Marine heat wave and multiple stressors tip bull kelp forest to sea urchin barrens

Extreme climatic events have recently impacted marine ecosystems around the world, including foundation species such as corals and kelps. Here, we describe the rapid climate-driven catastrophic shift in 2014 from a previously robust kelp forest to unproductive large scale urchin barrens in northern California. Bull kelp canopy was reduced by >90% along more than 350 km of coastline. Twenty years of kelp ecosystem surveys reveal the timing and magnitude of events, including mass mortalities of sea stars (2013-), intense ocean warming (2014–2017), and sea urchin barrens (2015-). Multiple stressors led to the unprecedented and long-lasting decline of the kelp forest. Kelp deforestation triggered mass (80%) abalone mortality (2017) resulting in the closure in 2018 of the recreational abalone fishery worth an estimated $44 M and the collapse of the north coast commercial red sea urchin fishery (2015-) worth $3 M. Key questions remain such as the relative roles of ocean warming and sea star disease in the massive purple sea urchin population increase. Science and policy will need to partner to better understand drivers, build climate-resilient fisheries and kelp forest recovery strategies in order to restore essential kelp forest ecosystem services.

Temperate kelp forests in northern California (Fig. 1) were particularly vulnerable to the MHW and other concurrent ecological stressors. This region, which was historically very productive, supported robust fisheries including the recreational red abalone, Haliotis rufescens, fishery (valued at $44 M yr −1 36 ) as well as the commercial red sea urchin, Mesocentrotus franciscanus, fishery (valued at $3 M yr −1 ). The bull kelp forests in this region (>350 km) were the first along the west coast of North America to show severe impacts to kelp productivity. The long-term kelp forest monitoring program was critical for tracking and understanding the biological responses to these multiple climate-related stressors and resulting degradation of fisheries and other ecosystem services 37 . Similar impacts seem to be developing in kelp forests from Baja California to Alaska (personal communications), so that the dynamics described from this northern California case study will be critical for tracking and understanding the biological responses to these multiple climate-related stressors and resulting degradation of fisheries and other ecosystem services 37 .

Results
The region north of San Francisco to the Oregon border ( Fig. 1) historically supported extensive, nearly pristine, productive, and persistent bull kelp, Nereocystis luetkeana, forests 39 . Human population densities and development are low in the region, so no abrupt anthropogenic impacts to ocean conditions and ecosystem health were anticipated. A series of perturbations 40 including a loss of sea star predators of urchins 41 , prolonged warm-water conditions, and a population explosion of purple sea urchins occurred prior to and concurrently with an abrupt shift from bull kelp forest to persistent urchin barrens (Fig. 2).
Bull kelp. Bull kelp canopy area declined dramatically in 2014 (Fig. 3) throughout the historically-persistent region of bull kelp forest (>350 km of coastline) in northern California. Maximum historic extent of kelp canopy (available data: 1999, 2002, 2003, 2004 and 2008) in the region exceeded 50 km 2 , with a range of 2.4 to 14.9 km 2 observed in any given year. Nearly 95% of the historic kelp canopy area was observed in Sonoma and Mendocino counties, a 250 km region of coastline dominated by contiguous rocky reef habitat. Bull kelp forests continued to be productive in 2009-2013, growing extensive thick beds throughout Sonoma and Mendocino counties ( Fig. 2a; personal observation). In 2014-2016, the kelp canopy area declined to <2 km 2 , with no appreciable recovery observed in the core region of the kelp forest in 2017-2019 (personal observation).  (Fig. 4c). Starting in 2015, the purple sea urchins shifted to a more aggressive feeding behavior associated with food limited urchin barren conditions, grazing down stipes of subcanopy kelps and fleshy algae (Fig. 2e), then grazing through the calcified crustose coralline algal cover (Fig. 2f). Since 2015, purple urchin densities have continued to increase at most of the sites (2018 range: 9.2-24.1 urchins m −2 ).

