Fig. 4: Predicted relationships with air temperature during fishing for clam reburrowing, clam mortality, and predator activity.

Curves generated from applying a Loess fit (step = 1) to raw burrowing and mortality proportions, at 48 h, for the average air temperature during fishing for each experimental trial (n = 30 per temperature; see Supplementary Data 5 for source data), and to raw individual values of relative predator activity (i.e., predator activity index as in Fig. 3, relative to the maximum observed index value; n = 1 per temperature; see Supplementary Data 4 for source data). Shaded areas are 95% confidence bands.