Trying to bend the rules: precipitation-productivity relationship in desert grassland

The relationship between precipitation (PPT) and aboveground net primary productivity (ANPP) has intrigued ecologists for decades because of its fundamental importance to the global carbon cycle. Across gradients from grassland to forest, the PPT-ANPP relationship is statistically well-defined and non-linear. Temporal patterns within a site over time, however, are generally weaker than spatial patterns and nearly always linear despite statistical attempts to bend the line. Linear relationships are inconsistent with positive asymmetry occurring when the increase in ANPP in a wet year is greater than the decline in ANPP in a comparably dry year. This led to the double asymmetry model which predicts that non-linear, concave down responses will occur when extreme wet and dry PPT years occur in a time series. Using long-term ANPP data from ambient plots, along with rainfall addition and reduction experiments in a Chihuahuan Desert grassland we tested the hypothesis that extending the range of precipitation would lead to a non-linear relationship between PPT and ANPP as predicted by the double asymmetry model. Based on a historical precipitation calculator for the past 2000 years, our experimental rain addition treatments matched the wettest years in the record, whereas our extreme drought experiments reduced precipitation ~43mm below historic lows. By extending the precipitation gradient to extremes through drought, water and nitrogen addition treatments we found support for the double asymmetry model with one important exception. The response was concave up under high precipitation under nitrogen fertilization. Without N addition, the response under high precipitation was linear. By experimentally extending the range of monsoon precipitation we were able to bend the rules and generate a non-linear PPT-ANPP relationship in this desert grassland, but only when nutrient limitation was alleviated. Because of strong water limitation, dryland ANPP is highly sensitive to year-to-year variability in precipitation. Water limitation under drought and nutrient limitation in wet years govern non-linear responses of ANPP to precipitation variability in this dryland ecosystem.