Random species loss underestimates dilution effects of host diversity on foliar fungal diseases under fertilization.

With increasing attention being paid to the consequences of global biodiversity losses, several recent studies have demonstrated that realistic species losses can have larger impacts than random species losses on community productivity and resilience. However, little is known about the effects of the order in which species are lost on biodiversity-disease relationships. Using a multiyear nitrogen addition and artificial warming experiment in natural assemblages of alpine meadow vegetation on the Qinghai-Tibetan Plateau, we inferred the sequence of plant species losses under fertilization/warming. Then the sequence of species losses under fertilization/warming was used to simulate the species loss orders (both realistic and random) in an adjacently novel removal experiment manipulating plot-level plant diversity. We explicitly compared the effect sizes of random versus realistic species losses simulated from fertilization/warming on plant foliar fungal diseases. We found that realistic species losses simulated from fertilization had greater effects than random losses on fungal diseases, and that species identity drove the diversity-disease relationship. Moreover, the plant species most prone to foliar fungal diseases were also the least vulnerable to extinction under fertilization, demonstrating the importance of protecting low competence species (the ability to maintain and transmit fungal infections was low) to impede the spread of infectious disease. In contrast, there was no difference between random and realistic species loss scenarios simulated from experimental warming (or the combination of warming and fertilization) on the diversity-disease relationship, indicating that the functional consequences of species losses may vary under different drivers.