Project title: Knowledge integration and Management Strategy Evaluation modelling
Modelling the future of the Kimberley region
Ecopath with Ecosim (EwE)
EwE is used to characterise the trophic structure, ecosystem attributes and impact of fishing and climate change in the region. It consists of a number of modules. Ecopath mass-balance model accounts for trophic interactions among organisms at multiple trophic levels by describing matter and energy flows. Ecosim and Ecospace uses the mass-balanced model generated by Ecopath to simulate dynamical changes due to human activities (including fishing), climate change and other time varying processes, as well as the effects of different management options, such as controls on fishing effort and spatial closures, on both fished species and the trophic interactions in the Kimberley marine ecosystem. More details about these modules can be found at http://www.ecopath.org/. An extensive list of references, showcasing the use of EwE on a number of ecological applications can be found here.
So far, the project has focussed on defining EwE spatial domain, foodweb structure (species considered) and foodweb interactions which we describe below.
EwE model domain
The geographical extent of EwE model domain can be seen in Figure 1 and the rationale for the choice of the domain boundaries is included in Table 1. At the core of these choices lies a distinction between two scales: i) the ‘jurisdictional’ scale at which Management Strategies apply and ii) the ‘geophysical/ecological’ scale at which Model Specification and Development Scenarios apply. To properly evaluate the impact of actions taken at the ‘jurisdictional’ scale, processes at the ‘geophysical/ecological’ scale need to be accounted for, since these will affect, and potentially invalidate, management interventions.
Figure 1 shows the EwE model domain. The rationale for this choice is more easily understood by analysing both the bathymetry and main ocean currents as in Figure 2 below.
Defining EwE foodweb involves two main steps: i) a decision on what species should be included in the foodweb and ii) a characterisation of these species in terms of abundances, biological characteristics (growth, mortality, consumption, etc) and feeding relations. As for any modelling choice, the final foodweb needs to reconcile the two competing needs of accounting for ecosystem complexity and simplifying our analysis to make it manageable. Finding a workable compromise requires both scientific knowledge and modelling experience.
Step (i) needs to be carried out first and is commonly done by selecting a minimum number of functional groups which satisfactory describe the behaviour of the overall foodweb. Here, a functional group is defined as a number of species of comparable ecological or feeding behaviour which can be treated as ‘functionally’ similar. Which functional groups are selected is also determined by the purpose of the analysis as functional groups can at times include individual species of specific commercial or conservation interest. Table 2 shows the chosen functional groups and the species they include.
Once the functional groups and their composition have been chosen, step (ii) needs to characterise these in terms of abundances, biological characteristics and diets. This is a time-consuming effort, since this information is rarely at hand and needs to be sought from disparate sources, often hard to locate. The output of this process, as discussed above, is a snapshot of what we know about the system and the data gaps highlight what we do not know about the system.
Data gaps can be partly filled by expert knowledge and by mass balancing the model, that is by ensuring that basic energy and mass conservation laws and functional groups stability are respected when the model runs (mass balance will be carried out in the second year on the project). Remaining data gaps will provide management relevant information by highlighting what type of data collection/survey plan could provide the largest impact on management-relevant knowledge. In Table 2, for each functional group, we include the available information about biomass, production, consumption and Ecotrophic efficiency. Table 3 includes the qualitative diet composition which expresses feeding relation between functional groups. Both Table 2 and Table 3 will be updated as a result of EwE mass balancing and when more information becomes available (including from other WAMSI projects).
Finally, Table 4 shows the information sources of the main EwE input parameters, while Table 5 shows the geographical origin of the data sets.
Preliminary model mass balancing
As a first step in the second year of the project, we have already carried out a first run of the EwE model. At this stage, this should be considered as an attempt to gain a first impression of the quality of the foodweb input data.
We have used some of the Ecopath outputs (ecotrophic efficiencies) as a preliminary diagnostic of the equilibrium assumption of the Ecopath model (see Table 6). The ecotrophic efficiency (EE) is defined as the fraction of total production of one functional group that is consumed by other groups. As a result, EE values should be between 0 and 1. Given a functional group, EE=0 indicates that the group is neither consumed by another group nor exported outside the model domain. Conversely, EE˜1 indicates that the group is heavily consumed (either preyed, grazed and/or fished), which allows no individuals to reach old age. The whole range of EE values can be found in Nature. However, values of EE>1 clearly indicate that the model dynamics is unrealistic. In the next modelling stage, Ecopath will be mass-balanced by focussing first on those groups with EE>1.
The trophic levels, estimated by the Ecopath model from the weighted average of prey trophic levels, varied from 1.0 for primary producers, detritus and mangrove to 4.7 for top predators such as pelagic sharks, and other piscivorous fish (e.g. lizardfish). The results presented in Figure 3 from the trophic levels and links estimated by the model at this stage should be seen as a diagnostic representation of changes in biomass and energy flows of a steady state of the Kimberley. In the next stage, we will mass balance and calibrate the model and it is expected that trophic levels and trophic links could change in the final version of the model.