key: cord-0309128-m6b6lwty
authors: Olán-González, Manuel; Reyes-Bonilla, Héctor; Arreola-Alarcon, Isabel Montserrat; Uribe, Regina Valdovinos; Olivier, Damien
title: COVID-19 lockdown reveals fish density may be much higher in marine reserves
date: 2022-05-19
journal: bioRxiv
DOI: 10.1101/2022.05.17.492376
sha: a06335bfdfe99673e26c82ac9e39206874a648c8
doc_id: 309128
cord_uid: m6b6lwty

Marine reserves generally allow ecotourism to offer an alternative income to fishing. However, we need to assess its impact on wildlife to make this activity sustainable. The COVID-19 lockdown provided a unique opportunity to evaluate wildlife diversity in the absence of human activity. In a Mexican reserve, we monitored fish assemblages before, during, and just after the lockdown. We show that ecotourism activities alter the behavior of fishes by finding a 2.5-fold density rise during the lockdown. We suggest that the noise pollution generated by the numerous recreational vessels is a significant factor of perturbation. In the absence of noise pollution, some fishes may be bolder (less hidden) and others can come back to the reserve from usually quieter areas (e.g., deeper waters). Our results represent a great worldwide incentive to improve the health of marine reserves by establishing concrete measures in managing plans to mitigate noise pollution. Open Research statement All data and code necessary to reproduce the results of the paper are enclosed in the submission for review purposes, and will be published on Zenodo following the acceptance of the paper.

However, disentangling the specific effect of the presence of people from other disturbance 52 sources (natural or anthropogenic) is challenging to achieve. Indeed, the closure of natural 53 reserves, which would allow diversity to be assessed in the absence of human activity, even 54 temporarily, is generally not economically viable for the local communities whose 55 livelihood depends on them.

In this study, we used the case of the Cabo Pulmo National Park ("Cabo Pulmo"; 

Each month, hundreds of boats carrying thousands of visitors entered the reserve to practice 69 scuba-diving and snorkeling activities (Figure 1a, b) . 

First, we ran linear mixed models (LMMs) to test the lockdown effect on the three 103 diversity metrics of the fish assemblages (by transect). We log-transformed fish density to 104 fit a normal distribution, and we used raw values for species richness and evenness. Sites i.e., we compared the lockdown period to the pre-lockdown period. We focused on the most 110 common families for statistical robustness, i.e., those observed in at least 20% of transects.

For each family, we ran six models: generalized linear models (GLM) with a Poisson or 6 negative binomial distribution, either with or without random variables (sites and depth 113 category), and zero-inflated (ZI) models with a Poisson or negative binomial distribution.

For each family, we selected the best of these models. We did not consider models with 115 high residual overdispersion values (>1.5). When multiple models met the overdispersion 116 condition, we selected the model with the lowest Akaike Information Criteria (we 117 considered irrelevant differences lower than two and chose the simplest model in that case).

Based on the selected models, we calculated the fitted fish density by family during and 119 before the lockdown by bootstrapping with 1,000 iterations. We used the difference 120 between the mean of the fitted abundances during each period and the standard error of this 121 difference to calculate a t-value, as follows:

,

where DL and BL refer to "during" and "after" lockdown, respectively. Those t-values 124 equal to or higher than two in absolute value indicated significant results at a 95% 125 confidence interval (CI). We used a bootstrapping method to be able to compare the 126 parameters (e.g., t-value) between distinct models (GLM(M) and ZI). We ran all analyses 

The species richness also increased during the lockdown period (WebTable 1). It is 146 noteworthy that during the lockdown period we observed some species rarely seen on the 208

However, such disturbances should be monitored. The density of highly mobile and large-209 sized species also increased during the lockdown. It is difficult to believe that these hide 

Contrary to global change problems, noise pollution is a temporal disturbance that we can WebTable 1. Comparison of the three fish diversity metrics during and after the lockdown to before the lockdown. Linear mixed models with the sites and the depth as random variables were used. The estimate (difference between during/after the lockdown to before the lockdown), the standard error (SE), the t-value, and the random standard deviance (RSD) of the two random variables are shown 

WebTable 2. Outputs of the models run with 1000 bootstraps to compare the fish density by family between before and during the lockdown. The estimate (difference between the two groups), the standard error (SE) and the t-value are shown. For each family, residuals overdispersion (disp.stat) and Akaike (AIC) values were indicated for the different tested models. Only models selected at least for one family are represented. 1.02/601 NB = Generalized linear model with negative binomial distribution, NB 2 = Generalized linear mixed model (site and depth as random variables) with negative binomial distribution, ZINB = Zero-inflated model with negative binomial distribution. The estimates are on log-scaled as negative binomial distribution was used for each model.

Large recovery of fish biomass in 294 a no-take marine reserve

Oil exposure alters social group 296 cohesion in fish

Performance of a novel 298 COVID-19 Quietus

How many fish could be vocal? 349 An estimation from a coral reef (moorea island)

Reducing vessel noise: An 351 example of a solar-electric passenger ferry

Soundscape Ecology: The 353 Science of Sound in the Landscape

An overview of fish bioacoustics and the impacts of 355 anthropogenic sounds on fishes

R: A language and environment for statistical 357 computing

Ambientales de Arrecifes Coralinos: El Caso del Parque Nacional de Cabo Pulmo

Desarrollo regional en Baja California Sur. Una 362 perspectivade los servicios ecosistémicos

When good intentions are 365 not enough … Insights on networks of "paper park" marine protected areas

COVID-19 lockdown allows researchers to 368 quantify the effects of human activity on wildlife

Spawning aggregations and reproductive 21

Linear mixed models (considering the sites as a random variable) estimated the fish 418 densities thanks to the data of two different sites (El Cantil and Las Navajas) consistently 419 monitored since

The yellow and purple circles represent the periods 421

The grey rectangle represents the period 422 studied in the main manuscript, i.e., before, during, and after the lockdown. The black star 423 represents the fish density estimated during the lockdown period. Although monitoring is 424 usually not performed during the June-July period, it does not seem to present higher fish 425 density than the

WebFigure 2. The relationship between the gains in fish density (difference between 428 during and before the lockdown) and a) the number of boat trips, b) snorkelers (log scale

and c) divers per year (averaged for the period of 2017-2019) in the seven surveyed sites

The lines are the result of third-order polynomial regression. Each site is numbered as on 431 the map of Cabo Pulmo