Temperate forests are considered most fragile hence need to recognize their vulnerability owing to continuous climatic changes and anthropogenic activities. In this study, we assessed soil respiration (SR) by using the chamber method in a natural Rhododendron (Rhododendron arboreum) forest which is recognized as the world’s largest forest type located at Annapurna Conservation Area in the temperate region of Nepal. We evaluated the consequences of multiple ecological parameters mainly climatic and biotic factors on SR variations during the month of October in 2016 and 2017. Our results confirmed that SR well corresponded with soil temperature (ST) variables represented with the highly significant () exponential curve (y = 1.049e0.529x, 2016 and y = 26.34e0.284x, 2017). And the variation in SR was mediated by a short-range (2-3°C) of ST difference in the month of October during autumn season. However, the effect of soil water content (SWC) on SR was scattered and the photosynthetic photon flux density (PPFD) stood weak to represent the SR variation. The seasonal trend of SR was compatible with the PPFD and litter input with having accountable temporal, diurnal, and interannual variations of SR, ST, SWC, and litterfall. The SR over the entire measurement period were averaged at 269.9 mg CO2 m−2·h−1 in 2016 and 295.1 mg CO2 m−2·h−1 in 2017. Our study manifested that temperate forests could store maximum soil carbon with limited emission through SR and become a larger sink of atmospheric carbon dioxide even though SR is very sensitive to environmental changes and interactively affected by multiple ecological factors. Thus, our finding is an appreciable measure for the temperate forest to understand the regional carbon balance and suggested temperate forests are valued to incorporate them in evaluating global carbon budget.

1. Introduction

Ecosystem carbon (C) assessment is so intricate to understand the global C balance depending on the major appearance of forests [1]. Forests accomplished to segregate C to become a sink and majority of that C is ultimately stored in soils [2, 3]. Evaluation of C in the soil through the estimation of C budget, storages, and fluxes is hence decisive for recognizing the ecosystem C capacity [46]. Soil release C in the form of carbon dioxide (CO2) and is represented as soil respiration (SR). SR is considered the largest C efflux from the terrestrial ecosystem back to the atmosphere [7]. To deal with prevailing climate change, understanding the dynamics of SR is becoming critical [8] from regional to global scale [9]. The belowground C should also be taken into account [10] which alters SR in response to predicted climate change scenarios as it is hypothesized that the climatic warming increase the rates of SR potentially fueling further increases in global temperatures [1]. Thus, accumulating records of SR from multiple ecoregions of the Earth is essential for a reliable estimate of soil C emission on a global scale [11].

Anticipating the response of SR is crucial owing to varying environmental circumstances, exclusively as SR is the integrate flux of respiration of the plant roots known as autotrophic respiration and the respiration of microorganisms in the decomposition of organic matters in soil known as heterotrophic respiration [12]. Global climate data manifested that forest ecosystems predominantly sequester a large quantity of anthropogenic CO2 emissions of terrestrial ecosystems [2]. Warming due to the increase in air temperature and changing precipitation or soil water dynamics [13, 14] is most dominantly effective to become common determinants of variability in the forest C balance [15]. However, litter decomposition and soil respiration in C cycle have different appearance in the temperate forest than that of the tropics [16, 17]. Although soil temperature (ST), soil water content (SWC), and the light activity (photosynthetic photon flux density (PPFD)) are considered major environmental criterion in influencing SR, seasonal fluctuations alter its variability in temporal scale [18] is different in temperate ecosystems [14, 19, 20] than that of the tropical [21, 22]. The temperate forests sequestrate and emit huge volume of atmospheric CO2, and that amount could literally contribute a significant role in the global C cycle [2]. The recently published data however indicated that temperate forests are an important sink of atmospheric C [23]. Thus, the research interest in a temperate forest is growing under the context of climatic warming [20, 24].

