How E ﬀ ective Were the Beach Nourishments at Cancun?

: Beach nourishment is generally seen as the preferred means of rectifying coastal erosion, due to its low environmental impact and natural evolution. The largest beach nourishment project ever carried out in Mexico took place on Cancun beach in 2006, as a response to the most intense hurricane season ever registered in Mexico, in 2005. After Hurricane Dean, in 2009, a second nourishment was conducted, which evidenced ﬂaws in the design and execution of the ﬁrst project. Previous investigations report that the need for beach re-ﬁlls directly correlates with wave energy. However, following a thorough revision of the extreme climatic events that occurred between 1978 and 2018, it has been found that the amount of erosion also depends on the frequency and duration of high energy events. The ﬁndings also show that the apparent success of the second nourishment is mainly associated with a decline in the number of extreme wave power events impacting the beach. In the conclusion to this paper, we share the knowledge gained, but not yet applied, in Mexico or elsewhere, regarding beach use, urbanization, and protection in beach planning.


Introduction
It has long been recognized that sandy beaches around the world are being eroded, e.g., [1,2] estimated that 70% of beaches were undergoing erosion and, according to [3], the problem was much more critical in the USA, where the figure was 90%. In the Caribbean Basin, [1,4,5] have estimated severe erosion rates for various beaches. In areas where there is potential for tourist or urban development, and erosion problems are detected, scientifically based engineering solutions are expected to control or mitigate these phenomena, e.g., [6].
Over time, the methodologies used for these solutions have evolved, e.g., [7]. In the past, the most frequently applied techniques were based on hardening the coast, by means of dikes, breakwaters, groins, etc. Despite initial criticism, the use of softer solutions (i.e., artificial beach nourishment) has gained popularity, becoming the preferred alternative for mitigating erosion [8][9][10]. The documented benefits [11][12][13] of artificial nourishment include: • The extension of the berm and beach profile, combined with protective dunes, dissipates wave energy, thus reducing expected damage from storms.

•
The berm is aesthetically valuable for tourism.

•
The extension of the beach increases its lifespan by delaying long-term chronic erosion.

•
The profile gradually re-shapes to fit the hydrodynamic conditions of the area, promoting stable conditions. • If the correct source of sediment is chosen (sorting, shape, quality and density), the nourishment will not induce drastic changes to the water current circulation patterns: color, transparency, temperature, pH, Dissolved Oxygen, Biological Oxygen Demand, Biochemical Oxygen Demand. Nor will it impact organisms, such as nekton, plankton and benthos. • Structural erosion (scouring) is avoided. • Implementation and maintenance costs are lower than those corresponding to rigid engineering works.
These benefits mean that when a beach has a sand deficiency, is important in terms of economic impact and job creation, and the environmental impacts are not very serious, beach nourishment can be seen as a feasible alternative to remedy problems of erosion.
In 1923, over a million cubic meters of sand were placed on a beach near Coney Island, New York, making this the first cited case of artificial beach nourishment in modern times [13,14]. From then until the 1950s some 72 beaches in the USA were artificially nourished. However, not all of these projects were based on scientific grounds, and most were carried out empirically. Many of the projects had a short life, arguably due to the inappropriate selection of the sediment used. In the United States and Europe, from the 1950s to the 1970s, many authors, including [15][16][17][18][19], offered technical criteria for the design of beach nourishment projects. These works formed the scientific basis for various compilations, such as [8,[20][21][22][23], from which most artificial sand nourishment projects worldwide have taken their evidence.
In Europe, the Netherlands leads the field in coastal protection, with 50% of its territory being below sea level. In 1984 the Netherlands introduced regulations for coastal protection based on artificial nourishment of their beaches and sand dunes [24]. The research carried out there and the experience accumulated confirmed the functional efficiency of this technique, as well as its adaptability and lower costs, compared to alternative methods [25]. Researchers in the Netherlands have contributed considerably to the knowledge of the criteria for designing sand nourishment projects, including the dune, berm, and submerged beach. Between 1952 and 1989, 60 million cubic meters of sand were placed on the coast of the Netherlands in over 50 projects. From the experience gained, in 1990 the Dutch government stated that beach nourishment was the first option in adapting to climate change, and an innovative large-scale beach nourishment project was begun, using the so-called Sand Engine strategy [26].
Important artificial nourishment projects have also been executed in Spain, Germany, the United Kingdom, France, Portugal, Georgia, Japan, and Australia. In the Caribbean area, some projects in Mexico, Cuba, and the Dominican Republic stand out, because of the large amount of biogenic sand required in the nourishments.
As with any engineering work, beach nourishment projects have a limited lifetime. This lifespan depends upon the sediment transport induced by the power and persistence of extreme sea states and the resilience of the system [27]. According to [28], an artificially nourished beach will have to be re-nourished, depending on the characteristics of its design and evolution. Artificial sand nourishment is simply a mitigation procedure, which needs periodic maintenance. If the environmental conditions are favorable, infilling intervals can be extended, with corresponding financial savings. On the other hand, it is accepted that if the intensity and frequency of the storms in the area increase, the erosion rates will also rise, as will associated costs of sand refills, perhaps making the whole procedure unsustainable. However, together with the intensity, the duration of heavy sea states is a variable which determines the lifespan of a beach nourishment. In any case, alternative measures should always be considered, including the use of rigid structures. If that also fails, a managed retreat of the infrastructure would be the only alternative.
Assessing the effectiveness of sand nourishment projects is important from both economic and technical perspectives. This is done through a monitoring program, by periodically measuring the morphological changes of the beach. The aim of this work is to assess the effectiveness of two beach nourishments carried out at Cancun, Mexico, in 2006 and 2009, using a total of 8 million cubic meters of sand. This assessment uses satellite images, photographs, and model-derived wave power records to examine the morphological evolution of the coastline over the last 40 years.

