boettiger, CA.
Recipient Organization
UNIVERSITY OF CALIFORNIA, BERKELEY
(N/A)
BERKELEY,CA 94720
Performing Department
Insect Biology
Non Technical Summary
This project focuses on the development of novel approaches and software tools which allow conservation researchers and decision-makers to better leverage both the rich trove environmental and ecological data now available and the powerful but complicated algorithms for making decisions in a complex and changing world. Today, rich data and powerful algorithms in the hands of relevant experts have transformed decisions about marketing and advertising, but not those decisions about how we save our planet. It is time to understand how these approaches can be applied to conservation challenges, and to invent the tools which will allow conservation researchers and policy-makers to better leverage these approaches so that conservation planning and decisions will be based on _all_ available data.
Animal Health Component
50%
Research Effort Categories
Basic
50%
Applied
50%
Developmental
(N/A)
Classification

Knowledge Area (KA)Subject of Investigation (SOI)Field of Science (FOS)Percent
12372991070100%
Knowledge Area
123 - Management and Sustainability of Forest Resources;

Subject Of Investigation
7299 - Research equipment and methods, general/other;

Field Of Science
1070 - Ecology;
Goals / Objectives
This project focuses on the development of novel approaches and software tools which allow conservation researchers and decision-makers to better leverage both the rich trove environmental and ecological data now available and the powerful but complicated algorithms for making decisions in a complex and changing world. Today, rich data and powerful algorithms in hands of relevant experts have transformed decisions about marketing and advertising, but not those decisions about how we save our planet. It is time to understand how these approaches can be applied to conservation challenges, and to invent the tools which will allow conservation researchers and policy-makers to better leverage these approaches so that conservation planning and decisions will be based on _all_ available data.
Project Methods
Research will proceed following my research workflow as documented elsewhere (e.g. http://www.carlboettiger.info/2012/05/06/research-workflow.html). Each project shall be assigned a repository on GitHub to coordinate research and collaboration following the R package structure to ensure portability and transparency. A Docker container environment will be maintained to ensure portability and replicability of the computation involved. Research notes and publication material will also be kept in the GitHub repository. Research repositories will be archived at the time of publication and following subsequent updates using the Zenodo data archive or similar appropriate archive, assigning a DOI to the work and provides robust long-term preservation and discovery. Mature software packages will be released to the central R repository, CRAN, and maintained on GitHub in public repositories with the use of unit testing and continuous integration software to ensure quality. Research will make use of computational resources on campus (Savio cluster) and NSF XSEDE facilities including the Chameleon & Jetstream cloud computing platforms, as well as Amazon Web Services. The use of cloud-based platforms keeps costs down through economy-of-scale and avoiding under-utilization while permitting rapid scaling when required.?

Progress 10/01/16 to 09/30/21

Outputs
Target Audience:Other scientists Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?The project continues to provide important training for three PhD student researchers over this reporting period. How have the results been disseminated to communities of interest?Four papers have been published disseminating our results to the broader research community, as detailed in the publication section. What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? This project realized several substantial accomplishments in the past year, leading to the publication of four additional research papers in leading scientific journals including Ecology Letters, One Earth, Theoretical Ecology, and Current Opinion in Environmental Sustainability. This year of the project saw the introduction of an important new element in understanding the role powerful algorithms can play in conservation decision-making, in which our work explored ethical and political consequences arising from the rapidly expanding role algorithms now have in ecological management and environmental policy. Two of our papers focused on exploring these issues with an interdisciplinary team. In a high-profile article in the journal, One Earth, a paper led by a PhD student under my supervision explores these issues with respect to governance of marine fisheries management. In a second paper, (Scoville et al 2021), our team explores these issues in the context of climate policy. Our two further pieces this year continue to advance the technical, statistical, and modeling aspects of our work in improving analysis of systems with alternative stable states -- that is, systems that can exhibit "tipping points" by undergoing sudden and rapid change.

Publications

  • Type: Journal Articles Status: Published Year Published: 2021 Citation: Chapman, Oestreich, Frawley, Boettiger, Diver, Santos, Scoville, Armstrong, Blondin, Chand, Haulsee, Knight, Crowder (2021). Promoting equity in the use of algorithms for high seas conservation. One Earth doi:10.1016/j.oneear.2021.05.011.
  • Type: Journal Articles Status: Published Year Published: 2021 Citation: Karatayev, Baskett, Kushner, Shears, Caselle, Boettiger (2021). Grazer behavior can regulate large-scale patterning of community states. Ecology Letters doi:10.1111/ele.13828.
  • Type: Journal Articles Status: Published Year Published: 2021 Citation: Reimer, Arroyo-Esquivel, Jiang, Scharf, Wolkovich, Zhu, Boettiger (2021). Noise can create or erase long transient dynamics. Theoretical Ecology doi:10.1007/s12080-021-00518-6.
  • Type: Journal Articles Status: Published Year Published: 2021 Citation: Caleb Scoville, Melissa Chapman, Razvan Amironesei, Carl Boettiger (2021). Algorithmic conservation in a changing climate. Current Opinion in Environmental Sustainability 51, 30-35, doi:10.1016/j.cosust.2021.01.009.


