Mycorrhizal type
Using the FungalRoot database (Soudzilovskaia et al. 2020), plant species were grouped according to their mycorrhizal types: arbuscular mycorrhizal (AM); ericoid mycorrhizal (ErM); ecto-mycorrhizal (EcM); orchid mycorrhizal (OM); or non-mycorrhizal (NM). Scores of 0, 0.5 or 1 were attributed to each plant species for each mycorrhizal type, depending on whether they were recorded as either associating (1) or not (0) with a certain mycorrhizal type, or associating with two different types (0.5) (e.g., AM and NM; there were no species associated with more than two types). We included plant species which were referred to in the FungalRoot database as “AM, others not addressed” and “EcM, other not addressed” simply as “AM” and “EcM”, respectively. Indeed, this statement means that they were confirmed to associate with the respective mycorrhizal type, but that uncertainty remained regarding their potential to form associations with more than one mycorrhizal type. The potential resulting error is minimal, however, as amongst all the species for which mycorrhizal types were assessed, only 0.3% were found to associate with more than one type of mycorrhizal fungi, indicating that such instances of multiple associations are negligibly rare.
We could extract mycorrhizal types at the plant species level for 65.0 % of all observations, ranging from a minimum of 30.2% in Argentina to a maximum of 93.2% in Norway. For the remaining species the dominant mycorrhizal type at the genus level was used instead, resulting in 98.4 % of all observations being assigned a mycorrhizal type, ranging from a minimum of 94.2% in Argentina to a maximum of 99.8% in India. This categorization by genus was done by Soudzilovskaia et al., 2020, who assigned a mycorrhizal type or NM status to a genus if more than 2/3 of the available observations in that genus converged.
We multiplied the cover percentage for each species in a plot by its score for each mycorrhizal type (0, 0.5 or 1). The resulting percentages of total vegetation cover were added up for each mycorrhizal type and then divided by the total vegetation cover of the plot obtained by adding up the cover values of all species present. This resulted in a percentage of total plot vegetation cover associated with each mycorrhizal type (for example: 65% of all vegetation in a plot is associated with AM fungi) which was then converted to proportions between 0 and 1 for further analysis; this value is henceforth referred to as ‘mycorrhizal type cover’. It is worth noting that this value is a percentage of a plot total vegetation cover and not of the plot total surface. OM mycorrhizal type cover was excluded from further analysis as it represented less than 0.1% of total cover so it is unlikely to be relevant when looking at changes in mycorrhizal type distribution.