Impact of Thinning and Contour-Felled Logs on Overland Flow, Soil Erosion, and Litter Erosion in a Monoculture Japanese Cypress Forest Plantation

Abstract

This study investigated the impact of thinning and felled logs (random- and contour-felled logs) on overland flow, soil erosion, and litter erosion in a Japanese cypress forest plantation (2400 tree ha⁻¹) with low ground cover, from 2018 to 2023 in central Japan. Monthly measurements of overland flow and soil and litter erosion were carried out using small-sized traps across three plots (two treatments and one control). In early 2020, a 40% thinning (tree ha⁻¹) was conducted in the two treatment plots. Overland flow increased in the plot with random-felled logs during the first year post-thinning (from 139.1 to 422.0 L m⁻¹), while it remained stable in the plot with contour-felled logs (from 341.8 to 337.1 L m⁻¹). A paired-plot analysis showed no change in overland flow in the contour-felled logs plot compared to the control plot from the pre- to post-thinning periods (pre-thinning Y = 0.41X − 0.69, post-thinning Y = 0.5X + 5.46, ANCOVA: p > 0.05). However, exposure to direct rainfall on uncovered ground areas post-thinning led to increased soil and litter erosion in both treatment plots. These findings suggest that thinning combined with contour-felled logs effectively stabilizes overland flow. Therefore, thinning with contour-felled logs can be considered a viable method for mitigating overland flow in monoculture plantations with low ground cover.

1. Introduction

Monoculture plantations are vulnerable to overland flow and soil erosion, often due to insufficient ground cover. This issue has been observed in rubber plantations in south China [1,2], vineyards in Spain [3], and Japanese cypress forest plantations in Japan [4,5]. The lack of ground cover in these plantations leads to the detachment of unstable soil aggregates by the kinetic energy of raindrops [6], soil crusting [7], and reduced soil surface water infiltration [8]. As a result, overland flow and soil erosion are common occurrences in these monoculture plantations, especially in regions characterized by steep terrain and heavy rainfall [9,10].

In Japan, commercial monoculture forest plantations were established in the 1950s for timber production [11]. However, the import of cheaper timber from other countries led to the neglect of these forest plantations by the 1970s [12]. Consequently, Japanese monoculture forest plantations grew increasingly dense, leading to a significant reduction in ground cover [12]. The dense canopy reduced sunlight penetration to the forest floor, suppressing understory vegetation [13]. This issue is especially pronounced in Japanese cypress forest plantations. The leaves of cypress trees break into small fragments that are easily washed away by overland flow, preventing their accumulation on the forest floor [14,15]. Hence, Japanese cypress forest plantations are more vulnerable to overland flow and soil erosion due to their lack of ground cover compared to other forest plantations in Japan, such as Japanese cedar and larch [16,17].

The lack of ground cover in Japanese cypress plantations caused rainfall to flow (i.e., overland flow) directly into streams, especially during heavy rainfall, rather than infiltrating into the soil. In the early 2000s, dense Japanese cypress forest plantations were identified as sources of disasters such as flooding and landslides in various regions of Japan [18]. In response, forest managers began taking control of these plantations for hydrological purposes [19]. Non-commercial thinning was introduced as a practical method to promote understory development and reduce overland flow and soil erosion. However, many studies reported an immediate increase in overland flow following thinning in dense Japanese cypress forest plantations [20,21]. This was primarily because thinning did not significantly change the ground cover but increased canopy openness, allowing more direct rainfall to reach the forest floor [22]. Additionally, canopy closure occurs a few years after thinning and prevents sufficient understory development [23].

