Academic conferences urgently need environmental policies

Only 4% of the conferences we assessed offered carbon-offset options for participants, and only 9% advertised any action that reduced the meeting’s environmental impact. You can find out more reading our correspondence article:

Holden, M., Butt, N., Chauvenet, A., Plein, M., Stringer, M. & Chadès, I. (2017). Academic conferences urgently need environmental policies. Nature Ecology & Evolution, 1. doi:10.1038/s41559-017-0296-2 ( http://rdcu.be/uOoO ).

Send me an email if you would like a copy.

 

An efficient formulation to select an optimal set of complementary species

In Tulloch, A. I. T., Chadès, I., Dujardin, Y., Westgate, M. J., Lane, P. W. and Lindenmayer, D. (2016), Dynamic species co-occurrence networks require dynamic biodiversity surrogates. Ecography, 39: 1185–1196. doi:10.1111/ecog.02143, we provide an efficient optimisation formulation to select complementary set of species. This paper was published in December 2016.

My original formulation and solution methods although correct were inefficient for the large datasets that Ayesha was dealing with. Luckily, Yann saved the day with a mathematical programming formulation. This efficient formulation is worth understanding as it can be used to solve many complementary problems in optimal monitoring, conservation planning, project prioritisation, value of information or priority threat management.

Ayesha uses this formulation to show that co-occurence species networks require dynamic set of surrogates.

We also provide the code. As usual, send me an email if you’d like to receive the paper.

 

If extinction could be reversed, how would you decide which species to manage?

We have published an opinion piece on how decision science can provide guidance if de-extinction was an option. It has been very stimulating to work on that piece led by Gwen Iacona (UQ). I was particularly interested in the consequences of switching our thought process from managing a non-renewable resource to a renewable resource … it’s fascinating and it works for so many applications e.g. coal generated power vs solar power.

Iacona, G., Maloney, R. F., Chadès, I., Bennett, J. R., Seddon, P. J. and Possingham, H. P. (2016), Prioritising revived species: What are the conservation management implications of de-extinction?. Funct Ecol. Accepted Author Manuscript. doi:10.1111/1365-2435.12720

Abstract:

“De-extinction technology that brings back extinct species, or variants on extinct species, is becoming a reality with significant implications for biodiversity conservation. If extinction could be reversed, there are potential conservation benefits and costs that need to be carefully considered before such action is taken.

Here, we use a conservation prioritization framework to identify and discuss some factors that would be important if de-extinction of species for release into the wild were a viable option within an overall conservation strategy.

We particularly focus on how de-extinction could influence the choices that a management agency would make with regards to the risks and costs of actions, and how these choices influence other extant species that are managed in the same system.

We suggest that a decision science approach will allow for choices that are critical to the implementation of a drastic conservation action, such as de-extinction, to be considered in a deliberate manner while identifying possible perverse consequences.”

Timing of critical habitat protection matters (open access)

 

 

The latest addition to my research interest on how time influences our decision-making process just came out in Conservation Letters (Martin et al, 2016, Free access). We demonstrate once again, that time spent gathering more information to make better decision is beneficial to a point. Aside from the massive modelling effort we had to go through (see lessons learned below), our conclusion summarizes our main point:

It may be tempting to assume that more information is of value for its own sake, in a decision-making context information has value only when it leads to a change in actions taken, specifically, a change with enough benefit to species protection to outweigh the cost of obtaining the information. In the often contentious environment of endangered species decision making, parties who benefit from delay in taking action often lobby strategically for more information, not because they are concerned for the efficacy of protective actions but because their interests are best served by delaying protection as long as possible. In this environment, reminding everyone that more information does not always translate into more efficacious action may help strike a better balance between action and research. When it comes to species conservation, time is the resource that matters most. It is also the resource we cannot get more of.

