Development of the StrathE2E model

Published: 13 February 2017

By MERP scientist, Prof. Mike Heath


Dreamstime | Eric GevaertSince March 2016, we have been consolidating a raft of new developments into the StrathE2E model, some arising from improved understanding of ecological processes, and others to meet specific needs in MERP. At the easy-end of the scale, these include the addition of different types of sediment detritus, changes in the way in which food consumption by birds and mammals is represented, and in how birds and mammals respond to temperature. At the more difficult-end of the scale have been changes in how the model captures the essential aspects of bathymetry and spatial variation in seabed sediment properties and the addition of early-life-history stages for benthos categories. A new group of migratory fish has also been added to represent mackerel, which do not live in UK shelf waters all-year round but migrate in at certain times of year, like a raiding party, and then leave. The latter is an important addition because the mackerel fishery is by far the most valuable in the UK, so it is important that we represent its impacts.

Two further developments are work-in-progress. These are the addition of kelp to the model, and exploration of how to do a better job of representing the dynamics of top predators (birds and mammals). On the former, there are published models of individual kelp plant growth, but very few, if any, precedents for a kelp forest model are simple enough to be incorporated into such a broad-scope ecosystem model as StrathE2E. So, we have been working to develop two new models – first, a very detailed model of a kelp forest based on modelling of individual plants and how they compete and interact in a forest. Second, a much simpler bulk-forest model, which we want to behave like our very detailed individual plant-based model and to eventually incorporate into StrathE2E. We have been talking to our MERP colleagues in SAMS, QUB and PML about the ecology of kelp in the course of building these models, and look forward to developing some joint publications.

In relation to top predators, we have been working with our MERP colleagues in CEH, Bangor and Glasgow, to develop a new top-predator model, which resolves six coarse groups of birds, seals and cetaceans. For several of these, the idea of a seasonal ‘raiding party’ of some of these groups which visits our model domain to feed and/or breed, is central to the model design. We have a working code for the new model, which we are driving with prey data from a run of the existing StrathE2E for testing and development. Once we are happy with this, we can merge it back into the main ecosystem model.

The final area of development is a completely new step, which paves the way for some exciting new prospects in the newly-funded Module 7. This involves 2-way coupling of our StrathE2E model of marine ecology, to a separate model of the activity and properties of fishing fleets. This is necessary because, up to now, the external driving data for StrathE2E has included values for the harvesting and discard rate on each fished resource group in the model. We have typically conducted experiments in which we vary these harvest and discard rates independently and study the impact on the ecosystem. But in reality, these harvest and discard rates are not independent because they are generated by fleets of fishing gears, which do not catch only one group of fish or shellfish at a time, but each takes a by-catch from a range of groups. To solve this problem, we have built a new model of fishing fleets that represents up to 12 different gears, their catching, discarding and seabed scarping properties, and is driven by their individual activity rates. This model outputs the total harvest and discard rates on each StrathE2E resource group as a result of the combined activities of all gears.

The game-changing step was to develop a mechanism for 2-way coupling between the fleet and ecology models, in which we can run the ecology for a period of time e.g. a year, then pause and extract indices of the state of the ecosystem from the model results, and use these to adjust the activity rates of the fishing gears, mimicking the fishery management process. Then we can pipe the updated harvest and discard rates back to the ecology model and resume from where it was paused. This ‘model management’ scheme opens the way to linking our ecology model to a range of ‘human society’ models, to represent the dynamics of monetary and non-monetary valuations, and fisheries and environmental management rules, which affect the human activities impacting on the ecology.

Other news articles:

01 March 2017