- Email: [email protected]
Coastal hazards from slope mass movements: Analysis and management approach on the Barlavento Coast, Algarve, Portugal
Ocean & Coastal Management 102 (2014) 285e293
Contents lists available at ScienceDirect
Ocean & Coastal Management journal homepage: www.elsevier.com/locate/ocecoaman
Coastal hazards from slope mass movements: Analysis and management approach on the Barlavento Coast, Algarve, Portugal ~o Braz Teixeira Sebastia Portuguese Environment Agency, Rua do Alportel 10, 8000-503 Faro, Portugal
a r t i c l e i n f o
a b s t r a c t
Article history: Received 19 May 2014 Received in revised form 12 October 2014 Accepted 14 October 2014 Available online
The Barlavento Coast, Algarve, Portugal, is dominated by rocky sea-cliffs, cut on Miocene calcarenites which evolves through intermittent and discontinuous events of slope mass movements, along a 46 km cliff front. Here, the main coastal geologic hazards result from the conflict between human occupation and sea-cliff recession. Most of the research on the dynamics of the cliffs has been directed to the risk with the aim of defining long term set-back lines, for a preventive planning of the cliff top occupation. Little attention has been given to the hazard associated with mass movements on bathing beaches backed by sea-cliffs. This article presents the results of a field inventory of 244 slope mass movements single events, collected in a rocky shore with tens of touristic pocket beaches, covering an nineteen year time span (1995e2014). Results show that landslides have seasonal pattern with higher incidence in the period between winter and early spring. More than 15% of movements occur during the Easter holidays (April) and 4% of landslides occur during the official bathing season (JuneeSeptember). The spatial distribution of landslides shows that only 22% of the mass movements occur in capes and headlands, while 78% occur on the beaches, which demonstrates that the beaches are real hot spots of risk. Based on the size distribution of slope mass movements runout ratio (the ratio between the radius of the base of the cone of and the height of the movement) a table of levels of security and hazard on beaches was built. Security levels enable the definition of cartographic hazard areas on beaches which can be provided to the beach users on information boards at the beach entrance. © 2014 Elsevier Ltd. All rights reserved.
Keywords: Rocky sea cliffs Hazard areas Beaches Runout ratio Algarve Portugal
1. Introduction Algarve coastal region, with 451,000 inhabitants in 2011, is the premier tourism destination of Portugal, with a mean annual of 15 million overnight stays through the 2002e2011 period (INE, 2012). Almost 90% of the visitors' reasons of staying in the Algarve coast are leisure, recreational and holiday (INE, 2012). The Barlavento Coast between Lagos and Albufeira (Fig. 1) is the core touristic area and receives 60% of the Algarve visitors attracted with the “sun and beach”, which is the Algarve' s top touristic product. About 50% of the Algarve visitors stay and bathe in beaches backed with rocky sea cliffs of the Barlavento Coast. Over the last three decades, there has been record of several accidents caused by the collapse of sea cliffs cut on Miocene rocks. On 22 March 1998, a Portuguese man was killed while fishing at the das Porcas site, when a sudden planar landcliff edge at the Mare slide dragged him down together with a volume of 2 104 m3 falling material; on 7 October 2000, three Swiss tourists were
E-mail address: [email protected]. http://dx.doi.org/10.1016/j.ocecoaman.2014.10.008 0964-5691/© 2014 Elsevier Ltd. All rights reserved.
injured by a block fall (volume 2 m3) on the Inatel beach; on 21 August 2009 an instantaneous topple (volume 1 103 m3) on a sea stack killed five Portuguese tourists and injured another two, resting on the cliff base on the Maria Luísa beach; on 26 May 2010 a four year Irish kid was slightly wounded on the Vau beach hit by the debris of a landslide (Fig. 2, volume 2 102 m3); on 11 October 2010 a German tourist was injured on the Beijinhos beach, hit by blocks of a small landslide (volume 1 m3). This record shows that the average number of accidents resulting directly from cliff collapse of sea cliffs is 2 event/decade, causing two fatalities and two injuries each decade. Searching for the accidents of the cliff top walkers, mostly sightseers and fishing anglers, in the decade 2003e2012, Teixeira and Dores (2013) identified a record of 50 accidents on the Barlavento rocky cliffs with 11 fatalities and 41 injured. 45% of the victims were foreigner tourists. Although statistically not very significant when compared to accidents resulting from the use of the top of the cliffs, accidents caused by landslides on beaches have great impact on public opinion. While accidents of suicide or death of fishermen by falling from the cliffs typically occupy small news in local newspapers, after the collapse recorded in Maria Luisa beach
286
S.B. Teixeira / Ocean & Coastal Management 102 (2014) 285e293
Fig. 1. Barlavento coast. Study area.
the accident was the subject of daily news in all national newspapers for more than a month. Five years on, the accident is still remembered every year at the start of the bathing season (JuneeSeptember). Slope mass movements on steep rocky sea cliffs are instantaneous phenomena with virtually no warning. The velocity of the movement is typically over 5 m/s, within the landslide velocity class 7, the extremely rapid events of the Cruden and Varnes (1996) classification. On the Barlavento rocky coast, where sea cliffs are subvertical with height varying from 5 to 40 m, the duration of the slope mass movements lies in a narrow time window of 1e2 s. The instantaneous nature of the phenomenon precludes any action to minimize the damage after the onset of the collapse process. On steep subvertical rocky sea cliffs prone to slope mass movements, the actions to minimize risk and damages are therefore exclusively based on prevention. The first step in the implementation of prevention is the knowledge and definition of the areas potentially affected by a mass movement, i.e. the spatially definition of hazard areas. Hazard areas, limited by hazard lines, correspond to areas parallel to the shoreline where, in a pre-defined period, it is likely that effects of slope mass movements will be felt (Fig. 3). Most studies and bibliography, oriented strand planning, focuses on the land hazard areas at the top of the cliffs worldwide
(see, for example, Hall et al., 2002; Lee and Clark, 2002; Moore and Griggs, 2002; Del Rio and Gracia, 2009; Stravou et al., 2011; Epif^ anio et al., 2013) and on Algarve (Marques, 1994, 1997, 2003; Teixeira, 2003, 2006; Bezerra et al., 2011; Marques et al., 2011; Nunes et al., 2009). Very few papers deal with sea hazard areas on sea cliffs (Marques, 2009), although studies on the identification of risk areas associated with damage caused by slope mass movements are very common in land areas (e.g, Copons et al., 2009; Michoud et al., 2012). In the case of touristic areas centered on beaches backed by cliffs, the sea hazard area has particular interest in that it is in this area where accidents occur with people affected by the debris of a mass movement. In Portugal, beach tourism is an important economic activity, all the coastal plans regulations include a seaward hazard area on beaches backed by sea-cliff extending seawards from the cliff toe, where beach support structures are interdicted (Marques, 2009). On the study area coastal plan regulations, in force since 1999, a sea hazard area is defined on beaches backed by sea cliffs with a width of 1.5 times the cliff height and corresponds to the maximum extent of cliff failures debris displacement near the toe. In this paper we present the results of slope mass movements inventories gathered on Barlavento Coast sea cliffs, for the last nineteen years (1995e2014), we assess the adequacy of the legal
Fig. 2. Rock fall on the Vau beach occurred in the 26 May 2010 (location on Fig. 1); mean width (Wm) ¼ 2.5 m; runout ratio (R) ¼ 0.8.
S.B. Teixeira / Ocean & Coastal Management 102 (2014) 285e293
287
Fig. 3. Hazard areas on rocky sea cliffs.
provisions of the hazard areas and we propose a solution to spatially define the seaward hazard lines on beaches backed by rocky cliffs prone to slope mass movements. 2. Study area The south coast of the Algarve, Portugal, located at South Western end of Europe, has a typical Mediterranean climate with hot dry summers and mild rainy winters. Average annual precipitation is 500e600 mm, 80% of which is concentrated in a wet season (OctobereMarch). Wave regime is moderate with mean annual significant height of 1 m (Costa et al., 2001) and 85% of sea storms (Hs 2.5 m) occur in the wet season Tide is semi-diurnal; mean tidal range is 2 m and spring-tide range is 3 m. The Barlavento Coast is a rocky coast with a continuous front of subvertical sea-cliffs cut in carbonate Mesozoic and Miocene rocks. s beach and The study area is the cliffed coast, between Porto de Mo Olhos de Agua beach (Fig. 1), cut in Miocene rocks, included in the ~o Formation (Cacha ~o et al., 1998). The total length of Lagos-Portima sea cliffs cut in Miocene is 46 km as measured using detailed topographic maps (scale 1:2000). Sea cliffs are composed of alternate decimetric layers of fine grained calcarenites and calcarenites with high content of macrofossils (Manupella, 1992). Carbonate content is 60e75% in fine-grained calcarenites and greater than 80% in fossiliferous calcarenites (Marques, 1997). Cliff height varies from 5 to 40 m and displays a sequence of strata, horizontal or gently sloping to the South or Southeast except in the vicinity of major tectonic features of the area (the Portim~ ao fault and Albufeira diapir; Terrinha, 1998) where the slope is greater (10e20 ). Coastal Miocene formations have a deep and well developed karst, with a dense network of caves, sinkholes and galleries covered by PlioQuaternary (Manupella, 1992) reddish silty-clayed sands. The combination of the network of karsic cavities with the action of coastal erosion provides a modeling of the coastline very diverse and irregular, very attractive for tourists that seek the region.
