blog

Pollinator movement through fragmented landscapes

beeMany pollinators live in complex and changeable environments. The location of of their food sources changes with time of year (as plants flower), time of day (as flowers open and close or nectar flows) and physical location (for although plants may not move very much during a flowering season, the places that flowers may be found may differ dramatically with time). Natural selection has shaped the behaviours of pollinators so that they have a suite of behaviours that allow them to exploit their environment: although it is unlikely that they know exactly when and where food will be available, they are able to couple clues from the environment with a repertoire of behaviours that will allow them to find food.

Agriculture and other human-generated change is altering the enviroment that pollinators live within, and it is very likely that the rules they are following are not ideal for the changed landscape. Because pollinators are essential for crop production, agricultural policies often dictate that there is some concession to the pollinators. This could be through leaving set-aside ‘wild’ land for nests and wild flowers, or by adding corridors of uncropped land or hedgerows where beneficial species can travel between environments. However, the pollinators will still be following their evolved rule sets, which means that we need to consider whether our concessions to them are of any use. This is something that is difficult to measure, and we need to use a wide range of techniques to ask whether particular manipulations are of benefit to some or many helpful species.

As a behavioural ecologist, I’m interested in how the behavioural decisions made by individual animals allow them to interact with the environment. For most animals, the environment that they live in is highly complex and frequently unpredictable, and it is often a challenge for us to understand how a particular decision gives the animal an advantage over an alternative behaviour. As well as conducting experiments and making observations of behaviour, we can also use theoretical techniques for exploring how simple behaviours could be the best solutions for dealing with complex environments. Simulation techniques are particularly useful for understanding how particular sets of decision allow an animal to cope with changes at the landscape level (bringing together two very different disciplines: landscape ecology and behavioural ecology).

In a paper that has just been published in PeerJ (Rands 2014), I describe a series of models describe a framework for considering how landscape alterations affect the foraging success of a pollinator nesting within the environment. These models build on earlier ideas presented by Rands and Whitney (2010, 2012, discussed in an earlier blog entry), where we simulated landscapes with simple geometries and allowed pollinators to forage within them. In the new PeerJ paper, I describe how hedgerow removal and set-aside field creation may affect the movement of pollinators. The models demonstrate that decreasing either landscape connectivity (be removing hedges) or wild land availability (through having lots of fields of unusable crops) affect how often pollinators have to switch between different environmental types. This may be important for how they find and collect food: for example, swapping between habitats may lead to a temporary reduction in nectar uptake if the pollinator has to work out how to collect it from a newly-encountered shape of flower.

The models are a first step, and are presented as a means of discussing how we can manipulate the environment in a reproducible way within a model. What needs to be done now is to identify a suitable set of behavioural rules to follow (for those presented in the models are basic, and are very likely to be improved upon!). Ongoing work should be able to demonstrate whether particular environmental manipulations are of value to some of our threatened pollinator species.

Further reading

Rands SA (2014). Landscape fragmentation and pollinator movement within agricultural environments: a modelling framework for exploring foraging and movement ecology. PeerJ 2: e269 | full text | pdf

Rands SA & Whitney HM (2010). Effects of pollinator density-dependent preferences on field margin pollination in the midst of agricultural monocultures: a modelling approach. Ecological Modelling 221: 1310-1316 | abstract | pdf (postprint version)

Rands SA & Whitney HM (2011). Field margins, foraging distances and their impacts on nesting pollinator success. PLoS One 6: e25971 | full text | pdf

Advertisements

Hacking together a cheap but effective infra-red camera

eyesMany animals annoyingly do things differently in the dark than in the light. This causes problems for both field- and lab-based behavioural biologists, as few are blessed with the power of night vision. However, there are ways around this problem. One old-tech method is to observe them using red light, making the assumption that the beastie you’re observing can’t see at these longer wavelengths, but you can. However, a number of studies suggest that this a flawed assumption (e.g. 1, 2). So, how do you see in the dark?