Discussion
A combination of large-scale environmental and ecological stressors led to dramatically reduced bull kelp canopy in northern California, starting in 2014. Climate-driven impacts of warm-water, including thermal stress and nutrient limitation, associated with the MHW suppressed bull kelp growth (and spore production) during the summer of 2014. These climate-driven impacts persisted for multiple years, and were exacerbated by a strong ecological impact of moderate sea urchin herbivory starting in 2014 and becoming intense in 2015-present. From field observations during subtidal monitoring work, we know that kelp was abundant prior to the impacts in 2014. The continued low bull kelp abundance after 2014 is likely due to the combination of unfavorable environmental conditions (warm water and low nutrients), intensive urchin grazing pressure, and limited spore availability due to multiple years of low production of this annual species.
Starting in 2014, sea urchin populations began to increase to higher densities than previously observed in the region. Populations increased at many sites to more than 30 times historic numbers by 2015, and have continued to increase. Despite widespread starvation conditions, spawning adults of purple urchins have been observed even at sites devoid of macroalgae, and young of the year (<20 mm) are abundant throughout the region. It is unknown if there was a primary driver of the urchin population increase, or if both top-down (sea star predation) and bottom-up recruitment of purple sea urchin processes were responsible. Similarly, the driver(s) of SSWS which led to the local extinction of the Sunflower star is unknown. The first observations of SSWS in the region were recorded during cold-water conditions in the summer of 2013, suggesting that this mass mortality was not initially driven by changes in ocean climate, however warm-water conditions may have later exacerbated the mortalities 44 . www.nature.com/scientificreports www.nature.com/scientificreports/ The large-scale ecosystem stressors leading to urchin barrens in northern California illustrates the vulnerability of our ecosystems and communities to climate-driven collapses. Many kelp forest ecosystem services have collapsed on a large scale throughout the region, with particularly severe economic impacts due to collapsed fisheries, kelp harvest, tourism opportunities, and loss of cultural resources. The northern California recreational red abalone fishery was the largest in the world, with 35,000 fishers landing 245,000 abalone (292 mt) yr −1 36 , however the California and Oregon fisheries were closed in 2018 due to abalone mass mortalities. Widespread abalone starvation and mortality was observed in the wild (Fig. 4d). From previous laboratory experiments, we showed that starvation conditions alone will impact red abalone health and reproduction, which will be exacerbated with warm water 46 . Similarly, the commercial red sea urchin fishery has collapsed due to starvation conditions leading to poor gonad production and unmarketable sea urchins. Small remnant kelp patches (<5%) observed since 2014 are not as capable of promoting kelp recruitment as intact kelp forests 47 . Further, this ecosystem shift to urchin barrens may persist as sea urchins can thrive in low food conditions on dissolved organics as both larvae 48 and adults 49 suggesting urchins barrens could be an alternative stable state.
Even if kelps recover from these multiple stressors, it may take decades before the complex biological communities, associates, and the ecosystem services provided by macroalgal forests (Table 1) rebound as has been www.nature.com/scientificreports www.nature.com/scientificreports/ observed in other parts of the world [50][51][52] . While the red sea urchin fishery may take only a few months to rebound after kelp recovery, red abalone populations have declined so low that population recovery will likely take decades after kelp populations recover. A host of economically important non-consumptive recreational opportunities, including scuba diving, kayaking, and nature photography, may also impact tourism as the broader nearshore kelp associated community slowly recovers ( Table 1).
The documented severe loss of kelp in northern California, starting in 2014, is remarkable because of the scale (>300 km), magnitude (>90%), and speed (within one year) of the impact in an area of historically persistent kelp forests. The severity of on-going ecological and economic consequences underscores the need to investigate the climate impacts and interactions of multiple stressors influencing the vulnerability of ecosystems, even in regions that are relatively pristine (minimal anthropogenic impacts). Identifying the relative impact of individual stressors on a natural system is frequently not possible with observational data alone, particularly when multiple stressors co-occurred or occurred in a rapid sequence. Here, we draw on the long time series of monitoring work and experience with the system, ecological knowledge and theory for kelp forest ecosystems to elucidate the timing of the strongest known stressors in the system.
