Covid-related surge in global wild bird feeding: Implications for biodiversity and human-nature interaction

a. Bird species richness and bird feeding We tested whether countries with greater bird species richness, our proxy for biodiversity, were more likely to have high search intensity for bird feeding during our sample period. Using data on bird species by country from BirdLife International and the cross-sectional search intensity data from Google Trends, we used two-tailed, two sample t-tests with unequal variances to compare the number of bird species in places with positive search intensity (labeled as "existence of search interest") in the Google Trends sample and those without enough search volume defined by Google (labeled as "absence of search interest") within the Google Trends sample. We also plotted the distributions of species for each bird grouping by these categories in box and whisker plots to compare the medians and assess the importance of outliers, which may be hard to detect in the t-tests. We note that these t-tests describe a correlation between high biodiversity, as measured by our proxy of bird species richness, and bird feeding. We do not make causal claims that it is bird species richness that drives bird feeding; instead, we describe an observed pattern in the data. Other factors, such as internet accessibility and GDP, could affect a country’s existence of search interest in bird feeding using Google. We use a linear probability model to examine the relationship between the existence of search interests and bird species richness while control- ling for each country’s internet accessibility.

Sc ¼ b0 þ b1Speciesc þ b2Speciesc þ b2Internetc þ b3GDPc þ Zc ð2Þ Where Sc is a binary variable that equals 1 if a country is labeled as "existence of search interest" and 0 otherwise. Spiecesc are binary variables for the tercile of the number of bird spe- cies. The omitted category is the country with the lowest tercile of bird species richness. The coefficients β1 and β2 indicate if, compared to countries in the bottom tercile, countries with more bird species are more likely to have search interests in bird feeding. Results are qualita- tively the same when using quantiles or terciles of bird species richness. We also note that we are not controlling for all other factors that could affect the likelihood of bird feeding in a country, including income and latitude. If bird species richness is corre- lated with these factors, then the relationship we describe may reflect income or latitude instead of the importance of species richness itself. IV. Results and discussion a. A surge in increase in global bird feeding interest Our first result demonstrates that there was a significant increase in bird feeding interest occurred globally. Using countries and territories with a sufficiently high search intensity for bird feeding, bird seed, and bird baths in the panel search intensity data, we estimated the change in search intensity before and after Covid-19 lockdowns relative to the week before the lockdown started. For a period of 52 weeks prior to lockdown, search intensity was, on aver-

age, similar to what it was in the week preceding lockdown (Fig 1). The coefficients remained close to zero and within the 95% confidence interval. This is an important internal validity check as it demonstrates that there was no independent, changing interest in bird-feeding that preceded Covid-19 lockdowns. After about two weeks of lockdown, a dramatic increase in bird feeding search intensity was evident (Fig 1). The result mirrors the interest in these topics found in the U.S. [28], a country with well-documented bird feeding interest [1]. Search intensity increased more quickly and dramatically for bird food and bird baths, doubling in intensity (see Table A2 in PLOS ONE Covid-related surge in global wild bird feeding PLOS ONE | https://doi.org/10.1371/journal.pone.0287116 August 2, 2023 5 / 15