Tutorial: Mass-balance
We’ll use the Anchovy Bay model, but special versions that have been made unbalanced by changing one or a few things. In total, the tutorial has five models, each of which you should try to balance on your own. Download the Anchovy Bay mass balance models.zip file from this link.
The errors that make the models unbalanced are very typical errors often made during model construction. First and foremost among such are unit conversion errors. Always be very careful with units!
Mass-balance cheat sheet
There is a Mass balance cheat sheet.pdf file included in the zip-file download, but don’t just read that file, it’s for checking what you did afterwards. Still, here’s the content of the cheat sheet.
- Seal biomass was increased an order of magnitude, and this resulted in too high predation mortality on seals’ prey, cod and whiting.
- The Q/B for cod was increased from 2.58 year-1 to 12.58 year-1. This made the consumption of whiting so high that this group wouldn’t balance, and it’s clear from the predation mortalities (M2) that the cod predation is too high, M2 exceeds the P/B. For anchovy, the increase makes cod the most important predator on anchovy. That’s not credible, mackerel should be more important.
Check the P/Q ratio for cod: it is very low, and this points to a problem with either the P/B or Q/B for cod. The P/B is reasonable (average life span of cod ~ 3 years), but the Q/B is not. - The Q/B for seal was to be estimated – which at first glance seems feasible as B, P/B and EE are given for the group. There is however a cycle, cod > whiting > cod, so when the consumption of cod isn’t known, it becomes impossible to calculate the EE for whiting, and without the EE, predation on whiting isn’t quantified, so we cannot estimate Q/B for cod. In this case the cycle leads to Catch 22.
Main lesson: stick to estimating EE or if needed B. There is no good reason for estimating P/B or Q/B based on mass balance: we have expectations for those parameters. - Here, the model has been changed to estimate biomasses for all but marine mammals. No errors when doing basic parametrization, so all good? No, look through the estimated biomasses, they are quite high, and the primary production (B x PB = 45 t km-2 x 120 year-1 = 5400 t km-2 year-1 makes Anchovy Bay a very productive system, almost twice as productive as the North Sea. That’s not right. The warning is: if you have nothing to constrain the model from below, you risk “blowing up” your model. One good primary production estimate would have made that clear, or, as we did here, a comparison to other areas.
- Problem: EE is high for cod and anchovy
- Check Mortalities: it is because of predation mortality
- Check Predation Mortality Rates: It is due to high predation by whiting on cod and anchovy.
Also notice the high cannibalism mortality rate for whiting. - Whiting P/Q ratio is 0.187
- So, for each unit that whiting eats, they produce 0.187
- But cannibalism is 0.15 – that is 80% of the whiting production
- So, to meet other demands for production (i.e. predation and catches), Ecopath
increases the biomass for whiting until the “consumption pie” is big enough to meet
all demands - This in turn causes the high EE for cod and anchovy
- The best solution here is to lower the cannibalism rate for whiting; go to diet
composition and lower it by 0.1, e.g., to 0.05, then increase the whiting diet for benthos by 0.1.
Main lesson: Be careful with cannibalism, it’s often because (slow turnover) old fish eat some (fast turnover) young fish, and the model is parameterized based on the older fish. Breaking it into stanza (can be done in Excel), the calculated weighted parameters will usually solve the cannibalism dilemma.