Using electromagnetic radiation that has wavelengths outside the visual spectrum of the animals is one solution. Night vision and other special cameras are able to detect infrared radiation, which has a longer wavelength than visible light. This IR may be radiated as thermal energy by the animal (this is what heat-detecting cameras register), or could be being reflected by the animals and their environment (which is what many night-vision and CCTV cameras do, using an additional source of IR such as a set of IR-emitting LEDs, like the ones found in TV remote controls).

We and most other animals are unable to see these longer IR wavelengths, so these cameras are essentially detecting ‘invisible’ information that is unavailable to us. Any extra information about the environment that could improve your chances of finding food or avoiding predators is going to be useful if it can be detected, and there are of course some animals that can detect IR3: vampire bats, several families of snake and a range of butterflies, beetles and bugs have well-researched abilities to detect IR. Similarly, some prey species have evolved counteradaptations to confuse species that are able to detect IR, as can be seen4 in the anti-pit viper tail-flagging displays of ground squirrels. However, given that many animals are insensitive to IR wavelengths (we assume – it doesn’t hurt to check and double-check if you’re working with a particular species!), using an IR camera is a good first step for observing their behaviour unobtrusively, and many bits of kit are commercially available that allow you to observe nesting behaviour or remotely capture images of animals in the wild (I have no intention of recommending any of the products out there, but you should be able to find hundreds of commercial types available by running a web search for ‘camera trap’, ‘nest cam’ or ‘trail camera’).

Making your own IR-sensitive webcam

However, commercially available kit ranges in price and can be expensive. If you’re piloting a bit of work and want to test things out before you write the big grant, buying one or multiple specially designed cameras may be that little bit too expensive. It is however extremely easy to build your own IR camera using a little bit of initiative. The sensors on webcams and other digital cameras are sensitive to IR, but usually have some sort of filter between the lens and the sensor that blocks out unwanted non-visible electromagentic radiation: it is easy to remove this in some cameras, but can involve some delicate and potentially destructive scraping in others. A quick hunt online will give you access to loads of text and video tutorials on how to do this (here’s a good one from the Naked Scientists). Once removed, there’s nothing else you need to do – the sensors will be registering IR, and probably displaying it as extra purplish light if you’re able to see the camera’s output.

I’ve played with a few different cheap webcams available, and acknowledge that they do differ in both quality and ease of modification. Because sensor types and the way light is filtered onto them differs from camera to camera, I’d recommend shopping around for a camera that has a bit of depth to it as it is likely to have a separate filter that can be easily chipped out, rather than one that has to be scraped directly off the sensor (which will probably damage teh sensor). Removing the filter will also affect the depth at which incoming visible and IR light is focussed on the sensor, and I’d suggest hunting for a camera where the lens is focussed manually rather than by having to use a software interface.

I ended up settling for the Tecknet® 1080P HD Webcam, which was very cheap, is very easy to take apart and alter, has a decent pixel size and image quality, and is easy and stable to focus. These webcams are extremely easy to modify. After removing the front, the focussing lens can be unscrewed (below, A), revealing the filter (the shiny square bit in B) that blocks IR wavelengths from reaching the sensor. There’s probably a neater way of doing this, but it is easy to carefully chip this out with the end of a screw-driver (CBlue Peter warning: make sure you’re wearing eye-protection and using appropriate safety equipment while you’re doing this, and get an adult to help you). Then, simply reassemble and plug into your favourite operating system. I’m fairly certain this voids the warranty though…

removing the IR filter

Hardware for running your IR camera

It’s not just the camera that’s important if you’re trying to build cheap functional kit – you also need something to run it from. You could technically run it on anything that has the correct webcam driver installed (hint: if you’re a Mac user and can’t find the right driver, try running the camera from within Skype, which may well be able to run it). However, since we’re aiming for budget kit here, I’ll give a quick description of a system I’ve put together that runs from a Raspberry Pi – a tiny, inexpensive (GB£20, €26, US$25) computer that runs open-source LINUX-based software*, which means that the system is incredibly portable and can be altered to run off a battery in the field (useful if you’re building a cheap and effective camera traps for IR and/or visible wavelengths, like those deployed by ZSL for monitoring black rhino). Furthermore, because you’re building it yourself, it doesn’t require system administrators to install things for you (a major time-lagging factor for many researchers working in larger institutions!).