Progress 10/01/19 to 09/30/20

Outputs
Target Audience:Scientific Researchers Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?Research over the past year provided valuable training and experience to three PhD students in my research group, as well as one post-doctoral scholar. One PhD student published her first first-author publication, and all have made imporant strides in the mastery of data science pratcies and their understanding of theoretical ecology. How have the results been disseminated to communities of interest?Results have been disseminated through peer-reviewed publications in internationally read academic journals. What do you plan to do during the next reporting period to accomplish the goals?Over the next reporting period, my group is making a major push into applications of deep reinforcement learning to difficult conservation decision-making problems. We will develop simulation environments based on influential models for several major global chance challenges, including climate change mitigation and adaptation, forest fire response, pandemic control, and invasive species management. We tune and evaluate the performance of leading artificial intelligence (AI) agents based on the latest research in deep reinforcement learning. Both the environments and the AI agents will be implemented open source software that is readily available for others to explore, extend, and compete against.

Impacts
What was accomplished under these goals? Significant progress has continued on both theoretical ecology and computational data science aspects of our research program. On the theory side, my latest paper in Theoretical Ecology highlights challenges to model inference driven by the interaction between stochasticity and a mechanism which drives long transients known as a "ghost attractor." My work showed how typical model estimation methods are easily mislead by these phenomena, which are thought to be common to many ecological systems. My work further showed how decision-theoretic approaches can nevertheless lead to successful management of such systems, despite high levels of uncertainty inherent in these dynamics. Meanwhile, our group continues to make progress in the further development of research software which enables the synthesis of complex ecological data. Our publication taxadb enables analyses of biodiversity data which may work across thousands or hundreds of thousands of species to efficiently resolve taxonomic synonyms to accepted names across a range of major naming authorities.

Publications

  • Type: Journal Articles Status: Published Year Published: 2020 Citation: Kari Norman, Scott Chamberlain, Carl Boettiger (2020). taxadb: A high?performance local taxonomic database interface. Methods in Ecology and Evolution. doi:10.1111/2041-210X.13440
  • Type: Journal Articles Status: Published Year Published: 2020 Citation: Carl Boettiger (2020). Ecological management of stochastic systems with long transients. Theoretical Ecology. doi:10.1007/s12080-020-00477-4
  • Type: Journal Articles Status: Published Year Published: 2019 Citation: de Aguiar, Newman, Pires, Yeakel, Boettiger, Burkle, Gravel, Guimar�es Jr, ODonnell, Poisot, Fortin, Hembry (2019). Revealing biases in the sampling of ecological interaction networks, PeerJ, doi:10.7717/peerj.7566.
  • Type: Journal Articles Status: Published Year Published: 2020 Citation: Pascal, Memarzadeh, Boettiger, Lloyd, Chad�s (2020). A Shiny r app to solve the problem of when to stop managing or surveying species under imperfect detection Methods in Ecology and Evolution. doi:10.1111/2041-210X.13501
  • Type: Journal Articles Status: Published Year Published: 2020 Citation: Carl Boettiger. (2020). [Rp] Fluctuation domains in adaptive evolution. ReScience C 6, 1, #15, https://rescience.github.io/bibliography/Boettiger_2020.html


Progress 10/01/18 to 09/30/19

Outputs
Target Audience:Scientific researchers Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?Post-doctoral researcher Allison Barner completed her term of employment and has successfully been recuited to a faculty position at Colby College, Maine. Graduate student Kari Norman continues to make excellent progress on her research program. Meanwhile two new graduate students have joined the lab to continue their academic and professional training with me: Mellisa Chapman, who is making rapid progress in development of mathematical models of decision-making in farming systems to understand the adoption of diversified farming practices and ecosystem services. Macus Lapeyrolerie is the newest student in the group, whose research is exploring the applications of machine learning to ecological forecasting, particularly in systems that may experience tipping points. How have the results been disseminated to communities of interest?Results have been communicated in academic journal publications. What do you plan to do during the next reporting period to accomplish the goals?Last year saw us wrap up recent work in ecological management issues in fisheries applications, and has seen us turn the bulk of our attention to agricultural systems and associated ecosystem services. Our goal in the next reporting period is to bring the preliminary models of agricultural decision-making into publication, while also pressing forward both with the development of more detailed models while also continuing to improve fundamental theory and computational infrastructure which powers our approach in these areas

