Soaring with Eagles, Swimming with Sharks: Measuring Animal Behaviour with Hidden Markov Models


Around the world there are concerns over the impacts of land use change and the developments (such as wind farms). These concerns have led to the implementation of tracking studies to better understand movement patterns of animals. Such studies have provided a wealth of high-resolution data and opportunities to explore sophisticated statistical methods for analysis of animal behaviour.

We use accelerometer data from aerial (Verreaux’s eagle in South Africa) and marine (blacktip reef shark in Hawai’i) systems to demonstrate the use of hidden Markov models (HMMs) in providing quantitative measures of behaviour. HMMs work really well for analysing animal accelerometer data because they account for serial autocorrelation in data. They allow for inferences to be made about relative activity and behaviour when animals cannot be directly observed too, which is very important.

In addition to this, HMMs provide data-driven estimates of the underlying distributions of the acceleration metrics – and the probability of switching between states – possibly as a function of covariates. The framework that we provide in ‘Analysis of animal accelerometer data using hidden Markov models‘ can be applied to a wide range of activity data. It opens up exciting opportunities for understanding drivers of individual animal behaviour.

The following images provide an inside view into the ecosystems in which the Verreaux’s eagle and blacktip reef shark reside.

Soaring with Veraux’s Eagles

Swimming with Blacktip Reef Sharks

To find out more, read our Methods in Ecology and Evolution article ‘Analysis of animal accelerometer data using hidden Markov models’.

Issue 8.2

Issue 8.2 is now online!

The February issue of Methods is now online!

This issue contains four(!) Applications articles and two Open Access articles. These six papers are freely available to everyone – no subscription required.

 Earth Mover’s Distance: The Earth Mover’s Distance (or EMD) is a method commonly used in image retrieval applications. The authors of this paper propose its use to calculate similarity in space use in the framework of movement ecology. This will be helpful for many questions regarding behavioural ecology, wildlife management and conservation.

 warbleR: The R package warbleR is a new package for the analysis of animal acoustic signal structure. It offers functions for downloading avian vocalisations from the open-access online repository Xeno-Canto, displaying the geographic extent of the recordings, manipulating sound files, detecting acoustic signals or importing detected signals from other software, and much more.

– meteR: The open-source R package, meteR directly calculates all of Maximum entropy theory of ecology’s (METE’s) predictions from a variety of data formats; automatically handles approximations and other technical details; and provides high-level plotting and model comparison functions to explore and interrogate models.

– Noise Egg: The Noise Egg is a device that can produce a low-frequency sound, which can be used as an experimental source of noise both in aquaria and in the field. It was developed to study the effects of noise on communication and behaviour in small aquatic animals; however, it could be used for other purposes, such as testing the propagation of certain frequencies in shallow-water habitats.

Continue reading

Biomonitoring Pollution in Wetlands: A New Method for More Reliable Interpretation of Chemical Data

Post provided by Mark Gillingham and Fabrizio Borghesi

Wetlands tend to accumulate considerable anthropogenic pollution.

Wetlands tend to accumulate considerable anthropogenic pollution.

All living organisms are dependent on trace elements (TEs), including metals, that are acquired in very small quantities through their environment or diet. Most TEs are essential for growth, development and physiology of the organism, but excessive intake can be detrimental for animals and plants.  Some TEs – especially heavy metals such as mercury, cadmium, lead and others – are generally toxic though. This toxicity occurs because species’ natural mechanisms fail to excrete excess TEs quickly enough for their metabolism to cope. TEs are present in the environment at different concentrations, either through natural processes or anthropogenic processes (i.e. pollution).

Since the industrial revolution, pollution from human activities has dramatically increased the concentrations of TEs in the natural environment. TEs from pollution tend to persist for a long time on the top layer of soils and sediments, because they do not undergo microbial degradation. As a consequence they tend to enter the food web quicker than the same elements of natural origin.

Some natural environments are more vulnerable to toxic effects of TEs. For instance, wetlands are geochemical endpoints of large river systems that often flow near or through cities, roads, factories, industries, cultivated lands, and/or mines, so they tend to accumulate considerable anthropogenic pollution. Vulnerable habitats like wetlands need to be closely monitored in order to assess the environmental health of these ecosystems. For this kind of monitoring we need reliable methods to measure TEs exposure, intake and bioaccumulation. Continue reading

Sounding Them Out: A Unique Conservation Tool for Monitoring Bush-Crickets

Below is a press release about the Methods paper ‘Potential for coupling the monitoring of bush-crickets with established large-scale acoustic monitoring of bats‘ taken from the British Trust for Ornithology.

Speckled Bush-cricket © Tom Housley

Speckled bush-cricket © Tom Housley

New research led by British Trust for Ornithology (BTO) and published today in the international journal Methods in Ecology and Evolution, shows how existing bat monitoring could improve our understanding of bush-crickets.

Bush-crickets are a little-known group of insects that inhabit our marshes, grasslands, woods, parks and gardens. Some may be seen in the summer when they are attracted to artificial lights, but as most produce noises that are on the edge of human hearing, we know little about their status. There are suggestions that some bush-crickets may be benefiting from climate change, while others may be affected by habitat changes. But how to survey something that is difficult to see and almost impossible to hear? Continue reading

Issue 8.1

Issue 8.1 is now online!

The January issue of Methods is now online!

All of the articles in this month’s issue of Methods in Ecology and Evolution are free for the whole year. You will not need a subscription to access or download any of them throughout 2017.

Our first issue of this year contains three Applications articles and two Open Access articles. These five papers will be freely available permanently.

– CDMetaPOP: Cost–Distance Meta-POPulation provides a novel tool for questions in landscape genetics by incorporating population viability analysis, while linking directly to conservation applications.

