Figure 1.
The Baltic Sea with the study area, the Baltic Proper (dark).
Figure 2.
The structure of the food-web model, also indicating the fishing pressure (F) for the respective fisheries on the three fish species, pp – primary producers, juv – juvenile stanza of given fish species.
Detritus pool is divided into two groups: detritus on the sediment (detritus (s)) and water column detritus (detritus (w)).
Figure 3.
Ecological indicators and ENA indices anomalies (note different scale) from 1974–2006.
Figure 4.
Results of the Principal Component Analyses, with the first and second principal component (PC1 and PC2).
The first column shows the dependencies between variable – (A and C) (for detailed values see Tab. S4 in File S1) the second column shows temporal trend PC1 and PC2 axis scores (B and D). Rows show the results of analyses of data sets including: model forcing (A and B), and modeled biomass (C and D), respectively. Vertical lines on PC components time trajectory represents shifts tested by the regime shift analysis. Please note that the scale differs between axes.
Figure 5.
Traffic light plots representing the applied forcing and development of the simulated biomass from the different groups.
The time series were transformed into quintiles and sorted according to the PC1 axis scores: (A) model forcing; (B) modeled biomass.
Table 1.
Timing of shifts detected using STARS, given time series of model forcing variables.
Table 2.
Shifts in PC1 index detected by STARS.
Table 3.
Shifts in given data sets detected by Chronological Clustering.
Table 4.
Timing of shifts detected using STARS, given time series of modeled biomass.
Table 5.
Timing of shifts detected using STARS, given time series of indices.
Table 6.
Coefficient of variation of used indices for given time period (regime).
Figure 6.
Time dynamics of redundancy (R) as percentage of capacity (C) in black and the red line represents the regime tested by the regime shift analysis for the period 1974–2006.
Figure 7.
The redundancy (R) versus the overall pressure index, which is the principal component 1 from the model forcing variables.