Impacts
What was accomplished under these goals? The past year has seen several major advances in our goals to improve conservation-decision-making theory and methods. In particular, our work in Partially Observed Markov Decision Processes (POMDPs) has led to a major theoretical advance which resolves a decades-old paradox on sustainable resource management under uncertainty (doi:10.1086/702704) and has also helped us revise crucial predictions about the future recovery of global fish stocks (doi:10.1073/pnas.1902657116).

Publications

  • Type: Journal Articles Status: Published Year Published: 2019 Citation: Milad Memarzadeh, Gregory L. Britten, Boris Worm, Carl Boettiger (2019). Rebuilding global fisheries under uncertainty. Proceedings of the National Academy of Sciences. doi:10.1073/pnas.1902657116
  • Type: Journal Articles Status: Published Year Published: 2019 Citation: Carl Boettiger, Ryan Batt (2019). Bifurcation or state tipping: assessing transition type in a model trophic cascade. Journal of Mathematical Biology. doi:10.1007/s00285-019-01358-z
  • Type: Journal Articles Status: Published Year Published: 2019 Citation: Milad Memarzadeh, Carl Boettiger (2019). Resolving the Measurement Uncertainty Paradox in Ecological Management. American Naturalist. doi:10.1086/702704.
  • Type: Journal Articles Status: Published Year Published: 2019 Citation: Dan Sholler, Karthik Ram, Carl Boettiger, Daniel S Katz (2019). Enforcing public data archiving policies in academic publishing: A study of ecology journals. Big Data & Society 6(1) 1-18. doi:10.1177/2053951719836258
  • Type: Journal Articles Status: Published Year Published: 2019 Citation: Carl Boettiger (2019). Ecological Metadata as Linked Data. Journal of Open Source Software, 4(34), 1276, doi:10.21105/joss.01276
  • Type: Journal Articles Status: Published Year Published: 2018 Citation: Katz, Allen, Barba, Berg, Bik, Boettiger, et al. (2018). The principles of tomorrow's university. F1000Research, 7:1926 doi:10.12688/f1000research.17425.1.


Progress 10/01/17 to 09/30/18

Outputs
Target Audience:Scientific researchers Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?Post-doctoral researcher Milad Memarzadeh has learned the fundamentals of ecological adaptive management, andsuccessfully moved on to a new position at the interface of civil engineering and ecology. Graduate student Kari Norman has passed her qualifying exam and advanced to candidacy. Post-doctoral researcher Allison Barner has accepted a faculty job offer in Fall 2018, though will continue to finish out her post-doctoral position through July 2019. How have the results been disseminated to communities of interest?Results have been communicated in academic journal publications and the release of open source software. What do you plan to do during the next reporting period to accomplish the goals?Our application focus has so far been primarily in fisheries, while in the next period we hope to build out the groundwork of applying our theoretical work and methods development to the context of decision-making in terrestiral agriculture systems, particularly in understanding the dynamics of farmer decisions and ecosystem services in the context of diversified farming practices.

Impacts
What was accomplished under these goals? A key paper illustrating how the algorithms we have been developing can be applied to adaptive management was published (Memarzadeh & Boettiger, 2018). I have also published an invited review on the stochastic phenomena in ecology and how they can lead to an additional source of information to better inform decisions about ecological systems. We have further released two software packages that play a foundational role in the cyberinfrastructure we use to streamline data management across diverse data and platforms.

Publications

  • Type: Journal Articles Status: Published Year Published: 2018 Citation: Carl Boettiger (2018). From noise to knowledge: how randomness generates novel phenomena and reveals information. Ecology Letters. doi:10.1111/ele.13085
  • Type: Journal Articles Status: Published Year Published: 2018 Citation: Milad Memarzadeh, Carl Boettiger (2018). Adaptive management of ecological systems under partial observability. Biological Conservation. 224, 9-15. doi:10.1016/j.biocon.2018.05.009.
  • Type: Journal Articles Status: Published Year Published: 2018 Citation: Carl Boettiger, (2018). Managing Larger Data on a GitHub Repository. Journal of Open Source Software, 3(29), 971, doi:10.21105/joss.00971.
  • Type: Journal Articles Status: Published Year Published: 2018 Citation: Karthik Ram, Carl Boettiger, Scott Chamberlain, Noam Ross, Maelle Salmon, & Stephanie Butland. (2018). A Community of Practice Around Peer-review for Long-term Research Software Sustainability. Computing in Science & Engineering, 9615(c), 11. doi:10.1109/MCSE.2018.2882753