Contour-felled logs are a ground eco-engineering method [24] used to restore natural regeneration of pine species in Mediterranean forests [25] and to control overland flow and soil erosion in forest plantations, especially in areas heavily disturbed by wildfires and heavy machinery harvesting [26,27,28,29]. Research in this area has noted that placing felled logs on slopes aids in water infiltration and sediment retention [25,26,27,28]. Lucas-Borja et al. [30] observed a long-term increase in soil organic matter and water storage 11 years after the application of log erosion barriers (i.e., contour-felled logs) as a post-fire treatment in a pine forest in Spain. In Japan, random- and contour-felled logs are traditionally applied after non-commercial thinning by either leaving felled logs where they fall (random-felled logs) or placing them along contour lines (contour-felled logs) [4]. Several studies have reported that contour-felled logs can enhance water infiltration into the soil in dense Japanese cypress forest plantations [4,31]. Farahnak et al. [4] observed that overland flow was reduced and remained low three years after thinning and contour-felled log placement, compared to thinning and random-felled logs and non-thinned Japanese cypress plots. Kuraji et al. [31] also reported a decrease in peak runoff after thinning and contour-felled log placement in a Japanese cypress plot compared to a grassland control plot. However, previous studies lacked data on soil and litter erosion [4,31]. Persistent soil erosion from headwater catchments reduces the capacity of rivers near residential areas to control flooding [32]. Litter erosion not only diminishes ground cover but also decreases the input of organic matter into the soil. Low soil organic matter weakens soil structure and the stability of soil aggregates [33], resulting in decreased water infiltration due to reduced pore space and increased surface crusting [34]. Both soil and litter erosion can exacerbate flooding and increase the risk of disasters during heavy rainfall events by increasing overland flow. Therefore, further research is needed to confirm the effectiveness of thinning combined with contour-felled logs in reducing not only overland flow but also soil and litter erosion in dense Japanese cypress forest plantations.

Miura et al. [17] emphasized the importance of ground cover (both understory and litter layer) in controlling soil erosion in Japanese cypress, cedar, and red pine forest stands. Therefore, to achieve effective forest thinning in dense monoculture forest plantations, it is crucial to promote understory growth, enhance the accumulation of the organic layer (litter), and reduce overland flow and soil erosion. On the one hand, a single thinning may not significantly increase understory development, and a second thinning may be required after several years (usually 10 years’ rotation in Japan). On the other hand, in regions like Japan with heavy rainfall during rainy seasons and typhoon periods, long-term thinning methods may not effectively control overland flow and soil and litter erosion, necessitating a short-term solution. Thinning combined with contour-felled logs could be an effective short-term method for managing overland flow [4], as well as soil and litter erosion in Japanese cypress forest plantations.

In Japan, most studies have primarily focused on the effects of forest thinning on overland flow and soil erosion [35,36,37]. Recently, a few studies have begun exploring the potential of felled logs as a short-term measure for mitigating these processes [4,31,38]. However, comprehensive assessments comparing different log arrangements remain limited. This study aimed to address this gap by evaluating the effects of thinning combined with different arrangements of felled logs—random- and contour-felled—on overland flow and soil and litter erosion. The main goal was to understand the impact of thinning and contour-felled logs on overland flow and soil and litter erosion in a Japanese cypress forest plantation. We also sought to compare the effects of thinning combined with contour-felled logs versus thinning combined with random-felled logs on overland flow and soil and litter erosion. Thus, we hypothesized that (1) thinning combined with contour-felled logs would stabilize overland flow more effectively than thinning combined with random-felled logs, and (2) thinning combined with contour-felled logs would mitigate the increase in soil and litter erosion more effectively than thinning combined with random-felled logs.

2. Materials and Methods

2.1. Study Site

This study was conducted in a monoculture Japanese cypress forest plantation of the Obora Experimental Forest (OEF) in northern Toyota city, central Japan (35°16′ N, 137°15′ E, approximately 600 m a.s.l.; Figure 1a,b). The Japanese cypress catchment, covering 2.27 ha, was planted in 1991 with a tree density of 2400 trees ha⁻¹. The study period spanned from 2018 to 2023, with a 40% thinning (trees ha⁻¹) conducted between January and March 2020. The mean annual air temperature and precipitation were 15.3 °C and 1470.4 mm, respectively, from 1991 to 2020 (in the Toyota meteorological station, 35°8′ N, 137°11′ E, 75 m a.s.l, approximately 17 km southwest of the study site). The rainfall pattern in the study area is characterized by two distinct periods: the Baiu (rainy season) from June to mid-July, marked by continuous rain (average monthly precipitation of 183.6 mm in June and 195.3 mm in July) and high temperatures (average 22.3 °C in June and 26.3 °C in July), and the typhoon season, which typically occurs from August to October, bringing heavy rainfall (monthly precipitation of 125.8 mm in August and 201.8 mm in September) and strong winds, with average wind speeds of around 1.4 to 1.5 m/s. In addition, the winter months from December to February typically experience lower precipitation (monthly precipitation ranges from 48.0 mm in January to 61.2 mm in February), often in the form of light rain or snow, with lower temperatures (average 3.6 °C to 4.5 °C) and reduced humidity [39].

