Table 1.
Summary table of studies (n = 22) examining change in fish abundance before and after cyclone.
Fig 1.
Major hurricane tracks over coastal North Carolina from 2010–2020.
Tracks are based on best track data from HURDAT2 [54]. Cyclone symbology represents hurricane location and Saffir Simpson category at 6-hour intervals (TD = tropical depression, TS = tropical storm). Inset map depicts study area with green circles indicating trawl survey locations and blue triangles indicating seagrass survey locations. Map sources: Esri, HERE, Garmin, SafeGraph, Meti/NASA, USGS, EPA, NPS, USDA.
Table 2.
North Carolina tropical cyclone and control year metrics 2010–2020.
Fig 2.
mBACI of short-term fish catches, species richness, and community structure across time periods and year type.
Only means are presented for short-term comparisons of catch per unit effort (A); catch per unit effort calculated sans L. rhomboides (B), and species richness (C). P-values indicate the significance of the ANOVA interaction term. Error bars represent standard error. Non-metric multidimensional scaling (D) of short-term seagrass-associated fish communities based on fourth-root Bray-Curtis extended distances of abundance. Gray coloration and circles indicate control years; blue coloration and triangles indicate impact years. Open symbols indicate before periods and filled symbols indicate after periods. Environmental correlates (p<0.1) are plotted as vectors in the direction of ordination influence. Orange symbols indicate group centroids.
Fig 3.
mBACI of seasonal fish catches, species richness, and community structure across time periods and year type.
Means (A-C), trend (D-F), and difference between control and impact trends (G-I) are depicted for seasonal comparisons. Confidence intervals and error bars represent 95% confidence and standard errors, respectively. Smoothed lines represent generalized additive models across the seasonal time frame (y ~ s(Days to Storm), k = 3) for both hurricane and storm-free years based on a cubic regression spline with shrinkage. Non-metric multidimensional scaling of seasonal seagrass-associated fish communities (J) are based on fourth-root Bray-Curtis extended distances of abundance. Gray coloration and circles indicate control years; blue coloration and triangles indicate impact years. Open symbols indicate before periods and filled symbols indicate after periods. Environmental correlates (p<0.1) are plotted as vectors in the direction of ordination influence. Orange symbols indicate group centroids.
Fig 4.
Change in fish metrics as a function of cyclone (tropical storms and hurricanes) intensity.
Short-term changes are depicted column 1, and seasonal-scale changes are represented in column 2. A & D) Proportional change in catch per unit effort. B & E) Proportional change in catch per unit effort sans L. rhomboides, and C & F) change in raw species richness. Models of CPUE and CPUE-Lr represent a tobit regression censored at -100%, and linear regressions are plotted for change in raw species richness. Model shading indicates 95% confidence intervals also censored at -100% for catch metrics. Point coloration (light blue to dark blue) indicates increasing month of impact.
Fig 5.
Trends in seagrass percent cover by month and BACI treatment for individual meadows.
Gray coloration and circles indicate control year treatments; blue coloration and triangles indicate hurricane years. Open symbols indicate measurements taken before hurricanes, and filled symbols indicate measurements taken after storms.