Canthigaster rostrata (Tetraodontiformes: Tetraodontidae) is distributed in the Central Western Atlantic Ocean, from North Carolina to Trinidad, including Bermuda, Bahamas, the Caribbean Sea and the Gulf of Mexico. It occupies a variety of tropical water habitats such as coral reefs, sea grasses, soft and rocky bottoms. It is considered an abundant species in coral reefs (Vázquez-Yeomans et al. 2017; ^{López-Rios, 2021}). It has mainly diurnal habits (^{Moura and Castro, 2002}), feeding on small invertebrates such as polychaetes, as well as grasses, algae, some sponges, crustaceans and soft corals. Due to its trophic level, it can play an important role in some food chains, by regulating the amount of algae and invertebrates in a locality (Moura and Castro, 2002). It breeds year-round (^{Lyczkowski-Shultz, 2003}) and has few predators. It is estimated that individuals of this species live up to 10 years, with the juvenile stages showing the highest mortality.

During the last 14 years, mass mortality events of this species have been reported throughout the Caribbean (Jordán-Garza et al., 2009; ^{Castillo and Pérez, 2014}; Gutiérrez, 2014; El Isleño, 2017; Piedra, 2017; ^{Piedra-Castro and Araya-Vargas, 2018}) (Figure 1). Possible causes of mass mortality have been hypothesized, such as: ectoparasitic copepods infections (^{Kirtisinghe, 1934}), accumulation of paralytic toxins by dinoflagellates (^{Ochoa et al., 1997}), disease outbreaks in populations of surgeonfish such as Ctenochaetus striatus (^{Stier et al., 2013}) and toxicosis (^{Work et al., 2017}). ^{Vail and Sinclair-Taylor (2011)} reported mass mortality of C. bennetti in Indonesia, associated with a schooling event. According to the authors, species within the Canthigaster genus typically occur in low population densities, so schooling was considered an atypical behavior.

Figure 1 Location of mass mortality events of C. rostrata in the Caribbean sea during 2008, 2014, 2017 and 2020. 

During October 2020, a mass mortality event of C. rostrata was observed in a sandy beach area of the Gandoca-Manzanillo National Wildlife Refuge (Regama), Manzanillo sector, Costa Rica (9.635238º N and -82.653470º W). At that time, 183 dead individuals were found on 150 m² of beach, which represented a density of 1.22 ind/m² Since there were already reports of this phenomenon in the area (Gutiérrez 2014; ^{Piedra-Castro and Araya-Vargas, 2018)}, the objective of this work was to report and explore possible causes for the phenomenon in the Regama.

During the 2020 event, collected individuals had their standard length (SL) and weight (W) measured in order to compare them with the values obtained by ^{Piedra-Castro and Araya-Vargas (2018)}. The 2020 event consisted of individuals with an average SL of 40.59 ± 3.82 mm and W of 0.91 ± 0.27 g, while those collected in 2017 presented SL of 36.73 ± 4.72 mm. and W of 0.90 ± 0.37 g (Figure 2). The SL of the 2020 individuals was significantly higher than those of 2017 (w = 4055.5; p < 0.05), while the weight did not significantly change (w = 7068.5; p > 0.05).

Figure 2 Standard length (A) and weight (B) of the C. rostrata individuals collected during the mass mortality events registered in Regama in 2017 and 2020. 

^{Sikkel (1990)} determined that the SL for C. rostrata in a locality in Panama was 54.1 ± 10.6 mm for males and 44.6 ± 4.2 mm for females. The author determined that males reach sexual maturity between 46 and 71 mm in size. Thus, based on Sikkel (1990), it is assumed that both the 2020 and 2017 events in the Regama consisted of sexually immature individuals, probably from the same demographic cohort (given the little variation between lengths and weights), a phenomenon also mentioned by Jordán-Garza et al. (2008).

The possible relationship between mortality events and sea surface temperature (SST) was explored additionally. At the time of obtaining the C. rostrata mortality records in the Regama, the average SST (Greenbelt, 2020) of the area between 9.52083° N; -83.1042° W and 10.3958° N; -81. 9375° W, was: 29.17 °C (May 2014), 29.59 °C (April-May 2017) and 30.34 °C (October 2020). Thus, it is observed that published mortality events in the Regama occurred during considered warm periods on the South Caribbean coast of Costa Rica (Fig 3).

Figure 3 Monthly behavior of sea surface temperature (°C) (SST) between 2003 and 2020 on the coast of the Gandoca-Manzanillo National Wildlife Refuge, Costa Rica. Each arrow indicates the month in which a C. rostrata mortality event was recorded in the region. 

Although the mortality events reported in the Regama occurred during months considered warm, at this stage of the study it is not possible to establish a cause-effect relationship between mortality and SST, given the few existing records. However, it must be taken into account that SST can be a facilitator of other ecological processes, which together may affect the mortality of this species. ^{Genin et al. (2020)} stated that the duration of sea warming events, as well as the separation between events, can cause the widespread death of hundreds of fish from different families. ^{Karniski (2020)}, on the other hand, points out that the sustained increase in sea temperature is a key element to take into account when monitoring impacts on fish, reef communities and seagrasses.

Regarding the after mentioned scenario, the need to establish regional monitoring of different environmental variables which may affect the biodiversity of marine ecosystems is highlighted. As proposed by ^{Garrabou et al. (2019)}, the region has a need to develop and maintain collaborative databases, functioning as platforms for monitoring fish communities and their relationship with environmental changes.