The quantitative assessment of cliff retreat rates in Miocene calcarenites was first performed by Marques (1994, 1997) based on identification and measurements of slope mass movements by comparative analysis of aerial photographs. The recession rate based on values of lost area of mass movements in the 1947e1991 period is 1e2 cm/year (Marques, 1997). Based on a continuous field inventory, of mass movements on Miocene cliffs, started in 1995, Teixeira (2003, 2006) calculated a mean recession rate of 1 cm/year. The maximum local retreat measured by Marques (1997) was 45 m, in an arch collapse in plunging cliffs that occurred between 1974 and 1980; in the 1995e2014 period the maximum local retreat measured was 25 m in a head collapse occurred in 6 October 1998 in Marinha beach (Fig. 4). The irregular shape of the coast promotes the formation of seventy small pocket beaches, with lengths of tens to hundred meters, and thin sand cover about 2 to 4 m thick, accumulated over cut shore platforms. Most beaches (80%) are accumulated in the irregularities generated by erosion of the cliffs. Only sixteen of the beaches are located in small river mouths, four of them in hanging valleys. One third of the cliff front (15 km) has a beach on their base. The total length of the fifty official bathing beaches extends for 9 km. 3. Methods The regional coastal management authority of Algarve (Portuguese Environment Agency e APA) has been conducting a continuous observation of the coast since 1995, which includes the systematic recording of slope mass movements occurring on the front of rocky sea cliffs cut on Miocene calcarenites. The very first source of information of the slope movements is the reporting of occurrences by public regional or local field authority staff, as well as by the tourism operators, private owners and fishermen. Complementarily, a periodic systematic observation in situ of the coast, by land, sea and air, is conducted several times a year,
~o Vistal 1992; mean width (Wm) ¼ 20 m; Runout ratio Fig. 4. Head collapse and generation of a new stack in Marinha beach, occurred on 6 October 1998. Postcard from Ediça (R) ¼ 0.5.
288
S.B. Teixeira / Ocean & Coastal Management 102 (2014) 285e293
Fig. 5. Topple occurred on 2 Out 2009, in Santa Eulalia beach; (location on Fig. 1) Wm ¼ 1.5 m; R ¼ 1.9.
especially after wave or rainy storms. At least, once a year a campaign of capture of oblique photographs either on a low altitude flight (since 2006) or on a boat tour (since 2002) is done in order to get a complete view of the area and identify recent slope mass movements. Fresh scars on the sea cliff face, detritus cones with fine sediment content, and a yellow plume in the water, resulting from washing out of the debris, provide the most obvious signs of a recent mass movement (Teixeira, 2006). The main source of the inventory comes from periodical observation of the sea cliff front that enabled the identification of a considerable amount of unreported slope mass movements (66% of the inventory). The gathered inventory covering a nineteen year time span (July 1995eJune 2014) has a total of 244 slope mass movements, with a mean length 1 m, which is the threshold of completeness of in situ inventories on Barlavento Miocene calcarenites (Teixeira, 2006). After each report of an occurrence or during the periodical field observation, a trip to the site is carried out and a characterization of mass movement type and geometry is performed in situ as described in Teixeira (2006). In the field inventory (244 events), mass movements were classified onto four types, according to the
generic mass movements proposed by Sunamura (1992) and Marques (1997): rock fall (Fig. 2), topple (Fig. 5), karst collapse (Fig. 6) and block fall (Fig. 7). For the purpose of the present study, the radius of the debris cone (the distance between the cliff base and the seaward limit of the debris), geometrically the same as the runout or travel distance used on landslide motions, (Hungr et al., 2005; Michoud et al., 2012) is an important parameter, crucial for computing the runout ratio (R) parameter. The runout ratio of a slope mass movement is defined as the ratio between the radius of the base of the cone of debris and the height of movement, a dimensionless scaling parameter. When the mass movement reaches the total cliff face, R is equal to the ratio between the radius of the base of the cone of debris and the cliff height (Fig. 8). A cliff collapse in which the cone of debris projecting from the base of the cliff at a distance equal to the height of the cliff has a runout ratio 1; a collapse in which the cone of debris occupies the sandy beach up to a distance equal to the height of 1.5 times the cliff height has a runout ratio 1.5. The measure of the runout is sometimes impossible in mass movements identified weeks after the occurrence as waves tend to
^ra beach, in February 2011 (location on Fig. 1); Wm ¼ 1.5 m; R ¼ 1.5. Fig. 6. Karst collapse in Armaç~ao de Pe
Fig. 7. Block fall on Albandeira beach, occurred on 12 September 2007 (location on Fig. 1); Wm ¼ 1.5 m; R ¼ 0.9.
S.B. Teixeira / Ocean & Coastal Management 102 (2014) 285e293
289
Fig. 8. Definition of the runout ratio of a slope mass movement on steep rocky sea cliff.
Fig. 9. Stack collapse occurred on 1 Dec 2010, in Olhos de Agua beach.
rapidly wash out and remove the debris. In movements recorded in plunging cliffs and promontories (22% of the field inventory) sometimes is not possible to calculate the runout as the cone of debris extends to the immersed area below the low tide line, making it difficult to measure the radius of the debris cone. For a total of 244 registered mass movements in the 1995e2014 period, the R was computed in only 170 (70% of the field inventory). 4. Results After each report of an occurrence or during the periodical field observation, a characterization of mass movement type and geometry was performed in situ, measuring mass movement mean and maximum linear parameters (length, height, width), directly on the cliff scar, as described by Teixeira (2006). The horizontal area of loss was computed as the product of mean length and mean width, and the volume of the mass movement was calculated as the product of horizontal area and mean mass movement height. The volume of the fresh debris produced by the mass movement provided a supplementary control of the evaluation of dimensional parameters. Between July 1995 to June 2014, covering a nineteen years span, 244 slope mass movements on sea cliffs, have been recorded and
Fig. 10. Cumulative absolute frequency size of mean width of mass movements occurred between July 1995 and June 2014 (n ¼ 244).
measured, spatially dispersed along the sea cliff front. Rock fall is the dominant movement (61%). Topples (21%) and karst collapse (16%) have approximately the same importance. Block fall is the less frequent mass movement type (3%) During the study period five stacks collapsed (Fig. 9) and two new ones were formed (Fig. 4). A total of 2100 m of the sea cliff front (4.5% of the total cliff front length) was altered by slope mass movements, during the nineteen year period of observation. The average value of the renewal of the cliff front of 0.24%/year gives an estimate for the cliffs mean life period of 425 years. The average loss of horizontal area was 400 m2/ year, the mean annual volume was 7400 m3, and the average recession rate was 9 mm/year, which are values similar to previous estimates obtained by Marques (1997) and Teixeira (2003, 2006). The field inventory showed an average frequency of 13 movements/year with a range of 2e42 annual movements. The series of mass movement mean width (the width of the instantaneous retreat) show a magnitude decay pattern, i. e. the wider the mass movements are, the less frequent they are (Figs. 10 and 11) as previously showed by Teixeira (2006). Increasingly, recent studies demonstrate that the dimensions (width, area or volume) of cliff failures fit to inverse power-law relationships (e. g. Marques, 2008; Lim et al., 2010; Young et al., 2011; Barlow et al., 2011, Katz and Mushkin, 2013).
Fig. 11. Return period distribution of mean width of mass movements occurred between July 1995 and June 2014 (n ¼ 244).
290
S.B. Teixeira / Ocean & Coastal Management 102 (2014) 285e293
Fig. 12. Annual distribution of maximum date uncertainty of slope mass movements. Year starts in July and ends in June.
At this stage, the destabilizing effect of the water seems to have an immediate effect. The second stage of failures that extends throughout the winter has the highest expression in early spring. Here the progressive increase in the frequency of collapses suggests the preponderance of the cumulative undermining effect of water. Storms and heavy rains in early spring seem more effective in triggering cliff failure than in autumn. In four movements it was possible to relate the collapse with direct human intervention. The most dramatic episode was the break in a swimming pool in June 1997 previously reported by Teixeira (2006). The other three are associated with the concentration of runoff resulting from poor sealing and channeling rainwater in May 1997, October 2008 and March 2010. In 18% of the monthly inventory the trigger is unknown, possibly associated with fatigue failure of the cliff material. Only 4% of mass movements occurred during the official bathing season (June to September) but three of the eight recorded cliff failure resulted in casualties (5 fatalities and 6 wounded).
