Introduction
Gardens have emerged as spaces that offer modern people a natural environment, as well as opportunities for leisure, experiential engagement, education, and food-related activities. Accordingly, therapeutic gardening activities and programs, often referred to as healing gardening in South Korea, has gained attention as potential solutions to the growing prevalence of social health issues (Choi, 2023; Park et al., 2022). Among various gardening-related activities, horticultural practices are particularly advantageous due to their accessibility, as individuals across age groups, including both older adults and younger populations, can participate without requiring specialized skills (Yoon and Kim 2009).
Indoor and outdoor horticultural activities involve a wide range of physical movements, spanning low to high levels of energy expenditure (Park et al., 2011, 2013a, 2014). These activities engage the muscles of the hands and both upper and lower extremities and may function as forms of weight-bearing exercise (Park et al., 2013b, 2014; Lee, 2017). However, many horticultural tasks require repetitive motions and physically demanding postures, such as squatting and bending, which can impose strain on the musculoskeletal system (Jeon 2009; Kim 2010; Kim et al., 2011). Squatting, in particular, is a common posture in gardening. Ahn (2022) reported that, among Korean farmers, musculoskeletal conditions such as knee osteoarthritis and varus deformity are associated with harvesting tasks performed in a squatting position. Prolonged and repetitive squatting may place excessive stress on the knee joint, potentially leading to a decline in physical function.
Previous research has primarily examined the exercise-related effects and rehabilitative potential of horticultural activities. For example, Lee (2017) analyzed horticultural movements using kinematic and kinetic approaches to develop horticultural therapy programs aimed at promoting both physical and psychological recovery in stroke patients. Similarly, Kim (2024) proposed participant-customized agro-healing (i.e., therapeutic agriculture, care farming, or social farming) programs based on kinematic and psychophysiological data collected during agricultural activities (Kim, 2024). However, there remains a lack of studies that quantitatively assess the muscular workload imposed on both upper and lower extremities during home gardening activities and that propose safe activity ranges to minimize physical strain.
To date, research in South Korea has largely focused on the upper body—particularly the neck, shoulders, arms, wrists, hands, and back—while relatively little attention has been given to the lower extremities. Consequently, quantitative investigations examining muscle load distribution during tasks such as handling heavy objects in a squatting posture or maintaining prolonged positions during gardening remain limited (Ahn, 2022). Therefore, the present study aims to quantitatively analyze the workload on upper and lower limb muscles in adults during home gardening activities by measuring electromyography (EMG) and exercise intensity, thereby providing foundational data for the development of safe gardening practices.
Research Methods
Participants
This study was conducted at the Institute of Sports Science at Dankook University, with an average indoor temperature and humidity of 22.9°C and 57.7%, respectively. 10 right-handed male participants in their twenties, with no restrictions on physical activity due to cardiovascular, metabolic, or musculoskeletal conditions, were recruited. All participants provided voluntary informed consent prior to participation. The exclusion criteria included a history of lumbar or knee disorders within the past six months, as well as the presence of clinically significant cardiovascular, metabolic, or musculoskeletal conditions. The participants’ physical characteristics were as follows: mean age 22.1 years (SD = 3.07), mean height 179.9 cm (SD = 4.91), mean body weight 75.35 kg (SD = 3.39), and mean body mass index (BMI) 23.32 (SD = 1.83) (Table 1). Prior to the experiment, all participants received a detailed explanation of the study’s purpose, methods, and procedures. This study was reviewed and approved by the Institutional Review Board (IRB) of Dankook University (DKU 2025-05-086-006).
Research Method
In this study, two representative home gardening activities—lifting pots and watering plants—were selected as experimental tasks. Each activity was performed under two postural conditions, resulting in four posture combinations: pot lifting in squatting and bending postures, and watering in squatting and standing postures. The plant pots used in the experiment measured 21 cm in diameter and 15 cm in height, with a weight of 10 kg. The watering cans had a capacity of 5 L, which was used to standardize task loading. The 10 kg pot weight was determined based on previous studies that evaluated changes in muscle activation, fatigue, and energy expenditure under repetitive task conditions (Ahn et al., 2025). The 5 L load for the watering task was established with reference to the manual handling guidelines of Waters et al., (1994), which indicate that even relatively light loads of 4–6 kg can impose cumulative musculoskeletal strain when tasks are performed repetitively or continuously.
