Extreme oceanographic forcing and coastal response due to the 2015–2016 El Niño

The El Niño-Southern Oscillation is the dominant mode of interannual climate variability across the Pacific Ocean basin, with influence on the global climate. The two end members of the cycle, El Niño and La Niña, force anomalous oceanographic conditions and coastal response along the Pacific margin, exposing many heavily populated regions to increased coastal flooding and erosion hazards. However, a quantitative record of coastal impacts is spatially limited and temporally restricted to only the most recent events. Here we report on the oceanographic forcing and coastal response of the 2015–2016 El Niño, one of the strongest of the last 145 years. We show that winter wave energy equalled or exceeded measured historical maxima across the US West Coast, corresponding to anomalously large beach erosion across the region. Shorelines in many areas retreated beyond previously measured landward extremes, particularly along the sediment-starved California coast.

T he El Niño-Southern Oscillation (ENSO) explains much of the interannual variability in sea surface temperature, sea-level pressure and atmospheric forcing across the equatorial Pacific, affecting global climate patterns 1 and economies 2 . For example, global economic losses associated with the extreme El Niño of [1982][1983] have been estimated at over US$11.5 billion 3 (in 2016 dollars), including significant losses along the coast. The extremes of ENSO oscillations, El Niño and La Niña, have been linked to elevated coastal hazards, particularly during boreal winter (December-February) for the Eastern North Pacific (for example, Hawaii, California and the Pacific Northwest (that is, Oregon and Washington 4,5 ) and Southwestern Pacific (for example, New Zealand 6 and Australia 7 ). El Niño events have also been associated with hazardous coastal conditions in Japan during the boreal fall 8 , greater frequency of tropical cyclone development in the Eastern Pacific 9 and rotational shifts of embayed beaches in Australia 10,11 . With seasonally elevated water levels, higher wave energy and southerly wave directional shifts common during El Niño, the North American west coast has historically experienced severe coastal erosion during El Niño winters, as reported during the 1982-1983, 1997-1998 and 2009-2010 events 4,5,12-16 .
By various metrics, the 2015-2016 El Niño winter was one of the three strongest events in the historical record 17 . For example, in the boreal winter of 2015-2016 the Oceanic Niño Index, a 3-month running mean of sea surface temperatures in the eastern tropical Pacific 18 , reached the highest value in its 66-year history (Fig. 1a). Based on a reconstruction that dates back to 1871 for the multivariate ENSO index 19 , a comprehensive assessment of conditions in the tropical Pacific Ocean 20 -which is significantly correlated with wave energy flux across the Eastern North Pacific 5 -the 2015-2016 winter was only exceeded by the similarly powerful El Niño events of 1982-1983 and 1997-1998 (Fig. 1b). However, a detailed record of coupled oceanographic forcing (that is, waves and water levels) and coastal response during these powerful events is limited primarily to anecdotal reports for the 1982-1983 event 13 , and a few discrete published data sets from the winter of 1997-1998 (refs 12,14,21). Further, climate change projections suggest a possible increase in the frequency of extreme El Niño and La Niña events 22,23 , which would affect coastal communities across the entire Pacific Basin margin 5 , making it critical to document the forcing and response of historically strong events as a possible proxy for future coastal vulnerability.
Here we provide a detailed assessment of wave conditions, water levels and coastal response during one of the most significant El Niño events of the Industrial Age: the 2015-2016 El Niño. The study analyzes two decades of winter oceanographic forcing across the US West Coast, focusing on the response of 29 beaches along the California, Oregon and Washington coasts, fronting a population of B25 million. The region experienced substantial increases in coastal hazards during previous El Niño winters, and has been shown to broadly represent conditions across the Central and Eastern North Pacific 5,15 . Both short-and long-term planning needs of coastal communities rely on assessments of the impacts of extreme El Niños due to the temporal scales of coastal hazard vulnerability, ranging from interannual storm hazard fluctuations to multi-decadal wave climate evolution and accelerating sea-level rise.