Cancun Before the First Beach Nourishment
Cancun is located on the Caribbean coast of the Yucatan Peninsula, in Mexico ( Figure 1). The beach is a prime example of an environment which has been progressively degraded by human activities linked to tourism. The urbanization of Cancun started in the late 1960s on a pristine barrier island between the Caribbean Sea and a system of coastal lagoons (Nichupté). The lagoons are supplied with fresh water, via continental ground water discharge, and with sea water, which arrives with the tides (micro-tidal regime), through two inlets. The lagoons used to be surrounded by a very rich mangrove forest, which is now greatly reduced. The beach of Cancun lacks the protection of coral reefs and the sediment is composed mostly of biogenic sediments. The main anthropic alteration to the Cancun coast has been the construction of very dense infrastructure (hotels, roads, gardens, golf courses, etc.) on the dunes. The inlets were also rigidized, causing modifications to the natural breaching of the sandbar during storms, and thus limiting the interaction between the sea and lagoon. These changes, together with a decline in the amount of natural sediment availability, brought about intense erosion. Beach loss was estimated at an annual rate of 1.8 m for 1967-2005 [29]. Figure 2 shows Cancun in 1947, and the slow, but continuous urbanization of the beach between 1978 and 1988. As documented by [30] and [31], until 1988 the urban planning of Cancun did not take into consideration the natural dynamics of the coast. In September 1988, Cancun received the full force of hurricane Gilbert. The waves and storm surge moved massive quantities of sand and, for the first time, the hoteliers there experienced the effects of substantial erosion on the beach. Several studies were carried out, but no action was taken to recover the beach. The individual hotel owners implemented uncoordinated, inadequate coastal protection schemes (e.g., vertical walls, detached breakwaters, and groins), accelerating the problem and transferring the effects to neighboring beach segments (see Figure 3). Even so, the beach partially recovered until hurricanes Ivan (2004) and Emily (2005) hit the coast [32]. Then, in October 2005, Hurricane Wilma hit Cancun, removing 8 million m 3 of sand [33]; almost all of the beach was left without sediment ( Figure 4).  It can be seen that in front of many hotels the beach totally disappeared; only rocks were left behind (panels (a), (c) and (f)), while in other places, waves were hitting the hotels walls and facilities directly (panels (b), (d) and (e)).

