Home

Secure areas - passwords and usernames required

Thames tideway project

Solent oyster transport

Research projects


Free access areas

Sequential Bayesian decision making simulation

Sequential Bayesian decision making tool


Present work

I am working with Turnpenny Horsfield Associates. The highest profile recent project has been modelling fish for the Swansea Tidal Lagoon proposal. Here's an example on the right.


click for MP4

Past projects

Here are a few examples of the types of project with which I have been involved...

 

Wake sorting and cooperative schools


dff
I have started researching the possibility that small oceanic fish which form very large schools might be creating turbulence to sort their prey. The idea is that the wake formed behind a fish may be adapted to push certain sizes of prey toward different zones behind the fish, which a following fish could exploit. It could have huge ramifications on understanding of why fish school and why it is best to keep the populations of these fish locally high. It looks promising but these are early days for this one. I've published a paper on this one, see pubs.


Native oyster larvae


Native oyster used to be common along the South coast of the UK. The population is now much reduced. I made these models to help with questions around the best way to manage the existing fisheries in the Solent. We experimented with simple behavioural rules which had significant impacts on the final settlement patterns.

some alternative text


Salmon smolt on the Severn

This project was part of the Strategic Environmental Assessment of tidal power schemes in the Severn in 2010. The main question addressed by these fish models was related to how many times it was likely that a migrating smolt would be drawn through the turbines. I have recently published a review paper summarising the methods used.

• Scientific references

 


 

Models of swimming larvae

 

Even very small (a few mm) larvae of aquatic animals swim. Some swim laterally, such as coral larvae, which swim toward the sounds of reefs. While others swim vertically to take advantage of tides around complex coastlines. These models can be very useful to understand environmental impact, fisheries management, or the benefits of marine protected areas. I have made many models of larvae swimming vertically in complex dynamic tidal systems. The picture on left is from a recent paper on Manila clam larvae around Poole Harbour.

• Scientific references

• Link to this video


 

Track analysis and tuna observations

 

Here is a screen shot of an application I developed to analyse the tracks of tuna from archival tags. Sometimes the tags were in the fish for over a year and each 4 mins they recorded depth, light and temperature. It is a challenge to locate the fish from these data.

• Scientific references


 

Theory of school formation

 

This modelling was extended from a simple grid based model of how gas aggregations formed. The basic concept is that particles move around randomly until they bump into each other when they join together to form a group. Then the group continues to move around like a singleton. When two groups meet, they combine, and so on. If you start with many singletons they eventually form one big group. That's the end point - the equilibrium, but - like life - the really interesting stuff happens in the dynamic zone away from equilibria. The graph here shows how groups of 2 or 3 build up and diminish as the model progesses from many groups of one to one group of many. This graph shows how perfectly the patterns from a dynamic model (red diamonds) fit a theoretical mathematical model (line).

 

• Scientific references



Thames tunnels


These models were made to examine the impacts of proposed new developments on the migration of small fish up the Thames. The fish use selective tidal stream transport to get up the river by sheltering near the banks when the tide is going out and moving into the middle when it's going upriver. There were a number of other challenges associated with keeping the model fish behaving like real fish.

model
fish


Simulated fish

 

click for MP4

These fish models are designed to investigate the way fish swim in schools and the balance of school fidelity over swimming efficiency. This is a frame from a video which is available to view.

• Link to video of these fish.


 

Cellular Automata used for modelling sandwaves and megaripples

 

These models are made one cell at a time, and when simple rules are applied complex dynamic and realistic shapes can appear. I added to the basic design by putting in complex current fields and multiple grain types, which led to some very realistic looking shapes - the ultimate was one which resembled chocolate chip ice cream melting - but it had a serious purpose in that it also resembled sand formations after dredging works.

• Link to this video


 

Pigeon tracks in Google Earth

 

We made a very small GPS tag for use with Manx shearwaters but tested it on pigeons, and this is an example of a track from one of those tests. The pigeon in this case is flying back  along the A34, although it appears to prefer the quieter lane alongside. Perhaps it likes to be near isolated leafy trees next to the smaller road, looking for potential food and protection.


 

Interfaces and theoretical models for learning track analysis

 

Here is an example of an application I made to contrast various theories relating to albatross movements. It was designed to be used by students.

 

Krill and whales

 

• Scientific references

 

I made an attempt to understand how whales maintain a high abundance of krill. Because evolutionary theory makes this a highly likely scenario, and the lack of a 'krill surplus' after most whales in the Southern Ocean were killed was a paradox which could potentially be answered through this type of work. I made a model of krill and whales which investigated if a risk aversion strategy by krill was consistent with current understanding of their life history. The model suggests that it is possible that whales force krill into 'grow fast, die young' habitat, when, left to their own devices, they would choose a 'grow slow, live long' habitat.

 
All these works were products of my own efforts and while I'm very pleased if you would like to copy these images, I'd be equally pleased if you'd give credit.
Jay Willis 2012