I’m assuming here that you’ve managed to get your Raspberry Pi up and running, formatting your SD card with something like Raspbian, and are happy with using a command-line tool (if you’re running a graphic-based interface, you can get at this using one of the ‘Terminal’ applications such as LXTerminal). I’m also assuming that you have managed to connect your system to the internet, as you’ll need to download some software. If you’re intending to set this up using a monitor rather than remotely, and want to be able to remove the monitor at some point during the camera’s use, it’s worthwhile getting the Model B Raspberry Pi with two USB ports, which means that you don’t need to detach the camera at any point.

Software for running your IR camera

Some bits of easily obtainable software are useful if you’re running this off a Raspberry Pi. Firstly, if you want to have a direct feed from the camera which is visible on a monitor in front of you, try something like Camorama. Assuming you have an internet connection, you can install this on your computer by entering

sudo apt-get install camorama

and answering ‘yes’ at appropriate moments. To run it, simply type ‘camorama’ into the terminal: as well as a direct feed and a point-and-click interface that allows you to play with the visual balances, you can take jpeg images too. If you want to record mpeg-format videos, you could try a program such as LUVCView instead, which you install and run in a similar manner (by replacing the word ‘camorama’ with ‘luvcview’ in the commands described).

If instead you want to either take still photos at regular intervals, or use your camera as a motion-triggered device (useful for camera traps), I recommend starting off with Motion, which you need to initially install using

sudo apt-get install motion

To run this at the default setting (where the camera is triggered by motion), just type ‘motion’ into the terminal once you have installed the software. As this package is run from the command line, you can create a text file that details exactly how you want the camera to be configured. For example, if you want to run the camera so that it doesn’t react to motion, but instead captures an image every quarter of a second, you can set up a configuration file using a word processor such as nano:

sudo nano motion.config

and typing the following:

framerate 4
output_all on

which you then save a the file ‘motion.config’ (using ctrl+x). The first command above sets the maximum number of frames per second that the device captures, and the second tells the system to turn off the motion-detection capability of the software, and instead take continuous images.

Having created the motion.config file, you then run motion by entering

motion -c motion.config

It may take a few seconds to start up, and you should then get a display whenever a file is written. To stop the program when it’s running, open another terminal window and type

killall motion

Motion is relatively simple to use, and has a good list of configurations that you can play with: simply reopen the config file you’ve created using the same commands, and alter the text. There’s the option of creating a timelapse video within this as well, but you are limited to only being able to use images that are a second or more apart.

A final word, and some caveats

A note for the coding purists out there: the description given here is written to enable non-coders to put together something that works with a minimum of poking. I am fully aware that there are other more elegant ways of doing this, and many other forms of software that can be used, but this should give a first step to enable a stressed lab/field scientist to have something functioning quickly. I am not willing to give any advice on this or similar applications, and accept no responsibility or liability if you follow these instructions and end up damaging your equipment or yourselves in any way: you follow them entirely at your own risk.

Having made your IR camera and worked out how to run the software behind it, you can then deploy it – you will need an IR light source too, and there are many available out there designed for CCTV systems (make sure you’re using them safely though). It looks like things may soon be made even easier by with the introduction of a super-cheap (US$25) IR camera specifically designed for the Raspberry Pi, which will hopefully be well supported within the Raspberry Pi user community. I’ve currently got some undergraduate project students trying out this kit in the lab, on a neat system that may be very nice for observing social and group behaviour – some more on this soon!

Further reading

1. Gibson G (1995). A behavioural test of the sensitivity of a nocturnal mosquito, Anopheles gambiae, to dim white, red and infra-red light. Physiological Ecology 20: 224-228. doi:10.1111/j.1365-3032.1995.tb00005.x

2. Heise BA (1992). Sensitivity of mayfly nymphs to red light: implications for behavioural ecology. Freshwater Biology 28: 331-336. doi:10.1111/j.1365-2427.1992.tb00591.x

3. Campbell AL, Naik RR, Sowards L & Stone MO (2002). Biological infrared imaging and sensing. Micron 33: 211-225. doi:10.1016/S0968-4328(01)00010-5

4. Rundus AS, Owings DH, Joshi SS, Chinn E & Giannini N (2007). Ground squirrels use an infrared signal to deter rattlesnake predation. Proceedings of the National Academy of Sciences of the USA 104: 14372-14376. doi:10.1073/pnas.0702599104

Metrics for dominance interactions 1: Zumpe and Michael’s ‘Dominance Index’ (1986)

dominance_iconThis is the first in a series of blog entries exploring the metrics used for assessing dominance hierarchies. The previous introductory post explains the rationale behind doing this. An index page will give detailed links to other metrics within this blog.