4.1. Date uncertainty The effort for permanent observation of the 46 km long coast and the level of dating of the slope mass movements is such that 77% of the events are monthly dated and less than a half of the events are daily dated (46%). In 94% of the occurrences the date of the event is within a tree month period. Maximum uncertainty of the date of the movements is less than a half year (Fig. 12). 4.2. Monthly distribution The dataset dated monthly (monthly inventory, 188 events, Fig. 13) enable more detailed analysis of the annual distribution of mass movements and their relationship to the beach use. The seasonal distribution of monthly dated movements shows clear seasonal pattern with predominance of cliff failures during the wet season, where 83% of cases occurred between November and April. In half of the movements in which the exact day is known the collapse occurred under sea storms (Hs 2.5 m) or daily rainfall exceeding 10 mm. 78% of failures occurred within five days after storms or heavy rain. A closer and detailed look at the monthly distribution of the collapses suggests a two stage dynamics. The forerunner stage happens at the beginning of autumn when a first peak of collapses occurs in October triggered by the first heavy rains and sea storms.
4.3. Runout ratio size distribution The amplitude of the runout ratio covers the 0.1e1.9 range and varies with the type of mass movement (Fig. 14). The most frequent type of movement (rock fall, 98 records) presents a normal type distribution with a mean value of 0.90. In the case of topples (34 records), with more pronounced horizontal component, the distribution also resembles a normal distribution with mean value 1.11, but with greater variance. The karst collapse failures produce the lowest R (mean 0.49, 30 records). Values show great dispersion, with a mode centered in the lower values associated with the prevalence of movements with almost exclusive vertical component. In the case of block fall (8 records) values have very tiny dispersion, with all values in between 0.75 and 1.0. Aggregating the result set (Fig. 15) it is apparent that in all the results fit a normal distribution (mean 0.88), with a secondary mode on values of the lower runout ratio, associated with movements with predominance of the vertical component. 5. Discussion 5.1. Beaches as hot spots Of the 244 mass movements recorded in field inventory, 190 (78%) occurred in the cliffs of the beaches. The remaining 54 (22%)
^ncia Portuguesa do Ambiente); sea storm data from Fig. 13. Monthly distribution of slope mass movements (n ¼ 188), rain and sea storms. Precipitation data from Algoz station (Age ^s do Mar e Atmosfera (IPMA). Instituto Portugue
S.B. Teixeira / Ocean & Coastal Management 102 (2014) 285e293
291
Fig. 14. Runout ratio distribution according to slope mass movement type (n ¼ 170).
were recorded in plunging cliffs and promontories. Over the observation period the field inventory gives a ratio of 12 movements/km on beaches whereas in the remaining area that ratio is 1.7 movement/km. Rocky coast sheltered areas (where beaches accumulate) are sevenfold more prone to slope mass movements than exposed sites (headlands and promontories). From the point of view of coastal erosion Barlavento beaches are real hotspots. Results show that the most vulnerable areas are not located in the areas most exposed to wave attack despite wave action is one of the most effective triggers on the evolution of sea cliffs. Plunging cliffs are permanently exposed to direct wave action, promontories are daily exposed (at flood tide), while in cliffs backed by beaches wave action acts episodically under storm conditions and spring tides. This dynamic pattern of the Barlavento cliffs coincides with the evolutionary pattern of rocky cliffs defined by Sunamura (1992) in which plunging cliffs occur on resistant rocks and shore platforms develop on weaker rocks. Bezerra et al. (2011) used a wave propagation model to quantify breaking wave conditions and wave energy to estimate the relative role of wave action and lithology in controlling cliff erosion and evolution along a 12 km of Miocene calcarenites of Central Algarve. They found that lithology represents the dominant control on mass movement occurrence. When the distribution of mass movements along the coast is analyzed without considering the lithological variation, there is no relationship between the number and displaced volumes of mass movements and wave energy for each sector, with the majority of the movements and the greater volumes occurring in the least energetic sector. Although subject to higher wave energy, capes and promontories are more resistant to erosion than the adjacent areas. On the other hand, the most sheltered areas (where beaches accumulate) are more prone to slope mass movements.
Fig. 16. Distribution of the probability of exceedance of the runout ratio (n ¼ 170).
5.2. Hazard and security zones Data collected in situ at 170 movements between 1995 and 2014 have resulted in the size distribution of runout ratio illustrated in Fig. 15. The distribution of the probability of exceedance shown in Fig. 16 allows setting security levels (inverse of hazard levels) for users depending on the runout ratio value (Table 1). The security level of 95% (risk level of 5%) is reached for R ¼ 1.5, that is to say that if a mass movement occurs on a beach backed by cliffs, a user resting on the sand at a distance of 1.5 height of the has a 95% probability of not being struck by debris that collapse (and 5% probability of being hit). This probability drops to 63% if the user is at a distance from the base equal to the height of the cliff and falls to 20% if he stands at a distance from the base equal to half the height of the cliff. The hazard level equals the level of security for a runout ratio value of 0.86.
5.3. Reducing risk by improving information Given the described concentration of mass movements on the beaches backed by sea cliffs cut on Miocene calcarenites and taking into account the intense occupation of beaches by tourists is the immediate conclusion that in the Algarve coastal management must focus on risk. On beaches, risk can be reduced in two ways: reducing exposure (removing users from areas prone to slope mass movements) or decreasing the vulnerability of the cliffs to slope mass movements acting with measures of active intervention. Provide information on the risk to the users of the cliffs and beaches is basic low-cost preventive action. After the accident occurred at Maria Luisa beach the Algarve coastal authority proceeded to significant enhancement of rock fall hazard signaling. Information boards with the definition of risk areas were placed on every access to the beaches. This initiative aims to provide all tourists with information and knowledge of hazard areas at each beach. The distribution of the runout set enabled to draw a board model that discriminates two bands with different hazard levels,
Table 1 Hazard and security level as a function of runout ratio value.
Fig. 15. Runout ratio distribution of all mass movement types (170 events).
Exceedance probability (hazard level)
Security level
Runout ratio
50% 25% 10% 5% 0%
50% 75% 90% 95% 100%
0.86 1.11 1.38 1.50 2.0
292
S.B. Teixeira / Ocean & Coastal Management 102 (2014) 285e293
Fig. 17. Model of sea cliff hazard sign.
limited by the line of R ¼ 0.86 (Fig. 17). The area of high hazard contains the fringe between the cliff base and a distance of 0.86 the cliff height, which corresponds to the zone in the event of a mass movement occurs, the probability of being occupied by debris is greater than the probability of not being occupied. The range of moderate risk is contained in the range of 0.86e1.5 runout ratio, the area in which, in the event of a mass movement occurs, the probability of being hit by mass movement debris is less than 50%. 6. Conclusions The pattern of occupation of the Algarve tourism and the natural geodynamics of the rocky sea cliffs, reflected in the discontinuous and intermittent occurrence of slope mass movements, determining the existence of risk to users of beaches backed by rocky cliffs. The continuous inventory of mass movements collected during nineteen years in the Barlavento Coast (46 km of cliffs that support 15 km of pocket beaches) shows that on average 13 mass movements occur per year, mainly rock fall type, concentrated during the winter and early spring. However, during periods of increased beach use occur a considerable number of movements: during Easter holidays (April) about 16% of annual slope mass movements occur and during the official bathing season (JuneeSeptember) occur 4% of annual movements. The spatial distribution of slope mass movement shows that 78% of movements occur in cliffs that support beaches, in contrast in natural resistant headlands and promontories occur only 22% of mass movements. The analysis of the size distribution of the runout ratio discussed in this article confirms that the coastal plan regulations are well adjusted to the field reality. The width of sea hazard area as defined in those regulations corresponds to a security level of 95% for users of the beaches. The size distribution of the runout ratio permits the design of continuous distribution of levels of security and hazard
along the sandy beaches, which may be reproduced cartographically. As a preventive measure of risk, the information collected may be provided to users, on boards at the beach entrance, in order to contribute to a more aware and safe use of the beach area available. Mapping of hazard areas should be updated whenever a mass movement that will significantly change the configuration of the cliff occurs. Signaling is the cheapest action to prevent risk on beaches backed by rocky cliffs. On beaches with increased use or where the cliffs show higher activity other measures progressively more expensive can be used, like beach nourishment. Beach nourishment of pocket beaches, besides the increased protection of the cliffs under the action of the sea, provides extra area available for sunbathers out of the cliffs hazard zones. The hazard mapping presented in this article is very relevant information in the design of any sandy replenishment project. Acknowledgments The author thanks to Marcos Rosa, Celso Pinto, Ricardo Gomes and Fernando Engr acia for their assistance in fieldwork. The author gratefully thanks the anonymous reviewers for their helpful comments and suggestions that significantly improved the manuscript. This is a contribution of SHORE Project (PTDC/MAR-EST/3485/2012) funded by the Portuguese Foundation for Science and Technology (FCT). References ^ncia Portuguesa do Ambiente (APA) www page, http://snirh.inag.pt. Age Barlow, J., Lim, M., Rosser, N., Petley, D., Brain, M., Norman, E., Geer, M., 2011. Modelling cliff erosion using negative power law scaling of rockfalls. Geomorphology 139e140, 416e424. http://dx.doi.org/10.1016/ j.geomorph.2011.11.006.