Martin T.G., Camaclang A.E., Possingham H.P., Maguire L., Chadès I. (2016) Timing of critical habitat protection matters. Conservation Letters In Press, DOI: 10.1111/conl.12266 (OPEN ACCESS, PDF)

Lessons learned: 

This paper was about 5 years in the making, along the way I have learnt a big deal about using AI reinforcement learning tools for this problem. Once more I had to give up using RL and opted for an exhaustive search to find the optimal stopping time – which was really disappointing considering the amount of time I spent on it. As painful as it sounds, I was using the wrong approach. On top of my head, the hurdles were:

1) the matrix population model of the northern abalone species exhibit some time lag, making the process non-Markovian;

2) the Q-Learning approach took way too long to find the optimal stopping time considering the amount of different configurations I had to go through;

3) the near optimal strategies of the Q-Learning approach were not consistent due to lack of convergence;

4) it was way faster to perform an exhaustive search, and this should have been my first solution for a decision problem that was quite simple to solve.

I am glad this paper is out in Conservation Letters for everyone to enjoy. Well done to all my co-authors for their support and hard work on this piece – especially Tara, for pushing it through the line.

Comparing adaptive management and real options: slides and PDF

Adaptive management and real options approaches for sequential decisions making have undergone significant evolution over the last two decades. Both approaches are based on stochastic optimal control and Markov decision processes. They evolved independently from each other and their developments were motivated by different needs.

Adaptive management was specifically developed to handle decision problems with imperfect knowledge of the dynamics of the system, and is known as ‘learning by doing’. On the other hand, real options analysis was introduced specifically to value the flexibility to change actions over time in response to the evolution of uncertainty, and represents both optimal sequential decisions under uncertainty and a capital budgeting methodology. Because of these different purposes, different analytic and numerical methods were developed to solve these problems.

In our recent MODSIM paper (Chades et al, 2015), we review and compare the concepts, applications and recent advances in the numerical and analytic techniques in adaptive management and real options methodologies. A large body of knowledge accumulated in both fields makes a comprehensive review impractical in the context of this paper. Therefore, our review focuses on the most recent developments, with the purpose to identify potential areas of new developments that would address new challenges in the environmental decision area.

I. Chadès, T. Tarnopolskaya, S. Dunstall, J.Rhodes, and A.Tulloch (2015). A comparison of adaptive management and real options approaches for environmental decisions under uncertainty. In Weber, T., McPhee, M.J. and Anderssen, R.S. (eds) MODSIM2015, 21st International Congress on Modelling and Simulation. Modelling and Simulation Society of Australia and New Zealand, December 2015. ISBN: 978-0-9872143-5-5. (PDF)

Simple interactive networks to represent complex strategies

I am fascinated by networks, what a great way of graphically representing anything you can think about that interacts somehow! In decision-making, networks can represent complex decision trees (or strategy), i.e. they graphically tell you what to do under specific conditions. The issue is, these networks can become too large to make sense of! I’ve investigated interactive ways of representing decision trees, and I was fortunate to stumble upon Gephi. Few days of programming led me to some interesting results, that I happily share with you. Here the nodes represent springs, that can be occupied or empty. There are about 10 springs, so if you follow a path from the root (bottom node) to a top node, you will find the optimal action to perform. It is a work in progress, and more can be done. I believe this is the future of publication by the way, interactive figures/graph/networks, they make communication so effective.

In summary, don’t be shy, use network! I have noticed that network theory communities can be protective of their tools. With a tendency of making simple things such as networks and their metrics seem overly complex. But really networks are quite simple and fun to play with and they can be whatever you wish, you just have to define them. 

A toolbox to solve stochastic dynamic programming problems in R, Matlab, SciLab or Octave

Our MDPToolbox is now published in Ecography. Thank you for supporting freely available programs. Please spread the word! The MDP/ SDP toolbox is now available in R, Matlab, SciLab and Octave. No excuses!

Stochastic dynamic programming (SDP) or Markov decision processes (MDP) are increasingly being used in ecology to find the best decisions over time and under uncertainty so that the chance of achieving an objective is maximised. To date, few programs are available to solve SDP/MDP. We present MDPtoolbox, a multi-platform set of functions to solve Markov decision problems (MATLAB, GNU Octave, Scilab and R). MDPtoolbox provides state-of-the-art and ready to use
algorithms to solve a wide range of MDPs. MDPtoolbox is easy to use, freely available and has been continuously improved since 2004. We illustrate how to use MDPtoolbox on a dynamic reserve design problem.