Given the loss of ecosystem services associated with the shift to an unproductive alternative state, it is important to understand the perturbations that disrupted the marine ecosystem 53 and its ability to rebound from perturbations (resilience to phase shifts) 4,54 . Identifying the relative importance of factors influencing climate vulnerability is the focus of ongoing research and will be critical for informing recovery potential. A plan for bull kelp recovery in northern California, developed in 2018-2019 with broad scientist and stakeholder input, identifies actionable recovery strategies aimed at enhancing ecological understanding of the drivers that will inform climate ready restoration actions and build resilience for the future 55 .
Science-based management action plans must be initiated to bolster resilience in vulnerable and impacted ecosystems 51 as all indications are the urchin barrens will persist. In the future, MHW are predicted to continue 56 increasing in frequency and intensity globally 31 with the NE Pacific a regional hot-spot 34 . This threat provides strong motivation for developing climate-ready action plans to increase ecosystem resilience to major climate stressors and identify recovery bright spots 57,58 . Such plans should focus on tracking resilience such as within favorable microclimates 59 , enhancing recovery of ecosystem engineers and keystone species, as well as identifying opportunities for economic incentives to support climate resiliency. For kelp forests in California, solutions may include developing economic opportunities to reduce urchin grazing pressure by supporting emerging purple sea urchin restorative fisheries and shifting away from fisheries being the sole support for ecosystem monitoring. Climate-ready resource management will require garnering support and building broad partnerships between science, industry and nonprofits, to develop new monitoring and restoration approaches that enhance resilience of foundational species and their ecosystem services into the future. Methods northern california region. We present monitoring data from the nearshore kelp forest ecosystem at sites in rocky subtidal habitats in northern California (San Francisco to the Oregon border), with particular focus on Sonoma and Mendocino counties, from 2003-2018 (Fig. 1). Kelp communities in this region are on rocky reefs dominated by bull kelp, Nereocystis luetkeana (Fig. 2a). The understory is comprised of short fleshy red and crustose coralline algae as well as subcanopy kelps, such as Pterygophora and Laminaria (Fig. 2b). These subtidal rocky reefs in northern California support a diverse assemblage of macroalgae and marine invertebrates.
Kelp canopy cover. Total kelp surface canopy area was assessed in 2008, 2014-2016 by aerial surveys from San Francisco to the Oregon border. Kelp canopy was quantified using low-flying aircraft to photographically survey the nearshore coastline. Cameras were mounted on the aircraft to capture the images. Image frames were auto-georeferenced using customized software, and manually shifted as needed. ERDAS IMAGINE software was used to mosaic the frames and run them through classification in ERDAS IMAGINE. Maximum extent of the These data are used to detect the magnitude of the temperature and the frequency and duration of exceedance above 12 °C, an important metric for bull kelp growth as NO 3 concentrations are low at this temperature and warmer 42 .
Subtidal scuba surveys. The nearshore kelp forest ecosystem monitoring program 60 conducts scuba surveys at sites in rocky subtidal habitats along Sonoma and Mendocino counties in northern California. These surveys of the nearshore rocky reefs were initiated in 1999 and allowed for photographic documentation of communities before and after multiple stressors impacted the region. Subtidal surveys were conducted by the California Department of Fish and Wildlife (CDFW) by motor boat at twelve sites along the Sonoma and Mendocino county coasts. The sites ranged in coastal length from 2.4 to 3.2 km. The sites in Sonoma County from south to north include: Fort Ross, Timber Cove, Ocean Cove, Salt Point, and Sea Ranch. In Mendocino County the sites from south to north include: Point Arena, Albion, Van Damme, Russian Gulch, Point Cabrillo (State Marine Reserve), Caspar Cove and Todd's Point (Fig. 1). The surveys are conducted along band transects 30 × 2 m, located randomly within four depth strata (random stratified) from 1 to 20 m depths. The density estimation for each species is determined by averaging the densities within each of the four depth strata and then calculating the average of the four densities from each depth. The error bars represent standard error of the mean densities across four depth strata. The sites are surveyed to enumerate abalone, sea urchins, sea stars, macro-invertebrate densities as well as percent cover of algae and substrate type. All size classes observed are recorded. At each site 15-55 transects were surveyed. Transects were located in areas with >50% rocky reef.
permissions for protected areas. Underwater surveys were conducted inside two marine protected areas.

Data availability
The data that support the findings of this study are available from the corresponding author upon request. Published: xx xx xxxx