Shrub and grass ANPP
The ANPP of each shrub species varied differently among treatments and years. In the wet year (2017), ANPP of A. prolifera in the nitrogen addition plots doubled ANPP in the control plots (p =0.04) (Fig. S6). Moreover, the ANPP of A. prolifera was higher in 2017 than in 2016 in the +N treatment (p <0.001) and was higher in 2017 than in the others studied years in the +NW treatment (p <0.05). The water addition treatment only affected A. volckmannii such that its ANPP increased in the +W plots (p =0.005) and, in addition, this species increased ANPP during the wet year, being more than double in 2017 than in the rest of the years (p <0.01) (Fig. S6). The ANPP of S. filaginoides decreased in +NW compared to C (p =0.001) and was higher in 2017 than in 2019 in all treatments, with intermediate values in 2016 and 2018 (p <0.05) (Fig. S6).
Total shrub ANPP in the +N, +W and +NW treatments were not significantly different compared to the control (Table 3, Fig. 4A). Shrub ANPP was about two folds higher in 2017 as a consequence of the anomalous precipitation in late March of that year than in the other study years, regardless of treatments (p≤0.01) (Table 3, Fig. 4A). Grass ANPP was not affected by the treatments until two years after the start of the experiment (Fig. 4B). From 2015 to 2019, grass ANPP was higher in +N and +NW than in C (p <0.02) (Table 3, Fig. 4B). There was no difference between grass ANPP in +W and C (p >0.5). When comparing within-treatment grass ANPP across years, we found no interannual differences in the control (Fig 4B). However, grass ANPP in nitrogen addition plots was higher in 2017 and 2018 than in 2014 (p <0.02) and ANPP in +NW treatment was higher in 2017 than in 2014 (p =0.03) (Fig. 4B). Community-level ANPP (shrubs + grasses) increased significantly in the treatments with nitrogen addition (p <0.001) compared to the control (Table 3, Fig. 4C). The total ANPP was higher in the wet year (2017) than in 2016 (p =0.001) and 2019 (p <0.001) (Fig. 4C).