Chadès, I., Chapron, G., Cros, M.-J., Garcia, F. and Sabbadin, R. (2014), MDPtoolbox: a multi-platform toolbox to solve stochastic dynamic programming problems. Ecography. doi: 10.1111/ecog.00888

Prioritising the management of threat affecting the Pilbara species: conversation article and report available

Here’s the reason of my long silence from this blog. The report of our 2-year project is now available online (PDF, 10Mo)(The Conversation). The report assesses the cost-effectiveness of 17 feasible strategies for managing threats to the 53 most threatened Pilbara species. Key outcomes are that management likely to provide all species with a >50% chance of persistence costs less than $5 million/year. Amongst the most cost-effective strategies are managing introduced species and fire regimes.

Apart from being the project manager – responsible to deliver on time and on budget – I’ve had a lot of fun developing a method to find the complementary strategies that would minimize the biodiversity loss and the cost (manuscript under review). I hope you will enjoy it! It has been a true collaborative project across multiple organizations and has required the involvement of the whole team from start to the end.

Carwardine J, Nicol S, van Leeuwen S, Walters B, Firn J, Reeson A, Martin TG, Chades I (2014) Priority threat management for Pilbara species of conservation significance, CSIRO Ecosystem Sciences, Brisbane.

Pilbara_report_snapshotPilbara_report_snapshot1Pilbara_report_snapshot2

Recovering Interacting Species: Are sea otters a threat for abalone?

And the answer is no, but poaching is.

I’ve recently been contacted to talk about our paper in Conservation Biology (v.26/6), that forced me to read again our paper and assess what information was really valuable in this study. I thought that Jessica Jonhson – science writer for The Wildlife Society’s magazine The Wildlife Professional – did a great job at explaining our findings:

Recovery targets for endangered species often ignore interactions between species. For ones that are tightly linked in a predator-prey relationship—such as the northern sea otter (Enhydra lutris kenyoni) and the northern abalone (Haliotis kamtschatkana), both endangered— management plans that focus on only one species can sometimes put the other at risk of extinction. As reported in Conservation Biology (v.26/6), Iadine Chadès of CSIRO Ecosystem Sciences in Australia and colleagues with Fisheries and Oceans Canada developed a computer model to predict the outcome of proposed recovery strategies that address both sea otters and abalone at once. To make its predictions, the model incorporates the population dynamics of each species, their interactions, and how management decisions affect their abundance.  For example, sea otter populations are recovering well, and abalone could be in danger of overpredation by sea otters. However, the model revealed that even if sea otter predation somehow ceased, the abalone populations would not rebound. Instead, the model identified poaching of abalone as the most significant threat to that species, and calculated that poaching would have to be cut in half in order for populations to grow again. Such models cannot guarantee that a management action will succeed, but can help managers make more informed decisions when complex species interactions are involved.

Thanks Jessica,
Chadès, I., Curtis, J.M.R., and Martin, T.G., 2012. Setting realistic recovery targets for interacting endangered species. Conservation Biology 26, 1016-1025. (PDF)

Apply now: 3-year postdoctoral fellow on optimizing adaptive management!

I’m delighted to announce that our 3-year postdoctoral fellow position on optimizing adaptive management is now open for application!

We are seeking a highly motivated and dynamic postdoctoral research fellow to join CSIRO Ecosystem Sciences’ conservation decisions team led by Dr Iadine Chades, to undertake research on optimizing adaptive management decisions under imperfect detection. The postdoctoral research fellow will be supervised by Iadine Chades, Andy Sheppard (CSIRO) and Pr Tom Dietterich (Oregon State University).

Resources to halt global biodiversity decline are still inadequate.  Managers of threatened species therefore need guidance on how to best invest their scarce resources to maximise the chance of saving species in the long term.  Decision theory is now helping decision-makers prioritise biodiversity threat management across time and space but a major drawback with current decision approaches is their need for “data-hungry” models that simulate how a species or system will behave in the future under different management decisions.

