Table 2. Results of linear mixed effects models of water properties using initial and final samples for each drift at five sites in Washington State, USA, in summer 2014.

Each column contains a distinct water property: chlorophyll-a (Chl), and proportion organic. Each row provides F value (P value) for main effects of habitat and total suspended solids (TSS), for their interaction, and for planned contrasts between biogenic habitats (eelgrass, oyster) and bare when habitat was significant. Random effects were site, and subsite in site. Samples were not used in analysis unless 0 < proportion organic < 0.5. The interaction effect (habitat × ln(TSS)) was close to P = 0.05 in this mixed effects model, therefore interpreted cautiously.

	Response
	ln(Chl) n = 261	ln(proportion organic) n = 262
Habitat	24.2 (<0.0001)	0.10 (0.90)
ln(TSS)	57.3 (<0.0001)	146.1 (<0.0001)
Habitat ×ln(TSS)	3.16 (0.044)	0.49 (0.61)
Bare vs. Eelgrass	E > B P < 0.0001
Bare vs. Oyster	O = B P = 0.03