About 23% of Nepal’s land area has been designated under various categories of protected area which includes national parks, reserves, conservation areas, and buffer-zones [25]. Middle mountain region (110 m to 3300 m a s l) extends over 4,309,396 ha; forest covers the greatest proportion (52.30%; 2,253,807 ha) including temperate forest cover representing 44.7% across all regions [26]. The mean C stock of the forests of Nepal including above- and belowground biomass and soil C is 176.9 t/ha with 61.5% of this in the tree component and 37.8% in forest soils [27]. Temperate regions represent the most common ecoregions of Asia. Nepal occupies 12% of the total land coverage comprises of 2,000 to 3,000m altitude in which the temperate forest is distributed from east to the larger area in the west mostly dominated by the broad-leaved trees such as Quercus lamellosa, Q. semecarpifolia, Tsuga dumosa, and Rhododendron spp., in pure/mixed stands. The common forest types of the temperate zone include Rododendraon arboreum, Rohododendron barbatum, Lyonia spp., Pieris formosa, and Tsuga dumosa [27]. Rhododendron is the largest genus ranges from the tiny structure to the largest evergreen and semideciduous trees distributed in different soil composition [28, 29] throughout Asia, Europe, North America, and Australia [30, 31]. Due to the consequences of current issues of climate change, this keystone Rhododendron forest is under vulnerable condition due to lack of proper conservation and management practices with imbalance carbon; hence, need urgent protection in their habitat especially in the temperate region [32].

In Nepal, the exploration of soil C emission via SR and its regulating factors of the forest in the temperate region is yet to be focused on. This vulnerable region is even not included in the global dataset of the SR measurements [33]. The temperate regions of Himalaya are subsidized as the most fragile and prioritized regions recognizing their vulnerability due to natural factors such as increasing levels of atmospheric CO2, continuous warming by climatic change, and posing of pressure over the resources through anthropogenic activities. Only measurable research works are being approached limiting the restricted boundaries to focus on the soil C emission and its influencing parameters [34]. The research initiation is most urgent to understand the C cycle in the forests of temperate region that contribute the considerable C addition to the global C budget and play a significant role [35, 36]. Hence, this study thus aimed to assess the soil C efflux via SR in a natural temperate Rhododendron arboreum forest located in Annapurna Conservation Area (ACA) in Nepal. Further, the study specifically desired to evaluate the associated multiple ecological or environmental components of the forest such as climatic and biotic factors that could firmly regulate the SR variations and compute its influences and sensitivity towards SR with the measurements of closed chamber technique by using an infrared gas analyzer.

2. Materials and Methods

2.1. Study Area Description

The study was conducted in a temperate forest (N 28 23′40.7″, E 083˚46′ 07.0″, and altitude 2675 m a.s.l.) dominated by Rhododendron arboreum trees, in the Annapurna Conservation Area (ACA) of Central region in Nepal (Figure 1). This is the natural and primary forest of the ACA which is the first largest (7629 km2) protected area with the total forest (1029.76 m2) of Nepal. This is constructed with the major purpose of natural resources and biodiversity conservation in management principles under the national trust for nature conservation (NTNC) and the Annapurna Conservation Area (ACA) project [37, 38]. It is situated at the Annapurna range of Himalayas across Manang, Mustang, Kaski, Myagdi, and Lamjung districts. The study area has a warm-temperate climate, i.e., hot and wet summer season and cold and dry winter season. This receives most (about 80%) of the annual precipitation during the summer/monsoon season (June–September), and most rain events occurred in July and August. The winter (December–February) is dry and cold with very rare rain events. In October, the area receives little autumn/postmonsoon rain similar in character to the spring/premonsoon but the frequency is low and rapidly decreases and increases with the progress of the autumn and spring season. Similarly, highest temperatures were recorded during summer mostly from July to August and the lowest in winter from December to January (Figure 2).

Green vegetation of the study area consisted of dominated tall Rhododendron arboreum (Nepali name: Laligurans, family: Ericaceae) trees from small (>1 m) to large (<5 m) diameter at breast height forming a dense Rhododendron forest. The associated other plant species occupies sparsely at the plot were Acer laevigatum, A. carpinifolia, Quercus semecarpifolia, Q. lamellosa, Lyonia spp, Pieris formosa, and Daphne bholua.

2.2. Environmental Parameter Measurements

Air temperature (°C) and precipitation (mm) of the study area from 2005 to 2017 were generated with the data recorded in meteorological station (Lumle) of the Department of Hydrology and Meteorology (DHM), Government of Nepal. The soil temperature (ST, °C) at 5 cm soil depth and the air temperature (°C) at 1.5 m height from the forest floor within the study site were measured by a digital lab stem thermometer (MEXTECH Multi-Stem Thermometer, ST9283B, India). The soil water content (SWC, Vol%) at 5 cm soil depth was measured by a time-domain-reflectometry (TDR) soil moisture probe (TRIME-PICO64) and recorded with the HD2-TRIME-FM (Imko, Germany) mobile data logger. The ST, air temperature, and SWC of the forest were recorded in each measurement of soil respiration (SR) at three different points near and around the soil chamber. The PPFD (light) was measured by using a LI-190SA quantum sensor and LI-1400 data logger of LI-COR, Inc. Lincoln, Nebraska, the USA. During measurements, the light sensor was placed on top of the soil chamber at three different points (n = 3) and recorded data for each point during the SR measurements.

