Contact Author: Thomas Lamy ()

Electronic supplementary material (ESM)forLamy T, Reed DC, Rassweiler A, Siegel DA, Kui L, Bell TW, Simons RD, Robert J. Miller RJ. Scale-specific drivers of kelp forest communities.

contact author: Thomas Lamy ()

1

Table S1. List of the 45 fish species and 29invertebrate and algal species surveyed across the four monitoring programs.We provide information on the sampling method used to survey each species across the four monitoring programs: the National Park Service Kelp Forest Monitoring program (KFM), the Partnership for Interdisciplinary Studies of Coastal Oceans (PISCO), the Santa Barbara Coastal Long Term Ecological Research Project (LTER), and the San Nicolas Island monitoring program (SNI). KFM surveyed fish communities based on a 30-minutes random dive within a permanent 100 m x 20 m transect. PISCO surveyed fish communities in eight to sixteen randomly positioned 30 m ×2 m transects per site. The SBC-LTER surveyed fish communities in two to eight 40 m × 2 m permanent transects, while SNI surveyed fish communities in one to two 50 m × 2 m transects per site.

1

Table S2. Multiscale analysis for each year for fish assemblages (left table) and for benthic assemblages of invertebrates and algae (right table). The broad spatial scale (BS)explained on average 27% of the total variation in fish assemblages and 19% of the total variation in the benthic assemblages of invertebrates and algae. The medium scale (MS) explained on average 8% and 14%, while the small scale (SS) explained an additional 10% and 12% of the total variation in fish assemblages and in invertebrate and algal assemblages respectively. All of these spatial effects are significant (P < 0.05) across years. The table provides the average contribution of an environmental variable to the variation inassemblage variation at a given spatial scale. The total contribution of a variable is decomposed into a unique contribution and a total contribution. See fig. 3 for comparison. P refers to the number of years for which this contribution is significant (P < 0.05).

1

1

Table S3.Partitioning the relationship between each environmental variable and the variation in fish assemblages (upper table) and to the variation in benthic assemblages of invertebrates and algae (lower panel) at different spatial scales. The total contribution of each environmental variable is decomposed into an individual fraction and a fraction that is shared with the other variables. The contribution of each fraction is computed as adujted-R2and its significance istested using a ANOVA-like permutation test. d.f. is the number of degrees of freedom and F stands for the pseudo-F values. See also fig. 2.***: P < 0.001; **: 0.01 < P < 0.001; * 0.05 < P < 0.01.

1

Fish assemblages

1

Benthic assemblages of invertebrates and algae

1

Fig. S1. Space-time surveys of the fish, invertebrate and algal species in the Santa Barbara Channel. 86 sites were surveyed for a ten-year period, between 2005 and 2014. While most sites were annually surveyed (green dots), some of them were not in some particular years (red dots).

1

Fig. S2.Distribution of the 135 ROMS connectivity cells in the entire Southern California Bight, that were used to quantify potential propagule delivery strength. Black points represent the 86 sites used for community survey.

1

Fig.S3. Small scale of variation in fish assemblages(a-b) and in benthic assemblages of invertebrates and algae (c-d).Each panel represent the variability along thefirst (a and c) and second (b-d) significant RDA axes of the small scale of variation in fish assemblages and of the small scale of variation in invertebrate and algal assemblages respectively.

1