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

Moreover, this pattern holds for each bird classification, including breeding, non-breeding, passage, and resident. As a visual analog to the t-test, we compare the bird species richness of countries with an absence of interest in bird feeding to those with a presence of interest in bird feeding in Fig 3. This figure shows the distribution of species richness across countries where search interest is present or absent for "bird feeder" for each bird classification. While there is an overlap in the lower range of species richness, countries with search interest for "bird feeder" consistently have a higher median, upper quartile, and a maximum number of species for each bird classifi- cation category. It seems the distribution of species richness is shifted toward greater diversity in countries and territories where search interest is present for "bird feeder" relative to those where search interest is absent. Results from a linear probability model indicate a positive correlation between the existence of search interest for "bird feeder" and bird species richness (Table 2 Column (1)), which con- firms the finding from the t-test discussed above. Compared to countries in the bottom tercile, countries in the middle tercile are more likely to have search interests in bird feeding. Such a positive correlation still holds after we control for each country’s internet accessibility. The results in Table 2 Column (2) suggest that conditional on internet access, the existence of search interest is still positively associated with bird species richness with a slightly larger mag-

nitude. In addition, we further analyze the relationship between these two by controlling for both internet access and GDP. The finding from this model again confirms the positive corre- lation between the existence of search interests and bird species richness (Table 2 Column (3)). It is important to note that the magnitude of the effects for the middle tercile gets smaller after controlling for GDP. This pattern, where areas with greater species richness and, by proxy, biodiversity have more bird feeding, is new and noteworthy for two reasons. First, areas with greater species Table 1. T-test comparing bird species richness by presence or absence of search interest for "bird feeder". PRESENCE SEARCH INTEREST ABSENCE SEARCH INTEREST T STATISTICS P-VALUE TOTAL Mean 511.2 294.9 -4.97 *** SD 400.5 288.6 N 107 144 RESIDENT Mean 352.7 215.8 -3.33 ** SD 392.9 256.6 N 107 144 BREEDING Mean 99.99 25.05 -7.88 *** SD 103.2 39.1 N 107 140 NON-BREEDING Mean 123.9 62.67 -7.35 *** SD 78.3 53.6 N 107 144 PASSAGE Mean 55 21.99 -7.46 *** SD 43.6 23.2 N 106 132 Notes: ** P < 0.01, *** P < 0.001 Columns report the average number of species for the presence or absence of search intensity data for "bird feeder" from Google Trends. The P-value corresponds to a two-tailed t-test with unequal variance. https://doi.org/10.1371/journal.pone.0287116.t001 PLOS ONE Covid-related surge in global wild bird feeding PLOS ONE | https://doi.org/10.1371/journal.pone.0287116 August 2, 2023 9 / 15