Understanding the Impacts of Forest Management in Sal (Shorea robusta) Dominant Forest Stands in the Western Lowlands of Nepal

Sal (Shorea robusta) is an economically and ecologically important tree species found in Nepal. Since 2003, Nepal's government has been managing the Sal forest under the Scientific Forest Management (SciFM) scheme. Due to newly implemented approaches, the information regarding the Sal forest condition under regeneration felling and thinning is low. This study aimed to understand the regeneration status after regeneration felling and compare crop stands between thinned and unthinned plots. We selected three years of harvested, thinned, and unthinned blocks in the Tilaurakot collaborative forest. Vegetation sampling was carried out in 63 concentric circular plots. The results show that the number of seedlings and saplings in the harvested blocks was 14,000 and 3368 per hectare, respectively. The growing stock and basal area per hectare in the thinned blocks were lower than in the unthinned blocks. The numbers of trees and poles per hectare were lower in the thinned block than in the unthinned block. Sal's importance valve index (IVI) was higher than other species in all three block types. The study suggests that the regeneration condition was better after the canopy's opening, and thinning promotes the growth of trees and undergrowth vegetation.


Introduction
Sal (Shorea robusta) is one of the important timber species that grows between the longitudes of 75° and 95° E and the latitudes of 20° to 32° N (Gautam 2001). Within this range, the distribution of Sal is controlled by climate and edaphic factors. (Gautam and Devoe 2006). In Nepal, Sal is dominant in the Bhabar Zone and covers most of the Siwalik Hills and the dunes between them (Paudyal 2016). Sal regenerates from seed, coppicing, and sprouting from root suckers is common (Gautam and Devoe 2006). According to the guideline of scientific forest management, the optimal rotation of Sal is 80 years. However, it depends upon the regeneration method. Coppice-with-standard systems are managed under a rotation of 40-60 years (Tewari 1995). In addition, the study suggested that Sal trees reach financial maturity at an average of 60 years (Magarath et al. 2013). Sal's current Government royalties for round logs is approximately $125 per cubic meter, while sawn timber price is more than $1400-1800 per cubic meter (Magarath et al. 2013). Because of its silvicultural characteristics and economic importance, the government attempted to manage Sal forests for commercial timber production to increase revenue (Gautam and Devoe 2006). In addition to timber and fuelwood, Sal produces fodder (Upadhyay 1992;Shakya and Bhattarai 1995;Gautam and Devkota 1999), leaves for various purposes (e.g., umbrellas, plates, and livestock bedding), resin (FRIB 1947), and oil (Sharma 1981).
Historically, the Sal forests were managed solely by the ruling elite's interests (Gadgil and Guha 1993). The aims of forest management were limited to timber production and economic benefits (Wiersum 1995). This excluded the local community from utilizing forest resources and, in turn, resulted in the leakage of forest products that degraded the forest condition. As a solution to forest degradation, the government of Nepal began the implementation of "Scientific Forest Management" (SciFM) which primarily follows a shelterwood silviculture system (Poudyal et al. 2020). The SciFM aims to enhance the productivity of the forest products that will ultimately increase the government's revenue and fulfill the necessities of the local users and maintain or enhance biological diversity (Jayasawal and Bishwokarma 2016). These SciFM practices were begun in Nepal using collaborative forest management (CFM) principles. While implementing the SciFM, the forest areas are divided into different blocks, compartments, and sub-compartments. This is followed by various silvicultural operations, such as regeneration felling, thinning, and improvement felling. In regeneration felling, only 20-30 mature mother trees per hectare are retained (Poudyal et al. 2020). In addition, thinning activities are carried out in adjacent sub-compartments. The timber and fuelwood harvested from silvicultural operations are distributed to the planned beneficiaries (Jayasawal and Bishwokarma 2016). In 2012, the CFM guideline defined that 50% of the benefits from SciFM go to the local community group and the remaining 50% go to the government's treasury. Because of this win-win mechanism, it can be said that thinning is one of the most prominent silvicultural activities in SciFM.
Thinning plays a critical role in promoting natural regeneration (Zhu et al. 2003). Mostly it is carried out in dense natural stands with seeding from natural dispersal, and adequate spacing is not present. Thinning helps to remove the poor-quality trees and create a favorable environment for targeted species. Due to thinning, environmental conditions such as light availability, soil temperature, and moisture content are significantly affected (Ritter et al. 2005), which helps in breaking seed dormancy (Jia et al. 2008). As a result, thinning changes stand structure (Zhou et al. 2013) as the Sal forest is dominated light-demanding tree species, the canopy opening in a forest stand promotes regeneration and the growth of understory seedlings and saplings (Troup 1986;Gautam 1990). Thinning also creates a favorable condition for seedling growth (Rautiainen and Suoheimo 1997).
Several studies have been conducted to determine the effect of thinning in community-managed forests (Timilsina et al. 2007;Oli and Subedi 2015;Baniya et al. 2019). However, the joint practices of regeneration felling and thinning in collaborative forest management regimes are still unknown. Since 2003, the government and local community have jointly managed the large extent of Sal forests in the Terai region using CFM with the objectives of enhancing social, economic, and environmental benefits for local and distant users. In Nepal, most of the studies on CFM have focused on subjects such as benefits sharing mechanism (Rai et al. 2017), policy (Bampton et al. 2007), and implementation strategies. However, knowledge and information on the effect of forest management, specifically regeneration felling and thinning, on forest vegetation in the CFM are very low. Thus, this study intends to fill the gap by assessing regeneration conditions under different types of silviculture treatments in Nepal. The study's findings will quantify the effects of thinning on seedlings, saplings, pole-size trees, and the tree in Sal forests. The information can be used to facilitate a longitudinal study, predicting the effect of thinning on growing stock and forest health. As a result, the study will contribute to achieving SciFM objectives, creating a healthy forest and continuing the supply of forest products to the local and distant users.