2.3. Soil Respiration (SR) Measurements

In October 2016, the study plot area of size 100 m × 70 m was established for measurement that was located at the northern slope (about 15°) of the forested area. A multiple of three different sets of SR measurements were conducted each day started from the morning at 7 : 00 am to the evening at 5 : 00 pm to observe the temporal variations of the SR. The measurements were carried out in October during the morning (7 : 00 am–9:00 am), afternoon (11 : 00 am–1:00 pm), and evening (3 : 00 pm–5:00 pm) with the time interval of 2 hours in each set of measurements each day between the dates 16th and 18th in 2016 and 27th and 30th in 2017 in 2 years. Chambers (n = 10) made of polyvinyl chloride with a diameter of 18 cm and height of 16 cm were arranged randomly in the experimental area. The chambers were installed below the ground surface at 2 cm into the soil taking care that they were perfectly air tight without any leakage between the soil and the edge of the chambers. This chamber is composed of two parts: a lid and a body. The lid was equipped with an IRGA for the measurement of CO2 and gas temperature of the chamber (body). Vaisala CARBOCAP CO2 probe GMP343 (Helsinki, Finland) was used for the measurement of CO2 concentration and gas temperature inside the chamber. This method involves placing a chamber over the soil surface and increase in concentration of CO2 within the chamber is measured as a function of time and data logger (VAISALA humicap hand held device, Helsinki, Finland) was used to record the measured data. The density of CO2 was calculated through recording the air temperature of the chamber. The first day of the measurements of SR was operated one day after placement of the chamber on the forest floor and continued for the rest of the days to avoid the variability of data due to the installation effects [34].

2.4. Data Analysis

To prevent any systematic errors, a multiple of three cycles of SR measurements in the forest were made on each soil chamber and the average of three measurement values was used for each chamber [34]. The mean value of SR from all chamber measurements (n = 10) recorded in each date was considered as the representative of daily SR and compared integrate the diurnal variations. Similar measurements of ST, SWC, and PPFD were operated for the analysis and comparison. Throughout field measurements, few SWC data could not be obtained due to some technical glitches caused by the instrument malfunction and unexpected rainfall. The SR rate of the forest was calculated using the following equation [39]:where F is soil respiration (SR, mg CO2 m−2·h−1), V is volume of air within the chamber (m3), A is area of the soil surface within the chamber (m2), and Δc/Δt is the time rate of change of the CO2 concentration in the air within the chamber (mg CO2 m−2·h−1).

The model to fit for the SR rate and ST was developed by an exponential regression. The dependence of the SR rate on ST was modeled using the following regression equation [21]. The fit of models were estimated by using the coefficient of determination (R2) and residual analysis.where F is the predicted SR (mg CO2 m−2·h−1), Ts is the ST (°C) at a depth of 5 cm, and α and β represent the regression coefficients.

The statistical analysis was made by using R software. GIS software of 10.4.1 versions is used for mapping of sampling site and the study area arc. The relationship between the variables were examined using the correlation and significance of correlation, i.e., statistical analysis.

2.5. Litter Biomass Measurements

The five random subplots (n = 5) within the study area in SR measurements plot were selected for litter biomass sampling and collected the sample at the ground level of the forest floor. The sampling of litter was made once in each period of SR measurements considering that the amount of litter remains constant between the days of measurements in the same season of a year. The litter samples were collected inside the SR chamber and the area was calculated for the square meter. Initially, the samples were oven dried at 70°C for 48 hours and weighed the dried samples with an electronic balance. The dry weight of litter biomass was calculated using the following formula.