– Rphylopars: An R implementation of PhyloPars, a tool for phylogenetic imputation of missing data and estimation of trait covariance across species (phylogenetic covariance) and within species (phenotypic covariance). Rphylopars provides expanded capabilities over the original PhyloPars interface including a fast linear-time algorithm, thus allowing for extremely large data sets (which were previously computationally infeasible) to be analysed in seconds or minutes rather than hours.

– ggtree: An R package that provides programmable visualisation and annotation of phylogenetic trees. ggtree can read more tree file formats than other software and allows colouring and annotation of a tree by numerical/categorical node attributes, manipulating a tree by rotating, collapsing and zooming out clades, highlighting user selected clades or operational taxonomic units and exploration of a large tree by zooming into a selected portion.

Continue reading

Lichens and the “health” of ecosystems: we are closer to a global ecological indicator

Below is a press release about the Methods paper ‘Tracking global change using lichen diversity: towards a global-scale ecological indicator‘ taken from the University of Lisbon.

Candelaria pacifica. © Paula Matos

Candelaria pacifica. © Paula Matos

For the first time, it is possible to integrate at the global scale the results obtained with the most widely used methods to evaluate the “health” of ecosystems using lichens. This is the result of a study now published in the journal Methods in Ecology and Evolution, and represents a fundamental step for this indicator to be considered at the global scale and included in the list of indicators of the United Nations.

Lichens have long been successfully used by scientists as ecological indicators – a kind of environment health thermometer. These complex organisms – the yellow or green taints we often see on the surface of tree trunks – are very sensitive to pollution and changes in temperature and humidity. Evaluating how many lichens, of what kind, and their abundance in a certain ecosystem allows scientists to understand the impact that problems like climate change or pollution have on those ecosystems.  Continue reading

How Should Biologists Measure Climate Change?

Post provided by Christopher Nadeau

Climate change could cause the extinction of one in six species and change the abundance and distribution of those that remain (Urban, 2015). This doesn’t necessarily mean that one in six species in your backyard will go extinct though. Climate change impacts will vary greatly around the globe, with some regions seeing disproportionate effects.

The degree to which climate change will affect species in your region depends on many factors (e.g., land use and species traits), but the amount of climate change that species experience in your region – known as climate change exposure – will certainly be important. For that reason, measuring and mapping climate change exposure is critical for predicting where climate change will have the biggest impacts. Yet, biologists have no agreed upon method to measure exposure and different methods can produce dramatically different results.

A Simple Measure of Exposure and its Limitations

Climate can be defined as a statistical description of weather (e.g., temperature, precipitation) over the course of a long time period, usually 30 years. Most often climate is reduced to the average value of a particular weather variable over a 30-year period of interest. Climate change is then measured as the difference between the averages in two time periods; say the predicted average between 2070-2099 minus the average between 1971-2000.

Projected changes in annual average temperature between 1971-2000 and 2070-2099

Projected changes in annual average temperature between 1971-2000 and 2070-2099.

For example, the map to the left shows projected exposure to changes in average annual temperature. This map suggests that species in the arctic will be exposed to the most temperature change while species in the southern hemisphere will experience the least change. However, there are many problems with this interpretation. Continue reading

How to Synthesize 100 Articles in Under 10 Minutes: Reviewing Big Literature Using ACA

Post provided by Gabriela Nunez-Mir

filesA search of almost any topic on Google Scholar promises to return tens of thousands of hits in less than a second. The first step in any research endeavour is to wade through the titanic amounts of articles available to become acquainted with the existing knowledge. For many people it’s one of the most dreadful and tedious parts of the scientific process.

But what if we could streamline/facilitate this step by automatizing parts of it? Automated content analysis (ACA) gives us the opportunity to do just that. ACA – a text-mining method that uses text-parsing and machine learning – is able to classify vast amounts of text into categories of named concepts. It can then quantify the frequency of those concepts and the relationships among them. Continue reading

Why Do We Need Digital Elevation Models to Infer the Local Adaptation of Alpine Plants?

Post provided by Kevin Leempoel

dsc_4214-crest-flight-27-06-11It’s not easy to characterise the local environment of species living in mountains because these habitats are highly heterogeneous. At a large scale, we typically assume that temperature varies with altitude, but at a local scale, we understand that exposure to wind or being in the shade has a great influence on climatic conditions. If you go from the south-facing to the north-facing side of a mountain, it can be easily 5°C colder. If we can feel that, so can the organisms that live up there. Plants in particular are submitted to tremendous climatic variations over a year. What we want to know is: how did they adapt to these climatic variations and how localised is their adaptation?

Overcoming the Challenges of Measuring Local Adaptation

We don’t know much about how organisms adapt locally because it’s so difficult to measure the environmental conditions that these plants are facing. Existing weather stations can’t capture micro-habitat conditions because they are few and far between. What we can do instead, is use topographic models of mountains to model their environment. After all, if orientation, slope or shade have an impact on climatic conditions, why couldn’t we use them to model local variations in temperature for example? Continue reading

Biogeography at a Global Scale: The Benefits of Distributed Experimental Networks

Post provided by Elizabeth Borer



I have always loved the Blue Marble image of Earth from the Apollo 17 mission, yet a large part of my science is focused on experimental responses at the scale of meter squared grassland plots or even individual grass plants. While I spent my early career wanting to be able to say something important about regional or global processes, I found myself feeling like generating any experimental insights into processes and ecosystem responses at larger scales would be an impossible fiction.

As a postdoc, I had the opportunity to do a multi-site study across a north-south precipitation gradient in California and jumped at it. Among other questions, I decided to ask about whether plants and insects varied similarly across sites in response to replicated experimental treatments. Yet, the idea of actually sampling – and then processing samples from – more than about four sites for more than a year or two was utterly daunting. Continue reading