Wind directions of volcanic ash-charged clouds in Ecuador – implications for the public and flight safety

ABSTRACT Ecuador has about 20 active continental volcanoes, with a volcanic explosivity index (VEI) between 2 and 7. In the last 17 years, five of these volcanoes (Sangay, Guagua Pichincha, El Reventador, Cotopaxi and Tungurahua) have manifested their activity with thousands of eruptions, some of them even capable of compromising public health, closing airports and affect main national and international air routes. Therefore, these eruptions have been evaluated taking into account wind directions of the ash-charged clouds using the archives of the NOAA’s Ecuador Satellite Imagery of the Satellite Services Division of the National Environmental Satellite, Data, and Information Service. The main wind direction of these ash clouds demonstrate a clear uniform E–W direction for the months between April and September, while other months vary slightly from this main direction. As early warning systems for volcanic activity prior and even after eruptions have frequently failed for the last 17 years in Ecuador, the evaluated statistics of wind directions of the ash-charged clouds will improve the hazard evaluation and assessment. These statistics with a complementary interpretation of remote sensing will most likely be an enhanced tool in volcanic hazard assessment in Ecuador than it has been so far.


Introduction
Volcanoes and associated hazards have been responsible for the death of hundreds of thousands of persons in the last two centuries worldwide (Peterson 1988;Tanguy et al. 1998). They destroyed a variety of strategic infrastructure throughout the world, and changed the local and global climate (Hofmann & Rosen 1983;Self & Rampino 1988;Pinatubo Volcano Observatory Team 1991;Hansen et al. 1992;Briffa et al. 1998) as well as nearby landscapes (Blong 1984;Jago & Boyd 2005;Scott et al. 2010). Expelled pyroclastic material such as ash, pumice and bombs are one of the most underestimated yet most hazardous volcanic phenomena (Miller & Casadevall 1999;Self 2006;Barsotti et al. 2010;Dingwell & Rutgersson 2014). While bombs and other pyroclastic ejecta have a limited radius of action due to their size, density and weight, pumice and especially fine-grained ash can reach dozens or even thousands of kilometres from the emission centre causing damage or severely affecting agriculture and potable water systems and are threatening downwind the health and life of livestock and humans alike, causing respiratory problems derived from inhalation of fine ash among other affects (Wakisaka et al. 1978;Schiff et al. 1981;Akematsu et al. 1982;Martin et al. 1984;Buist & Bernstein 1986;Bernstein et al. 1986;Martin et al. 1986;Uda et al. 1999;OPS 2000;Baxter 2003;Forbes et al. 2003;Horwell & Baxter 2006;Naumova et al. 2007;Carlsen et al. 2012).
Pending of the vicinity of the volcanic ash fallout, damages can include roof collapses by accumulation of precipitated material becoming even heavier when it becomes wet by following rains (Newhall & Punongbayan 1997), road traffic interruptions and interruption and damage of electric power lines and of communication systems (Wilcox & Coats 1959;Sarkinen & Wiitala 1981;Bebbington et al. 2008;Guffanti et al. 2009;Wilson et al. 2012). Airport activities are also disrupted and air travel safety jeopardized (Witham et al. 2007). In fact, airborne ash is able to interfere dramatically with aerial navigation as well as landing and take-off operations (Casadevall et al. 1996;Self 2006;Guffanti et al. 2010;McCallie et al. 2011;Ulfarsson & Unger 2011;Budd & Ryley 2012). Airplanes inadvertently crossing ash-loaded clouds have experienced in-flight loss of engine power due to the accumulation and solidification of ash in turbine nozzle guide vanes, failure of high-bypass turbine engines, severe damage to turbine blades and to other fundamental external parts of the airplanes (Casadevall 1993;Przedpelski & Casadevall 1994;Casadevall et al. 1996;Miller & Casadevall 1999).