Specifically you will:

  • Develop innovative concepts, theories and techniques to facilitate optimal adaptive management over time for hard to detect invasive and threatened species populations.
  • Contribute to the development of adaptive management recommendations to help practitioners protect biodiversity.
  • Publish findings in high impact journals, present finding at both national and international conferences and participate in interdisciplinary working groups.
  • Contribute to a dynamic, innovative and effective research team working with CSIRO Ecosystem Sciences.
  •  Participate in CSIRO’s postdoctoral training program.

Location:   Dutton Park, Brisbane, QLD, Australia
Salary:      AUD$81K – AUD$88K plus up to 15.4% superannuation
Tenure:     3 year specified term
Reference: Q13/03434

To be successful in this position you will need:

  • A PhD in artificial intelligence, ecology, conservation, computational sustainability or related field of decision theory (e.g. applied mathematics, computer science, economics or related discipline).

Note:  Owing to the terms of CSIRO Postdoctoral Fellowships, you must not have more than 3 years relevant post doctoral experience.

  • Demonstrated research achievement in decision theory, optimal resource allocation, adaptive management or ecological modelling. In particular, demonstrated research achievement in one or more of Markov decision processes (MDP), partially observable Markov decision processes (POMDP), stochastic dynamic programming, reinforcement learning and adaptive management.
  • Demonstrated ability to initiate research characterised by originality, creativity and innovation. Publish the findings from research in international peer reviewed journals or selective conference proceedings.
  • Enthusiasm for applying advanced computational and decision theoretic tools to ecological problems.
  • High-level written, oral and interpersonal communication skills, including demonstrated experience in preparing briefings for a range of audiences, and ability to work effectively in a team.

Position Details – Q13/03434

Harnessing artificial intelligence for biodiversity conservation: the slides!

I’ve been asked to provide the slides of my talk today at the Global Change Institute. Feel free to send me your feedback/comments. Here they are! Thanks for coming to the seminar today, and thanks to Chris Brown for inviting me.

photo

Seminar at the Global Change Institute, University of Queensland

I’m excited to present my work at the Global Change Institute tomorrow (26/09/2013). A good opportunity to communicate and reflect on my work so far. I can promise that the slides will have no equations :-). Here’s the abstract:IMG_1115

At the forefront of linking conservation science with quantitative tools from the field of artificial intelligence (AI), Dr Iadine Chadès will introduce the process of making smart conservation decisions under imperfect knowledge and resource constraints. During her PhD, Dr Chadès developed new methods to tackle complex optimisation problems for mobile robots using Markov decision processes (MDP). She discovered that these models can also be used to improve decision-making in modern conservation science – teaching a robot to navigate utilizes the same mathematics as choosing the best conservation actions to save threatened species under uncertainty. Eager to contribute to conservation science, she changed career and turned towards decisions in ecology. Combining expertise in AI with ecological and economic models, this seminar will look at complex applied conservation problems and the solutions that can be applied to efficiently eradicate invasive weeds, control mosquito-borne diseases and protect threatened species from extinction.

Accounting for complementary to maximize monitoring power for species management

Congratulations to Ayesha Tulloch. We have a new paper in Conservation Biology that addresses how to monitor management actions. I really like that paper and I hope it will become a good reference, check it out! Note that we also provide the Matlab code.

One challenge faced by researchers and conservation practitioners is designing and implementing effective monitoring programs particularly when funds are limited. Decisions about how to monitor are hindered by uncertainty in management outcomes. This research demonstrates a new framework for addressing the uncertainties in selecting species for monitoring change due to a management action or policy, using network theory and decision analysis.

Tulloch A.I.T., Chadès I., Possingham H.P. (2013) Accounting for Complementarity to Maximize Monitoring Power for Species Management. Conservation Biology 27, 988-999. Abstract

Adaptive management of migratory birds

We have a new paper published at IJCAI (top Artificial Intelligence conference, ranked A*, probably the most selective conference, congratulations to Sam and team!).

