5 Conclusion
(1) The hydrodynamic zone of the study area could be divided into runoff
area and stagnant area. They had different hydrodynamic conditions and
hydrochemical characteristics. The stagnant area had higher reservoir
pressure, gas content and ion concentration than the runoff areas.
(2) The microbial functional community structure was different between
runoff area and stagnant area. Genes involved in several important
anaerobic respiration processes, such as N cycling genes (e.g., nifDKH,
amoB, narGHI, napAB, nirK, norC and nosZ), methanogenesis genes (e.g.,
mcr, fwd, mtd, mer and mtr) and S cycling genes (e.g., dsrAB, sir, cysN,
sat, aprAB and PAPSS), were increased in the stagnant area. The machine
learning model shows that these significantly different genes could be
used as an effective index to distinguish runoff area and stagnant area.
(3) Increased genes involved in nutrient cycling, including organic
matter decomposition, methanogenesis, denitrification and sulfate
reduction, contributed to the increase of CO2 and
reduction of sulfate and nitrate from runoff area to stagnant area.
(4) Carbon and hydrogen isotopes indicate that methane in the study area
was thermally generated. The main reason for the lack of biogenic
methane in the study area was that methanogens were inferior to other
anaerobic heterotrophic bacteria in the substrate competition, biogenic
methane was consumed by methanotrophic bacteria and was not enough to
support the enrichment of a large amount of biogenic methane in the
study area.