Paddy soil indigenous N supply is often poorly related to N status and our aim was to assess its linkage to reduction of Fe3+ and Mn4+, primary terminal electron acceptors in submerged soils. Transplanted rice was grown in the greenhouse in four Bangladeshi paddy soils with distinct SOC to Feox ratio (1.6 to 4.9) for 72 days firstly under continuous and then intermittent flooding. Solution buildup rates of Fe2+ and Mn2+ in the first 2–3 weeks of flooding correlated negatively with soil organic carbon (SOC) to NH4-oxalate extractable Fe (Feox) and Mn (Mnox) ratios (p < 0.01). Mössbauer analysis detected ferrihydrite and goethite in all soils and with Mn3O4 these are the likely source minerals. An electron (e−) balance calculated from soil C-emission rates suggested reductive Fe and Mn dissolution to relevant e−-accepting processes, probably responsible for no >50% of e− capture, though. Reduction of abundantly present octahedral Fe3+ in chlorites and vermiculite and their interstratified forms in these floodplain silty Inceptisols is hypothesized to also support microbial activity. Notwithstanding, a close temporal synergy existed between solution Fe and soil mineral N and their build-up rates were correlated (r: 0.77 to 0.90; p < 0.01) and with that of dissolved OC (DOC) (r: 0.84 to 0.96; p < 0.01), C emission rate (r: 0.99; p < 0.01) and SOC:Feox (r: −0.71; p < 0.01). These correlations suggest Fe3+ reduction to be a relevant intermediary step in soil N mineralization, possibly through release of associated DOC, N or both. After switching to intermittent flooding dissolution of Fe, Mn and DOC were decoupled from mineral N release but since Eh remained in the Fe3+-reduction range in three out of four soils, possibly clay-Fe3+ alternated with O2 as e−-acceptor. Most importantly, in all soils N release slowed or halted after only 2 weeks of flooding but recommenced with intermittent flooding. As a next step, field experiments could verify if indigenous soil N supply also benefits from non-continuous irrigation management. Lastly, experimental proof is pending for release of clay-bound N and interlayer NH4+ following reduction of octahedral Fe3+ with consequent increased negative charge or structural destabilization, possibly an important process in floodplain paddy soils in Bangladesh.

The availability of alternative electron acceptors like Fe3+ and Mn4+ may form a bottleneck to anaerobic SOM mineralization and thereby NH4+-release in flooded paddy soils. We assessed the influence of availability of soil Fe and Mn on anaerobic N mineralization in lab incubation experiments. Collected paddy soils from Bangladesh either untreated, amended with Fe2O3, or with Mn/Al mixed oxides were anaerobically incubated. In a first 8weeks incubation with 5 treatments from a long-term field experiment (control, N, NP, NPK and N+FYM) we found no considerable differences in evolution of soil solution Fe and Mn between the control and Fe2O3 treated soils. Whereas, the soil solution contents of Fe were lower and Mn were higher in Mn/Al mixed oxide treated soils. Similar observations were made for dissolved Fe and Mn in a second 10-week incubation experiment with four farmers field soils. Evolution of KCl-extractable NH4+ was not affected by Mn4+ or Fe3+ application and we therefore conclude that availability of electron acceptors was not limiting release of NH4+ in the studied soils. The large and rapid increase of exchangeable-NH4+ at the onset of the incubations provoked the question whether part of it derived from release of fixed-NH4+. A third experiment revealed, however, instead a small significant increase of the fixed-NH4+ within four weeks. In addition, the microbial biomass carbon already plateaued after two weeks. Both results suggest that released mineral N was mainly derived from biotic anaerobic N mineralization and not from defixation of NH4+. Finally, while not directly dependent on Fe and Mn application, there was a remarkable convergence in the buildup of soil exchangeable-NH4+ and soil solution Fe concentrations. This warrants further investigation and still suggests involvement of reductive Fe and Mn-oxide dissolution in NH4+-release, e.g. through release of bound organic N after reduction of these oxides.

Introduction: The emission of green house gases (CH4, CO2 and N2O) from rice paddy fields is influenced greatly by several factors like soil redox potential (Eh), availability of labile organic carbon and oxidants for soil organic matter (SOM) degradation.The alternate aerobic and anaerobic cycling of rice fields declines CH4 emission and enhances N2O emission, often in contrast to continuously flooded fields. In irrigated rice soils, soil moisture level, Eh and levels of dissolved organic carbon (DOC) and reduced species like NO3-, Mn2+, Fe2+ etc. differs strongly between depth increments soil layers. But relation of these dynamic depth profiles with bulk soil CH4 and N2O emissions have not been resolved. Objectives: The evidence of an integrated beneficial effect of water saving irrigation management on GHG emission is still lacking for young floodplain paddy soils in Bangladesh. Our objective is to understand how depth distributed Eh, moisture, DOC, Fe and Mn content drive CH4, N2O and CO2 emissions. We set up a paddy field experiment at the Bangladesh Agricultural University in the 2015 Boro season. We hypothesize that green house gas emission from paddy soils are mostly governed by soil moisture profiles and Eh profiles, rather than bulk soil levels of DOC or NO3-. Materials and methods: The cultivated rice variety (BRRI dhan28) was grown for 14 weeks under three water management practices (continuous flooding (CF), Alternate wetting and drying (AWD) and dry seeding (DS)). To explore our main hypothesis, we compared global warming potential (GWP) based on cumulative emissions of CH4, CO2 and N2O at several sampling events (closed chamber/GC).The soil Eh at four different depths was measured continuously by a HYPNOS III data logger an Eh-probes system. Soil moisture tension was monitored regularly via Electronic Tensiometers with hypodermic needle (SMS 25000 S). DOC, Fe and Mn in soil solution, collected from three depths with rhizon solution samplers, were measured at regular interval. Results: The Eh(mV) decreased sharply in both AWD and CF plots within 1st week and stabilized near -600mV in both CF and AWD treatments over all soil depths. Eh peaked for several days close to 0mV at drainage events in AWD mostly at 5cm and several times also at 12.5 and 20.5cm depth. After 70DAT, reductive conditions (-300 to -400mv) in AWD and CF and even in DS (so as to 84DAS) due to continuous flooding resulted in CH4 fluxes (10-30 mg m-2 h-1) in case of the AWD and CF treatments. Between 14-70DAT CH4 emission rate of 0-20 mg m-2 h-1 was irregular in the AWD plots and progressive under CF. During the drainage and re-flooding events both CH4 emission and Eh at different depths fluctuates abruptly in AWD treatments. Mean CH4 fluxes in DS plots are lower than AWD and CF plots throughout the season. Analysis of N2O, CO2, DOC, Fe and Mn and volumetric soil moisture levels are on-going and will be presented in detail at the workshop. Conclusion: Most strikingly, both rate and magnitude of the drop of Eh in the Mymensingh floodplain soil was unusually large. Even under AWD, strong reducing conditions are seen at times even at 5cm depth and continuously at 30cm depth. We therefore also anticipate fast preferential reduction of NO3-, Mn4+ and Fe3+, preceding methanogenesis under both CF and AWD. The young age of non-calcareous Bangladeshi floodplain soils developed from Ganges and Brahmaputra alluvial sediments with low pedogenic Fe content may explain these observations. Our partial results already indicate that GWP-benefits of water reducing management may be less than anticipated in these floodplain soils, when compared to studies in other areas. The analysis of remaining soil chemical parameters will provide a more process-oriented insight in the biogeochemistry of non-continuously flooded paddy soil systems.