Progress 10/01/16 to 09/30/17

Outputs
Target Audience:Scientific researchers Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?Post-doctoral scholars: - Milad Memarzadeh - Allison Barner - Daniel Scholler Graduate Students: - Kari Norman Training: Students and post-docs participate at least weekly in group and individual meetings to discuss research, review scientific literature, learn and trouble-shoot technical methods. A monthly group meeting discusses and shares issues in both hard/technical and soft/social skills for building a successful in career in scientific research, while separate individual meetings are used to formulate, review and reflect on each student's goals and progress towards them. Students and post-docs are encouraged to participate in appropriate course-work, teaching, academic meetings, workshops, and other experiences that contribute to their research and professional development. How have the results been disseminated to communities of interest? - Ecological Society of America Annual Meeting, Portland, OR, 2017 - CROSS Symposium Speaker, UC Santa Cruz, 2017. - Imagining Tomorrow's University Chicago, IL, 2017. - rOpenSci unconference, Los Angeles, CA, 2017 - Prov-a-thon: Practical Tools for Reproducible Science, Tamaya, NM, 2017. - NSF Translational Data Science Workshop, Berkeley Institute for Data Science, Berkeley, CA, 2017. - Force16 Codemeta Workshop, Portland OR (organizer), 2016 - CodeMeta NSF Workshop Portland, OR (organizer), 2016 - rOpenSci unconference, San Francisco, CA, 2016 What do you plan to do during the next reporting period to accomplish the goals?Continue research following the project plan.

Impacts
What was accomplished under these goals? ? Novel Approaches: The recent publications in Ecology Letters on "Making ecological models adequate" & BioScience on "Skills and Knowledge for Data Intensive Research" (see Products page) review a wide array of the novel approaches for ecologists and conservation researchers to better leverage our rapidly expanding data resources. Software tools: Recent publications in The American Statistician on "Packaging data analytical work reproducibly using R" and in the R Journal on "An Introduction to Rocker: Docker Containers for R," describe the creation of new software tools that improve a researcher's ability to implement, communicate, reproduce, and scale computationally intensive approaches requiring complex and highly customized software environments. Additional Software created in the period and described in the Journal of Open Source Software, "Generating Codemeta Metadata for R Packages," presents a community standard metadata system to improve discovery and attribution of research software. All of these make an important part of the software ecosystem required to bring modern algorithmic approaches to ecological and conservation challenges.

Publications

  • Type: Journal Articles Status: Published Year Published: 2017 Citation: Hampton, Jones, Wasser, Schuldhauer, Supp, Brun, Herandez, Boettiger, Collins, Gross, Fernandez, Budden, White, Teal, Labou & Aukema. (2017) Skills and Knowledge for Data Intensive Research. BioScience. doi:10.1093/biosci/bix025.
  • Type: Journal Articles Status: Published Year Published: 2017 Citation: Ben Marwick, Carl Boettiger, Lincoln Mullen (2017). Packaging data analytical work reproducibly using R (and friends). The American Statistician. doi:10.1080/00031305.2017.1375986.
  • Type: Journal Articles Status: Published Year Published: 2016 Citation: T Alex Perkins, Carl Boettiger, Benjamin L. Philips. (2016) After the games are over: life-history trade-offs drive dispersal attenuation following range expansion. Ecology and Evolution 6 (18) 6425-6434. doi:10.1002/ece3.2314.
  • Type: Journal Articles Status: Published Year Published: 2017 Citation: Getz, Marshall, Carlson, Giuggioli, Ryan, Roma�ach, Boettiger, Chamberlain, Larsen, D'Odorico, O'Sullivan, D. (2017). Making ecological models adequate. Ecology Letters. doi:10.1111/ele.12893
  • Type: Journal Articles Status: Published Year Published: 2017 Citation: Carl Boettiger (2017). Generating Codemeta Metadata for R Packages. The Journal of Open Source Software 2 (19), 454, doi:10.21105/joss.00454
  • Type: Journal Articles Status: Published Year Published: 2018 Citation: Carl Boettiger, Dirk Eddelbuettel (2018). An Introduction to Rocker: Docker Containers for R. The R Journal. https://journal.r-project.org/archive/2017/RJ-2017-065/index.html.