3. Results

3.1. Air Temperature and Precipitation

The recorded (2005–2017) meteorological data (source: Department of Hydrology and Meteorology, DHM, Lumle) representing the air temperature and precipitation of the R. arboreum forest area (Figure 2) showed that the area mostly acquired higher air temperature in July and August and lower during winter in December and January which has reached the highest at 22.3°C in August and lowest at 7.9°C in January 2012. The forest area received maximum precipitation as rainfall during the month of summer in July and August; however, some of the years’ June and September as well received the higher amount of precipitation due to early and delay rain events. The winter season also known as dry season received the least amount of rain and sometimes no precipitation occurred during that period. The highest rainfall was recorded at 1818.1 mm in July 2010. The mean (2005–2017) annual-average air temperature of the forest area was recorded at 16.5°C with the highest in July and August at 21.0°C and the lowest was recorded in January at 9.9°C (Table 1). Similarly, the mean (2005–2017) annual-average precipitation of the forest area recorded was 434.7 mm with the maximum and minimum monthly mean precipitations were 1523.9 mm in July and 8.1 mm in December, respectively. The mean (2005–2017) annual precipitation of R. arboreum forest was recorded at 5216.1 mm y−1. The ACA is located in the central part of Nepal receives the highest amount of rainfall compared to all regions of Nepal.

3.2. Soil Respiration (SR) and Soil Temperature (ST)

The temperature dependency of SR variations of this R. arboreum forest was assessed from the ST measured at 5 cm soil depth that was plotted against the SR (equation (1) with optimal regression using the exponential form of equation (2) over the different periods and dates in October 2016 and 2017 for the entire experimental period (Figure 3). The result exhibited increasing ST boosting the SR exponentially was explained positive statistically significant () curve (Day1: D1; R2 = 0.21, y = 0.572e0.580x, Day2: D2; R2 = 0.13, y = 0.567e0.585x, and Day3: D3; R2 = 0.23, y = 1.248e0.525x). Similarly, positive exponential statistically significant (; Day1: D1; R2 = 0.20, y = 12.12e0.379x and Day2: D2; R2 = 0.24, y = 3.789e0.500x) and statistically insignificant (; Day3: D3; R2 = 0.19, y = 6.688e0.448x and Day4: D4; R2 = 0.18, y = 14.86e0.379x) curves were detected in 2017. The highly significant () exponential relation with the ST were derived when different dates in a season were integrated (R2 = 0.18, y = 1.049e0.529x) in 2016 and (R2 = 0.17, y = 26.34e0.284x) in 2017 (Figure 4 a1, b1). Within a short-range even at 2–3°C of the ST difference, i.e., from 9.2°C to 11.3°C in 2016 and 6.4°C to 9.3°C in 2017 in a season, the variations of the SR in this forest were well recognized to explain the ST effect on SR of the forest.

3.3. Soil Respiration (SR) and Soil Water Content (SWC)

The scattered relation was found for SWC dependence on SR variations of the forest while assessing them in different periods of two consecutive years (Figures 4 a2, b2). The SWC was ranged between 27.0% and 40.6% in 2016, and 22.3% and 44.8% in 2017. The peak values increase in SWC after the unexpected rain events that were very common in the study area. A comparatively much higher range of SWC was noticed in 2017 than the year 2016 as the measurement dates in 2016 were half a month earlier to the summer season, i.e., much near to the rainy season than the year 2017.

3.4. Soil Respiration (SR) and Photosynthetic Photon Flux Density (PPFD)

To elucidate the dependence of SR on the variability of light in the forest, SR was plotted against the PPFD in different periods, dates, and years of the measurements showed that the light parameter, PPFD, was not accomplished to establish the relationship with SR variations to define independently the PPFD effect on SR in the R. arboreum forest (Figures 4 a3, b3). The values obtained in measurements of PPFD were assessed much randomly in different chamber and even constitute in the same chamber at different points. The range of the PPFD observed was between 2.7 μ·mol·m−2·s−1 and 96.2 μ·mol·m−2·s−1 in 2016 and 1.3 μ·mol·m−2·s−1 and 761.0 μ·mol·m−2·s−1 in 2017. In two years, the values of the PPFD were most likely concentrated within 100 μ mol·m−2·s−1 that were correctly measured in a different time period of a day for the consecutive 3–4 days in a year. The high and low values of the PPFD were most likely recorded in a different period sunny to the cloudiest weather.

