Previous PageTable Of ContentsNext Page

Soil nitrogen mineralisation as affected by size of rainfall events

Victor Sadras and Jeffrey Baldock

CSIRO Land & Water, PMB 2 Glen Osmond 5065, Australia, victor.sadras@csiro.au

Abstract

Power laws describe the relationships between the number N (s) and the size s of daily rainfall events, i.e. N (s) ~s; higher τ corresponds to greater frequency of small rainfall events. We tested the hypothesis that soil nitrogen is mineralised faster in sites or periods with high τ. Daily rates of nitrogen mineralisation on a sandy loam soil were calculated using simulation models for (a) uncropped soils in six Mallee locations, and (b) wheat crops in 39 locations in an E-W transect in the southeastern wheat-belt. Rates were calculated using actual rainfall, and variable or fixed temperature and evaporative demand. In the Mallee, the annual pattern of mineralisation rate, calculated as a function of rainfall and variable temperature and evaporative demand, was bimodal with peaks in April and November. These peaks disappeared and differences among locations were reduced when the effects of temperature and evaporative demand were removed. Under constant temperature and evaporative demand, mineralisation rates between April and November were 68% greater than rates between December and March. In the former period, characterised by a high frequency of small rainfall events, monthly mineralisation rate was a direct function of the amount of rainfall. In contrast, mineralisation was independent of the amount of rainfall during the period of larger, less frequent rainfall events from December to March. Parameter τ accounted for 75% of the variation in mineralisation rate in the period December-March and accounted for a substantial part of the variation between periods. In the E-W transect, mineralisation rate was higher than expected from rainfall in locations with high τ indicative of high frequency of small rainfall events.

Introduction

In environments with low and erratic rainfall, every rain event generates a pulse of increased soil water content that, depending on event size and evaporative demand, can last from a few hours to several weeks (1). These pulses of soil water can stimulate plant growth directly by improving plant water status, and indirectly by enhancing microbial activity and availability of nutrients (2). Frequency and size of rainfall events thus modulate a range of physical and biological processes relevant to a range of terrestrial ecosystems (2). We explored the links between the seasonal patterns of rainfall and mineralisation, and quantified the relative importance of the main components of the rainfall pattern, including frequency and size of events.

Power laws i.e. N ~s, fitted the relationships between number (N) and size (s) of daily rainfall events in a range of environments (3). High τ indicates high frequency of small rainfall events, and this parameter accounted for significant part of the variation in soil evaporation and other components of the water budget of wheat crops in eastern Australia. Here, it is proposed that the rate of soil nitrogen mineralisation increases with increasing τ, as affected by both spatial and temporal sources of variation. Simulation models were used to disentangle the effect of rain and other variables with marked seasonality.

Method

Rates of nitrogen mineralisation on a sandy loam soil were calculated using long-term weather data in two simulation studies. In the first study, we used the CropSyst model for uncropped soils in six Mallee locations (rainfall 260-360 mm year-1). In the second, we used the APSIM model for wheat crops in 39 locations in an E-W transect in the south-eastern wheat-belt (176-374 mm year-1). Daily rates of mineralisation were calculated using actual rainfall, and variable or fixed temperature and evaporative demand.

Results

In the Mallee, the annual pattern of mineralisation rate, calculated as a function of rainfall and variable temperature and evaporative demand, was bimodal with peaks in April and November. These peaks disappeared and differences among locations were reduced when the effects of temperature and evaporative demand were removed. Under constant temperature and evaporative demand, mineralisation rates between April and November were 68% greater than rates between December and March. In the former period, characterised by a high frequency of small rainfall events, monthly mineralisation rate was a direct function of the amount of rainfall. In contrast, mineralisation was independent of the amount of rainfall during the period of larger, less frequent rainfall events from December to March. Parameter τ accounted for 75% of the variation in mineralisation rate in the period December-March and accounted for a substantial part of the variation between periods.

In the E-W transect, mineralisation rate averaged 50 kg N ha-1 season-1, and was closely related to seasonal rainfall (Figure 1a). Mineralisation rate was higher than expected from rainfall in locations with high τ indicative of high frequency of small rainfall events (Figure 1b). We conclude that parameter τ can be used to summarise temporal and spatial variation in size of rainfall events influencing the rate of nitrogen mineralisation.

Figure 1. Influence of (a) amount of seasonal rainfall and (b) relative size of rainfall events, as quantified by τ (dimensionless) on the simulated rate of nitrogen mineralisation during the growing cycle of wheat crops in 39 locations.

Acknowledgements

We thank the Grain Research and Development Corporation of Australia for financial support (grants CSO212, CSO209), and Daniel Rodriguez and José Paruelo for useful discussions.

References

(1) O. E. Sala, W. K. Lauenroth, W. J. Parton, M. J. Trilca, Oecologia 48, 327-331 (1981).

(2) O. E. Sala, W. K. Lauenroth, Oecologia 53, 301-304 (1982).

(3) V. O. Sadras, 11th Australian Agronomy Conference 2003.

Previous PageTop Of PageNext Page