2.2. Study Hillslope Soil Characteristics

The selected hillslope soil is classified as cambisol [40], originating from weathered granite bedrock [41], with a bulk density of 0.6 g cm⁻³ in upper 5 cm soil depth [42]. The hillslope has a soil depth of 1.6 m, covered with a thin organic layer (litter and root mat approximately 3 cm). The soil shows low near-saturated hydraulic conductivity, measured at 2.37 ± 1.27 mm h⁻¹ (mean ± standard deviation) using a mini disk infiltrometer (Decagon Devices, Inc., Pullman, WA, USA) at five points on the hillslope. Both root mat surface and soil surface showed strong water repellency during rain-free periods, as measured at a nearby Japanese cypress catchment [43].

2.3. Study Plots

Three open-type study plots were selected on an approximately 25 m wide hillslope on a northeast-facing slope (Figure 1c and Figure 2). Open-type refers to a plot without borders. Five small-sized traps were placed at the bottom of each plot (15 traps in total, Figure 1c,d and Figure 2). The distance between each trap was 1 m in each plot (Figure 1d). Plots A and B were designated as treatment plots, and plot C served as the control. In plot A, 11 logs (18 cm in diameter and 2 m in length) were randomly placed on the ground after thinning (random-felled logs or not aligned with contour lines). In plot B, 20 logs (19 cm in diameter and 2 m in length) were securely staked parallel to contour lines after thinning (contour-felled logs). The condition of plot C remained unchanged before and after thinning. The general statistics of the study plots are shown in

Table 1. General statistics of study plots.

Period	Plot	Tree Height (m)	DBH 1 (cm)	Tree Density (Tree ha−1)	Average Slope Angle
Before thinning	A	15.7 ± 0.7	20.0 ± 3.7	1800	34.8°
	B	16.81 ± 1.0	20.4 ± 2.7	2000	31.9°
	C	16.4 ± 0.5	20.2 ± 2.5	1520	30.1°
After thinning	A	15.79 ± 0.8	20.6 ± 4.0	1100	34.8°
	B	17.1 ± 1.1	21.0 ± 2.9	1200	31.9°
	C	16.4 ± 0.5	20.2 ± 2.5	1520	30.1°

Notes: ¹ DBH: Diameter at breast height. Tree height and DBH values are mean ± standard deviation.

.

2.4. Small-Sized Trap Design

The small-sized traps allowed us to simultaneously collect overland flow, soil erosion, and litter erosion (Figure 3). Traps with dimensions of 0.15 × 0.25 × 0.20 m (height × width × depth), made of stainless steel, were installed by horizontally inserting the plate into the root mat and top soil (5 cm, Figure 3). The outlet side of each trap was wrapped with a 1 mm mesh to capture soil and litter (Figure 3c). A gutter was installed below the outlet to collect overland flow, which was connected to a tank by a hose (Figure 3b,c). A plastic roof prevented direct rainfall into the gutter (Figure 3c). Since we used open-type plots (each trap without borders), where the contribution area is uncertain, overland flow and soil and litter erosion were reported as transport rates per unit width (L m⁻¹ and g m⁻¹). Overland flow included both the Hortonian overland flow and subsurface flow through the root mat and top soil layers [44], facilitated by inserting the plate into these layers.

2.5. Data Collection and Processing

Rainfall data were recorded using a 0.5 mm tip⁻¹ tipping bucket rain gauge (OW-34-BP, Ota Keiki Seisakusho Co., Ltd., Tokyo, Japan) at a nearby weather station (500 m from the study plots). Rainfall data were recorded every 10 min and converted to monthly data. Overland flow and soil and litter erosion data were also collected at monthly intervals. Data from October 2019 to April 2020 were missing due to the removal of small-sized traps, the thinning operation, and the re-installation of the traps. The small-sized traps were removed and re-installed carefully to minimize disturbances. Thinning was performed using a chainsaw, ensuring minimal disturbances to the forest floor. Thinning was conducted by prioritizing the removal of trees with poor growth conditions, such as those with crooked trunks or split crowns. To achieve the target thinning ratio of 40% per hectare, some healthy trees were also cut if necessary.