3.5. Comparisons of SR and ST, SWC, PPFD, and Litter Biomass Variations

The temporal, diurnal, and interannual variations of the SR, ST, SWC, and PPFD of the forest represented the degree of correlations between the SR and the climatic parameters. Comparatively higher SR (n = 10) was observed in the afternoon (11 : 00 am–1 : 00 pm) and the maximum value was recorded at 306.3 mg CO2 m−2·h−1 on 18th October, 2016, but the values became random in 2017 and the maximum value was observed during the evening at 348.9 mg CO2 m−2·h−1 on 27th October (Figure 5). The minimum SR (n = 10) was observed in the morning on 17th October at 201.6 mg CO2 m−2·h−1 in 2016 and on 29th October at 232.5 mg CO2 m−2·h−1 in 2017. With acquiring the temporal variations of the SR in each day of measurements, the values of SR were not compatible with each other. However, the difference in values between the maximum and minimum SR was most likely compatible in both years. The ST (n = 10) was recorded maximum at 10.7°C on 16th October, 2016 and 8.9°C on 27th October, 2017 during the evening between 3 : 00 pm–5 : 00 pm among different days of the measurements. The minimum ST was observed during the morning at 9.6°C on 17th October, 2016 and 7.1°C on 27th October, 2017. The ST followed the similar trend of the SR, and it was always lower during the morning than that of the afternoon and evening.

The maximum SWC (n = 10) was recorded during the evening at 36.7% on 16th October, 2016 and 36.9% on 27th October, 2017. The minimum SWC was observed during the afternoon at 31.8% and 31.4% on 16th October, 2016 and 30th October, 2017, respectively. The maximum and minimum SWC of the forest was very close and compatible for both years but it varied among the dates and time of measurements each year. The temporal and diurnal variations of the SWC were not compatible to the SR of the forest. The SR was increased during the afternoon when the SWC decreased and higher SR was observed while the SWC was decreased in the evening (Figure 6).

Correlating to other climatic parameters, the PPFD of the forest was noticed much fluctuating in different dates and years in measurements. The maximum PPFD (n = 10) was recorded at 41.0 μ·mol·m−2·s−1 during the morning on 18th October, 2016 and 253.2 μ·mol·m−2·s−1 during the afternoon on 30th October, 2017, where the minimum values of PPFD were recorded at 9.0 μ·mol·m−2·s−1 on 18th October, 2016 and 7.3 μ·mol·m−2·s−1 on 27th October, 2017 during the evening (Figure 7). The range of maximum and minimum values of the PPFD of this forest was found much higher than that of the other related parameters. Comparing the temporal variations of PPFD to the SR, it was much similar to each other in the days of sunny weather; however, the trend of diurnal changes in the PPFD was not visible in a cloudy day and they did not had direct effect on the temporal changes in SR of the forest.

The diurnal variations of the SR, ST, SWC, and PPFD were detected in both years (Table 2). The daily average SR was 278.2, 242.0, and 289.5 mg CO2 m−2·h−1 on 16th, 17th, and 18th October in 2016 and it was 332.1, 301.6, 273.3, and 273.5 mg CO2·m−2·h−1 on 27th, 28th, 29th, and 30th October in 2017. The rates of SR varied among the days of measurements; however, the values were compatible with each other and not unexpectedly different. Similarly, the ST was 10.4, 10.0, and 10.1°C in 2016 and 8.5, 8.6, 8.0, and 7.4°C in 2017. The minimal diurnal variations of the ST were detected in both years (upto 0.4°C in 2016 and 1.2°C in 2017), to detect its effect. The SWC was 34.2, 34.2, and 34.5% in 2016 and 35.7, 32.8, 33.6, and 33.4% in 2017. The diurnal variations of SWC were minor (0.3%) in 2016; however, it was higher (2.9%) in 2017. Much more variations between the days were observed for the PPFD were 17.1, 14.1, and 23.7 μ·molm−2·s−1 in 2016 and 12.3, 39.0, 20.0, and 90.4 μ mol·m−2·s−1 in 2017. Compared to the rest of the climatic parameters, the PPFD varied more to the diurnal changes of the season.