I’ll start with giving details for how to calculate the Dominance Index metric described by Zumpe & Michael (1986), which uses counts of agonistic encounters to generate individual scores, which can then be used to suggest a hierarchy. Using this technique is possible with pen and paper, so I may be giving a bit more detail of the nuts and bolts than I will with some of the more complex metrics. This technique is intended to give the user a ‘cardinal ranking’ – rather than just sorting the interacting individuals into a ranked order, this technique provides a way of assigning each individual a weighting statistic. The authors suggest that this could be useful for assessing how individual dominance changes over time, if different datasets are used.

The agonistic data required for this statistic are the counts of aggressive/dominant and submissive behaviours between all possible pairings of the group members. These should be collected in two tables. So, for an example group with four individuals (identified as A, G, H, and K), we tally the number of aggressive behaviours committed by each individual to each of its three group members:

table_1
Table 1

For example, A directs aggression towards H on seven occasions. Note also that no aggressive interactions are observed between G and K.

Similarly, we tally the number of submissive behaviours directed towards each individual by the other three group members:

Table 2
Table 2

For example, H displays submissive behaviours to G on 11 occasions. Note also that no submissive behaviours were recorded between H and K.

Having collected this information, we now calculate the percentage of the aggressive actions between pairs of individuals that each individual directs at the other. For example, within the pairing of A and G, nine aggressive acts are recorded (five by A, and four by G). A is aggressive towards G in 55.6% of their aggressive interactions (= 5 / 9), and G is aggressive towards A for 44.4% (= 4 / 9). If we work out these two percentages within each pairing, we can build up a table giving the percentages of aggressive behaviours given by each individual. If no aggression is seen within a pair, the two corresponding table entries for the pair should be marked as ‘null’, as is given for the two entries between G and K here:

Table 3
Table 3

Similarly, the percentages of submissive actions received by individuals within each pairing should also be calculated. Again ‘null’ values should be recorded where no submissive actions within a pair were observed, as seen between H and K here:

Table 4
Table 4

The aggression/submission percentages are then combined by calculating an average aggression/submission score for each possible pairing of group members. For example, the average score for A when it interacts with G is

65.3% = (55.6% + 75.0%) / 2.

If no aggressive actions are recorded for a pair, this average is simply given the value of the percentage of submissive actions (calculated in table 4). So, the average score for G when it interacts with K is 90.9%. Similarly, if no submissive actions are recorded between pair members, the average is assumed to be the percentage of aggressive actions committed by an individual (recorded in table 3). So, the average score for H when it interacts with K is 71.4%. Calculating all possible pairing, we get:

Table 5a
Table 5a

Finally, the Dominance Index for each of the group members is calculated as the mean of the averages calculated for each focal individual, as given in table 5b. For example, the dominance index for A is calculated as 73.4% = (65.3% + 69.2% + 85.7%) / 3.

Table 5b
Table 5b

From this, we can use the Dominance Index rankings to construct a hierarchy. In this case, A > G > H > K.

The metric falls apart when there are no aggressive or submissive acts recorded within a pairing, which means that no average score can be recorded in table 5a. This could potentially be remedied by observing the interacting individuals until some agonistic interaction is recorded, but it may be that the non-interacting individuals are able to assess each other without needing to interact (using alternative cues, or through recognising each other from earlier unrecorded interactions). This metric is therefore not ideal if some individuals do not interact with others.

Similarly, a dataset which records few interactions between individuals may be biased by a few anomalous recorded encounters. However, using mean percentages (as calculated in table 5b) removes biases that could be introduced by simply scoring the overall number of ‘wins’ in dyadic agonistic encounters for each individual, which may be incorrectly inflated by many interactions with a small subset of the group members. I’m also curious to see what happens when a group consists of two or more subgroups where interactions tend to be within rather than between subgroups.