S.B. Teixeira / Ocean & Coastal Management 102 (2014) 285e293 Taborda, R., 2011. Influence of wave action and Bezerra, M.M., Moura, D., Ferreira, O., lithology on sea cliff mass movements in Central Algarve coast, Portugal. J. Coast. Res. 27 (6A), 162e171. http://dx.doi.org/10.2112/JCOASTRES-D-11-00004.1. ~o, M., Boski, T., Moura, D., Dias, R., Marques da Silva, C., Santos, A., Cacha Pimentel, N., Cabral, J., 1998. Integrated framework of Neogene and Quaternary sedimentar units of the Algarve (Portugal). Comun. Inst. Geol. Min. 84 (1), A169eA172 (in Portuguese). Costa, M., Silva, R., Vitorino, J., 2001. Contribution to the study of the wave climate on the Portuguese coast. In: Com. 2ª Jornadas Portuguesas de Engenharia ria. Associaç~ Costeira e Portua ao Nacional de Navegaç~ ao, Sines. CD-ROM, 20p (in Portuguese). Cruden, D.M., Varnes, D.J., 1996. Landslide types and processes. In: Turner, A.K., Schuster, R.L. (Eds.), Landslides: Investigation and Mitigation. Transport Research Board, pp. 36e75. Special Report, 247. Copons, R., Vilaplana, J.M., Linares, R., 2009. Rockfall travel distance analysis by using empirical models (Sol a d' Andoraa la Vella, Central Pyrenees). Nat. Hazards Earth Syst. Sci. 9, 2107e2118. http://dx.doi.org/10.5194/nhess-9-2107-2009. Del Rio, L., Gracia, F.J., 2009. Erosion risk assessment of active coastal cliffs in temperate environments. Geomorphology 112, 82e95. http://dx.doi.org/ 10.1016/j.geomorph.2009.05.009. ^nio, B., Ze ^zere, J.L., Neves, M., 2013. Identification of hazardous zones Epifa combining cliff retreat with landslide susceptibility assessment. In: Conley, D.C., Masselink, G., Russell, P.E., O'Hare, T.J. (Eds.), Proceedings 12th International Coastal Symposium (Plymouth, England), Jour. Coastal Res., SI 65, pp. 1681e1686. http://dx.doi.org/10.2112/SI65-284.1. Hall, J.W., Meadowcroft, I.C., Lee, E.M., Van Gelder, P.H.A.J.M., 2002. Stochastic simulation of episodic soft cliff recession. Coast. Eng. 46, 159e174. http:// dx.doi.org/10.1016/S0378-3839(02)00089-3. Hungr, O., Corominas, J., Eberhardt, E., 2005. Estimating landslide motion mechanism, travel distance and velocity. State of the Art Paper #4, Estimating landslide motion mechanism, travel distance and velocity. In: Hungr, O., Fell, R., Couture, R., Eberhardt, E. (Eds.), Landslide Risk Management. Proceedings, Vancouver Conference. Taylor and Francis Group, London. ^s do Mar e da Atmosfera (IPMA). Daily Weather Forecast. www Instituto Portugue page, http://www.ipma.pt. Instituto Nacional de Estatística (INE), 2012. Statistical Yearbook of Algarve Region, Lisbon. Katz, O., Mushkin, A., 2013. Characteristics of sea-cliff erosion induced by strong winter storm in the eastern Mediterranean. Quat. Res. 80, 20e32. http:// dx.doi.org/10.1016/j.yqres.2013.04.004. Lee, E.M., Clark, A.R., 2002. Investigation and Management of Soft Rock Cliffs. Thomas Telford, London, p. 382. Lim, M., Rosser, N.J., Allison, R.J., Petley, D.N., 2010. Erosional processes in the hard rock coastal cliffs at Staithes, North Yorkshire. Geomorphology 114, 12e21. http://dx.doi.org/10.1016/j.geomorph.2009.02.011. Manupella, G., 1992. Geological Map of the Algarve Region. Scale 1/100.000. Serv. Geol. Port., Lisbon (in Portuguese).
293
Marques, F.M.S.F., 1994. Sea cliff evolution and related hazards in miocene terranes of Algarve (Portugal). In: 7th Int. Cong. of the I.A.E.G. Lisbon, Proc, vol. 4, pp. 3109e3118. Marques, F.M.S.F., 1997. The Sea Cliffs of the Coast of Algarve, Dynamics, Processes and Mechanisms (PhD. thesis). University of Lisboa, p. 556 (in Portuguese). Marques, F.M.S.F., 2003. Landslide activity in Upper Paleozoic shale sea cliffs: a case study along the western coast of the Algarve (Portugal). Bull. Eng. Geol. Environ. 62, 299e313. http://dx.doi.org/10.1007/s10064-003-0204-7. Marques, F.M.S.F., 2008. Magnitude-frequency of sea cliff instabilities. Nat. Hazards Earth Syst. Sci. 8, 1161e1171. http://dx.doi.org/10.5194/nhess-8-1161-2008. Marques, F.M.S.F., 2009. Sea cliff instability hazard prevention and planning: examples of practice in Portugal. J. Coast. Res. 856e860. SI 56. Marques, F., Matildes, R., Redweik, P., 2011. Statistically based sea cliff instability hazard assessment of Burgau-Lagos coastal section (Algarve, Portugal). J. Coast. Res. 927e931. SI 64. Michoud, C., Derron, M.H., Horton, P., Jaboyedoff, Baillifard, F.J., Loye, A., Nicolet, P., Pedrazzini, A., Queyrel, A., 2012. Rockfall hazard and risk assessment along roads at a regional scale: example in Swiss Alps. Nat. Hazards Earth Syst. Sci. 12, 615e629. http://dx.doi.org/10.5194/nhess-12-615-2012. Moore, L.J., Griggs, G.B., 2002. Long-term cliff retreat and erosion hotspots along the central shores of the Montery Bay National Marine Sanctuary. Mar. Geol. 181, 265e283. http://dx.doi.org/10.1016/S0025-3227(01)00271-7. Shaefer, M., Clifton, J., Baily, B., Moura, D., Loureiro, C., 2009. Nunes, M., Ferreira, O., Hazard assement in rocky cliffs at Central Algarve (Portugal): a tool for coastal management. Ocean & Coast. Manag. 52, 506e516. http://dx.doi.org/10.1016/ j.oceaman.2009.08.004. Sunamura, T., 1992. Geomorphology of Rocky Coasts. John Wiley & Sons, p. 302. Stravou, A., Lawrece, R.N., Mortimore, R.N., Murphy, W., 2011. A geotechnical and GIS based method for evaluating risk exposition along coastal cliff environments: a case study of the chalk cliffs of southern England. Nat. Hazards Earth Syst. Sci. 11, 2997e3011. http://dx.doi.org/10.5194/nhess-11-2997-2011. Teixeira, S.B., 2003. Climatic Agents Activity and Slope mass Movements in the ^ncias da Terra (UNL), Lisbon n esp. Miocene Cliffs on the Algarve (Portugal). Cie V, CD-ROM, L61eL64. (in Portuguese). Teixeira, S.B., 2006. Slope mass movements on rocky sea-cliffs: a power-law distributed natural hazard on the Barlavento Coast, Algarve, Portugal. Cont. Shelf Res. 26, 1077e1091. http://dx.doi.org/10.1016/j.csr.2005.12.013. Teixeira, S.B., Dores, A., 2013. Characterization of accidents in sea cliffs dissociated of natural geodynamics in the Algarve (Portugal) coast during the last decade ~es VII Congresso sobre Planeamento e Gesta ~o das (2003-2012). In: Comunicaço zonas Costeiras dos Países de Express~ ao Portuguesa. Assoc. Portug. Rec. Híd., Maputo, Moçambique, p. 3 (in Portuguese). Terrinha, P.A.G., 1998. Structural Geology and Tectonic Evolution of the Algarve Basin, South Portugal (Ph.D. thesis). Imperial College, London, p. 430 (unpublished). Young, A.P., Guza, R.T., O'Reilly, W.C., Flick, R.E., Gutierrez, R., 2011. Short-term retreat statistics of a slowly eroding coastal cliff. Nat. Hazards Earth Syst. Sci. 11, 205e217. http://dx.doi.org/10.5194/nhess-11-205-2011.