The growing market of air travel and the subsequent increase of the number of civil airports and air travellers led to the increase of the encounters between airplanes with volcanic ash clouds. Accumulated ash in the route of airplanes causes reduction and even loss of visibility, threatening the continuance and maintenance of the planned route reduces flight power and is also able to cause standstill of turbine engines (Kienle 1994;Bliss 2010). Closure of airports due to ash clouds generates high amount of losses in local and regional economy, as demonstrated in Argentine due to the eruption of the Chilean volcano Chaiten in 2008 or in Western Europe due to the Icelandic volcano Eyjafjallaj€ okull in 2010 (Folch et al. 2008;Martin et al. 2009;Guffanti et al. 2009;O'Dowd et al. 2012 Dingwell, andRutgersson 2014). In 20% of the reported cases of airport closure, airports were located as far as 500 km away from the active volcano ). In Ecuador, many airplanes were covered by ash in the airport of Quito and were unable to be used for a week due to the lack of warning by the national monitoring institute, although ash from El Reventador volcano travelled 5.5 hours until precipitation took place in that area (Reischmann et al. 2003;Baxter 2003;Hall et al. 2004;Johnston et al. 2005). The main purpose of our study is to analyze the wind directions of ash-charged clouds in the last 17 years taking into account different seasons in order to be able to predict with a better probability the routes of such ash clouds. Subsequently, with this information we might be able to suggest how to divert rapidly domestic and international flights from and to Ecuador and the Gal apagos. Thus, these statistics appear as an important tool in the establishment of contingence and emergency plans for public and aircraft safety.

Geodynamic setting and the Ecuadorian ash problem
The northern Andes in Ecuador are part of the 7000-km long classical example of an active continental margin with several volcanic sequences of Mesozoic and Cenozoic ages (Ramos 2009). The Ecuadorian volcanic arc appear in an NNEÀSSW strike as a result of the perpendicular collision between the Miocene oceanic Nazca plate which is subducted with an angle slightly oblique to the southern American continental lithosphere, which itself is composed of the Caribbean and South American continental plates (M eGard 1987;Colmenares & Zoback 2003;Dumont et al. 2005;Egbue & Kellogg 2010;Toulkeridis 2013). The Nazca Plate incorporates the aseismic Carnegie Ridge, which was formed by the passing of the ESE moving Nazca Plate over the Galapagos hot spot (Johnson and Lowrie 1972;Freymuller et al. 1993;Werner et al. 2003;Toulkeridis 2011). The up to 250 volcanoes on the volcanic front, main, back and rear arc are distributed along four distinctive zones, namely the Western, the Interandean, the Eastern and the Subandean volcanic cordilleras (Toulkeridis 2013) and belong to the Andean Northern Volcanic Zone (Barberi et al. 1988;Bryant et al. 2006;. Near 20 Ecuadorian volcanoes are considered to be active for the last century, five of these erupted in the last 17 years, therefore the airport and airplane safety is a fundamental matter and subject of serious hazard assessment (Figures 1 and 2). The domestic flights have routes, which cover almost the whole country and the Islands of Gal apagos. The major airports, however, are located within the Ecuadorian volcanic arcs or on their western side. The major airports with their domestic and international flight routes are illustrated in Figure 3.

Methodology
There are a number of tracking models to refine and simulate the movement of airborne ash in near real-time providing rapid notification of eruptive activity (Krueger et al. 1994;Schneider et al. 1995;Searcy et al. 1998;Lacasse et al. 2004;Marzano et al. 2006;Folch et al. 2008;Mastin et al. 2009;Folch 2012). The idea to monitor and forecast the trajectories of ash clouds is fundamental, as the crossing of a volcanic ash-charged cloud may lead to failure of the airplane and the loss of air traveller's life. Therefore, to some extent such tracking models are useful as a further supporting tool in making improved hazard assessments for aviation safety as well as for public health and air quality ).
We have used available data from the Ecuador Satellite Imagery of the Satellite Services Division of the National Environmental Satellite, Data, and Information Service (NESDIS) for the period between September 1999 and September 2014. Some 4672 images in "jpg" or "gif" format (e.g. http://www.ssd.noaa.gov/VAAC/ARCH99/gifs/guag0147.gif) from the website http://www.ssd.noaa. gov/VAAC/archive.html were evaluated for their wind direction and were weighted ( Figure 6) for the main direction of the ash-charged cloud and their respective flanks. A maximum potential error in such plotting may reach 10 degrees or equivalent 2.8%. This way to evaluate wind directions of ash emissions is novel and the most likely case may be of Papp et al. (2005), while the most regular way to evaluate ash dispersal bases simulation studies (Suzuki 1983;Connor et al. 2001;. The data were based on winds generated from Global circulation model (GFS/NAM) on a 24/7 base and reported every few hours, pending on the gravity of the emissions and are projected for the following 6, 12 and 18 hours, respectively.