Overland flow was measured in a tank (Figure 3b) using a ruler, and the measurements were converted to water volume. Addionally, water samples (250 mL) were collected from the tank to account for any fine soil that had passed through the mesh and entered the tanks. After measuring the water height and collecting the water samples, all tanks were drained and cleaned to prepare for the following month’s data collection.

Samples comprising a mixture of soil and litter (inside each small-sized trap) were air-dried after collection (Figure 3c). The air-dried samples were suspended in a bowl of water and well mixed. After several minutes, the floating litter was scooped out, and any remaining litter on the bottom was picked up. The water and soil were allowed to settle for 24 h, after which the clear water was removed. The soil samples and removed litter were then placed in an oven (DX602, Yamato Scientific Co., Ltd., Tokyo, Japan) for 72 h at 60 °C, and the dried mass of soil and litter were weighed. Following this protocol [42], no large gravel (>2.0 mm) was found, and the litter consisted of leaves, twigs from Japanese cypress trees, and fine organic debris.

Fine soil in the sampled water was obtained by filtering the water samples through a 1 µm glass-fiber filter (GS-25; ADVANTEC, Tokyo, Japan), which was then dried at 60 °C in an oven (DX602; Yamato Scientific Co., Ltd., Tokyo, Japan) for 72 h and weighed. The fine soil concentration was multiplied by the volume of water in each tank. The total soil erosion data were calculated as the sum of the soil collected in the trap and the fine soil in the sampled water.

2.6. Statistical Analyses

Fortran 98 (ISO/IEC 1539-1:1997 [45]) was used for data arrangement and compiling. R version 4.1.3 (R Development Core Team, Vienna, Austria) was used for data visualization and statistical analyses. The average data from five traps in each plot were used to illustrate general results. If water overflow was observed in a tank on the data collection day, the overland flow data for that month was considered a data gap. We employed two different approaches for the statistical analyses: paired-plot and paired-trap analyses.

For the paired-plot analysis, we used the average data of five traps in each plot, comparing treatment plots to the control plot during pre- and post-thinning periods. We utilized analysis of covariance (ANCOVA) to compare the slopes of regression lines between treatment plots and the control plot in the pre- and post-thinning periods. Post hoc comparisons with Tukey’s honestly significant difference (HSD) test were performed to determine whether the slopes of the regression lines differed between the pre- and post-thinning periods.

In the paired-trap analysis, we began by identifying the paired-trap with the highest coefficient of determination in the treatment plot with the control plot using pre-thinning data. This paired-trap was then used to compare pre- and post-thinning data with the ANCOVA test, followed by the post hoc Tukey’s HSD test. The paired-trap analysis was conducted for each trap within the treatment plots to minimize the influence of spatial variability and ensure that differences observed in the paired-plot analysis were due to the treatment effects rather than inherent plot variability.

We also performed cumulative frequency distribution to indicate treatment effects (observed subtract estimated values). Based on pre-thinning data, we developed calibration regression equations between data (overland flow and soil and litter erosion) in the control plot (C) and treatment plots (T) as follows:

$${\mathrm{C}}_{\mathrm{p}\mathrm{r}\mathrm{e}}=\mathrm{a}{\mathrm{T}}_{\mathrm{p}\mathrm{r}\mathrm{e}}+\mathrm{b}$$

(1)

where a and b are the regression coefficients. Using these parameters (i.e., a and b), we calculated the estimated data in treatment plots during the post-thinning period (${\mathrm{T}}_{\mathrm{p}\mathrm{o}\mathrm{s}\mathrm{t}}^{\mathrm{e}\mathrm{s}\mathrm{t}}$) as follows:

$${\mathrm{T}}_{\mathrm{p}\mathrm{o}\mathrm{s}\mathrm{t}}^{\mathrm{e}\mathrm{s}\mathrm{t}}=\mathrm{a}{\mathrm{C}}_{\mathrm{p}\mathrm{o}\mathrm{s}\mathrm{t}}^{\mathrm{o}\mathrm{b}\mathrm{s}}+\mathrm{b}$$