The First Beach Nourishment, 2006 to 2009
In 2006, the problems of erosion had come to a head; the situation was critical economically as the all-important tourist industry in Cancun was in jeopardy. An urgent solution that could restore the beach immediately and activate the economy was necessary and a beach nourishment was undertaken to achieve this goal. From January to April 2006, 2.7 million cubic meters of sand were placed on the beach, with a cost of USD 19 million. The sediment was borrowed from two nearby sand banks of La Ollita and Megarrizaduras ( Figure 1). La Ollita sand bank is 12 km north-east of Cancun, off the northern end of Isla Mujeres. Around 1.7 million cubic meters of sand were extracted from here, from an average depth of 25 m [33]. The Megarrizaduras sand bank is 15 km north of Punta Cancun, in shallow waters (7-10 m depth) between Cancun and Isla Mujeres. Around one million cubic meters of sand were extracted from this bank.
Immediately after the sand nourishment, the beach had an average width of 60 m ( Figure 5). However, the beach rapidly began to narrow, losing an average 8 m in width, from May to September 2006, and even more, 29 m, from September 2006 to September 2007 [34]. In the same period, the sand on the beach formed a dramatic, pronounced scarp, making it very difficult for tourists to enjoy the beach (see upper panels of Figure 6). These scarps remained, due to the lack of sediment in the submerged part of the beach profile. The beach was dramatically damaged by hurricane Dean in August 2007.   From exhaustive monitoring of the beach evolution in 2006-2009, [38] concluded that during extreme wave conditions, longitudinal currents are highly dynamic and determine sand availability at the northern and southern ends of Cancun beach. In the central part of the system, offshore sediment transport is dominant and the sediment balance is negative. This sediment imbalance prevents the natural self-regulation of the coastal system; causes very low sediment input into the system; and leads to large amounts of sediment being transported out of the system by hurricane-induced high wave energy. These conclusions have been very important in understanding the beach behavior at Cancun.

The Beach Nourishment of 2010, until 2020
Given that the beach had not reached a dynamic equilibrium, it continued eroding and consequently was not attractive for tourists (Figure 8). A second beach nourishment was therefore performed, from December 2009 to January 2010. The methodology was similar to that of 2006. However, this time the amount of sand used to replenish the beach was almost twice that of 2006 (5.2 million cubic meters). The width of the beach at the end of the process was 80 m and the thickness of the sand layer in the beach profile was substantially higher than that of 2006 ( Figure 9). Additionally, a 305 m long breakwater crowned 2.5 m above the mean sea level was built on Punta Cancun, to prevent the transport of the sediment to the north.  For the second nourishment project, 2.5 million cubic meters of sand were extracted from La Ollita II sand bank, next to the sand bank used in 2006. The sand was extracted from an average depth of 25 m. The remaining 2.8 million cubic meters of sediment were taken from the sandbank at Punta Norte, in the shallow waters north of the island of Cozumel, 48 km away, at a depth of 12-29 m [33]. While all the borrowed sand was of marine biogenic origin, the sand used had different mechanical characteristics from the sand native to Cancun beach (see Figure 10). The differences in the sand characteristics may explain the unpredictable behavior of sand transport on the beach and therefore the beach evolution after the second nourishment. The estimated lifespan of the second nourishment was 3 years [33], a period calculated on the basis of the trends in severe storm impacts previously observed. Nevertheless, after the beach nourishment process, the system never attained the expected profile (see Figure 11). The wave action caused a steep beach scarp to form; an unpleasant 2 m vertical step ( Figure 12). A very thorough analysis of this process on the beach in Cancun was made by [38].  In 2013 all the scarps had disappeared ( Figure 13) and, 10 years after the second nourishment, the beach at Cancun has managed to conserve a nearly stable width of approximately 30 m (Figures 14 and 15). Therefore, no further beach nourishments have been required.

