This study was conducted in Fusong County, Jilin Province in the northwest of Changbai Mountains (127°29′ to 128°02′E; 42°20′to 42°40′N; 600 to 800 m A. S. L., Figure 1). It is characterized by north temperate continental climate with prolonged and severe winters and hot and humid summer: long-term average annual precipitation recorded at the Baihe weather station ranged from 800 to 1,040 mm and the average annual temperature is 4.5–7.8°C. The forests are characterized by the original broad-leaved Korean pine forest and natural secondary forests regenerated after devastation with complex canopy, which is dominated by Pinus koraiensis, B. platyphylla, L. gmelinii, and Acer pictum. The soil is characterized as a mountain dark brown forest soil, and its average depth is about 50 cm (Deng et al., 2023b).

In the summer of 2015, two types, namely, understory removal plot (UR) and understory left intact plot (CK), of monitoring large plots with three repetitions (50 m × 50 m) were set up in the mixed broad-leaved P. koraiensis (BKF), B. platyphylla (BF), and L. gmelinii (LF) forests, respectively, and according to the forest inventory data from the local forestry bureau, the ages of the three stands were 195, 42, and 22 years, respectively. To minimize the edge effects, the large plots of each plot were replicated at least 20 m away from each other. The understory removal experiment in each stand was performed by a kind of machine that was used to cut the above-ground part of the understory shrubs, vines, and tall herbs that hinder the growth of seedlings, saplings, and trees. As a result, the plant residues were evenly stacked in the forests with some spacing. In July 2020, we randomly set up three subplots and their control plots with an area of 20 m × 20 m with similar abiotic conditions of topography, geology, soil, and climate in each large plot. The characteristics of each stand with two kinds of understory treatments are shown in Table 1.

Five soil cores (0–20 cm depth) of each subplot were collected randomly with an auger after removing the litter layer in July 2020 and were mixed with one composite sample. Soil samples of four plots from each treatment in each forest of each site were divided into two parts. One fresh part was dried and weighed to test its water content (SWC). The other part was air-dried for the determination of soil total carbon (TC), total nitrogen (TN), total phosphorus (TP), available phosphorus (AP), ammonium nitrogen (NH4+-N), and nitrate nitrogen (NO3--N) content, as well as pH, resulting in a total of 24 soil samples.

Soil pH was determined using a pH meter (PHS-3C) in a 1:2.5 soil:water solution (weight/volume). Soil TC and TN contents were measured using a C/N analyzer (Elementar vario MACRO cube, German). Soil TP and AP contents were obtained by using a spectrophotometer (UV-9000S, China) with H₂SO4-HClO₄ digestion and 0.5 mol/L NaHCO₃ extraction, respectively. Soil NH4+-N and NO3--N contents were extracted by 1 mol/L potassium chloride solution and analyzed using dual-wavelength ultraviolet spectrophotometry.

Understory removal significantly decreased the understory biomass of BKF and LF (P > 0.05) but can promote the increase of BF, and there were no significant differences among the three stands (Figure 4). Five years after understory removal, the total understory biomass of BKF and LF had decreased by 0.94 t·hm⁻² and 1.32 t·hm⁻², respectively, where there was a more than 25% decrease both in the biomass of the shrub and herb layers. In contrast, the total understory biomass of BF increased by 1.31 t·hm⁻², and understory biomass increased by 82.63% and 21.77% in the shrub and herb layers, respectively, in understory removal stands as compared to that in the understory left stands.

Understory removal had a greater impact on the soil's physical and chemical properties of BF and LF (Table 2). Five years after understory removal, there were no significant variations in soil properties in BKF (P > 0.05); the content of soil NO3--N and available phosphorus (AP) of BF; while the water content (SWC), total nitrogen (TN), and total phosphorus (TP) content of LF increased significantly (P < 0.05). For the three stands, the total nutrient content (TN, TP, and total soil carbon (TC)) of BF in soil was the highest and significantly higher than BKF, followed by LF after understory removal; there were no significant differences in soil pH, SWC, and NH4+-N among the three stands.