The experiment was conducted under four postural conditions presented in randomized order. Participants performed 25 repetitions of the 10 kg pot-lifting task in both squatting and bending postures and completed the watering task for 30 seconds in both squatting and standing postures. To minimize muscle fatigue, participants rested for six minutes in a seated position between measurements under each postural condition (Fig. 1).
The number of repetitions for the pot-lifting task was determined to assess workload and exercise intensity characteristics. According to the American College of Sports Medicine’s Guidelines for Exercise Testing and Prescription (ACSM, 2009), performing 15 or more repetitions under low-load conditions falls within the range of local muscular endurance (LME). Therefore, 25 repetitions—exceeding the minimum threshold—were selected to ensure a stable assessment of task load and intensity. The execution time for the watering task was determined based on the results of a pilot test conducted prior to the experiment. When the task was performed using a 5 L watering can, it took an average of approximately 30 seconds to empty the can; therefore, the execution time for the watering task in this study was set to 30 seconds.
Surface electromyography (EMG) was used to assess workload in the upper and lower limb muscles. Based on previous studies, the following muscles were selected as primary movers during squatting postures: upper trapezius (UT), anterior deltoid (AD), rectus femoris (RF), biceps femoris (BF), and erector spinae (ES) (Fig. 2).
Muscle activity for each task was analyzed by defining the entire duration under each postural condition as a single continuous interval. The mean percentage of maximal voluntary contraction (%MVC) was calculated using root mean square (RMS) values obtained from this interval. Surface electrodes were attached to the midpoint of each target muscle in accordance with SENIAM guidelines. All EMG signals were analyzed using EMGworks Analysis software (version 4.8.0, Delsys Inc.), sampled at 2,100 Hz, and preprocessed using a second-order Butterworth bandpass filter (20–450 Hz) (De Luca et al., 2010).
Prior to the experiment, the maximal voluntary contraction (MVC) of each muscle was measured. Each muscle’s MVC was recorded three t imes for 5 seconds, with a 30-second rest period between trials to minimize muscle fatigue. The mean RMS value from the middle 3 seconds of each 5-second recording (excluding the first and last seconds) was used as the MVC value for each muscle. To eliminate the influence of muscle fatigue caused by MVC measurement, the measurement was completed 24 hours prior to the experiment. (Fig. 3).
For each task, RMS values were calculated from the EMG signals recorded during task performance. These values were normalized to the MVC and expressed as a percentage of MVC (%MVC) to enable inter-individual comparisons (Mirka, 1991). %MVC was calculated using the following equation:
where EMGtask represents the RMS value recorded during the task performance, EMGMVC represents the RMS value obtained during MVC measurement, and EMGrest represents the RMS value recorded at rest prior to MVC measurement. The resulting %MVC values were used to analyze the mean muscle activity for each task.
All statistical analyses were conducted using SPSS version 28.0. Differences among the four postural conditions were evaluated using the Friedman test. When significant differences were identified, post-hoc analyses were performed using the Wilcoxon signed-rank test. The significance level for post-hoc comparisons was adjusted using the Bonferroni correction.
Results and Discussion
Exercise Intensity Across Activities and Postures (METs)
The metabolic equivalents (METs) for the four experimental conditions (Table 2) indicated that pot lifting in both the squatting posture (3.84 ± 0.56 METs) and the bent-over posture (3.22 ± 0.60 METs) was classified as moderate-intensity activity (3.0–5.9 METs). In contrast, watering plants in the squatting and standing postures yielded 1.56 ± 0.45 METs and 1.37 ± 0.26 METs, respectively, both of which fall within the light-intensity range (< 3.0 METs).
These results are consistent with previous classifications of gardening task intensity reported by Park et al., (2011, 2014), suggesting that even when the same tool is used, body posture can result in significant differences in energy expenditure. During the pot-lifting task, the squatting posture elicited a higher MET value than the bent-over posture, potentially reflecting greater activation of lower-limb musculature and, consequently, a higher level of physical exertion.