Further reading

Zumpe D & Michael RP (1986). Dominance index: a simple measure of relative dominance status in primates. American Journal of Primatology 10: 291-300 | doi: 10.1002/ajp.1350100402

Metrics for dominance interactions: introduction

dominance_iconMany interactions between group-living individuals can be influenced by hierarchies that exist between the interactors. These interactions can be measured in lots of different ways, and once measured, whatever has been scored needs to be processed to give a reproducible estimate of the shape of these interactions.

What this means in practice if you’re starting a new project with a new study organism is that you spend a lot of time thinking about what behaviours to record, and how to record them, but don’t really give consideration to the means of crunching these numbers down to something meaningful at the end. Good experimental design implies that the analysis has been considered during the design of the experiment, but this intermediate stage of generating ‘raw’ information about any hierarchies that are in place may be left out, meaning that something has to be cobbled together post hoc after the work has been done. This is never ideal!

Having supervised a fair number of projects where exactly this has been done, I’ve decided to try and get my head around the various statistics out there that are designed for assessing and ranking hierarchies. Some of these are fairly straightforward, and some are slightly more involved, dipping into social network analysis and other emerging fields in animal behaviour. To make this a useful exercise, I’ll attempt to put together a how-to guide for using them, aimed at researchers with a mixed range of skills in manipulating numbers, and where time permits, I’ll try and add in some practice examples. How well this works depends upon both my own understanding, the time I have available, and the limitations of inputting maths into a WordPress blog!

What I won’t be doing (at least, initially) is being particularly critical about which techniques work best: this is a voyage of discovery for me too! I also won’t be focussing on what dominance is for, why it exists, and how it does or doesn’t drive particular group behaviours (but I do discuss how leadership decisions don’t necessarily depend upon the hierarchy present in Rands et al. 2008 and Rands 2011). This series of blog postings will take a little time to put together (An index page will give detailed links to other metrics within this blog), so if you’re looking for general advice on the sort of indices that are out there, I strongly recommend hunting down a copy of the excellent book by Hal Whitehead (pages 186-195 in particular).

Further reading

  • Rands SA, Cowlishaw G, Pettifor RA, Rowcliffe JM & Johnstone RA (2008). The emergence of leaders and followers in foraging pairs when the qualities of individuals differ. BMC Evolutionary Biology 8: article 51 | abstract | pdf | full text
  • Rands SA (2011). The effects of dominance on leadership and energetic gain: a dynamic game between pairs of social foragers. PLoS Computational Biology 7: e1002252 | full text | pdf
  • Whitehead H (2008). Analyzing animal societies: quantitative methods for vertebrate social analysis. Chicago: University of Chicago Press

How do pollinators respond to the shape of agricultural landscapes?

beeThe global debate rumbles on about pollinator decline. In the UK, the recent European Commission directive banning neonicotinoid pesticides has at least partly been a catalyst for some very public debate on why decline is happening, and what could be done about it (with the BBC rushing out a nicely-balanced edition of their science programme Horizon, exploring a few of the factors that may be driving the disappearance of pollinators).

This posting ties in with my talk at INTECOL 2013 in London (if you’re there, it’s in the Ecosystem Services session in Capital Suite 13 on Wednesday 21st at 2.15pm).

Aside from disease and poisoning, one factor that is frequently pointed to is the huge changes that have been made to the landscape in recent years.  The intensification of agriculture has meant that the ‘wild’ bits of the landscape have been taken away through changes in field management, and the steady creep of urbanisation.  These wild bits, even if they’re simply hedgerows and the other untidy bits at the edges of fields, are hugely important for providing nesting sites, refuge and food for wild pollinators and the other beasties that contribute to making agricultural systems work.  If we take these messy little spaces away, not only do we remove the resources that these beneficial species use, but we also make it much more difficult for those existing beneficial species already present to gain access to the parts of our managed agricultural species that are not close to these refuge areas.