(2)

where ${\mathrm{C}}_{\mathrm{p}\mathrm{o}\mathrm{s}\mathrm{t}}^{\mathrm{o}\mathrm{b}\mathrm{s}}$ is the observed data in the control plot during the post-thinning period. We then calculated the residual between the observed and estimated post-thinning values for cumulative frequency distribution of plots A and B as follows:

$$\mathsf{\Delta }{\mathrm{T}}_{\mathrm{p}\mathrm{o}\mathrm{s}\mathrm{t}}={\mathrm{T}}_{\mathrm{p}\mathrm{o}\mathrm{s}\mathrm{t}}^{\mathrm{o}\mathrm{b}\mathrm{s}}-{\mathrm{T}}_{\mathrm{p}\mathrm{o}\mathrm{s}\mathrm{t}}^{\mathrm{e}\mathrm{s}\mathrm{t}}$$

(3)

The relationships between monthly overland flow and 1 h maximum rainfall, as well as between overland flow and soil and litter erosion, were investigated using Pearson’s correlation analysis. This analysis was performed to understand the mechanisms of overland flow and soil and litter erosion generation in each plot during the pre- and post-thinning periods. A 95% confidence interval was used to assess the significance and reliability of the correlation results.

3. Results

During the study period, the average annual rainfall at the study site was higher than that recorded at nearby meteorological stations (

Table 2. Annual rainfall measured at Obora Experimental Forest (OEF) site, Obara Meteorological Station (OMS), and Toyota Meteorological Station (TMS).

Period	Year	OEF (mm)	OMS	TMS
Before thinning	2018	2091.5	1862.5	1599.5
	2019	1713.5	1605	1623.5
After thinning	2020	1782	1976.5	1696.0
	2021	2593	2044.5	1719.0
	2022	2098	1749.0	1476.0
	2023	2110.6	1885	1522.5
Mean		2064.8	1853.7	1606.08

). The annual overland flow and soil and litter erosion data are presented in

Table 3. Annual overland flow, soil erosion, and litter erosion in study plots.

Period	Year	Plot	Overland Flow (L m−1)	Soil Erosion (g m−1)	Litter Erosion (g m−1)
Pre-thinning	2018	A	186.0	83.2	143.7
		B	231.8	93.8	180.1
		C	288.5	70.3	150.6
	2019	A	139.1	56.2	129.4
		B	341.8	62.7	139.3
		C	273.0	49.7	129.2
Post-thinning	2020	A	422.0	131.2	165.3
		B	337.1	146.7	192.7
		C	678.3	43.6	102.7
	2021	A	458.3	187.4	273.5
		B	556.2	250.9	314.6
		C	876.9	71.9	157.9
	2022	A	579.9	206.3	22.7
		B	358.4	287.0	207.7
		C	503.5	51.7	83.4
	2023	A	573.3	90.0	161.3
		B	347.8	146.1	170.2
		C	465.3	46.5	86.8

. Overland flow increased in the first year after thinning in plots A and C, whereas plot B remained unchanged (Table 3). Plot B had the lowest annual overland flow in each year in the post-thinning periods (except 2021) compared to plot A and C (Table 3). In 2021, which had the highest annual rainfall (2593 mm) during the study period, the annual overland flow was higher in the control plot (876.9 mm) than in the treatment plots (A: 458.3 mm, B: 556.2 mm, Table 2 and Table 3). Soil and litter erosion increased in both treatment plots in the first year after thinning, whereas plot C showed no changes (Table 3).

3.1. Monthly Seasonal Variability

Overland flow did not show any seasonal variability but showed high fluctuations during both the pre- and post-thinning periods, regardless of plot type (Figure 4a,b). In contrast, soil and litter erosion showed distinct seasonal variability (mostly during the rainy season), with more pronounced changes during the post-thinning periods in both treatment plots (Figure 4a,c,d). The seasonal variability of soil erosion showed an increasing trend until the third year after thinning, regardless of the treatment plots (Figure 4c). Litter erosion had an increasing trend in the first two years after thinning (2020 and 2021), then started decreasing until the end of the monitoring period in both treatment plots (Figure 4d).

3.2. Paired-Plot and -Trap Analyses

The paired-plot analysis showed that overland flow increased only in plot A in post-thinning period compared to pre-thinning (Figure 5a). In plot B, overland flow remained consistent with pre-thinning levels during the post-thinning period (Figure 5b). Soil and litter erosion increased in both plots A and B from the pre- to post-thinning periods (Figure 5c–f).