The biomass and diversity of understory vegetation were different due to different soil properties (Figure 5, Tables S2, S3). Redundancy analysis (RDA) was used to explore correlations of understory vegetation characteristics with soil properties (Table S2). Soil N/P was the primary factor impacting understory vegetation diversity and biomass prior to understory removal (Table S3), and as its value increased, shrub diversity and understory biomass increased, whereas herb diversity decreased (Figures 5A, C). Five years after understory removal, soil NH4+-N and TP were the key factors regulating understory vegetation diversity and biomass, respectively (Table S3), and as these values increased, shrub and herb diversity reduced while understory biomass increased (Figures 5B, D).

Disturbance and the availability of resources, particularly light and soil conditions have been recognized as key factors influencing plant diversity in the understory of temperate forests (Hofmeister et al., 2009; Forster et al., 2017). After understory removal, the vertical light distribution changed dramatically, and 50% of the light absorbed by the higher canopy was released (Matsuo et al., 2021), resulting in an increase in photosynthetically active radiation reaching the forest land, as well as an increase in the availability of temperature, humidity, and nutrients in the forest (De Lombaerde et al., 2021). Furthermore, understory species or species groups that demand different environmental resources may respond in various ways to the canopy opening (Forster et al., 2017; Deng et al., 2023b). Although previous studies have confirmed that understory management decreased species richness in the semi-arid ecosystem (Jimenez et al., 2015), our study in temperate forests demonstrated that its impact varies by forest type and understory vegetation layers (Hart and Chen, 2008). In this study, the species diversity of understory increased rapidly in the three stands, especially for B. platyphylla forest (Figure 2). The response of herbaceous plants to interference is more sensitive than shrub plants (Wang et al., 2021), the dominance of strong adaptability and reproductive ability and heliophiles herbaceous species increased significantly after understory removal, such as Impatiens noli-tangere, F. palmata, and E. hyemale, which caused significant differences in three forest types. Because the standground was not fully expanded by the shrub plants, it directly reduced the dominance of sciophiles herbaceous species (Table S1). The growth cycle of shrubs is usually longer than that of herbaceous species, therefore shrub diversity was not affected by understory removal, and their recovery process still needs to be studied on a longer time scale (Copeland et al., 2019). Because broad-leaved forests (BF) have higher canopy light transmission (Messier et al., 1998), their soil water and light conditions are superior to coniferous forests (LF) (Hart and Chen, 2008), resulting in increasing vegetation diversity and resource availability, which was in line with other research studies, that is, increase of broad-leaved trees benefitted the understory plant richness via improved nutrient availability in the soil (Salemaa et al., 2023). Shrub plants of broad-leaved forests grow faster and have more plant species than coniferous forests. The increase in dominance of the shrub layer would lead to the loss of species diversity under shrubs (Pajunen et al., 2011). Therefore, after understory removal, BF has the highest shrub diversity and the lowest herb diversity, which is in contrast to LF (Figures 2, 3).

Biomass compensation for understory removals occurs significantly in BF, but not in BKF and LF after treatment. In an arid region forest, the researchers found the degree of biomass compensation was dependent on restoration term (Canteiro et al., 2011) and a fertilized environment (Pires and Xavier, 2010) after the removal, and they observed that there was a compensation for removals under high fertilization during the first 3 years but was reduced compensation after 9 years. Moreover, biomass compensation for plant removals can occur relatively rapidly in temperate ecosystems (Gonzalez et al., 2019). However, in contrast with our findings, biomass compensation after removals in the short term only appeared rapidly in the broad-leaved pure forest (that is BF) with a fast litter turnover and decomposition rate (Zhao et al., 2022b). After a 5-year removal, the litter's turnover mechanism (Qiao et al., 2014) and nutrient content (Peng et al., 2020) changed, the litter in broad-leaved forests increased and accumulated faster than in coniferous forests, and its biological return was greater (Zhou et al., 2018a), which promoted the rapid accumulation of soil nutrient content such as organic matter in soil in BF (Zhang et al., 2022). In addition, the canopy density of BF is higher than that of coniferous forests and coniferous and broad-leaved mixed forests, and the light environment of the lower layer of the forest has improved after understory removal (Wang et al., 2021), thus the species with light-demanding and strong reproductive ability (such as F. Palmata and A. ukurunduense) enter the forest first and occupy a large area under the forest. Finally, light-demanding and shade-tolerant shrubs exist at the same time, resulting in the recovery of the species in the short term (Table S1). However, soil nutrient contents in BKF and LF were lower than in BF (Table 2), and some shade-tolerant and hygrophilous plants grew slowly, which limited understory vegetation restoration in a short time. Overall, understory removal appeared to promote shrub diversity and biomass of broad-leaved forests more than that of conifer forests (Figure 4).