Working with Heather Whitney (University of Bristol), I’ve done some work looking at how the shape of the agricultural environment affects the ability of pollinators to access it. In a paper published in Ecological Modelling, we considered a simple case where the environment was considered to be a square grid of hedgerows, with pollinators nesting in the hedgerows. The pollinators were considered to only fly a set distance from their nest (realistic, since many solitary bees fly a maximum  of about a kilometre from their nest), and the model demonstrated that if this distance was small, and kind of environmental manipulation that increased the size of fields beyond a certain point may have a detrimental effect upon the amount of wild space available to a pollinator.

However, the model was extremely simplistic.  Although I believe very strongly in keeping exploratory models as simple as possible, it felt like there was too many rigid assumptions made when we assumed that the landscape was a square grid.  In order to make the landscapes more realistic, we took two approaches: firstly, simulating random landscapes filled with hedgerows, and secondly, using landscape data from the UK, where there is a large amount of variation in wild refuge space within the landscape, as you can see from the four sample landscapes given below.

Examples of British field structures used within the model
Examples of British field structures used within the model

These landscape-informed models, published in PLoS One, demonstrated again that pollinators that only fly short distances from their nest (less than about 125 metres, which is relevant for some solitary bees such as Andrena hattorfiana) are affected heavily by landscape manipulations, but are unlikely to benefit from having wild land added to the environment unless it is targetted specifically for them (the equivalent of trying to help an isolated island community by building a new hospital for them on the mainland.  For species travelling more than 125 metres, adding wild space into the (British) landscape is a good thing, regardless of the exact distance the species travels.

So, we should maybe consider how far specific pollinators are able to travel when we are considering their conservation.  Lots of work is being conducted by  research groups across the world to quantify and observe the lengths of these commuting distances, and many research teams are finding that pollinators are thriving in response to many unexpected resources such as urban gardens. We still have a lot of work to do to explore how different species choose to move through the environment, and how this can be manipulated to benefit them and us.

Further reading

Rands SA & Whitney HM (2010). Effects of pollinator density-dependent preferences on field margin pollination in the midst of agricultural monocultures: a modelling approach. Ecological Modelling 221: 1310-1316 | abstract | pdf (postprint version)

Rands SA & Whitney HM (2011). Field margins, foraging distances and their impacts on nesting pollinator success. PLoS One 6: e25971 | full text | pdf

Mission statement

I’ve been thinking of opening up a blog for a few years now, but have constantly been held back due to a fairly standard professional uneasiness with the prospect of opening up candid and uncensored opinions to the general reader.  We (at least, the scientific community that I most often interact with, who may now howl back in denial) like to have our ideas poked, prodded and checked over multiple times before we unleash them on the world.

However, having been playing on the web since some time around Mosaic 1.0, and being a ‘passive user’ of usergroups, blogs, and countless other social content, it does feel strange not to have anything other than a static presence in a few poorly designed personal pages in a dusty corner of an academic server. What we often forget is that it’s fun to talk about the things that we’re passionate about.  We all have our own world views, so why not take the opportunity to unleash our own skewed take on what’s important, or interesting, or just plain strange? Social media gives us the opportunity to do just this. In a recent issue of Ideas in Ecology and Evolution, Bertram and Katti give some very succinct arguments for why the social media is important for scientists, both as a work tool and as a means of communication. If you’re a ‘passive user’, like I have been, read the article, then decide how to progress.

So, expect irregular postings about random things, all lumped together under the common theme of ‘interaction biology’. I realise this vague phrase could be seen as being close to ‘systems biology’ in its overarching vagueness, but I promise I have a very clear idea of what it means to me (honest!). As well as blatant self-advertisement of my group and my colleagues, I’ll be looking at papers and trends that are catching my eye, talking about (well-formed) ideas as they progress, become over-enthusiastic at conferences, and even, just very occasionally, rant against the system. I hope to get my students and colleagues to contribute as well (Sarah Giles, currently in the third year of her PhD, has already dipped her toes into the blogosphere during her recent three month policy internship at the Royal Society). There may even be some dancing weasels, if you’re good!

Further reading

Bertram SM & Katti M (2013). The social biology professor: effective strategies for social media engagement. Ideas in Ecology and Evolution 6: 22-31. doi:10.4033/iee.2013.6.5.f