Nicol S, Buffet O, Iwamura T, Chadès I (2013). Adaptive management of migratory birds under sea level rise. Proceedings of IJCAI-13, Beijing, China. (PDF);

Or read the blog version on the computational sustainability website.

In this paper we are posing an adaptive management challenge to the AI community:

  • Why do we care? Because solving adaptive management problem is a complex optimisation problem and efficient methods are lacking!
  • What are we hoping? We hope that future AI research will account for the specific description of adaptive management problems using our problem as a classic benchmark problem.
  • How can you help, what’s next? If you have a complex problem feel free to submit a challenge to the AI community!

 

Sam at IJCAI 2013
Sam at IJCAI 2013 – Photo: O. Buffet

The conservation decisions team has a new team leader!

That is now official, I am the new team leader of the conservation decisions team and we will be hiring this year! So stay tuned if you are looking for a wicked postdoc position in adaptive management/computational sustainability!

IMG_1113

New paper out: Growing biodiverse carbon-rich forests

Congratulations to JB Pichancourt and team: An excellent work in an excellent journal! Feel free to contact JB directly if you require additional information on our paper.Pichancourt, JB; Firn, J,; Chades, I.; Martin T.G. 2013. Growing Biodiverse Carbon-Rich Forests. DOI: 10.1111/gcb.12345
Abstract:

Regrowing forests on cleared land is a key strategy to achieve both biodiversity conservation and climate change mitigation globally. Maximizing these co-benefits, however, remains theoretically and technically challenging because of the complex relationship between carbon sequestration and biodiversity in forests, the strong influence of climate variability and landscape position on forest development, the large number of restoration strategies possible, and long time-frames needed to declare success.Through the synthesis of three decades of knowledge on forest dynamics and plant functional traits combined with decision science, we demonstrate that we cannot always maximize carbon sequestration by simply increasing the functional trait diversity of trees planted. Continue reading New paper out: Growing biodiverse carbon-rich forests

My ICCB talk is available online

A small post from Baltimore, I gave my talk this morning and I have been asked for the slides. I had many interesting feedback for this talk. I will try to account for as many as I can. Thanks!

 

Eliciting expert opinion and the 4-point estimates method

I’m currently involved in 3 projects where data is not available but we still need to provide guidance to managers on what action will be most efficient. In such cases, we have no choice but eliciting information from experts. There are many ways of proceeding, and you can find relevant information on google, but I still find that the details of how exactly doing it isn’t written anywhere. I feel that there is a big part of non-written way of proceeding that would benefit many of us. So if you are an expert in expert elicitation, please write us a guide – e.g. not another review!

For example we had trouble using 4-point estimates* data, and explaining to our experts what the confidence value represented. We did explain it many times, but we still get errors when we analyze the data. I do feel sorry for our experts that constantly have to rethink their values.

Beta distribution corresponding to the 4-point estimates
Beta distribution corresponding to the 4-point estimates

With internship student Martin Peron, we have developed a program to fit beta distribution to 4-point estimates. We are hoping to submit this program to MATLAB exchange very soon (and GNU Octave). So stay tuned if you are looking for such a program!

* 4 point estimates: best guess, min, max and confidence that the true value of the parameter we are estimating lies in this interval.

Good news

I had good news yesterday.

1) I received a Julius Career Award to help me doing my research on adaptive management over the next 3 years and in particular a 6-month sabbatical.

2) We made substantial progress on our research project where we are trying to find the best social network for a given ecological network. We spent the day brainstorming and programming with Sam Nicol, Shaun Coutts and Angela. Jesse Hoey has been very helpful too and kindly updated SPUDD for the purpose of this study. I feel that this project has the potential to be a kick-ass paper. It feels great to do some cool fresh science!

3) I’m also the new Team Leader of our Conservation Decisions team. I hope I can keep managing my time efficiently so that I can produce good science and mentoring to the team.

I’m off to ICCB in 15 days (Baltimore) and then France for more adaptive management research!