Microbially mediated nitrogen conservation and loss in low and high nitrogen input rice paddies

High rates of fertilizer nitrogen (N) application in rice paddies is a common practice in the majority of the rice-producing countries. However, more than 60% of the applied fertilizer-N is lost to the atmosphere and other terrestrial ecosystems. This causes serious environmental pollution and economic loss. On the other hand, un-fertilized rice paddies have shown maintained soil N status and stable N supply to the rice plant for hundreds of years. Microbial N2 fixation is known to contribute N to un-fertilized rice paddies. However, N2 fixation alone cannot fully explain the maintained N nutrition in rice paddies, where favourable conditions for N loss by denitrification exist. In largely anaerobic rice paddies, caused by submerged conditions, nitrification takes place in the oxic rhizosphere and soil-water interface which continuously supply nitrate (NO3-) to the anaerobic soil. Fertilizer-N input increases the nitrification rate, NO3- concentration and the denitrification rate in rice paddies. Rice paddies under long-term no or low N input have also shown continuous nitrification activity and NO3- production. Despite the nitrification activity and NO3- production, there is minimal N loss by denitrification in long-term low N input rice paddies. This implies that microbes in low N input rice paddies have evolved to efficiently recycle mineral N within the system, allowing minimal N loss. However, the mechanisms are still unknown. Dissimilatory nitrate reduction to ammonium (DNRA), a largely overlooked N-cycling process, has been found to compete with denitrification for loss-prone NO3- and retain it as ammonium (NH4+) and could help to explain the sustained N nutrition in low input paddies. The nrfA gene in soil microbes encodes cytochrome c nitrite reductase which catalyses the DNRA process. DNRA has been found to be a major NO3- consumption pathway (up to 98% of the available NO3-) in several forest soils and sediment ecosystems and the NO3- partitioning between DNRA and denitrification is thought to be affected by the soil organic carbon (OC) and NO3- availability. This means DNRA can limit NO3- loss from soils where a favourable environment for the process exists. However, scant research has been conducted to explore the significance of DNRA and factors affecting the process in arable soils. There are no studies which have explored the effect of N fertilization regimes on the balance between microbial N conservation and loss in rice paddies. This study used 15N tracing technique, an acetylene reduction assay and quantitative PCR assays to quantify DNRA, denitrification, anammox and N2 fixation rates and the relevant microbial gene abundances, in laboratory incubation experiments using paddy soils from Australia and Myanmar. The first study used soils from three high N input (>150 kg N ha-1) rice paddies from Southeast Australia for a glasshouse pot experiment and subsequent laboratory experiments. The pot experiment was conducted with two levels of N input, zero and 150 kg urea-N ha-1. The rhizosphere and bulk soils (separated by nylon bag) in the pots were collected separately for the laboratory experiments before the panicle initiation stage of rice. Effects of the history of N fertilization on the N-cycling processes were investigated in the second study using three paddy soils from long-term high N input Australian sites (>150 kg N ha-1) and three paddy soils from low N input (<25Kg N ha-1) Myanmar sites. Long-term no or low N (16% compared to N added paddy soils. DNRA rates were similar or higher than denitrification rates in paddy soils without N addition. But DNRA consumed less than 50% of the NO3- than by denitrification in paddy soils with N addition. Nitrogen omission did not have a clear effect on the nrfA gene abundance and the N2 fixing activity in long-term high N input paddy soils. DNRA and N2 fixation did not vary between the rhizosphere and bulk soils but denitrification was lower in the former than the latter. This study confirmed that DNRA plays a significant role in retaining N in rice paddies which do not receive fertilizer-N. Comparison of the effect of long-term N fertilization regimes on microbial N loss and retention in Myanmar and Australian rice paddies in the second and third study revealed more serious negative consequences of the long-term high N input on microbial N retention. Microbial N2 fixation was found to add more than 4 kg N ha-1 day-1 in long-term no or low N input paddies, whereas N2 fixation was less than 1.8 kg N ha-1 day-1 in long-term high N input paddies. Furthermore, DNRA was able to retain around 60% of the soil available NO3- as NH4+, allowing less than 15% to be reduced to N2 by denitrification, whereas denitrification reduced around 30% of the NO3- to N2 in long-term high N input rice paddies, where DNRA only retained approximately 10% of the NO3- as NH4+. These results confirmed that microbes can efficiently retain and recycle N in no or low N input paddies, allowing maintained soil N status and stable N supply to rice plants. However, in high N input rice paddies, microbes have less incentives to invest energy in retaining and recycling N. These results suggest that, under high N input systems, microbes (N2 fixing and DNRA) in rice paddies lose the ability to retain N which exists in low N fertilised systems, in the order of ~ 3 kg N ha-1 day-1 in our study, which needs to be recompensed with fertilizer-N input. Soil OC and NO3- concentrations were the most important environmental determinants of the fate of NO3- in rice paddies. Rates of DNRA consistently showed a positive correlation with the soil OC:NO3- ratio and a negative correlation with the soil NO3- concentration. In accordance with the previous findings from chemostat experiments using pure bacterial culture, higher soil OC:NO3- ratios in this study promoted NO3- partitioning to DNRA over denitrification. Artificially manipulating the soil OC:NO3- ratio by adding NO3- and different levels of labile OC confirmed that maintaining higher soil OC:NO3- ratio can improve N retention when there is high NO3- concentration in rice paddies. This study, for the first time, provides a comprehensive dataset of the effect of fertilizer-N input on the balance between microbial N conservation and loss in rice paddies. These findings suggest that long-term no or low N input paddy systems have unique microbial N regulation mechanisms to maximize N retention and minimize N loss, thus supporting primary productivity in the systems, whereas these mechanisms were not apparent in high input systems. This new knowledge is important for fertilizer-N management in low input paddies such as in Myanmar, where fertilizer-N use is starting to increase. In such a context, fertilizer-N dose should be established based on the examination of the effect of different fertilizer-N rates on the balance between N conservation and loss.