In this study, we found that before the understory removal, soil N/P promoted the increase of understory vegetation biomass and shrub plant diversity (Figure 5). The effect of soil N/P on the diversity of the shrub layer and the herb layer was opposite, indicating that the shrub layer was more stable in response to changes in the soil's physical and chemical properties, which may be due to the understory canopy limiting the distribution of herbaceous plants. The shrub layer and the herb layer have a competitive relationship with environmental conditions such as light, water, and nutrients. Therefore, with the change in the soil's physical and chemical properties, the diversity of the shrub layer and the herb layer showed an opposite trend of change (Zhou et al., 2021).

After understory removal, the understory canopy closure decreased (Gurlevik et al., 2004; Matsushima and Chang, 2007), which increased light transmittance and, combined with the plant residues, increased the amount of rainwater in the coniferous forest (LF). This increase facilitated the availability of more nutrients to the soil with rainwater and significantly increased the soil water content of the forest land, which improved the utilization efficiency of nutrients in the ecosystem (Wang et al., 2023). Soil moisture in broad-leaved forests was comparable to those in coniferous and broad-leaved mixed forests, indicating that soil water content had little effect on the biomass and composition of the forest (Table 2). In addition, the plant residues from understory management could be used as biomass energy, which is conducive to the realization of carbon neutralization (Wang et al., 2023). The management process generates a significant number of branches, plant residues and other residues, which causes a rapid increase in organic matter on forest land, stimulates litter decomposition, and enhances soil nutrient input. Because of the strong demand for plant growth and its limited nutrients, increasing soil TP concentration promotes the development of plant roots, which is the primary determinant for the expansion of early biomass of understory plants (Zhao et al., 2022a). It promotes the growth of understory biomass in BF in particular, which shows the quickest microbial activity, decomposition rate, nutrient accumulation rate, and energy cycle (Li et al., 2016; Zhao et al., 2022b). Furthermore, some studies suggested that plant diversity decreased with the decrease in soil pH, that is, the increase of soil acidity (Zarfos et al., 2019), which is consistent with our findings on the shrub layer (Figure 4). The species richness and diversity of the herb layer increased, because the tolerance of the shrub layer to soil pH was higher than that of the herb layer, and the herbaceous plants were mostly shallow roots, and the competitiveness of soil nutrients was poor (Zhao et al., 2022a). NH4+-N is the basis of plant growth and development. This study found that with the increase of soil NH4+-N, the diversity of shrubs and herbs decreased and the biomass increased. It may be because the increase in plant biomass will increase the competition between species, thus reducing species diversity. In general, the coupling relationship between the diversity of the shrub layer and the herb layer and the physical and chemical properties of soil is complex and quite different. This may be because the structure and the growth pattern of shrubs and herbaceous plants are different (Keeley and Johnson, 1977), and the absorption of different elements in the soil is different (Salas-Luévano et al., 2017).

The diversity and biomass of understory vegetation is an important indicator to measure the effectiveness of forest management measures. Our study only analyzed the impact of soil properties on understory plant characteristics; however, for a full appreciation of the impact of understory removal on understory diversity and biomass, environmental parameters, such as light environment, should be taken into account to guide forest management and promote the restoration of understory vegetation in the future.