Results
Oceanographic forcing during the 2015-2016 El Niño. The wave climate in the Eastern North Pacific varies seasonally, with larger waves in the fall and winter months driven by the development and passage of extra-tropical cyclones across the midlatitudes, as well as episodic Eastern Pacific tropical storms in the summer and fall. High pressure dominates in the spring and summer months, with prevailing northwesterly winds and southern hemisphere storms typically resulting in lower wave energy conditions 4,24,25 . Winter wave energy flux/direction and water-level anomalies were determined from 1997 to 2016 for six wave buoys and six tide gauges, respectively, representing conditions across a 2,000-km section of the west coast of North America, and co-located with beach surveys grouped into six distinct geographic regions ( Fig. 2; Supplementary Data 1).
As a key driver of coastal change, mean and elevated (that is, top 5%) wave energy flux (a function of wave height and period, see Methods), were B50% above normal averaged across all regions during the 2015-2016 El Niño winter. During the 19 years of analysis, mean wave energy flux was only exceeded by the 1997-1998 El Niño (61% above normal), but elevated wave energy flux in 2015-2016 was the highest on record ( Fig. 3a; Supplementary Data 2: note the top 0.1, 0.5, 1 and 2% of wave energy flux is also included in this table, yielding results consistent with the top 5% metric for elevated wave conditions, but with an even greater discrepancy in elevated wave energy flux for the winter of 2015-2016. In the text hereafter, however, we refer exclusively to the top 5% as 'elevated'). The elevated winter waves of 2015-2016 brought two to four times more wave energy flux than the preceding anomalously low-energy winters of 2013-2014 and 2014-15. Further, one of the most energetic single wave events in the history of the regional wave buoy network struck on 10-11 December 2015, with significant and maximum wave heights off the California and Oregon coasts ranging from 8 to 11 m and 12 to 19 m, respectively 26 .
An unusual aspect of the 2015-2016 oceanographic conditions was the lack of a regionally consistent wave direction anomaly typical of prior El Niño winters 5 . Elevated wave energy flux in particular approached from more southerly angles during the 1997-1998 and 2009-2010 events, ranging from 4°to 13°south of the mean for the California and Washington regions. In contrast, mean and elevated wave energy flux direction in the winter of 2015-2016 was relatively close to the 20-year mean at most sites along the US West Coast, although the Southern California region did experience a marked northerly shift in elevated wave energy flux direction of 18°and 24°relative to the 1997-1998 and 2009-2010 winters, respectively, while Oregon recorded a southerly shift of 10°relative to the mean (Fig. 3b).
Seasonal water-level anomalies averaged 11 cm above the mean across the study area during the winter of 2015-2016, with the highest anomaly ( þ 17 cm) measured on the Oregon coast (Fig. 3c). The anomalies were significantly less than in 1997-1998 across all regions, particularly in Northern California and in the Pacific Northwest (averaged þ 23 cm in 1997-1998). In California, the water-level anomalies approximated those recorded during the 2009-2010 El Niño and the winter of 2014-2015, where the latter non-El Niño-related water-level anomaly was driven by a high-amplitude upper level ridge that persisted for several years in the Gulf of Alaska, promoting high pressure and unusually high sea surface temperatures 27  Coastal response during the 2015-2016 El Niño. Beach morphology responds to the seasonal modulation in forcing across the Eastern North Pacific, with beaches tending to build seaward (prograde) during the low wave energy summer months and retreat landward (erode) in the stormier winter months 21,28,29 . As coastal populations and infrastructure are most susceptible to storm hazards (for example, flooding, cliff failures and structural damage due to elevated water levels and wave attack) when beaches are depleted, we use the relative movement of a representative shoreline contour (a proxy for beach volume change 15 ) to assess the magnitude of coastal response and vulnerability.