Contents lists available at ScienceDirect
Ocean & Coastal Management journal homepage: www.elsevier.com/locate/ocecoaman
Coastal hazards from slope mass movements: Analysis and management approach on the Barlavento Coast, Algarve, Portugal ~o Braz Teixeira Sebastia Portuguese Environment Agency, Rua do Alportel 10, 8000-503 Faro, Portugal
a r t i c l e i n f o
a b s t r a c t
Article history: Received 19 May 2014 Received in revised form 12 October 2014 Accepted 14 October 2014 Available online
The Barlavento Coast, Algarve, Portugal, is dominated by rocky sea-cliffs, cut on Miocene calcarenites which evolves through intermittent and discontinuous events of slope mass movements, along a 46 km cliff front. Here, the main coastal geologic hazards result from the conflict between human occupation and sea-cliff recession. Most of the research on the dynamics of the cliffs has been directed to the risk with the aim of defining long term set-back lines, for a preventive planning of the cliff top occupation. Little attention has been given to the hazard associated with mass movements on bathing beaches backed by sea-cliffs. This article presents the results of a field inventory of 244 slope mass movements single events, collected in a rocky shore with tens of touristic pocket beaches, covering an nineteen year time span (1995e2014). Results show that landslides have seasonal pattern with higher incidence in the period between winter and early spring. More than 15% of movements occur during the Easter holidays (April) and 4% of landslides occur during the official bathing season (JuneeSeptember). The spatial distribution of landslides shows that only 22% of the mass movements occur in capes and headlands, while 78% occur on the beaches, which demonstrates that the beaches are real hot spots of risk. Based on the size distribution of slope mass movements runout ratio (the ratio between the radius of the base of the cone of and the height of the movement) a table of levels of security and hazard on beaches was built. Security levels enable the definition of cartographic hazard areas on beaches which can be provided to the beach users on information boards at the beach entrance. © 2014 Elsevier Ltd. All rights reserved.
Keywords: Rocky sea cliffs Hazard areas Beaches Runout ratio Algarve Portugal
1. Introduction Algarve coastal region, with 451,000 inhabitants in 2011, is the premier tourism destination of Portugal, with a mean annual of 15 million overnight stays through the 2002e2011 period (INE, 2012). Almost 90% of the visitors' reasons of staying in the Algarve coast are leisure, recreational and holiday (INE, 2012). The Barlavento Coast between Lagos and Albufeira (Fig. 1) is the core touristic area and receives 60% of the Algarve visitors attracted with the “sun and beach”, which is the Algarve' s top touristic product. About 50% of the Algarve visitors stay and bathe in beaches backed with rocky sea cliffs of the Barlavento Coast. Over the last three decades, there has been record of several accidents caused by the collapse of sea cliffs cut on Miocene rocks. On 22 March 1998, a Portuguese man was killed while fishing at the das Porcas site, when a sudden planar landcliff edge at the Mare slide dragged him down together with a volume of 2 104 m3 falling material; on 7 October 2000, three Swiss tourists were
E-mail address: [email protected]. http://dx.doi.org/10.1016/j.ocecoaman.2014.10.008 0964-5691/© 2014 Elsevier Ltd. All rights reserved.
injured by a block fall (volume 2 m3) on the Inatel beach; on 21 August 2009 an instantaneous topple (volume 1 103 m3) on a sea stack killed five Portuguese tourists and injured another two, resting on the cliff base on the Maria Luísa beach; on 26 May 2010 a four year Irish kid was slightly wounded on the Vau beach hit by the debris of a landslide (Fig. 2, volume 2 102 m3); on 11 October 2010 a German tourist was injured on the Beijinhos beach, hit by blocks of a small landslide (volume 1 m3). This record shows that the average number of accidents resulting directly from cliff collapse of sea cliffs is 2 event/decade, causing two fatalities and two injuries each decade. Searching for the accidents of the cliff top walkers, mostly sightseers and fishing anglers, in the decade 2003e2012, Teixeira and Dores (2013) identified a record of 50 accidents on the Barlavento rocky cliffs with 11 fatalities and 41 injured. 45% of the victims were foreigner tourists. Although statistically not very significant when compared to accidents resulting from the use of the top of the cliffs, accidents caused by landslides on beaches have great impact on public opinion. While accidents of suicide or death of fishermen by falling from the cliffs typically occupy small news in local newspapers, after the collapse recorded in Maria Luisa beach
286
S.B. Teixeira / Ocean & Coastal Management 102 (2014) 285e293
Fig. 1. Barlavento coast. Study area.
the accident was the subject of daily news in all national newspapers for more than a month. Five years on, the accident is still remembered every year at the start of the bathing season (JuneeSeptember). Slope mass movements on steep rocky sea cliffs are instantaneous phenomena with virtually no warning. The velocity of the movement is typically over 5 m/s, within the landslide velocity class 7, the extremely rapid events of the Cruden and Varnes (1996) classification. On the Barlavento rocky coast, where sea cliffs are subvertical with height varying from 5 to 40 m, the duration of the slope mass movements lies in a narrow time window of 1e2 s. The instantaneous nature of the phenomenon precludes any action to minimize the damage after the onset of the collapse process. On steep subvertical rocky sea cliffs prone to slope mass movements, the actions to minimize risk and damages are therefore exclusively based on prevention. The first step in the implementation of prevention is the knowledge and definition of the areas potentially affected by a mass movement, i.e. the spatially definition of hazard areas. Hazard areas, limited by hazard lines, correspond to areas parallel to the shoreline where, in a pre-defined period, it is likely that effects of slope mass movements will be felt (Fig. 3). Most studies and bibliography, oriented strand planning, focuses on the land hazard areas at the top of the cliffs worldwide
(see, for example, Hall et al., 2002; Lee and Clark, 2002; Moore and Griggs, 2002; Del Rio and Gracia, 2009; Stravou et al., 2011; Epif^ anio et al., 2013) and on Algarve (Marques, 1994, 1997, 2003; Teixeira, 2003, 2006; Bezerra et al., 2011; Marques et al., 2011; Nunes et al., 2009). Very few papers deal with sea hazard areas on sea cliffs (Marques, 2009), although studies on the identification of risk areas associated with damage caused by slope mass movements are very common in land areas (e.g, Copons et al., 2009; Michoud et al., 2012). In the case of touristic areas centered on beaches backed by cliffs, the sea hazard area has particular interest in that it is in this area where accidents occur with people affected by the debris of a mass movement. In Portugal, beach tourism is an important economic activity, all the coastal plans regulations include a seaward hazard area on beaches backed by sea-cliff extending seawards from the cliff toe, where beach support structures are interdicted (Marques, 2009). On the study area coastal plan regulations, in force since 1999, a sea hazard area is defined on beaches backed by sea cliffs with a width of 1.5 times the cliff height and corresponds to the maximum extent of cliff failures debris displacement near the toe. In this paper we present the results of slope mass movements inventories gathered on Barlavento Coast sea cliffs, for the last nineteen years (1995e2014), we assess the adequacy of the legal
Fig. 2. Rock fall on the Vau beach occurred in the 26 May 2010 (location on Fig. 1); mean width (Wm) ¼ 2.5 m; runout ratio (R) ¼ 0.8.
S.B. Teixeira / Ocean & Coastal Management 102 (2014) 285e293
287
Fig. 3. Hazard areas on rocky sea cliffs.
provisions of the hazard areas and we propose a solution to spatially define the seaward hazard lines on beaches backed by rocky cliffs prone to slope mass movements. 2. Study area The south coast of the Algarve, Portugal, located at South Western end of Europe, has a typical Mediterranean climate with hot dry summers and mild rainy winters. Average annual precipitation is 500e600 mm, 80% of which is concentrated in a wet season (OctobereMarch). Wave regime is moderate with mean annual significant height of 1 m (Costa et al., 2001) and 85% of sea storms (Hs 2.5 m) occur in the wet season Tide is semi-diurnal; mean tidal range is 2 m and spring-tide range is 3 m. The Barlavento Coast is a rocky coast with a continuous front of subvertical sea-cliffs cut in carbonate Mesozoic and Miocene rocks. s beach and The study area is the cliffed coast, between Porto de Mo Olhos de Agua beach (Fig. 1), cut in Miocene rocks, included in the ~o Formation (Cacha ~o et al., 1998). The total length of Lagos-Portima sea cliffs cut in Miocene is 46 km as measured using detailed topographic maps (scale 1:2000). Sea cliffs are composed of alternate decimetric layers of fine grained calcarenites and calcarenites with high content of macrofossils (Manupella, 1992). Carbonate content is 60e75% in fine-grained calcarenites and greater than 80% in fossiliferous calcarenites (Marques, 1997). Cliff height varies from 5 to 40 m and displays a sequence of strata, horizontal or gently sloping to the South or Southeast except in the vicinity of major tectonic features of the area (the Portim~ ao fault and Albufeira diapir; Terrinha, 1998) where the slope is greater (10e20 ). Coastal Miocene formations have a deep and well developed karst, with a dense network of caves, sinkholes and galleries covered by PlioQuaternary (Manupella, 1992) reddish silty-clayed sands. The combination of the network of karsic cavities with the action of coastal erosion provides a modeling of the coastline very diverse and irregular, very attractive for tourists that seek the region.
