3 Result
N addition and drought significantly changed the relationship between herbivore and plant communities. For instance, the relationship between plant community and plant suckers was found to be significant, and it was driven primarily by plant biomass under natural condition (βbalgra = 168%) (Fig.1a) but by plant species turnover under experimental drought (βsimsne = 188%) (Fig.1b&4). Moreover, omnivores were subject to both species turnover and plant N content under drought condition but not to them under any other experimental treatment. Drought engendered significant effects of plant species turnover on herbivores regardless of N addition, and N addition stimulated the significant effects of plant N content on herbivores (plant suckers and detritivores) whether with or without drought.
There were three remarkable points in the interaction of arthropod functional groups in networks (Fig.1): (a) the interacting functional group pairs (i.e., plant suckers vs. vegetation hunters, plant suckers vs. parasitoids, detritivores vs. parasitoids) were observed in all the treatments including ambient condition. Both web spiders and ground hunters were strongly correlated with herbivore/detritivores (including both plant suckers and detritivores) under ambient condition and N addition, yet part of links of web spiders and ground hunters has been lost under experimental drought; (b) Under N addition, further application of experimental drought did not impact the relative abundance of vegetation hunters, but exerted a negative effect on the interaction between plant chewers and vegetation hunters, and a positive effect on the interaction between vegetation hunters and parasitoids. (c) omnivores appeared a key functional group as both its relative abundance and number of interactions with other arthropod functional groups increased by experimental drought regardless of N addition.
Finally, by quantifying ecological network properties, we found that both experimental drought and N addition decreased CFG, ID and IS of the interactions between arthropod functional groups and their sum except for increased ID and IS of carnivores-carnivores and ID of herbivores-carnivores (Fig.2). The IS of environmental factors on arthropod functional groups has been significantly increased by experimental drought, and the change of micro-habitat exerted significant effects on much more arthropod functional groups under experimental drought than ambient condition (Fig.1&3).Our manipulation of snowmelt effectively maintained snow cover for 1.7 days (in 2019) and 10.8 days (in 2020) longer. Compared with drought and N addition, delaying snowmelt only exerted a modest impact on plant and arthropod communities in the peak of growing season (Table.1&2&3; but see Fig.S1). Thus, we ignored the influence of delaying snowmelt when analyzing the ecological network at the peak of the growing season.