Description of the Study Area
The study was carried out in natural forests in the Tilaurakot Collaborative Forest Management area of Kapilvastu, Nepal (Fig. 1). The study site has a tropical monsoon climate with an elevation from 200 to 300 m above sea level. The latitude of this area is 27°35′8.13″ to 27°45′32.98″ N with longitude of 83°1′56.78″ to 83°9′6.49″ E. Tilaurakot forest lies along the East-West highway in between the Balhundra River and Gorusinghe spreading in a North and South direction. Because of its strategic location near the national highway, it makes potential to manage forests scientifically by involving stakeholders and people as per the collaborative management for the social and economic development of forest products at the national level. It consists of various tropical tree species, including Sal (Shorea robusta) as a dominant species, followed by Sissoo (Dalbergia sissoo), Asna (Terminalia

Data Collection and Analysis
Within the Tilaurakot Collaborative Forest Management unit, we have selected three blocks (viz: harvested, thinned, and unthinned blocks) for the study. The sizes of the blocks were different. The thinned and unthinned blocks were 30 ha each, whereas the harvested block was 6 ha. Approximately 15 mother trees per hectare were retained in the harvested block for seeding. The harvesting and thinning activities were carried out in the blocks in 2015. The study data were collected three years after the silviculture treatments in 2018. We assumed that the three years period was a sufficient time period to observe the impact of thinning and regeneration felling on regeneration status.
To measure the regeneration status and crop stand condition, the individual trees were grouped into four different categories (CFM OP 2009). Individual plants were categorized into seedlings (height [ht] ≤ 1.3 m), sapling (diameter at breast height -[dbh] < 10 cm and ht > 1.3 m), pole-size tree (10 < dbh < 30 cm) and trees (dbh ≥ 30 cm). A boundary survey was carried out using a geographical positioning system (GPS), and sample plots were laid out on the map using ArcGIS 10.2.1 software (ESRI 2011). Forest inventories with systematic random sampling were conducted 11 concentric sample plots were taken from the harvested block and 26 from each of the thinned and unthinned blocks (Fig. 2). A sampling intensity of 4% was used for thinned and unthinned blocks, whereas a 7% sampling intensity was used for the harvested block. The height and dbh of trees, pole-size trees, and saplings were measured using a range finder (Apresyes Pro550) and diameter tape, respectively. Plot sizes of 400 m 2 with radii of 11.29 m were established in the forest to measure the dbh and height of trees (Fig. 2). For pole-size trees, 100 m 2 plots with radii of 5.64 m were used, and dbh and height of tree species were measured (Fig. 2). We only measured the dbh for saplings within a radius of a 2.83 m plot. We counted the number of seedlings within 10 m 2 plots with radii of 1.78 m.
Growing stock and basal area were calculated by using the following formulas: In addition, the Importance Value Index (IVI) was calculated to identify the relative abundance of the species in the study site. The IVI was found by adding the percentage values of relative frequency, relative density, and relative dominance. Mathematically, where RF is Relative Frequency and is equal to the number of sample plots where certain species are distributed, which is divided by the number of total sample plots and multiplied by 100. RD is Relative Density and equal to the total number of certain species across all sample plots divided by the total number of all species found on the sample plots multiplied by 100.
RDm is Relative Dominance and equal to the sum of dbh of a certain species across all sample plots divided by the sum of dbh of all species found on the plots and multiplied by 100.
The secondary data related to the location of the forest, forest tree species, and other relevant data were collected using secondary sources, such as publications from the divisional forest office, sector forest offices, libraries, and the operational plan of the CFM. For the statistical analysis, both descriptive and inferential statistics were used. We used the Analysis of Variance (ANOVA) test to evaluate the significance differences in regeneration status among the three blocks (viz; harvested, unthinned, and thinned). In addition, a t-test was used to compare the mean of growing stock, basal area, and species count per hectare between the two blocks (viz; unthinned and thinned).

