FOOD WEB ANALYSIS
After the stable isotope analysis for macroinvertebrates and their two basal resources using mass spectrometers (CF/IRMS; Conflo II & Delta S, Finnigan MAT, Germany, analytical precision: ±0.1‰ both for δ13C and δ15N; Conflo III and Delta plus XP, Thermo Fisher Scientific, USA, analytical precision: ±0.1‰ for δ13C and ±0.2‰ for δ15N), we estimated the TP of each taxon at each site based on the following stable isotope mixing model (Okuda et al., 2017; Okuda et al., 2020):
f 1 + f 2 = 1 (2)
f 1δ13C1 +f 2δ13C2 + Δδ13Cef × (TP–1) = δ13Ctaxon (3)
f 1δ15N1 +f 2δ15N2 + Δδ15Nef × (TP–1) = δ15Ntaxon, (4)
where f 1 and f 2 denote the proportion of reliance of POM and EOM, which are regarded as basal resources of planktonic and benthic food webs at each coastal site, respectively. δR 1, δR 2, and δR taxon (R = 13C or15N) are carbon (δ13C) and nitrogen (δ15N) stable isotope ratios of POM, EOM, and each macroinvertebrate taxon, respectively. The isotope values were averaged for each taxon at each site to consider individual variation and for time-series samples of POMs and EOMs at each site to consider temporal variation, excluding POM data from February 2006 when temporary enrichment of its N15 occurred due to winter mixing-driven advection of profundal POM with a higher δ15N than primary consumers (Yoshimizu et al., 2008). Trophic enrichment factors, Δδ13Cefand Δδ15Nef, for macroinvertebrates were set to 0.31‰ and 2.49‰ for δ13C and δ15N, respectively, based on metadata for invertebrate body tissues (Caut et al., 2009).
Next, we calculated the i TP according to (Ishikawa et al., 2017). The i TP of a coastal macroinvertebrate community containingn taxa is expressed as follows:
i TP =\(\sum_{i=1}^{n}{(\text{TP}_{i}\times\ \frac{B_{i}}{B_{T}})}\) (5)
where TPi denotes the TP of taxon i , and Bi and BT denote the biomass of taxon i and the total biomass of the local community, respectively. Hence, i TP is a calculation of the average number of times x that assimilated biomass is transferred along TLs in a focal food web (i.e., i TP = x + 1). In the present study,i TP does not include primary producers (i.e., TP = 1), such as planktonic and benthic algae, and coastal predatory vertebrates, such as fish and birds, whose biomasses are difficult to estimate at the same spatial scale as benthic macroinvertebrates. During the winter sampling period, since most fish predators were absent or inactive in coastal areas due to cold WT (mean ± SD = 4.93°C ± 1.09°C, N = 33), we assume that top-down trophic cascading effects of fish predators were low in our focal communities.
Out of the 33 sites, 3 showed no macroinvertebrates, and 1 showed small biomass measurement of a few taxa that was insufficient for isotope analysis. For the other three muddy sites, since macroinvertebrate communities are dominated by hypoxia-tolerant deposit feeders (oligochaetes and chironomids) with δ13C and δ15N much lower than the two basal resources, thei TP could not be appropriately estimated, suggesting that these sites relied on methanotrophs in hypoxic sediment (Jones et al., 2008; Jones & Grey, 2011; Kiyashko et al., 2001). These communities were excluded from the i TP calculations. Biomass and abundance data were log-transformed for statistical analysis.