In contrast, watering plants was categorized as a light-intensity activity in both squatting and standing postures, with only minimal differences in MET values between postures. This finding suggests that, under the present experimental conditions, watering constitutes a low-load activity regardless of posture.
%MVC and Risk Levels by Muscle
An EMG analysis (Table 3) revealed that, during plant pot lifting in the squatting posture, the anterior deltoid (AD, 31.17%), erector spinae (ES, 36.27%), and rectus femoris (RF, 20.01%) were classified within the high-risk or risk categories. During pot lifting in the bent-over posture, the erector spinae (ES, 41.27%), biceps femoris (BF, 30.70%), and anterior deltoid (AD, 30.41%) were all categorized as high risk, indicating substantial loading on both the spine and lower-limb muscles. These results are consistent with previous studies demonstrating that forward-flexed postures, such as trunk bending, increase the load on the erector spinae (Ahn 2022; Kim et al., 2011).
Although watering plants was classified as a light-intensity activity based on METs, the upper trapezius (UT, 34.20%) and anterior deltoid (AD, 24.01%) in the squatting posture, as well as the upper trapezius (UT, 36.45%) and anterior deltoid (AD, 29.15%) in the standing posture, were classified as high risk. This suggests that even low-intensity activities can impose considerable strain on specific upper-limb muscles. In addition, the %MVC of the anterior deltoid (AD) exceeded 24% across all activities, indicating sustained loading on the shoulder musculature. Notably, the bent-over pot-lifting condition yielded the highest erector spinae (ES) value (41.27%), reflecting the greatest spinal load among the tasks evaluated.
Overall, these findings indicate concurrent high-risk exposure of both upper-limb and spinal muscles. Therefore, postures that impose substantial loads on the spine and shoulder musculature, such as squatting and bending, should be avoided for prolonged durations. Instead, adjustments to working height, the use of assistive devices, and stretching before and after tasks are recommended. Reducing the load on the anterior deltoid (AD) and erector spinae (ES), identified as high-risk muscles, may help mitigate the risk of musculoskeletal injury during home gardening activities. To further elucidate posture-specific differences, additional %MVC comparisons between muscles were conducted.
Analysis of Workload Differences in Key Muscles by Posture
An analysis of posture-related workload differences in muscles classified as high-risk for each task revealed distinct patterns depending on task characteristics and muscle type (Tables 4 and 5). During pot lifting, the anterior deltoid (AD) exhibited minimal differences in mean %MVC between the squatting and bent-over postures, with the Wilcoxon signed-rank test indicating a small effect size (r = 0.11) (Table 4). In addition, the proportion of high-risk exposure remained elevated under both postural conditions, suggesting that AD workload was largely comparable across postures. This finding implies that AD loading during pot lifting may be influenced more by repetitive lifting actions and object weight than by posture alone.
In the same task, the erector spinae (ES) demonstrated a difference in mean %MVC between the squatting and bent-over postures, with a medium effect size (r = 0.40) (Table 4). Although the high-risk proportion was elevated in both conditions, the bent-over posture resulted in a relatively higher mean %MVC, indicating that trunk-flexed postures may be associated with increased ES workload.
During plant watering, the upper trapezius (UT) showed a small effect size across postures (r = 0.27), and the high-risk proportion was somewhat lower in the standing condition; however, overall differences in workload between postures were limited (Table 5). This suggests that upper shoulder muscle loading may remain relatively consistent across postural conditions. This suggests that upper shoulder muscle loading may remain relatively consistent across postural conditions.
In contrast, the anterior deltoid (AD) in the same task exhibited a more pronounced posture-related difference, with a large effect size in the comparison between squatting and standing postures (r = 0.53) (Table 5). Notably, the forward-flexed posture condition yielded a 100% high-risk proportion, indicating a strong association between this posture and elevated AD workload.
In summary, the anterior deltoid (AD) consistently demonstrated relatively high workload across different postures and task types, whereas the erector spinae (ES) and upper trapezius (UT) showed variability in workload levels and high-risk proportions depending on postural conditions. These findings suggest that, in the design of agro-healing and home gardening activities, workload management for the shoulder musculature should prioritize task volume and repetition, while safety strategies for the back musculature should emphasize posture-specific ergonomic design of the work environment.