A total of 19620 data of the 4905 images were subdivided per month in order to determine the main wind directions for the different seasons in Ecuador (Table 1). Around 89.55% of the obtained data belongs to Tungurahua volcano, 4.18% to El Reventador volcano, 2.57% to Sangay volcano, 3.07% to Cotopaxi volcano and the remaining 0.63% to Guagua Pichincha volcano. This data set has been plotted in regular rose diagrams and it provides an excellent overview on the wind directions of ash clouds of representative active volcanoes in the Ecuadorian mainland (Table 1; Figure 7).

Results and discussion
To our best knowledge, a systematic analysis of the wind directions of ash-charged clouds based on the archives of the NESDID has not been used so far, neither by the authorities or response organizations nor the volcanic monitoring units so far. Based on our evaluation, wind directions of ashcharged clouds are relatively uniform from E to W from April to September and varies and changes slightly into EÀW, ENE to WSW but also to other less frequent directions such as WÀE and SEÀNW (Figure 7). Therefore, the period between April and September is the best predictable one, while the rest of the year has a relatively high probability to present the same direction with some variation of lesser extent. As the data set is based on the past volcanic activity of four volcanoes, these volcanoes namely Tungurahua, Guagua Pichincha and Reventador, but not Sangay and Cotopaxi were responsible for a variety of route changes of airplanes and the closure of airport activities (Smithsonian Institution 1999À2016). Remarkably are the explosions of Guagua Pichincha at the 5 October 1999 and 7 October 1999, 26 November 1999 and 10 December 1999, the eruptive activity of El  (Reischmann et al. 2003;Hall et al. 2004;Robin et al. 2008;Arellano et al. 2008). Other eruptions with lesser impacts followed from the Tungurahua and also Reventador volcanoes.
Closure of main airports lasted for about a week for the eruptions of Guagua Pichincha in October 1999 and few more days later the same year. A whole week of closure of Quitos international airport occurred as well for Reventador's activity in November 2002 and less activity in the following years. Closure of airports due to activity of the Tungurahua volcano occurred in the vicinity of this volcano (Riobamba, Ambato, Latacunga) but sometimes even reaching the coastal area (Guayaquil, Manta, Portoviejo, etc.) (Auker et al. 2013;Smithsonian Institution 1999À2016). People deceased by ash cleaning of the roofs in Quito in 1999 (2) and 2002 (1), while hundreds were injured by the same activity (Tanguy et al. 1998;Witham 2005;Auker et al. 2013;Smithsonian Institution 1999À2016). Inhalation of ash (and gases) has been a major health problem in cities like Quito (1999 and, Baños, Salcedo, Ambato and Riobamba since 1999 and in the coast mainly in Guayaquil in 2006, and 2010. Since 1977 the Instituto Geof ısico of the Escuela Politecnica Nacional, a regular state university is in charge of the volcanic monitoring. However, early warnings for volcanic events often fail in Ecuador for all monitored volcanoes. There are many prominent examples of such incidents, as report by the press, and the scientific community (Smithsonian Institution 1999À2016; Tobin & Whiteford 2002).
In 1999 for the first time of volcanic monitoring in Ecuador an orange alert has been declared for the volcano Guagua Pichincha, which lasted one week (from 27 September to 4 October). This alert has been declared due to the growth of volcanic domes in the Pichincha Volcanic Complex as evidenced by the daily increasing seismic signals (Garcia-Aristizabal et al. 2007;Toulkeridis 2013). The downgrading of the alert status has been given due to the decrease of seismic signals ignoring the fact of an implosion/explosion effect of such domes. Therefore, due to the collapse of that first dome and all following ones some enormous explosions occurred such as those of the 5 October 1999 (VEI D 3) and 7 October 1999 (Fig. Xa), 26 November 1999 and 10 December 1999, all during yellow alert status (Garcia-Aristizabal et al. 2007;Smithsonian Institution 1999À2016). All the mentioned eruptions ejected a high amount of pyroclastic material. Later eruptions have been of lesser extent and without any effect to the nearby and more distant airports.
In 2002, the second most active Ecuadorian volcano (El Reventador) reactivated after a 26 yearlong rest. The 3 November 2002 eruptive activity with a VEI D 4, resulted in the strongest volcanic explosion of the last 14 decades in the continental part of the country and appeared without any kind of alert status at all. Ash precipitated after 5.5 hours in Quito and some 36 hours later in Gal apagos (Smithsonian Institution 1999À2016;Reischmann et al. 2003;Hall et al. 2004;). During precipitation of ash in Quito an orange alert status has been declared. Later eruptive activity also occurred during yellow or orange alert always without any previous warning.
