SARS-CoV-2 spread rapidly from Wuhan, Hubei province, China, affecting nearly every country of the world. As of 1 April, 823,626 cases of COVID-19 were registered globally, with 72,736 deaths worldwide.1 According to the World Health Organization (WHO) estimates as to 1 April, in Europe Italy presented the highest number of cases (105,792), followed by Spain (94,417).1

Canary Islands are a Spanish archipelago located in the Atlantic Ocean in front of Morocco. They are the southernmost autonomous community of Spain and include eight islands: Tenerife, Gran Canaria, Lanzarote, Fuerteventura, La Palma, La Gomera, El Hierro, and La Graciosa. The population of the Canary Islands is 2,153,389 inhabitants (2019 figures), mostly concentrated in the two capital islands: around 43% in Tenerife (917,841) and 40% in Gran Canaria (851,231).2

Canary Islands’ epidemic situation could have reached worrying levels as, while the pandemic was raging in Europe, the archipelago celebrated the carnival, attracting thousands of visitors. Moreover, a noteworthy sandstorm obliged to close Tenerife and Gran Canaria airports and forced the inhabitants to a lock-down of at least two days (22-24 February). Locally known as calima, it is a relatively frequent meteorological phenomenon in the Canary Islands, although in the last 40 years there has not been such an intense one as that of February 22-24, 2020. A warm or hot east-wind from the Sahara Desert, Africa, lifts up and transports to the archipelago clouds of sand and dust, with increasing warm and humid air. Often associated with fog and patchy drizzle, calima is also a contributing source of particulate matter (PM).3

According to the Spanish Health Ministry estimates, as to 1 April, the cumulative incidence of COVID-19 was higher in Tenerife than in Gran Canaria.4 In this paper we address the hypothesis that the extraordinary sandstorm occurred on 22-24 February 2020 might have a role in the different cumulated incidence of COVID-19 cases between the two islands, since it obliged to reduce significantly air traffic and forced to suspend all major carnival street events in most locations, but not so in Santa Cruz de Tenerife where, despite the calima, important carnival events took place. With this retrospective analysis, we aim to investigate the impact of environmental and weather characteristics of the Canary Islands, in addition to other confounding factors, on the COVID-19 epidemic.

In this retrospective observational study, meteorological and weather-related data were obtained from the Canary Government website (Red de Control y Vigilancia de la Calidad del Aire de Canarias). A retrospective analysis of the cases who presented symptoms related to COVID-19 and tested positive to SARS-CoV-2 by polymerase chain reaction in Tenerife and Gran Canaria islands, as to 1 April 2020, was carried out. Data were collected and registered in ReVECa (Red de Vigilancia Epidemiológica de Canarias), a detailed database constantly updated. Moreover, aggregated epidemiological data are published daily by the Canary authorities detailing the situation of each island. Due to the anonymization process performed prior to data extraction from ReVECa database, no ethical issue is concurring in the present study. Independent samples z-test was performed. The two sample z-test approach is based on comparing the incidence rates on the natural log scale. Statistical analysis was performed with SPSS 23.0 software (IBM).

Canary Islands, located 96.6km off the coast of Morocco in Northwestern Africa, are one of the most popular tourist destinations in Europe during winter due to warm temperatures compared to the continent. Nevertheless, on 22-23 February 2020 temperature raised to 33°C, tourists and residents were forced indoors due to an intensive sand and dust storm blowing in from the Sahara Desert, leading to all ports and airports closure. PM reached extraordinarily high peaks worsening the risk of respiratory and cardiovascular symptoms.4

Due to this extraordinary incidence, the state of alert was declared since 22 February at 4:00 p.m.,9 and all the eight airports of the archipelago were closed by the Spanish authority for the air traffic management being over 800 flights affected, especially in Gran Canaria, as visibility was reduced to less than 400 meters. Air traffic resumed on the morning of February 24, 2020. Also, the authorities closed primary schools on Monday 24th and advised residents to stay indoors, particularly those with respiratory problems.

