In arid ecosystems, current year precipitation explains a small proportion of annual aboveground net primary production (ANPP). Precipitation that occurred in previous years may be responsible for the observed difference between actual and expected ANPP, a concept that we called legacy. Thus, previous year PPT may have an indirect effect on current year ANPP through two novel mechanisms that have not yet been explored. Our hypotheses predict that legacies in ANPP are a result of preceding precipitation affecting: 1) the density of tillers and, 2) the availability of inorganic nitrogen. We tested these hypotheses using a water and nitrogen manipulative experiment.

Our study was conducted in the Northern Chihuahuan desert (NM, USA), where 132 6 m² plots were subjected to one of five levels of water input during two growing seasons (-80% reduced precipitation, -50% reduced, ambient PPT, +50% increased, and 80% increased PPT). For the third growing season, treatments either remained the same or were reversed at -80%, ambient, or +80%. Ammonium nitrate was applied to half of the plots for the duration of the study. Tiller density was measured at the beginning and at the end of each growing season. We also sampled inorganic N pools, in situ net N mineralization, N leaching loss and plant N uptake during the growing seasons.
Results/Conclusions

We found that tiller density increased with increasing PPT during the first two years; it was 50% higher in irrigated than in drought plots. Precipitation treatments affected N loss and plant N uptake in different ways, but did not affect net N mineralization. As a result, inorganic N accumulated in shallow soil layers during drought, which was then lost through leaching when treatments were inversed. ANPP in the third year was significantly affected by current-year PPT (p<0.0001), previous-year PPT (p=0.02) and fertilizer addition (p=0.02). The model including these three variables explained 76% of the variance in ANPP. We concluded that the number of meristems and N availability might constrain ANPP response to an increase in water availability. Our research provides a better understanding of the mechanisms responsible for legacies, which is crucial to understanding the global carbon cycle, and to predicting changes in ANPP with changes in climate.