RESUMO
Under the background of global climate change, atmospheric nitrogen deposition and precipitation are undergoing substantial changes, which leads to an uncertainty on litter decomposition in desert grassland. The experiment was set up with a split-plot design. There were three precipi-tation treatments, including natural precipitation, an increase of 30% and a decrease of 30%, and four levels of nitrogen application, including 0 (N0), 30 (N30), 50 (N50) and 100 kg·hm-2·a-1 (N100). A two-year decomposition experiment aimed to examine how water and nitrogen manipulations interactively influence litter decomposition of three dominant species in the desert grassland, i.e., Salsola collina, Stipa breviflora and Kochia prostrata. The results showed that litter mass remaining rate decreased with time, which was consistent with Olson negative exponential decay model. Litter decomposition coefficient (k) was highest for S. collina, followed by S. breviflora and K. prostrata. The decomposition coefficient (k=0.028) was the highest under the treatment of increased precipitation 30% and N application level of 100 kg·hm-2·a-1. Under single factor treatment, litter decomposition was the fastest under increased precipitation 30% and N application level of 50 kg·hm-2·a-1. Under the combined water and nitrogen treatments, litter decomposition under increased precipitation 30% and N application level of 100 kg·hm-2·a-1 was the fastest. The initial nitrogen content was the greatest for S. collina, followed by S. breviflora and K. prostrata. The decomposition coefficients were positively correlated with initial nitrogen contents for S. collina and S. breviflora. The total carbon content, cellulose content, lignin content, C/N, lignin/N and cellulose/N were higher in K. prostrata than in S. breviflora and S. collina. The litter characteristics were negatively correlated with decomposition coefficient for S. collina. For S. breviflora and K. prostrata, the decomposition coefficients decreased with increasing C/N, lignin/N and cellulose/N. The results indicated that the decomposition rate was the highest in S. collina, and the lowest in K. prostrata. The findings suggest that appropriate amounts of added water and nitrogen will contribute to accelerating litter decomposition, promoting nutrient cycling, and will play an important role in the sustainability and ecological balance of the desert grassland .