An enormous amount of eruptions in Ecuador occurred during the re-awakening of Tungurahua volcano in 1999. Most eruptions of Tungurahua volcano appeared "suddenly" so that alert status, yellow, orange or even red did not correspond with the current volcanic activity. On 16 October 1999, yellow and later orange alert has been declared and the city of Baños and nearby villages were completely evacuated for about three months (Tobin & Whiteford 2002). During evacuation no ash precipitation took place, as Baños is located in the northern foothill of the dead part of Tungurahua I ) and therefore of some distance to the main wind direction of ash-charged clouds ( Figure 7). As the city did not suffer any damage nor has been affected at all by the volcano (e.g. ash precipitation), the citizens returned violently back to their homes. Since this incident, the volcano alert status has been declared mostly to be orange without accomplishing the predictions set up by the monitoring unit. Exactly after a three months period of time in which a declaration of emergency had been given by the authorities and the monitoring unit for the area surrounding the Tungurahua volcano, the very first pyroclastic flows took place on the western side of the volcano (14 July 2006). Ash clouds were regular for about a month until the night between 16 and 17 August a VEI D 3 eruption devastated five villages, killed six persons and ash precipitation caused the closure of a variety of airports including Manta and Guayaquil. No airport has been warned for about eight hours after eruption took place. Therefore, several flights from Quito to Guayaquil had to return to Quito while flying through the ash clouds putting passengers and flying crew in risk, as no warning was given in the early hours of August 2006. Upgrading of the alert status of most of the main eruptions of Tungurahua volcano after 2006 were declared two to four hours after the eruption initiated. Therefore, further closures of airports, which have been listed above, were given warnings just shortly before ash arrived to the airports (Toulkeridis, 2007;Smithsonian Institution 1999À2016).
Demonstrating the lack of precise or even close warnings of volcanic eruptions and associated ash precipitations of the few volcanoes which have been active during the last 17 years led to the preoccupation of what else could happen if the other currently dormant volcanoes awake. The VEI of the Ecuadorian volcanoes listed in Table 2 gives an approximate idea of the high potential in ash volume until September 2014 based on the archives of the NOAA archive data set. The area covered is continental Ecuador, with few exceptions where ash reached the Pacific Ocean. The altitude of the ash clouds is varying between FL180 and FL400, with two predominant heights being between FL200 and FL250 and a further between FL300 and FL400, generating together some 98% of all available data.
to be expelled by a variety of volcanoes. Some studies of previous and potential future eruptions of the volcanoes Guagua Pichincha (Barberi et al. 1992), Cotopaxi (Barberi et al. 1995;Biass et al. 2013Biass et al. , 2014 Gunkel et al. 2008;Gunkel et al. 2009), Pululahua (Papale & Rosi, 1993;Padr on et al. 2008;Volentik et al. 2010;Bonadonna & Costa 2012), and Chalupas (Hall & Mothes 2010;Toulkeridis 2013) demonstrate that a precise monitoring in field and by remote sensing (Ewert et al. 2005), together with the presented most probable wind directions of ash dispersion (per month), are the most constructive, useful and effective tools in this volcanic hazard assessment in order to be able to warn airports and public of ash precipitation of ongoing eruptions in real time.

Conclusions
(1) As demonstrated, the volcanic alert and warning system in Ecuador is not capable to warn the public, planes and airports in time. A different volcanic hazard assessment and modern real-time forewarning to improve air safety and prevent health issues must be implemented. This may be reached with a different and better instrument monitoring and a complementary standard operating procedure in order to warn airport operators of ongoing volcanic activity and associated ash dispersion.
(2) The statistics of the evaluated satellite images of ash-charged clouds and their wind directions are a useful tool in volcanic hazard assessment for the public, the domestic and international air routes and the infrastructure of airports. Effective early warnings and knowledge about the dispersion routes of ash clouds will give enough time to airports and the public for the implementation of corresponding mitigation activities. (3) Airports and air routes previously affected by the Ecuadorian volcanic activity will continue to be vulnerable to ash clouds and subsequent ash precipitation. Nonetheless, modelling of future eruptive activity of now dormant volcanoes and the distribution of ash precipitation will allow to predict better hazard areas and prevent public, air-traffic and airports of volcanic ash "surprises."