The quantitative assessment of cliff retreat rates in Miocene calcarenites was first performed by Marques (1994, 1997) based on identification and measurements of slope mass movements by comparative analysis of aerial photographs. The recession rate based on values of lost area of mass movements in the 1947e1991 period is 1e2 cm/year (Marques, 1997). Based on a continuous field inventory, of mass movements on Miocene cliffs, started in 1995, Teixeira (2003, 2006) calculated a mean recession rate of 1 cm/year. The maximum local retreat measured by Marques (1997) was 45 m, in an arch collapse in plunging cliffs that occurred between 1974 and 1980; in the 1995e2014 period the maximum local retreat measured was 25 m in a head collapse occurred in 6 October 1998 in Marinha beach (Fig. 4). The irregular shape of the coast promotes the formation of seventy small pocket beaches, with lengths of tens to hundred meters, and thin sand cover about 2 to 4 m thick, accumulated over cut shore platforms. Most beaches (80%) are accumulated in the irregularities generated by erosion of the cliffs. Only sixteen of the beaches are located in small river mouths, four of them in hanging valleys. One third of the cliff front (15 km) has a beach on their base. The total length of the fifty official bathing beaches extends for 9 km. 3. Methods The regional coastal management authority of Algarve (Portuguese Environment Agency e APA) has been conducting a continuous observation of the coast since 1995, which includes the systematic recording of slope mass movements occurring on the front of rocky sea cliffs cut on Miocene calcarenites. The very first source of information of the slope movements is the reporting of occurrences by public regional or local field authority staff, as well as by the tourism operators, private owners and fishermen. Complementarily, a periodic systematic observation in situ of the coast, by land, sea and air, is conducted several times a year,
~o Vistal 1992; mean width (Wm) ¼ 20 m; Runout ratio Fig. 4. Head collapse and generation of a new stack in Marinha beach, occurred on 6 October 1998. Postcard from Ediça (R) ¼ 0.5.
288
S.B. Teixeira / Ocean & Coastal Management 102 (2014) 285e293
Fig. 5. Topple occurred on 2 Out 2009, in Santa Eulalia beach; (location on Fig. 1) Wm ¼ 1.5 m; R ¼ 1.9.
especially after wave or rainy storms. At least, once a year a campaign of capture of oblique photographs either on a low altitude flight (since 2006) or on a boat tour (since 2002) is done in order to get a complete view of the area and identify recent slope mass movements. Fresh scars on the sea cliff face, detritus cones with fine sediment content, and a yellow plume in the water, resulting from washing out of the debris, provide the most obvious signs of a recent mass movement (Teixeira, 2006). The main source of the inventory comes from periodical observation of the sea cliff front that enabled the identification of a considerable amount of unreported slope mass movements (66% of the inventory). The gathered inventory covering a nineteen year time span (July 1995eJune 2014) has a total of 244 slope mass movements, with a mean length 1 m, which is the threshold of completeness of in situ inventories on Barlavento Miocene calcarenites (Teixeira, 2006). After each report of an occurrence or during the periodical field observation, a trip to the site is carried out and a characterization of mass movement type and geometry is performed in situ as described in Teixeira (2006). In the field inventory (244 events), mass movements were classified onto four types, according to the
generic mass movements proposed by Sunamura (1992) and Marques (1997): rock fall (Fig. 2), topple (Fig. 5), karst collapse (Fig. 6) and block fall (Fig. 7). For the purpose of the present study, the radius of the debris cone (the distance between the cliff base and the seaward limit of the debris), geometrically the same as the runout or travel distance used on landslide motions, (Hungr et al., 2005; Michoud et al., 2012) is an important parameter, crucial for computing the runout ratio (R) parameter. The runout ratio of a slope mass movement is defined as the ratio between the radius of the base of the cone of debris and the height of movement, a dimensionless scaling parameter. When the mass movement reaches the total cliff face, R is equal to the ratio between the radius of the base of the cone of debris and the cliff height (Fig. 8). A cliff collapse in which the cone of debris projecting from the base of the cliff at a distance equal to the height of the cliff has a runout ratio 1; a collapse in which the cone of debris occupies the sandy beach up to a distance equal to the height of 1.5 times the cliff height has a runout ratio 1.5. The measure of the runout is sometimes impossible in mass movements identified weeks after the occurrence as waves tend to
^ra beach, in February 2011 (location on Fig. 1); Wm ¼ 1.5 m; R ¼ 1.5. Fig. 6. Karst collapse in Armaç~ao de Pe
Fig. 7. Block fall on Albandeira beach, occurred on 12 September 2007 (location on Fig. 1); Wm ¼ 1.5 m; R ¼ 0.9.
S.B. Teixeira / Ocean & Coastal Management 102 (2014) 285e293
289
Fig. 8. Definition of the runout ratio of a slope mass movement on steep rocky sea cliff.
Fig. 9. Stack collapse occurred on 1 Dec 2010, in Olhos de Agua beach.
rapidly wash out and remove the debris. In movements recorded in plunging cliffs and promontories (22% of the field inventory) sometimes is not possible to calculate the runout as the cone of debris extends to the immersed area below the low tide line, making it difficult to measure the radius of the debris cone. For a total of 244 registered mass movements in the 1995e2014 period, the R was computed in only 170 (70% of the field inventory). 4. Results After each report of an occurrence or during the periodical field observation, a characterization of mass movement type and geometry was performed in situ, measuring mass movement mean and maximum linear parameters (length, height, width), directly on the cliff scar, as described by Teixeira (2006). The horizontal area of loss was computed as the product of mean length and mean width, and the volume of the mass movement was calculated as the product of horizontal area and mean mass movement height. The volume of the fresh debris produced by the mass movement provided a supplementary control of the evaluation of dimensional parameters. Between July 1995 to June 2014, covering a nineteen years span, 244 slope mass movements on sea cliffs, have been recorded and
Fig. 10. Cumulative absolute frequency size of mean width of mass movements occurred between July 1995 and June 2014 (n ¼ 244).
measured, spatially dispersed along the sea cliff front. Rock fall is the dominant movement (61%). Topples (21%) and karst collapse (16%) have approximately the same importance. Block fall is the less frequent mass movement type (3%) During the study period five stacks collapsed (Fig. 9) and two new ones were formed (Fig. 4). A total of 2100 m of the sea cliff front (4.5% of the total cliff front length) was altered by slope mass movements, during the nineteen year period of observation. The average value of the renewal of the cliff front of 0.24%/year gives an estimate for the cliffs mean life period of 425 years. The average loss of horizontal area was 400 m2/ year, the mean annual volume was 7400 m3, and the average recession rate was 9 mm/year, which are values similar to previous estimates obtained by Marques (1997) and Teixeira (2003, 2006). The field inventory showed an average frequency of 13 movements/year with a range of 2e42 annual movements. The series of mass movement mean width (the width of the instantaneous retreat) show a magnitude decay pattern, i. e. the wider the mass movements are, the less frequent they are (Figs. 10 and 11) as previously showed by Teixeira (2006). Increasingly, recent studies demonstrate that the dimensions (width, area or volume) of cliff failures fit to inverse power-law relationships (e. g. Marques, 2008; Lim et al., 2010; Young et al., 2011; Barlow et al., 2011, Katz and Mushkin, 2013).
Fig. 11. Return period distribution of mean width of mass movements occurred between July 1995 and June 2014 (n ¼ 244).
290
S.B. Teixeira / Ocean & Coastal Management 102 (2014) 285e293
Fig. 12. Annual distribution of maximum date uncertainty of slope mass movements. Year starts in July and ends in June.
At this stage, the destabilizing effect of the water seems to have an immediate effect. The second stage of failures that extends throughout the winter has the highest expression in early spring. Here the progressive increase in the frequency of collapses suggests the preponderance of the cumulative undermining effect of water. Storms and heavy rains in early spring seem more effective in triggering cliff failure than in autumn. In four movements it was possible to relate the collapse with direct human intervention. The most dramatic episode was the break in a swimming pool in June 1997 previously reported by Teixeira (2006). The other three are associated with the concentration of runoff resulting from poor sealing and channeling rainwater in May 1997, October 2008 and March 2010. In 18% of the monthly inventory the trigger is unknown, possibly associated with fatigue failure of the cliff material. Only 4% of mass movements occurred during the official bathing season (June to September) but three of the eight recorded cliff failure resulted in casualties (5 fatalities and 6 wounded).