Characterization of Extreme Events 1979-2018
To evaluate the response of the beach nourishment to the wave climate, three periods, namely Stage 0 from 1979 to 2005; Stage 1 from 2006 to 2009; and Stage 2 from 2009 to 2018, were analyzed separately. For this characterization, wave records for 21.0 • N and 86.5 • W were used. Hourly information of significant wave height (H s ), peak wave period (T p ), mean wave period (T m ), and wave direction (θ) were taken from the reanalysis dataset ERA 5 [40] of the European Centre for Medium-Range Weather Forecasts (ECMWF).
An extreme event was defined as when the significant wave height exceeds the general 90th percentile significant wave height (H s90 ) [41]. This value was 2.10 m for Stages 0 and 1, and 2.03 for Stage 2. If the difference in time between the events is more than 12 h, these are considered independent events [42]. The deep-water parameters for each extreme event found of the wave time series were plotted against the number of events, that is: event duration, storm cluster time intervals (IN), significant wave height (H s ), maximum significant wave height (H smax ), peak period (T p ), maximum peak period (T pmax ) and mean wave power by event (P wE ). The storm cluster IN is the time elapsed (days) between storms [43].
The results shown in Figure 16 give an overview of the intensity, persistence, and clustering of the extreme events which occurred in each of the three stages. All of these factors are important in the morphodynamic response of the beach. In Figure 16 the variables are presented in histograms, to facilitate comparisons between the stages.  Figure 16 shows quite similar distributions of the variables for all three stages. As the period is much longer, Stage 0 shows a greater number of events in all the histograms and also more events in the tail of the distributions; this is a natural phenomenon. The duration of the individual events shows no significant change between the stages. Of particular interest is the low values of IN found in Stage 2, which show that those storm seasons were more active and, when this time is less than the natural recovery period of the beach, the damage to the beach tends to accumulate. Panels (c), (d), (e) and (f) show that the most intense storms occurred in Stages 0 and 1, whilst in Stage 2 the reduction in the number and intensity of storms is clear.
These findings, analyzed for each stage, are in agreement with the levels of wave power (intensity, persistence) generated by the most intense events. From the information in Figure 16, the levels of erosion reported by [44] can be more clearly understood: in Stage 0 from 1983 to 1990, the eroded area was 218,000 m 2 ; during the 1990s, the erosion was lower; 19,600 m 2 , while the eroded surface from April to October 2005 was 265,000 m 2 . The hydrodynamic characteristics of extreme events and other factors, such as translation speed and trajectory, sea level, wind effects, and sediment properties, contribute to beach erosion. However, it has been reported elsewhere that episodic extreme storms and a series of lesser storms occurring close together could result in greater coastal impact than one single huge storm [45][46][47]. This means that longer periods of persistently high wave energy can lead to great changes in the beach profile [48,49]. Therefore, in this paper, a detailed analysis focusing only on wave power was performed.
For the study of the wave power, a time series was calculated using the formula for irregular waves (Equation (1)), see Figure 17: H s 2 T e (W/m).
where H s (m) is the significant wave height and T e is the energy period of the spectra, which can be expressed as a linear function of the mean wave period T 01 (s), where α = 1.08 [43].  Table 1. According to reports, these six hurricanes generated the waves which induced the most severe beach erosion. However, the orders of magnitude of the sand eroded were not directly correlated with the maximum wave power reached in these events. Looking at the dates of the greatest events (Figure 17), and considering their durations (Table 1), together with the largest beach losses reported, erosion seems to be closely related to the intensity and duration of extreme sea states. The arrival direction of the wave power to the shore can limit its erosiveness. For this reason, the dominant direction of the incoming waves from 1979 to 2018 was taken from the same ERA-5 database. The results, separated into Stages 0, 1 and 2, can be seen in Figure 18. From Figure 18 it can be inferred that, during Stage 0, the extreme events induced sediment transport to leave the littoral cell in all directions. In Stage 1, most transport was from the south to the north and during Stage 2 no events are found that could produce relevant sediment transport to cause beach loss. The Cancun beach, like any other, has seasonal and interannual variations in erosion-accretion. However, when the wave power exceeds a threshold of around 100 kW/m, the sand is exported out of the coastal cell and the system is not able to recover naturally