Analysis of Differences in %MVC Between Muscles
Pairwise comparisons of %MVC between muscles (Table 6) revealed relatively significant differences between AD and UT (p = .005), as well as between ES and UT (p = .007) in the squatting pot-lifting condition, indicating an imbalance in workload between the shoulder musculature and the erector spinae. In the bent-over posture, significant differences were identified between upper- and lower-limb muscles, including RF-AD (p = .005) and ES-RF (p = .005). These results suggest limited utilization of lower-limb musculature alongside increased loading of the back and shoulder muscles. During squatting plant watering, significant differences were observed for RF-UT (p = .005) and ES-RF (p = .005). In standing plant watering, significant differences were found for AD-UT (p = .005), RF-UT (p = .005), BF-UT (p = .005), and ES-RF (p = .005). These findings indicate that even in tasks primarily involving the upper limbs, a measurable level of lower-limb and spinal muscle activity is still required.
Overall, these findings suggest that task type and posture may lead to recurrent overloading of specific muscle groups, potentially increasing the long-term risk of musculoskeletal disorders.
Conclusion
This study analyzed exercise intensity (METs) and electromyographic activation (%MVC) of upper- and lower-limb muscles across four representative postures in home gardening activities: lifting plant pots while squatting and bending over, as well as watering plants while squatting and standing.
The results indicated that pot-lifting tasks corresponded to moderate intensity (3.0–5.9 METs), whereas watering tasks were classified as light intensity (< 3.0 METs), with significant differences in energy expenditure observed among postures.
During pot-lifting tasks, high-risk activation levels were frequently identified in the anterior deltoid (AD), erector spinae (ES), and lower-limb muscles, including the rectus femoris (RF) and biceps femoris (BF). Although watering was categorized as a light-intensity activity based on METs, the upper trapezius (UT) and anterior deltoid (AD) were still classified as high-risk muscles, suggesting that certain upper-limb muscles may be susceptible to overload even during light-intensity activities. Furthermore, the %MVC of the AD exceeded 24% across all tasks, indicating continuous exposure of the shoulder muscles to high-risk levels.
A comparative analysis of workload among high-risk muscles across postures revealed that the anterior deltoid (AD) consistently exhibited relatively high levels of workload regardless of posture or task type. In contrast, the erector spinae (ES) and upper trapezius (UT) demonstrated posture-dependent variations in both workload and high-risk proportion.
Comparisons of %MVC across muscles showed significant differences in workload among upper-limb, lower-limb, and spinal muscles during both bending-forward and squatting pot-lifting tasks. Even in upper-body-dominant watering activities, lower-limb and spinal muscles were moderately engaged, suggesting full-body muscular loading.
These findings suggest that the design of home gardening activities should minimize prolonged bending or squatting postures, optimize working height, incorporate assistive tools, and ensure adequate rest and stretching. In particular, preventive strategies should focus on reducing the workload of the anterior deltoid (AD) and erector spinae (ES). Furthermore, adjustments to activity intensity should consider not only MET-based classifications but also %MVC-based risk categories.
This study provides quantitative evidence on the physical workload associated with home gardening tasks and offers foundational data for the development of safe gardening programs through the assessment of muscle- and posture-specific loads. Future research should focus on individualized program design and the application of assistive tools by examining muscle loads concentrated in specific postures. Additionally, analyses including diverse age groups and both sexes, as well as measurements of various muscle loads, should be performed. Given the potential for excessive loading on the knees and lower back, appropriate adjustments in working duration and the use of assistive devices are recommended. Light-intensity tasks, such as watering, tend to be easily repeatable due to their low workload, and thus may serve as complementary activities when combined with other gardening tasks according to program objectives.
This study has the following limitations. First, the participant sample was limited to healthy males in their 20s, and therefore differences related to age, sex, and physical fitness were not considered. Second, the experiments were conducted under controlled laboratory conditions using standardized weights, which may not fully capture real-world variability, including differences in task duration, environmental conditions (e.g., temperature and humidity), and irregular movements.
PDF Links
PubReader
ePub Link
Full text via DOI
Download Citation