4.1. Date uncertainty The effort for permanent observation of the 46 km long coast and the level of dating of the slope mass movements is such that 77% of the events are monthly dated and less than a half of the events are daily dated (46%). In 94% of the occurrences the date of the event is within a tree month period. Maximum uncertainty of the date of the movements is less than a half year (Fig. 12). 4.2. Monthly distribution The dataset dated monthly (monthly inventory, 188 events, Fig. 13) enable more detailed analysis of the annual distribution of mass movements and their relationship to the beach use. The seasonal distribution of monthly dated movements shows clear seasonal pattern with predominance of cliff failures during the wet season, where 83% of cases occurred between November and April. In half of the movements in which the exact day is known the collapse occurred under sea storms (Hs 2.5 m) or daily rainfall exceeding 10 mm. 78% of failures occurred within five days after storms or heavy rain. A closer and detailed look at the monthly distribution of the collapses suggests a two stage dynamics. The forerunner stage happens at the beginning of autumn when a first peak of collapses occurs in October triggered by the first heavy rains and sea storms.
4.3. Runout ratio size distribution The amplitude of the runout ratio covers the 0.1e1.9 range and varies with the type of mass movement (Fig. 14). The most frequent type of movement (rock fall, 98 records) presents a normal type distribution with a mean value of 0.90. In the case of topples (34 records), with more pronounced horizontal component, the distribution also resembles a normal distribution with mean value 1.11, but with greater variance. The karst collapse failures produce the lowest R (mean 0.49, 30 records). Values show great dispersion, with a mode centered in the lower values associated with the prevalence of movements with almost exclusive vertical component. In the case of block fall (8 records) values have very tiny dispersion, with all values in between 0.75 and 1.0. Aggregating the result set (Fig. 15) it is apparent that in all the results fit a normal distribution (mean 0.88), with a secondary mode on values of the lower runout ratio, associated with movements with predominance of the vertical component. 5. Discussion 5.1. Beaches as hot spots Of the 244 mass movements recorded in field inventory, 190 (78%) occurred in the cliffs of the beaches. The remaining 54 (22%)
^ncia Portuguesa do Ambiente); sea storm data from Fig. 13. Monthly distribution of slope mass movements (n ¼ 188), rain and sea storms. Precipitation data from Algoz station (Age ^s do Mar e Atmosfera (IPMA). Instituto Portugue
S.B. Teixeira / Ocean & Coastal Management 102 (2014) 285e293
291
Fig. 14. Runout ratio distribution according to slope mass movement type (n ¼ 170).
were recorded in plunging cliffs and promontories. Over the observation period the field inventory gives a ratio of 12 movements/km on beaches whereas in the remaining area that ratio is 1.7 movement/km. Rocky coast sheltered areas (where beaches accumulate) are sevenfold more prone to slope mass movements than exposed sites (headlands and promontories). From the point of view of coastal erosion Barlavento beaches are real hotspots. Results show that the most vulnerable areas are not located in the areas most exposed to wave attack despite wave action is one of the most effective triggers on the evolution of sea cliffs. Plunging cliffs are permanently exposed to direct wave action, promontories are daily exposed (at flood tide), while in cliffs backed by beaches wave action acts episodically under storm conditions and spring tides. This dynamic pattern of the Barlavento cliffs coincides with the evolutionary pattern of rocky cliffs defined by Sunamura (1992) in which plunging cliffs occur on resistant rocks and shore platforms develop on weaker rocks. Bezerra et al. (2011) used a wave propagation model to quantify breaking wave conditions and wave energy to estimate the relative role of wave action and lithology in controlling cliff erosion and evolution along a 12 km of Miocene calcarenites of Central Algarve. They found that lithology represents the dominant control on mass movement occurrence. When the distribution of mass movements along the coast is analyzed without considering the lithological variation, there is no relationship between the number and displaced volumes of mass movements and wave energy for each sector, with the majority of the movements and the greater volumes occurring in the least energetic sector. Although subject to higher wave energy, capes and promontories are more resistant to erosion than the adjacent areas. On the other hand, the most sheltered areas (where beaches accumulate) are more prone to slope mass movements.
Fig. 16. Distribution of the probability of exceedance of the runout ratio (n ¼ 170).
5.2. Hazard and security zones Data collected in situ at 170 movements between 1995 and 2014 have resulted in the size distribution of runout ratio illustrated in Fig. 15. The distribution of the probability of exceedance shown in Fig. 16 allows setting security levels (inverse of hazard levels) for users depending on the runout ratio value (Table 1). The security level of 95% (risk level of 5%) is reached for R ¼ 1.5, that is to say that if a mass movement occurs on a beach backed by cliffs, a user resting on the sand at a distance of 1.5 height of the has a 95% probability of not being struck by debris that collapse (and 5% probability of being hit). This probability drops to 63% if the user is at a distance from the base equal to the height of the cliff and falls to 20% if he stands at a distance from the base equal to half the height of the cliff. The hazard level equals the level of security for a runout ratio value of 0.86.
5.3. Reducing risk by improving information Given the described concentration of mass movements on the beaches backed by sea cliffs cut on Miocene calcarenites and taking into account the intense occupation of beaches by tourists is the immediate conclusion that in the Algarve coastal management must focus on risk. On beaches, risk can be reduced in two ways: reducing exposure (removing users from areas prone to slope mass movements) or decreasing the vulnerability of the cliffs to slope mass movements acting with measures of active intervention. Provide information on the risk to the users of the cliffs and beaches is basic low-cost preventive action. After the accident occurred at Maria Luisa beach the Algarve coastal authority proceeded to significant enhancement of rock fall hazard signaling. Information boards with the definition of risk areas were placed on every access to the beaches. This initiative aims to provide all tourists with information and knowledge of hazard areas at each beach. The distribution of the runout set enabled to draw a board model that discriminates two bands with different hazard levels,
Table 1 Hazard and security level as a function of runout ratio value.
Fig. 15. Runout ratio distribution of all mass movement types (170 events).
Exceedance probability (hazard level)
Security level
Runout ratio
50% 25% 10% 5% 0%
50% 75% 90% 95% 100%
0.86 1.11 1.38 1.50 2.0
292
S.B. Teixeira / Ocean & Coastal Management 102 (2014) 285e293
Fig. 17. Model of sea cliff hazard sign.
limited by the line of R ¼ 0.86 (Fig. 17). The area of high hazard contains the fringe between the cliff base and a distance of 0.86 the cliff height, which corresponds to the zone in the event of a mass movement occurs, the probability of being occupied by debris is greater than the probability of not being occupied. The range of moderate risk is contained in the range of 0.86e1.5 runout ratio, the area in which, in the event of a mass movement occurs, the probability of being hit by mass movement debris is less than 50%. 6. Conclusions The pattern of occupation of the Algarve tourism and the natural geodynamics of the rocky sea cliffs, reflected in the discontinuous and intermittent occurrence of slope mass movements, determining the existence of risk to users of beaches backed by rocky cliffs. The continuous inventory of mass movements collected during nineteen years in the Barlavento Coast (46 km of cliffs that support 15 km of pocket beaches) shows that on average 13 mass movements occur per year, mainly rock fall type, concentrated during the winter and early spring. However, during periods of increased beach use occur a considerable number of movements: during Easter holidays (April) about 16% of annual slope mass movements occur and during the official bathing season (JuneeSeptember) occur 4% of annual movements. The spatial distribution of slope mass movement shows that 78% of movements occur in cliffs that support beaches, in contrast in natural resistant headlands and promontories occur only 22% of mass movements. The analysis of the size distribution of the runout ratio discussed in this article confirms that the coastal plan regulations are well adjusted to the field reality. The width of sea hazard area as defined in those regulations corresponds to a security level of 95% for users of the beaches. The size distribution of the runout ratio permits the design of continuous distribution of levels of security and hazard
along the sandy beaches, which may be reproduced cartographically. As a preventive measure of risk, the information collected may be provided to users, on boards at the beach entrance, in order to contribute to a more aware and safe use of the beach area available. Mapping of hazard areas should be updated whenever a mass movement that will significantly change the configuration of the cliff occurs. Signaling is the cheapest action to prevent risk on beaches backed by rocky cliffs. On beaches with increased use or where the cliffs show higher activity other measures progressively more expensive can be used, like beach nourishment. Beach nourishment of pocket beaches, besides the increased protection of the cliffs under the action of the sea, provides extra area available for sunbathers out of the cliffs hazard zones. The hazard mapping presented in this article is very relevant information in the design of any sandy replenishment project. Acknowledgments The author thanks to Marcos Rosa, Celso Pinto, Ricardo Gomes and Fernando Engr acia for their assistance in fieldwork. The author gratefully thanks the anonymous reviewers for their helpful comments and suggestions that significantly improved the manuscript. This is a contribution of SHORE Project (PTDC/MAR-EST/3485/2012) funded by the Portuguese Foundation for Science and Technology (FCT). References ^ncia Portuguesa do Ambiente (APA) www page, http://snirh.inag.pt. Age Barlow, J., Lim, M., Rosser, N., Petley, D., Brain, M., Norman, E., Geer, M., 2011. Modelling cliff erosion using negative power law scaling of rockfalls. Geomorphology 139e140, 416e424. http://dx.doi.org/10.1016/ j.geomorph.2011.11.006.