Seasonal beach behaviour was assessed for 29 beaches along a B2,000 km span of the US West Coast that have been surveyed using aerial Light Detection and Ranging (Lidar), global positioning system-based (GPS) topographic beach surveys with All-Terrain Vehicles and backpacks, and/or discrete measurements of sand levels ( . At the regionally averaged scale, every region except for Central California experienced the highest seasonal shoreline retreat ever measured, and beaches in Central and North-central California recorded the most landward/eroded shoreline positions ever measured. However, it should be noted that the full extent of erosion during the comparably powerful 1997-1998 event probably was not recorded due to two important factors. First, topographic survey coverage in 1997-1998 was not as spatially extensive as in more recent years, with some of the sections of coastline anecdotally most impacted (for example, California) having particularly poor or spotty coverage, or none at all. Second, the Lidar survey utilized to establish the post-El Niño shoreline for many of the California study sites was not collected until April 1998, when beaches were already rapidly recovering, aided by the greater availability of   (Fig. 4c). The shoreline retreat recorded in 2015-2016 represents a fourfold increase over the prior, mild wave energy winter of 2014-2015 in Southern California, a fivefold increase over the prior winter in Central California, a threefold increase in North-central California and a twofold increase in Northern California and Washington (that is, the Columbia River littoral cell, which includes a beach in northernmost Oregon). Seasonal erosion on Oregon beaches exceeded 2014-2015 levels by a factor of 1.3.

Discussion
During the winter of 2015-2016, highly elevated winter wave energy flux (B50% above normal), coupled with seasonally elevated water levels ( þ 11 cm), drove unprecedented levels of winter shoreline retreat (76% above normal), including the most landward shoreline positions measured for the majority of beaches in California since topographic data collection began 20 years ago. The historical significance of this El Niño can be determined by analysing the relatively consistent record of wave energy, water levels and beach behaviour across the study area (available since 1997), and buoy records that date back to the mid-1970s 4 . These historical records, with wave hindcasts that stretch back to the mid-20th century 31 Measured multi-decadal increases in wave heights for the Pacific Northwest relative to California 4,37-39 is evidence of this broader trend, as is the predicted poleward migration of storm tracks and correlative northerly shift in the focus of extreme wave impacts along the west coast of North America noted in global wave modelling projections for the 21st century 40,41 .
While projections of El Niño frequency and magnitude for the 21st century are variable 42,43 , one recent study suggests a potential doubling of extreme El Niños 22 , similar to the strength of the 2015-2016 event. Such a trend would result in more significant hazards risk to coastal communities, which would be compounded by anticipated sea-level rise 44 . In addition to providing insight into possible future conditions when extreme El Niños are more frequent, the 2015-2016 El Niño winter may have disrupted the dynamic equilibrium of many US West Coast beaches for years to come, much like the highly anomalous wave activity and coastal response along the Atlantic coast of Europe during the winter of 2013-2014, the most energetic since at least 1948 (ref. 45).
Although erosive conditions were clearly amplified in 2015-2016, large landward shifts in shoreline positions did not translate to pervasively severe erosion of the dunes and bluffs that back the beaches in the Pacific Northwest. This is likely due to the fact that these beaches have generally been accreting since the previous 2009-2010 El Niño, and were in a significantly prograded state in summer 2015 due to the previous two mild winters (Figs 3a and 4a). As a result, increased beach sand volumes moderated the landward erosion resulting from increased wave energy and water levels. Similarly, recent nourishments along several beaches in Southern California also prevented shoreline retreat from reaching landward extremes during the winter of 2015-2016, thereby providing more storm protection for dunes and adjacent coastal infrastructure 46 . Based on this recent behaviour, such naturally or artificially sediment-rich coastal settings are likely to be more resilient to future storm impacts.
The potential for even more extreme coastal erosion during the 2015-2016 El Niño was also moderated by the earlier onset of peak annual high tides and the seasonal water-level anomaly associated with El Niño. The fall 2015 peaks in California, for example, were significantly earlier than the winter peaks that occurred during the El Niños of 1982-1983 and 1997-1998, thereby reducing the probability for the coincident arrival of the largest waves and water levels during the 2015-2016 event 47 .
