Soil C and N and specific bacterial functions affect greenhouse gas emissions in dryland rice under water and fertilizer coupling management

Abstract

Background Dryland rice has obvious advantages in saving water and fertilizer and improving agricultural productivity. However, irrational application of fertilizer and irrigation by farmers is prone to cause a large amount of greenhouse gas (GHG) emissions. It is of great environmental significance to explore the greenhouse gas emissions in dryland rice under water-fertilizer coupling regimes. Methods In this study, we set up the three irrigation levels of W1 (200 m3·hm-2), W2 (300 m3·hm-2), and W3 (400 m3·hm-2), the four fertilizer levels of F1 (400·kg-hm-2), F2 (600·kg-hm-2), F3 (800·kg-hm-2), and F4 (1000 kg·hm-2) with a total of twelve water and fertilizer treatments in a completely orthogonal combination as well as a non-irrigated and non-fertilized CK as a control in a dryland rice farmland. The cumulative greenhouse gas (GHG) emissions, global warming potential (GWP), and greenhouse gas intensity (GHGI) of dryland rice in the growth period under different irrigation and fertilization treatments were monitored, the differences in soil physicochemical properties and bacterial structure and function under different water and fertilizer treatments were analyzed, and the effects of the representative C and N as well as the specific functions of bacteria in soil on the average cumulative GHG emissions through the structural equation model were revealed. Results The data showed that the increase in irrigation and fertilizer application increased the soil C and N content and yield of dryland rice. The CO2 cumulative emissions also increased by 62.40% compared with CK, but the CH4 and N2O cumulative emissions significantly decreased by 98.20% and 42.67%, respectively. It was worth noting that the GWP of the W3F4 treatment had a maximum value of 22085.64 kg CO2/ha, but the GHGI was lower. The relative abundance in Acidobacteriota was low. However, it had a great importance for water and fertilizer treatments. Although the functional relative abundance in Chemoheterotrophy and aerobic_chemoheterotrophy was more than 30%, the relative abundance in nitrate_ reduction, predatory_or_exoparasitic, and chitinolysis function was less than 3.33% which directly affected the GHG emissions. Moreover, the soil C and N indirectly affected the GHG emissions through soil bacterial functions. Conclusions This study can provide a reference basis for the response of the GHG emissions to soil C and N and specific bacterial function in dryland rice under different water and fertilizer regimes, as well as a practical guidance for attenuating the GHG emissions in farmers' production.