The paired-trap analysis revealed that overland flow increased in two out of five traps (No. 4 and No. 5) in plot A during the post-thinning period (Figure 6). In plot B, trap No. 1 showed a decline in overland flow after thinning and contour-felled log treatment, whereas the other traps remained unchanged (Figure 6).

Soil erosion data indicated that three out of five traps in plot A (No. 1, 2, and 5) showed higher soil erosion during the post-thinning period compared to the pre-thinning period (Figure 6). Additionally, two traps in plot A (No. 1 and No. 2) with higher soil erosion did not show a corresponding increase in overland flow during the post-thinning period (Figure 6). In plot B, two traps (No. 2 and No. 4) showed higher soil erosion during the post-thinning period, whereas overland flow in these traps remained unchanged before and after thinning and contour-felled log treatment (Figure 6).

Litter erosion increased only in trap No. 5 in plot A during the post-thinning period compared to the pre-thinning period (Figure 6). Trap No. 5 in plot A also showed high overland flow and soil erosion during the post-thinning period (Figure 6). In plot B, trap No. 1 showed a decline in litter erosion during the post-thinning period compared to the pre-thinning period, which also corresponded with a decline in overland flow (Figure 6). However, traps No. 3 and No. 4 in plot B showed an increase in litter erosion, whereas overland flow remained unchanged and soil erosion increased only in trap No. 4 during the post-thinning period compared to the pre-thinning period (Figure 6).

3.3. Cumulative Frequency Distribution

The cumulative frequency distributions in Figure 7 show changes in overland flow and soil and litter erosion due to thinning treatments in plots A and B. Plot A, with random-felled logs, showed an increase in both high-frequency and low-frequency overland flow after thinning (Figure 7a). In contrast, plot B, employing contour-felled logs, only showed an increase in high-frequency overland flow in the post-thinning period compared to the pre-thinning period (Figure 7b).

Both plots A and B experienced a higher frequency of soil erosion in the post-thinning period (Figure 7c,d). However, litter erosion remained unchanged from pre- to post-thinning periods in both treatment plots, indicating no significant impact of the thinning treatments on litter erosion dynamics (Figure 7e,f).

3.4. Driving Factors of Overland Flow and Soil and Litter Erosion

We considered 1 h maximum rainfall (mm h⁻¹) instead of monthly totals in our analysis, as not all the rain in a month contributed to overland flow generation (Figure 8a,b). The 1 h maximum rainfall showed a positive relationship with overland flow in the pre-thinning period across all plots (Figure 8a,b). In Plot B, a stronger positive relationship between overland flow and 1 h maximum rainfall was observed in the post-thinning period compared to the pre-thinning period, while plots A and C showed no change (Figure 8a,b).

Raindrop kinetic energy is a driving factor in soil and litter erosion, as it detaches unstable soil and litter aggregates. We examined the relationship between soil and litter erosion and overland flow, as the flow transports these aggregates downslope following their detachment by raindrops. The relationship between soil erosion and overland flow was relatively consistent across all study plots during the pre-thinning period, showing similar positive correlations (Figure 8c). In contrast, the treatment plots exhibited a stronger positive relationship between soil erosion and overland flow compared to the control plot in the post-thinning period (Figure 8d). Notably, plot B exhibited the strongest relationship during this period, in contrast to plots A and C (Figure 8d).

The relationship between litter erosion and overland flow was similar to that of soil erosion (Figure 8e,f). Plot B also demonstrated the strongest relationship between litter erosion and overland flow in the post-thinning period compared to plots A and C (Figure 8f).

4. Discussion

We examined the impact of thinning and felling logs on overland flow, as well as soil and litter erosion, in a Japanese cypress forest plantation. Thinning combined with contour-felled logs (plot B) proved more effective in controlling overland flow compared to thinning combined with random-felled logs (plot A). These findings are consistent with our previous study [4], conducted in a closed-type plot (i.e., plots with borders), which also demonstrated the efficiency of thinning and contour-felled logs in controlling overland flow compared to thinning and random-felled logs in a nearby Japanese cypress forest plantation with similar tree density. Bombino et al. [46] reported a significant reduction in runoff following the use of contour-felled burnt logs, combined with spontaneous herbaceous and shrubby vegetation, in a post-fire Mediterranean pine forest in Italy. The effectiveness observed in Bombino’s study [46] was likely due to the high soil disturbance caused by the wildfire, whereas in our study, the minimal soil disturbance during thinning resulted in no significant changes in overland flow in the contour-felled logs plot compared to the control plot.

