Supplementary material

Figure S1. Water level change during winter, Δh (cm), and sliding mean (3 years) in regulated Tehinselkä and Ruotsalainen, unregulated Southern Konnevesi, and moderately regulated Puula. Negative Δh = a decrease in water level.

Figure S2. Time series of standardized ln(D + 1)-transformed whitefish larval density and corresponding linear trendlines in Päijänne Tehinselkä, Southern Konnevesi, and Puula in the years 2000–2022 and Ruotsalainen in the years 2008–2022. The linear trend was significant only in Puula (p = 0.019) (Tab. 2).

Table S1. Correlation between the time series from different study areas for the water level in spring (hm_y, on white background) and water level in autumn (h_{Nov15 y–1}, on gray background). Pearson correlation coefficients (r), risk levels (p) according to the one-tailed hypothesis (H1: positive correlation) and the number of observations (n). Statistically significant correlation coefficients are in bold.

Table S2. Correlation between the time series from different study areas for interannual variation in wintertime water level changes (Δh). Pearson correlation coefficients (r), risk levels (p) according to the one-tailed hypothesis (H1: positive correlation), and number of observations (n). Statistically significant correlation coefficients are in bold.

Table S3. Sampling dates by study area from 2000 to 2022.

Table S4. Whitefish larval sampling scheme in the littoral and pelagic zones of the study lakes. Samples were collected in 4 strata in the littoral zone and one pelagic stratum in two vertical sampling depths. In the sublittoral zone there were differences in depth limits: ¹⁾ Ruotsalainen ²⁾ Tehinselkä ³⁾ Southern Konnevesi ⁴⁾ Puula. In the strata with bottom depth > 2 m, the lower limit of whitefish distribution was assumed to be 2 m from the surface.

Table S5. The water volume of sampling strata in the study lakes. For all strata with a bottom depth ≥ 2 m, the volume of the 0–2 m layer for the surface was included in the stratum volume.

Table S6. Significant mean differences in the ln(D + 1)-transformed larval density time series between the years. 2-ANOVA, multiple comparison: Tukey HSD. Mean difference and risk levels (p) according to the two-tailed hypothesis (H1: average larval densities are different) and 95 % confidence interval (CI).

Table S7. Parameter estimates, standard errors of the mean (s.e.), and significance (p) for linear (3a) and compensatory Ricker (5a) models fitted to the Päijänne Tehinselkä basin whitefish larval density–spawning stock data in 2000–2022. α = constant survival, β = compensatory density dependent mortality. One-tailed H1: α, β > 0.

Table S8. Parameter estimates, standard errors of the mean (s.e.) and significance (p) for linear (3b) model including environmental factors fitted to Päijänne Tehinselkä basin larval density-spawning population data 2000–2022. φ1 = effect of wintertime water level change (Δh), φ1.1 = effect of water level in autumn (h_{Nov15 y–1}), φ1.2 = effect of water level on spring (hm_y) and φ2 = effect of vendace population, year y, age groups 1+ and older. One-tailed H1: α, β, φ1 > 0, φ2 < 0.

Table S9. Parameter estimates, standard errors of the mean (s.e.) and significance (p) for linear (3b) model including environmental factors fitted to Päijänne Tehinselkä basin larval density-spawning population data 2000–2022. φ1 = effect of wintertime water level change (Δh), φ1.1 = effect of water level in autumn (h_{Nov15 y–1}), φ1.2 = effect of water level on spring (hm_y) and φ2 = effect of vendace population, year y-1, age groups 1+ and older. One-tailed H1: α, β, φ1, φ2 > 0.

Table S10. AICc analysis for different combinations of model 5b including the constant survival α, the compensatory density-dependent mortality β (Ricker 1954) and different non-correlated environmental factors fitted to Päijänne Tehinselkä basin larval density in year y. Data from the years 2000–2022. X = variable included in the model, Δh = wintertime water level change, h_{Nov15 y–1} = water level in Autumn y–1, hm_y = minimum water level in spring y, Ven y-1 = vendace population (ages 1+ and older) index in year y-1 and Ven y = vendace population (ages 1+ and older) index in year y, AICc = Akaike (1973) Information Criterion corrected for small sample sizes, ΔAICc = AICc – min(AICc), and wi = Akaike weight, the probability that the model is the best among the whole set of considered models.

© EDP Sciences, 2024