Regeneration Status
The status of regeneration in the harvested block seems better as compared to the thinned and unthinned blocks. Table 1 presents the seedling and sapling counts in Basal area = dbh 2 ∕4 Source ∶ Khanna and Chaturvedi (1994) Growing stock = Basal area * height * form factor Source ∶ DoF (2004) the three blocks. The number of seedlings per hectare was 14,000, and the number of saplings per hectare was 3368 in the harvested block. The regeneration status of seedlings was 14,461 per hectare, and the number of saplings was 1846 per hectare in the thinned block. The regeneration of seedlings was 11,000, and the number of saplings was 1320 per hectare in the unthinned block. The ANOVA test shows the significant difference in sapling count between harvested, thinned, and unthinned blocks, whereas no statistically significant difference was found for seedling counts.

Crop Stands in Thinned and Unthinned Block
The condition of the crop stands can be measured by the presence of growing stock and basal area. Our study compares basal area, growing stock, and crop stand [count per hectare] between the thinned and unthinned blocks ( Table 2). Not surprisingly, the results show that the per hectare basal area, growing stock, and crop stand in the unthinned block was higher than thinned block ( Table 2). The growing stock of poles between the thinned and unthinned block were statistically significant but not basal area nor crop stand. However, the growing stock, basal area, and crop stand (count per hectare) for trees between the thinned and unthinned block were statistically significant. The detailed output for the t-test is available in the Supplementary Materials (SM1 & SM2).

Importance Value Index
The study found the forest was dominated by Sal (Shorea robusta) species and followed by Asna (Terminalia tomentosa). Others species, such as Sissoo (Dalbergia Sissoo), Amala (Phyllantus emblica), and Bajhi (Anogeissus latifolia), were also found in the study plots. The importance value index (IVI) for Sal was highest compared to Asna and other species in all three blocks; harvested (266.26), thinned (176.48), and unthinned (256.10) (Table 3).

Discussion
The regeneration status in the harvested block is better compared to the thinned and unthinned blocks. The presence of quality seedlings and saplings is one of the indicators of regeneration status. The numbers of seedlings and saplings per hectare

