Forest understory plays an important role in the gross primary production (GPP) of some forest ecosystems. However, differences in understory GPP caused by obviously different overstory canopy structure have not been taken into consideration in previous studies, thus potentially over- or underestimating understory GPP. To estimate the understory GPP more accurately, we separated a forest into "canopy area", with closed-overstory canopy, and "gap area", with open-overstory canopy. The study was conducted in a mature deciduous forest dominated by beech and with an understory dominated by dwarf bamboo, Sasa senanensis. We measured S. senanensis GPP at the community scale (GPPSasa-community) using a static chamber system that covered the aboveground part of the plants and then upscaled it to the ecosystem scale (GPPSasa-ecosystem) by considering the proportions of canopy and gap areas within the forest. GPPSasa-community was 192 g C m-2 year-1 in the canopy area and 699 g C m-2 year-1 in the gap area. The large difference likely occurred because the photosynthetic ability and biomass of the S. senanensis community differed strongly between the two areas. The seasonal dynamics of GPPSasa-community also differed between the areas. The 10-day cumulative GPPSasa-community peaked from July to August in the gap area, whereas there was no clear peak of GPPSasa-community in the canopy area. Multiple linear regressions showed that light intensity and biomass were significant predictors of GPPSasa-community in the canopy area, whereas air temperature and biomass were significant predictors of GPPSasa-community in the gap area. GPPSasa-ecosystem during growing season in 2019 was 3.74 t C ha-1 year-1, which contributed between 16.37 and 19.85% of the entire forest ecosystem GPP. This study highlights the need to consider differences in overstory structure for the accurate estimation of understory GPP.
Keywords: Dwarf bamboo; GPP; Gap-mosaic structure; Mature forest; Sasa senanensis; Understory.