4.1. Changes in Overland Flow Post-Thinning

Thinning combined with random-felled logs resulted in increased overland flow compared to the control plot, indicating that this method negatively impacts overland flow control. These findings are consistent with Kubota et al. [19], where thinning followed by the removal of felled logs also resulted in similar increases in overland flow. This suggests that random-felled logs do not contribute to mitigating overland flow and are as ineffective as removing the logs entirely. Conversely, thinning combined with contour-felled logs did not result in changes in overland flow. Previous studies have reported an immediate increase in overland flow following thinning in Japanese cypress plantations, primarily due to increased canopy openness, which enhances throughfall reaching the forest floor and reduces transpiration, leading to higher overland flow [47,48]. However, our results indicate that despite increased canopy openness, contour-felled logs effectively prevented an increase in overland flow in the post-thinning period. Therefore, our first hypothesis was supported, as thinning combined with contour-felled logs proved more effective in mitigating overland flow than thinning with random-felled logs.

Despite expectations that contour-felled logs might reduce overland flow after thinning, our results did not show a significant reduction in overland flow with this treatment. While contour-felled logs may offer some benefits, they were not sufficient to effectively mitigate overland flow under the specific conditions of our study. Similarly, Kim et al. [49] observed the ineffectiveness of contour-felled logs in reducing runoff in a post-fire Japanese red pine forest in South Korea, ascribing it to the small diameter of the logs (7–10 cm). In contrast, the outcome in our study was related to specific areas contributing to overland flow, rather than the diameter of the felled logs (18 cm). Gomi et al. [23] identified a vital zone for generating overland flow at the lower part of a slope rather than the entire hillslope. In our study, most felled logs were placed in the upper and middle parts of the slope, while the lower part (near the small-sized traps) remained unchanged before and after thinning. In plot A, the random placement of logs increased overland flow by directing water downslope [4]. In plot B, with contour-felled logs, the logs effectively halted overland flow in the upper and middle parts of the slope, but the lower part still contributed to generating overland flow. This can be attributed to the flow paths and residence time of the water. In the upper and middle sections, the contour-felled logs slowed down the water, increasing its residence time and allowing more infiltration into the soil [31]. However, in the lower part of the slope, where the logs were less effective or absent, the water could accumulate and flow more rapidly, leading to continued overland flow. This suggests that the placement and distribution of contour-felled logs are crucial for effectively managing overland flow across the entire slope. Overall, contour-felled log placement was more effective in controlling overland flow than random-felled logs.

4.2. Changes in Soil and Litter Erosion Post-Thinning

Both treatment plots had a negative impact on controlling soil erosion, as soil erosion increased in the post-thinning period compared to the pre-thinning period. Consequently, we can conclude that our second hypothesis was rejected, as both treatment plots exhibited an increase in soil and litter erosion following thinning. Fernández and Vega [27] reported that erosion barriers (i.e., contour-felling with a trench upslope) were effective only for a relatively brief period (a few months) in controlling sediment after severe wildfires in Spain. Similarly, Adams et al. [50] and Cole et al. [51] observed the short-term effectiveness of log barriers in reducing erosion in post-fire forested areas in the USA. In all cases, the capacity of the trenches was exceeded after several months, rendering the erosion barriers ineffective [27,50,51]. In our study, with low soil disturbance, the increase in soil erosion could be attributed to the exposure of unstable soil aggregates following thinning. The disturbances caused by thinning operations on the forest floor, combined with canopy openings, exposed more ground area to direct rainfall, which facilitated the detachment of soil aggregates, particularly on the downslope near the small-sized traps. This finding was supported by the results, which showed increased soil erosion in traps where overland flow did not increase.

Soil and litter erosion increased until the third and second years after thinning in both treatment plots, indicating that neither random-felled nor contour-felled logs effectively controlled erosion. After this period, erosion rates began to decrease, suggesting that most unstable soil and litter aggregates had been removed, allowing erosion rates to return to pre-thinning levels. Miyata et al. [52] observed that soil erosion increased for several years following removal of the litter layer, but eventually stabilized. They explained that unstable soil aggregates were removed and exhausted after several years of litter layer removal, leaving behind stable aggregates that are resistant to detachment by raindrop kinetic energy [52].