3
Understanding the Impacts of Forest Management in Sal (Shorea… were higher in the harvested block than in the thinned and unthinned blocks in the study areas. A higher number of seedlings and saplings per hectare in harvested and thinned blocks suggests that an open canopy favors regeneration. A study by Rai et al. (1999) found that a dense forest with a high density of larger trees was not a favorable condition for natural regeneration. In addition, regeneration status is also determined by disturbance intensity (Sapkota 2009). Khanal and Adhikari (2018) found that, on average, the number of seedlings per hectare increased by 1.2-24 times due to regeneration felling in community forests of the Rupandehi district of Nepal. However, the increase in the number of saplings after regeneration felling was relatively less, which ranges from 1.3 to 8.8 times (Khanal and Adhikari 2018).
Our study shows that Tilaurakot CFM has a good status for seedlings and saplings. However, during the data collection time, regeneration in the harvested block was not found to be uniform. It was observed that regeneration status varied between the plots of each block. The success and failure of natural regeneration depend upon local factors. Negi et al. (2002) found that Sal borer (Hoplocerambyx spinicornis) coupled with intensive grazing, unmanaged harvesting, and frequent forest fires were serious problems damaging Sal forests in India. In such a condition, artificial regeneration, such as seed sowing and seedling planting, could be applied to assist natural regeneration. Also, regeneration can be established by protection against grazing and forest fire.
In the same way, crop stand (count per hectare) is one of the key indicators for forest conditions. This study reported that due to thinning practices, the number of trees and pole-sized trees was lower in the thinned areas. However, the number of trees in the thinned block was still greater in number. Lewis et al. (2010) recommended that generally well-managed forests for timber production should have lower stocking (generally less than 200 trees per hectare) and consist of good form trees of desirable species. The stocking ranges between 80 and 150 trees per hectare (≥ 10 cm dbh) are common for well-managed productive forests. The intent of thinning is to create a favorable environment for existing stands. Therefore, trees need to be harvested according to the harvestable diameter as regeneration has already been established, and for better growth of that regeneration, exploitable diameter trees should be harvested in both thinned and unthinned blocks.
Another important indicator of forest condition is growing stock (m 3 per ha). As our study showed, the average growing stock in the thinned block was less than that in the unthinned block. However, the growing stock of the study site was still higher in both thinned and unthinned blocks compared to the forest operational plan of the whole Tilaurakot CFM OP (2009) and the average growing stock of Nepal (164m 3 per ha) (DoFRS 2015). This result indicates that the ratio of forest fellings to increment is relatively low. Our results suggest that forest products are underutilized. Since the intent of the establishment of collaborative forest management was to fulfill the requirements of users, the CFM fails to fully meet this objective.
Similarly, the basal area in the thinned block was lower than in the unthinned block. In addition, the study indicates that the basal area of tree and pole sizes are the same in the thinned block, whereas the basal area of tree stands was higher than the pole-sized trees in the unthinned blocks. This result provides evidence that the number of pole-sized trees was higher than the trees in the thinned block. The basal area of the study area was lower as compared to the study carried out in community forests in Chitwan, Nepal (Paudel and Mandal 2019). This is because these community forests have been managed and implemented various silvicultural operations. However, the CFM did not apply any types of forest management practices. Basal area growth response was influenced by a reduction in competition (Gradel et al. 2017). The purpose of the execution of silvicultural operations is to improve forest health.
As in a previous study (Sapkota 2009), this study shows that Sal IVI was higher as compared to other species in all three blocks. It indicates that the study site was predominately by Sal because of its silvicultural characteristics such as light demander and establishment by suckers. Although for timber, the rotation age of this species is too long (approximately 80 years), the higher market value attracts the users and the government for the management of Sal forests. Also, Sal is a naturally grown species that supports tropical biodiversity and ecosystem. Therefore, for regeneration management, Sal should be given priority in order to maintain the forest's existing ecological services.
The condition of forest types is influenced by thinning. Our results indicate that thinning enables both diameter and height growth for Sal. The thinning favors the forest stands by reducing the competition, providing more sunlight, and reducing risks from insects, diseases, and fires. However, soil type, site index, topographic factors, slope, aspects, and climatic conditions are key factors in determining the forest's growth and development that must also be considered. Thinning plays the role of catalyst in improving forest conditions. The most important point of thinning in the community-managed forest is that thinning supplies forest products such as poles, firewood, and fodder for the local people's upliftment. Loss of biodiversity, soil erosion, water source damage, and habitat loss were some of the critical issues of thinning practices in natural forests. To address the issue, it is necessary to understand the implications of thinning and executing the CFM concept in natural resources.

Conclusions
In summary, this study illustrated the effect of thinning on regeneration and existing stand conditions in Sal-dominated forests of the western lowlands of Nepal. Regeneration status was good in harvested blocks. Crop stands were better in thinned blocks as compared with the unthinned blocks. Thus, the findings conclude that thinning enhances the growth of the trees and undergrowth vegetation in the regeneration layer. As the diversity index of Sal species was found more in all treatment blocks, it can be said that Sal was the dominant species of the Tilaurakot CFM forest. To promote good regeneration in the unthinned block, the trees with harvestable size need to be removed or thinned. Also, the protection of harvested blocks against intensive grazing and forest fire should be implemented.
To carry out this study, a three-year-old harvested block was used. So, it might not fully represent the actual situation for predicting future regeneration. Future research needs to be carried out to show the effects of different forest treatments to promote regeneration, growing stock, and height growth. Follow-up research is recommended to observe the actual effect of thinning on the growing stock and forest health. Following our findings, the study suggests that proper thinning should be carried out in order to enhance the growth and undergrowth of vegetation. As Sal is dense and dominant in the lowland of Nepal, the study recommends the removal of oversized and harvestable trees to promote good regeneration.

Study Limitations
Two major constraints limit this study's findings. First, lack of baseline data (data before the thinning operations), that we could not statistically compare the means between before-and after-events (thinning treatments). Second, the smaller size of the study area and the lack of replicates limits the generalizability of the results, although vegetation composition in a sampled area is an indicator of forest management conditions. Future studies should be designed with a significant number of sampling plots in a relatively across multiple blocks to increase the study's credibility.