Demography and dynamics of Giant Kelp cohorts across four decades: Lessons for conservation and resilience planning
Data files
Nov 24, 2025 version files 596.53 KB
-
AlgaeBoxData.RData
388.28 KB
-
README.md
3.96 KB
-
StipeData.RData
204.30 KB
Abstract
Kelp forests throughout many temperate zones are in decline due to various human stressors, chiefly marine warming. Conservation measures, including restoration, are presently of great interest and focus on both historical and novel methodologies. Of paramount importance for these efforts is an understanding of the mechanics of kelp decline to identify the factors and triggers leading to stepwise declines and thus support the development and spatial priorities of strategic intervention to facilitate resilience. Here, we utilized a unique dataset documenting the demographic dynamics of giant kelp, Macrocystis pyrifera, in response to multiple disturbances across >40 years off San Diego (California, USA). The recruitment and life history of >14,000 individuals were used to evaluate cohort structure, size, and longevity forced by algal community structure and disturbance. Cohort dynamics varied spatially by depth and study subregion, thus aiding the identification of areas to prioritize for intervention to foster resilience. Five algal assemblages were characterized, providing context for cohort dynamics in response to physical disturbances and sea urchin grazing. A trend of decreasing cohort size and resilience was observed over time, accentuated by the marine heat wave of 2014-15 (MHW), after which competition with understory canopies increasingly interfered with giant kelp cohort development and plant size structure. Cohort recruitment ranged on a continuum from discrete (‘pulsed’) to more gradual (‘trickled’) episodes. Pulsed cohorts mainly produced single cohort-dominated age stands punctuated by major disturbances. Pulsed events were more common than trickled recruitment, especially at deeper sites. Trickled cohorts resulted in relatively mixed age stands, especially when individual cohorts overlapped within sites. Trickled recruitment increased over time as understory dominance increased. Cohort longevity was highly variable among sites and among cohorts within a site, with high first-year mortality mostly due to warming, waves, or their combination. Longevity was inversely related to temperature and sea urchin density, and was greatest at deeper sites, especially after the MHW. The downward trend of single cohort dominance and individual plant size over time and its step downward after the MHW suggest that deeper areas should be prioritized for restoration. Regardless, understory canopies will increasingly dominate Southern California with continued warming.
Dataset DOI: 10.5061/dryad.fttdz096d
Description of the data and file structure
This dataset includes algal data collected across 20 sites off San Diego, CA, USA. The data were collected from 1984 to 2024 at depths between 8 and 21 meters along permanent leaded transect lines. The transect lines are described in the article text. Each site consists of 4 leaded lines (A, B, C, D) that are 25 meters long and 4 meters wide. Data are recorded in 10 quadrats (sampling units) that are 5 meters along the line and 2 meters on each side of the line. Therefore, 400 square meters are sampled at each site along 4 transects that are each 100 square meters in area. Transect lines are oriented perpendicular to the shore, and ‘A’ lines are located at the northern end, with lines ‘B’, ‘C’, and ‘D’ located progressively to the south. The zero meter mark is located at the offshore end, while the 25 meter mark is onshore. Quadrats ‘boxes’ are numbered 1-10 with the left side consisting of odd numbers (1, 3, 5, 7, 9 [offshore to onshore]), and the right side consisting of even numbers (2, 4, 6, 8, 10 [offshore to onshore]). Therefore, quadrats 1 and 2 are located next to each other, as is true for all other odd/even quadrat pairs.
The data consists of 2 R data files (*.RData) as follows.
Files and variables
File: AlgaeBoxData.RData
Description:
This data file consists of one object ‘tl’ with 19 fields as described below:
Site = Name of study site (20 sites)
Bout = numerically indexed sampling bout
BoutDate = Sampling bout date
Line = Transect line name (A, B, C, or D)
Box = Quadrat number (1-10)
MacroAd = Number of adult Macrocystis pyrifera plants (adults with at least 4 stipes)
MacroPA = Number of Macrocystis pyrifera pre-adults (plants with 2 stipes at least 1 meter tall)
Stipes = Number of Macrocystis pyrifera stipes
PteryN = Number of Pterygophora californica plants
CystoN = Number of Cystoseira osmundacea plants
EisN = Number of Eisenia arborea plants
DictyP = Fractional cover of Dictyopteris undulata plants
DesP = Fractional cover of Desmarestia ligulata plants
ArtCorP = Fractional cover of Articulated Coralline Algae
AgN = Number of Agarum fimbriatum plants
BT = fractional cover of all other Brown Turf Algae
RT = fractional cover of fleshy Red Turf Algae
Lam = scaled cover of Laminaria farlowii (0 = none, 1 = trace cover such as bare stipes, 2 = 5 % cover, 3 = 10% cover, 4 = 25% cover, 5 = 50% cover, 6 = 75% cover , 7 = 90%, to 8 >90% cover)
MacRecs = Number of Macrocytis pyrifera recruited juveniles (identifiable plants with single blades and early bifurcation)
File: StipeData.RData
Description: Contains one object named ‘stipes’ with 8 fields. Each row indicates a plant was found alive, and the number of stipes was recorded. NOTE: Each plant was followed over time and was identified in the data with an ID name that was unique within each transect line over the entire time series.
Site = Name of study site (20 sites)
Date = Survey Date
Season = quarterly season (‘F’, ‘W’, ‘Sp’, ‘Su’)
year.season = combined year and season (e.g., ‘2006F’)
Line = Transect line name (A, B, C, or D)
Box = Quadrat number (1-10)
Name = unique plant name ID within a transect line over time
numStipes = Number of stipes counted on the plant at the time of the survey
Code/software
Data files are provided as R (open source) data files: R Core Team (2021). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/.
All analyses were conducted with standard packaged libraries within R as described in the manuscript.