S.B. Teixeira / Ocean & Coastal Management 102 (2014) 285e293 Taborda, R., 2011. Influence of wave action and Bezerra, M.M., Moura, D., Ferreira, O., lithology on sea cliff mass movements in Central Algarve coast, Portugal. J. Coast. Res. 27 (6A), 162e171. http://dx.doi.org/10.2112/JCOASTRES-D-11-00004.1. ~o, M., Boski, T., Moura, D., Dias, R., Marques da Silva, C., Santos, A., Cacha Pimentel, N., Cabral, J., 1998. Integrated framework of Neogene and Quaternary sedimentar units of the Algarve (Portugal). Comun. Inst. Geol. Min. 84 (1), A169eA172 (in Portuguese). Costa, M., Silva, R., Vitorino, J., 2001. Contribution to the study of the wave climate on the Portuguese coast. In: Com. 2ª Jornadas Portuguesas de Engenharia ria. Associaç~ Costeira e Portua ao Nacional de Navegaç~ ao, Sines. CD-ROM, 20p (in Portuguese). Cruden, D.M., Varnes, D.J., 1996. Landslide types and processes. In: Turner, A.K., Schuster, R.L. (Eds.), Landslides: Investigation and Mitigation. Transport Research Board, pp. 36e75. Special Report, 247. Copons, R., Vilaplana, J.M., Linares, R., 2009. Rockfall travel distance analysis by using empirical models (Sol a d' Andoraa la Vella, Central Pyrenees). Nat. Hazards Earth Syst. Sci. 9, 2107e2118. http://dx.doi.org/10.5194/nhess-9-2107-2009. Del Rio, L., Gracia, F.J., 2009. Erosion risk assessment of active coastal cliffs in temperate environments. Geomorphology 112, 82e95. http://dx.doi.org/ 10.1016/j.geomorph.2009.05.009. ^nio, B., Ze ^zere, J.L., Neves, M., 2013. Identification of hazardous zones Epifa combining cliff retreat with landslide susceptibility assessment. In: Conley, D.C., Masselink, G., Russell, P.E., O'Hare, T.J. (Eds.), Proceedings 12th International Coastal Symposium (Plymouth, England), Jour. Coastal Res., SI 65, pp. 1681e1686. http://dx.doi.org/10.2112/SI65-284.1. Hall, J.W., Meadowcroft, I.C., Lee, E.M., Van Gelder, P.H.A.J.M., 2002. Stochastic simulation of episodic soft cliff recession. Coast. Eng. 46, 159e174. http:// dx.doi.org/10.1016/S0378-3839(02)00089-3. Hungr, O., Corominas, J., Eberhardt, E., 2005. Estimating landslide motion mechanism, travel distance and velocity. State of the Art Paper #4, Estimating landslide motion mechanism, travel distance and velocity. In: Hungr, O., Fell, R., Couture, R., Eberhardt, E. (Eds.), Landslide Risk Management. Proceedings, Vancouver Conference. Taylor and Francis Group, London. ^s do Mar e da Atmosfera (IPMA). Daily Weather Forecast. www Instituto Portugue page, http://www.ipma.pt. Instituto Nacional de Estatística (INE), 2012. Statistical Yearbook of Algarve Region, Lisbon. Katz, O., Mushkin, A., 2013. Characteristics of sea-cliff erosion induced by strong winter storm in the eastern Mediterranean. Quat. Res. 80, 20e32. http:// dx.doi.org/10.1016/j.yqres.2013.04.004. Lee, E.M., Clark, A.R., 2002. Investigation and Management of Soft Rock Cliffs. Thomas Telford, London, p. 382. Lim, M., Rosser, N.J., Allison, R.J., Petley, D.N., 2010. Erosional processes in the hard rock coastal cliffs at Staithes, North Yorkshire. Geomorphology 114, 12e21. http://dx.doi.org/10.1016/j.geomorph.2009.02.011. Manupella, G., 1992. Geological Map of the Algarve Region. Scale 1/100.000. Serv. Geol. Port., Lisbon (in Portuguese).
293
Marques, F.M.S.F., 1994. Sea cliff evolution and related hazards in miocene terranes of Algarve (Portugal). In: 7th Int. Cong. of the I.A.E.G. Lisbon, Proc, vol. 4, pp. 3109e3118. Marques, F.M.S.F., 1997. The Sea Cliffs of the Coast of Algarve, Dynamics, Processes and Mechanisms (PhD. thesis). University of Lisboa, p. 556 (in Portuguese). Marques, F.M.S.F., 2003. Landslide activity in Upper Paleozoic shale sea cliffs: a case study along the western coast of the Algarve (Portugal). Bull. Eng. Geol. Environ. 62, 299e313. http://dx.doi.org/10.1007/s10064-003-0204-7. Marques, F.M.S.F., 2008. Magnitude-frequency of sea cliff instabilities. Nat. Hazards Earth Syst. Sci. 8, 1161e1171. http://dx.doi.org/10.5194/nhess-8-1161-2008. Marques, F.M.S.F., 2009. Sea cliff instability hazard prevention and planning: examples of practice in Portugal. J. Coast. Res. 856e860. SI 56. Marques, F., Matildes, R., Redweik, P., 2011. Statistically based sea cliff instability hazard assessment of Burgau-Lagos coastal section (Algarve, Portugal). J. Coast. Res. 927e931. SI 64. Michoud, C., Derron, M.H., Horton, P., Jaboyedoff, Baillifard, F.J., Loye, A., Nicolet, P., Pedrazzini, A., Queyrel, A., 2012. Rockfall hazard and risk assessment along roads at a regional scale: example in Swiss Alps. Nat. Hazards Earth Syst. Sci. 12, 615e629. http://dx.doi.org/10.5194/nhess-12-615-2012. Moore, L.J., Griggs, G.B., 2002. Long-term cliff retreat and erosion hotspots along the central shores of the Montery Bay National Marine Sanctuary. Mar. Geol. 181, 265e283. http://dx.doi.org/10.1016/S0025-3227(01)00271-7. Shaefer, M., Clifton, J., Baily, B., Moura, D., Loureiro, C., 2009. Nunes, M., Ferreira, O., Hazard assement in rocky cliffs at Central Algarve (Portugal): a tool for coastal management. Ocean & Coast. Manag. 52, 506e516. http://dx.doi.org/10.1016/ j.oceaman.2009.08.004. Sunamura, T., 1992. Geomorphology of Rocky Coasts. John Wiley & Sons, p. 302. Stravou, A., Lawrece, R.N., Mortimore, R.N., Murphy, W., 2011. A geotechnical and GIS based method for evaluating risk exposition along coastal cliff environments: a case study of the chalk cliffs of southern England. Nat. Hazards Earth Syst. Sci. 11, 2997e3011. http://dx.doi.org/10.5194/nhess-11-2997-2011. Teixeira, S.B., 2003. Climatic Agents Activity and Slope mass Movements in the ^ncias da Terra (UNL), Lisbon n esp. Miocene Cliffs on the Algarve (Portugal). Cie V, CD-ROM, L61eL64. (in Portuguese). Teixeira, S.B., 2006. Slope mass movements on rocky sea-cliffs: a power-law distributed natural hazard on the Barlavento Coast, Algarve, Portugal. Cont. Shelf Res. 26, 1077e1091. http://dx.doi.org/10.1016/j.csr.2005.12.013. Teixeira, S.B., Dores, A., 2013. Characterization of accidents in sea cliffs dissociated of natural geodynamics in the Algarve (Portugal) coast during the last decade ~es VII Congresso sobre Planeamento e Gesta ~o das (2003-2012). In: Comunicaço zonas Costeiras dos Países de Express~ ao Portuguesa. Assoc. Portug. Rec. Híd., Maputo, Moçambique, p. 3 (in Portuguese). Terrinha, P.A.G., 1998. Structural Geology and Tectonic Evolution of the Algarve Basin, South Portugal (Ph.D. thesis). Imperial College, London, p. 430 (unpublished). Young, A.P., Guza, R.T., O'Reilly, W.C., Flick, R.E., Gutierrez, R., 2011. Short-term retreat statistics of a slowly eroding coastal cliff. Nat. Hazards Earth Syst. Sci. 11, 205e217. http://dx.doi.org/10.5194/nhess-11-205-2011.