Discussion and Conclusions
Between 1988 and 2012, six major hurricanes impacted Cancun and information was demanded. As a result, a large number of coastal and oceanographic studies were undertaken, focusing on Cancun, and paying attention to developing human and technical capabilities. Until 2010, the time of the second nourishment, the general perception was that the beach at Cancun had a clear erosion tendency. In this section we will show that this tendency was directly linked to the incidence of extreme hydrodynamic conditions and the scarcity of natural sediment sources. These findings could change erroneous perceptions, and present many challenges for future nourishment projects, includes the need for improving long term predictions of wave climate under global warming scenarios.
Previous research [37,38] explained the dynamics of Cancun beach in terms of wave energy fluxes with reasonable accuracy. Therefore, in this paper we opt for a long-term analysis, only evaluating the overall wave power that reaches the beach. With this information, we identified the events which produced the most erosion at Cancun beach. The effects of these climatological events were identified by a cluster analysis which suggested three stages, before the first nourishment , after this (2005)(2006)(2007)(2008)(2009), and after the second nourishment (2010-2018). As storms were more frequent in Stages 0 and 1, the recovery capacity of the beach was debilitated.
The analysis of wave power in the region indicates that for the periods before and after the 2010 sand nourishment, the severity of the hydrodynamic conditions was very different. In the period 1988-2009, the beach was repeatedly affected by intense storms, which induced very high erosion rates. In the period after the second beach nourishment (2010-2018), storm impact was very low. The analysis shows that for the last 8 years, no major hurricanes have impacted the region, and the beach of Cancun has been stable. In this sense, beach stability means that Cancun has maintained a beach width suitable for recreation and for the protection of infrastructure from inundation and damage (i.e., >40 m wide) for most of its length, most of the time (see Figure 14). In the absence of hurricanes, all work related to beach monitoring has ceased, despite the importance of Cancun in terms of the tourist industry. Decision makers and investors believe that the beach is stable and that the future of this resort was guaranteed with the second beach nourishment. However, Cancun has taught us that sudden erosion processes occur when we are least prepared.
The considerable amount of sand deposited on the beach (8 million cubic meters), along with the low storm impact in the last 13 years, has led the beach away from an erosion tendency. However, it is still too early to assert whether this balance will continue over time, given that the main causes of the erosion are still present.
In the last 50 years of touristic development at Cancun, the exploitation of the dunes, and the use of the beach for recreation, there have been various lessons that should have been learnt: 1.
The passage of Hurricane Gilbert in 1988 showed that the resulting erosion put at great risk the touristic infrastructure. 2.
Between 1988 and 2004, it was seen that without the presence of very severe waves the beach can slowly and partially recover.

3.
During this period, we also learned that the individual "coastal protection" efforts (by hoteliers) in fact increased structural erosion and that such uncoordinated efforts should be avoided. The need for comprehensive coastal management plans and solutions for the entire beach was made evident. 4.
The number of nearby sand banks available for borrowing sand is limited; these areas cannot be exploited indefinitely. As this sand is of biogenic origin, the production of these sediments is a very long-term process.

5.
Beach scarps are unsuitable for tourism. The design and techniques employed for beach nourishment has to be different, since for long periods the beach profile had a 2 m scarp. 6.
From the morphodynamic behavior of neighboring beaches in the same period, it is evident that the impact of urbanization on dune ecosystems reduces their resilience. 7.
The experience gained from the first beach nourishment and subsequent coastal monitoring programs, implemented by government and research bodies, enhanced the design of the second beach nourishment. 8.
Much to the contrary of what was previously thought, it has been seen that, in the last 10 years, the beach has not gradually eroded. The erosion events are episodic and the beach may recover again. Nevertheless, it is expected that an extreme event could drive Cancun rapidly to an unstable stage, since much of its natural resilience has been lost due to the urbanization of the coastal dunes.
Cancun is a vital asset to the economy of Mexico. In order to have a tourist destination with a healthy sustainable beach, a series of actions are necessary, including the re-establishment of coastal monitoring programs to have more accurate information to correctly diagnose the beach behavior. It is also important to continue comprehensive coastal planning work that allows the coexistence of natural processes and sea-sand-sun tourism.