While natural or artificial increases in beach volumes may reduce erosion-related hazard risk during extreme El Niños at some beaches, hazard risk on many US West Coast beaches may be worsened by historical and possible future reductions in watershed sediment supply to beaches. Even with major reductions in the coastal sediment supply due to dam construction, which has reduced pre-historical riverine sediment supply by 50% in Southern California 48 and by B80% down the Columbia River in the Pacific Northwest 49 , coastal watersheds remain an important source of sand for many US West Coast beaches 50,51 . However, 21st century climate projections clearly suggest a significantly warmer climate for California, coupled with precipitation changes that range from negligible to a 26% reduction, with the most severe potential temperature increases and precipitation decreases tied to the upper end emissions scenarios 52,53 , mirroring current trajectories 54 . Along with the historical trend of declining sediment supply, these 21st century climate projections would promote less runoff and reduced fluvial discharge rates 55 , likely further reducing the coastal sediment supply. In addition, the risk of extended drought in the Southwest United States is expected to increase significantly in the coming decades 56 , which, if punctuated by the predicted more frequent extreme El Niño events 22 , could increase coastal hazard threats. Reduced fluvial discharge would cause sand supply to beaches to be particularly depleted in the years leading up to these energetic winters, and the resulting narrower pre-El Niño beaches would provide even less protection than normal from increased El Niño wave attack.
In California, the 2015-2016 El Niño serves as proxy for this potential trend: a multi-year drought 57 44 . Therefore, the 2015-2016 El Niño also provides an indication of future background coastal water-level conditions and the associated beach hazards that will become more common during typical winters. The added potential for severe flooding and erosion will be compounded during El Niño winters with higher wave energy and seasonally elevated water levels, posing increasing threats to coastal populations across the US West Coast and beyond. was calculated as the difference between the summer/fall (August-November) maximum and subsequent winter/spring (January-April) minimum to coincide with oceanographic forcing fluctuations. The annual shoreline erosion anomaly for each region was calculated as:

Methods
where EA y is the erosion anomaly (%), E y is the erosion in year y and avg(E y ) is the mean of this quantity over the entire record. Hence, positive values of the anomaly correspond to erosion larger than the mean. Mean monthly sand height values from Isla Vista beach, within the Central California region, were calculated from near-daily observations taken at the vertical face of a concrete staircase in the intertidal zone (BMSL) from 1993 to 2016 (ref. 30). Individual sand height measurements were averaged by calendar month, smoothed using a 3-month running mean, and the average height for each calendar month computed over the entire 23-year record. Monthly height anomalies then were calculated for each month in the record as the difference between the average height in that month and the average for that month over the entire record. Negative values thus correspond to lower than average heights for a given month.
Wave and water-level statistics. For each of the six study regions, wave (that is, significant wave height, peak wave period and peak wave direction) and water-level data (that is, hourly measured and predicted) were used to assess interannual variability in wave forcing and water-level anomalies to characterize conditions across the US West Coast. Wave energy flux, F, was calculated using: where r¼1;025 kg m À 3 is the density of seawater, g, is the acceleration of gravity, H s is the significant wave height and T is the wave period. Wave directional anomaly was calculated as the number of degrees clockwise ( þ ) or counterclockwise ( À ) of the peak direction from the average peak direction. Water-level data were gathered from nearby tide stations, which are usually located in semi-enclosed harbours and sheltered from waves. All data were binned into boreal winter (December 1-February 28) and fall/winter (October 1-March 31) averages. A summary of the oceanographic forcing for each region is presented in Supplementary Data 2.
Code availability. The codes used to generate the results for this project are available upon request from the corresponding author.
Data availability. All relevant data used in the production of this manuscript are available upon request from the corresponding author.