The carnival of Gran Canaria and Tenerife usually takes place each year between February and March. It usually lasts many weeks and various festivities and activities take place during the day and the night10,11 (Table I in online Appendix shows the program of the celebrations). Tenerife's carnival started earlier than Gran Canaria's, being the most important celebrations held on 21-24 February 2020, while the great carnival event in Las Palmas (the Cabalgata), was held on Saturday the 29th. Since the pandemic was still not perceived as a major problem in the archipelago nor in Spain, the decision whether suspending the celebrations were only made based on of the weather conditions of those days and the importance of the carnival event.

According to the WHO,12 the incubation period for SARS-CoV-2 (time from exposure to symptom onset), lasts around 5-6 days, to a maximum of 14 days. So, the cancellation of important functions in Gran Canaria may be the explanation for the SARS-CoV-2 incidence delay between Gran Canaria and Tenerife. Indeed, these celebrations represented a way for COVID-19 spread, especially among young individuals who are more likely to be asymptomatic or present mild symptoms.

Among the variables that may have played a role in the incidence differences between the two islands, the number of tourists and their nationality needs to be taken into account. In fact, according to the Canary Institute of Statistics, during the month of February 2020, a total of 387,431 tourists entered in Gran Canaria by airway. Of these, 33% (128,276) came from Nordic Countries, 21% (81,126) form Germany, 15% (56,873) from United Kingdom, 2% (7,104) from Italy and 1,5% (5,883) from Belgium. Spanish residents who entered in Gran Canaria in February 2020 were 10% (37,738). During the same month, a number of 528,872 airway passengers entered in Tenerife. Of these 35% (184,208) from United Kingdom, 16% (84,332) form Germany, 12% (63,872) came from Northern Countries, 3% (17,297) from Italy and 3% (17,979) from Belgium, while Spanish residents who entered in Tenerife were 9% (49,853).13 Considering absolute numbers, these figures show that Tenerife received higher rates of tourists coming from Belgium, Italy and the United Kingdom (by then already presenting high rates of COVID-19) if compared to Gran Canaria.

On 14 March, a lockdown was compulsory in the whole country. It has been one of the hardest interventions to control the epidemic12 and included a strict limitation to travel among provinces. These measures played a critical role in the contention of COVID-19 spread, as proven by literature.14 In fact, in Spain, the estimated number of cases sharply increased until the lockdown, then it showed a slow down followed by a decrease after the full quarantine was implemented.15,16 Since 19 March 2020 Canary airports start operating at minimum levels.

In conclusion, the cancellation of carnival celebrations due to the sandstorm occurred 22-24 February 2020 in the Canary Islands, together with the different flow and origin of tourists visiting Tenerife and Gran Canaria, contributed to the differences of SARS-CoV-2 cumulated incidence between the two islands. Thus, it is likely that the “worst calima in the last 40 years” differentially affected COVID-19 spread through the Canary Islands, in particular during carnival celebrations involving multitudes of people.

Miguel Ángel Negrín Hernández.

The corresponding author on behalf of the other authors guarantee the accuracy, transparency and honesty of the data and information contained in the study, that no relevant information has been omitted and that all discrepancies between authors have been adequately resolved and described.

L. Tomaino and L. Serra-Majem had the original idea, conceptualization, and developed the investigation hypothesis. L. Tomaino wrote and redacted the first draft of the manuscript. J. Pinilla and P. Barber-Pérez carried out the statistical analysis, and prepared tables and figures. S. Rodríguez-Mireles, B. González López-Valcárcel, C. La Vecchia and A. Sierra, together with the rest of the authors, contributed substantially to the reviewing and editing, and to the supervision and validation of the final version of the manuscript. All results were discussed among the authors. All co-authors have read and approved the final version of the manuscript.

The authors desire to express their gratitude to the precious contribution of Lourdes Ribas-Barba (CIBER de Fisiopatología de la Obesidad y Nutrición [CIBERobn], Spain) and Domingo Núñez-Gallo (representing the Canary Government's Public Health Direction, Santa Cruz de Tenerife, Spain).

None.

None.