We also observed that increases in soil and litter erosion occurred in traps where overland flow did not increase in the post-thinning period. This discrepancy could be attributed to the detachment of soil and litter aggregates that were splashed into nearby traps by the kinetic energy of raindrops rather than being transported by overland flow [6].

4.3. Changes in Hydrological and Erosional Processes Post-Thinning

The relationship between overland flow and 1 h maximum rainfall in the plot with thinning and contour-felled logs (plot B) became stronger in the post-thinning period, whereas the other treatment plot with thinning and random-felled logs (plot A) and the control plot remained unchanged. These results suggest that thinning and contour-felled logs were successful in altering the hydrological processes. Furthermore, based on the results of the frequency distribution curve, we noticed that thinning and contour-felled logs were well-designed for controlling low-frequency overland flow but not high-frequency overland flow. In general, forest thinning leads to changes in hydrological processes. Sun et al. [36] observed significant increases in flow peaks and quick flow following high-intensity rainfall events after thinning in a Japanese cedar and cypress catchment. Robles et al. [53] found that thinning resulted in increased runoff only during years with higher winter precipitation (more than 230 mm) in a pine forest in central Arizona, USA. Farahnak et al. [4] found that a higher rainfall intensity was required to generate overland flow in post-thinning areas combined with contour-felled logs. Therefore, we concluded that thinning combined with contour-felled logs could result in more favorable outcomes in controlling overland flow, particularly when applied at hillslope or catchment scales.

4.4. Limitation of Our Study

One limitation of our study was the monthly interval of data collection, which made it difficult to precisely explore the mechanisms of change in overland flow and soil and litter erosion in the post-thinning period. Future studies could benefit from event-based overland flow and soil and litter erosion measurements to better elucidate these processes. Event-based data could help explore the relationship between rainfall characteristics (rainfall amount, intensity, and duration) and overland flow and soil and litter erosion.

4.5. Suggestions

We suggest placing felled logs not only on the upper and middle parts of a hillslope after thinning but also on the downslope to improve the control of overland flow and soil and litter erosion. Contour-felled logs can function as water diversion structures across a hillslope, not only for reducing overland flow but also decreasing the transport of eroded materials before they are deposited in downhill streams, check dams, and rivers [54]. Additionally, we recommend placing slash (branches from cut trees) between intervals of felled logs on the slope to increase ground cover and further reduce overland flow and soil and litter erosion. Jourgholami et al. [55] suggested that combining organic mulch, such as leaf litter, with contour-felled logs may yield acceptable results in reducing soil penetration resistance. Nainar et al. [56] highlighted the importance of the litter layer in controlling overland flow in secondary broadleaf forests. In our study, we observed that litter erosion was as significant as soil erosion in the treatment plots. Therefore, incorporating slash could be a practical approach to increasing organic matter entering the soil, enhancing soil structure, improving water infiltration, and reducing direct surface runoff.

5. Conclusions

Overland flow, soil erosion, and litter erosion were monitored over six years in three adjacent open-type plots within a monoculture Japanese cypress forest plantation. Two years into the monitoring period (2018–2019), a 40% thinning (trees ha⁻¹) was conducted in two treatment plots: one with thinning and random-felled logs, and the other with thinning and contour-felled logs. Supporting our first hypothesis, thinning combined with contour-felled logs showed better control results by maintaining unchanged overland flow from the pre- to post-thinning periods compared to the thinning and random-felled logs plot. However, our second hypothesis was rejected, as erosion increased in both treatment plots. The rise in soil and litter erosion was primarily attributed to thinning disturbances and uncovered forest floor areas in the post-thinning period. We suggest that additional ground cover is necessary to mitigate overland flow, soil erosion, and litter erosion, particularly by placing more contour-felled logs and slash in key areas, such as downslope regions that significantly contribute to overland flow. These findings emphasize that incorporating additional ground cover, such as contour-felled logs and slash, especially in high-risk areas, can help reduce soil and litter erosion. Implementing this approach can assist forest managers in controlling overland flow and erosion in monoculture forest plantations on steep terrain.