The measurements of SR, ST, SWC, PPFD, litter biomass, and air temperature in 2 years determined that the interannual variations of the soil C emission via. SR and the ecological parameters were much common in the R. arboreum forest; however, the intensity of variance were incompatible for each component (Figure 8). The SR over the measurement period of 2 years averaged at 269.9 mg CO2 m−2·h−1 in 2016 and 295.1 mg CO2 m−2·h−1 in 2017 and the SR rate in 2016 was lower than the year 2017 but were comperative. The 2-years averaged SR represented the seasonal SR of the R. arboreum forest was 282.5 mg CO2 m−2·h−1. The variations from the seasons to the years in carbon sequestration and emission were adopted in multiple years of estimation and were observed in a temperate ecosystem [19]. Similar variations of ST were observed in 2-years averaged throughout the measurement period was higher at 10.2°C in 2016 than at 8.1°C in 2017. The seasonal ST of the forest was calculated from two years’ records averaged were at 9.1°C. Compared to the other climatic parameter of the forest, less variation was detected for SWC records between the years at 34.3% in 2016 and 33.9% in 2017, and slightly higher SWC was observed in 2016 than the year 2017, but the values were compatible to each other. The seasonal SWC value estimated from two years’ data of SWC averaged was 34.1%. Similarly, the PPFD of the forest in 2016 was 18.3 μ·mol·m−2·s−1 and in 2017 was 41.4 μ mol·m−2·s−1 and the variance was noticeable between the years. The represented seasonal PPFD averaged both years’ measurement was 29.9 μ·mol·m−2·s−1. The dry weight of litter biomass collected from the forest floor was observed at 8.9 g·d. w. m−2 in 2016 and 15.3 g·d. w. m−2 in 2017 (Equation (3)). The higher litter was recorded in 2017 than in the year 2016 as the collected litter in 2017 was from late October and it was the season to leaf abscission of the seasonal trend of the forest. The amount of the seasonal litter of R. arboreum forest averaged 2-years was calculated at 12.1 g·d. w. m−2. The air temperature of the forest was noticeably varied between the years and it was higher at 7.1°C in 2016 than in the year 2017 at 5.6°C, and the average of 2 years was at 6.4°C. Overall this study showed that, comparatively higher SR, PPFD, and litter was detected in 2017 than in the year 2016 that was across the ST, SWC, and the air temperature of the forest owed to the measurements in 2016 were carried half a month earlier in the season than the year 2017 that was not much far from the wet summer season with high temperature and precipitation.

4. Discussion

Soil respiration (SR) is one of the major pathways of the carbon dioxide efflux from the soil, thus determining SR is the most effective measure to elucidate the carbon emission that could be made accurate in estimation of the global carbon budget [40]. Given the critical role of carbon dioxide efflux from the soil in the global carbon budget, SR measurement and carbon modeling have been kept in priority and well-studied [6]. Thus, different studies have focused on investigating the effects of environmental variables driving the SR rate in the temperate region forests of different climatic conditions [14, 20, 41]. In this study, the major environmental factor, the ST showed a statistically significant exponential effect on SR variations in 2 years (Figures 2; and 3 a1, b1) of observations illustrating that the temperate forest soil could be more proactive in C emission while temperature rises and warming could be alarming more in emission from the forest floor in future. These results comply with the previous studies focused on temperate regions [34, 41, 42]. It was more appreciable that the ST effect on SR was even detected within the narrow range (2-3°C) of the ST difference. However, the effect of ST on SR in two days of 2017 was found statistically insignificant (, Figure 2(b); Figure D3, D4) because the difference between the maximum and minimum ST was very narrow range normally within 2°C as the measurements were carried in a single season. Similar to this study, Cui et al. [43] has also detected no significant variations of SR due to narrow variation range of the ST in the tropical lowland rainforest of China during the spring season [43] when the variations of ST were very less as we observed in the autumn season. The compatible trend of SR and ST with significant exponential relation of this study was most likely noticed previously while observing the cutting effect on SR in Rhododendron forest in the National Park area in the UK [44]. The result hence revealed that the factor for determining SR might be primarily the ST in the temperate forest, however, the angle of possession could be altered in different ecosystems [34, 45] and the exponential effect of ST on SR was the most quoted one [4, 14, 46].

The SWC dependency of SR of this R. arboretum forest was not satisfied to define the effect and variations of SR under different soil moisture conditions (Figures 3 a2, b2). Higher range of the SWC (27%–40.6% in 2016 and 22.3%–44.8% in 2017) was noticed in both years during the entire measurement period which was not harmonized to the SR variations that might be the cause of undetectable relationship between the SR and SWC. The higher range of SWC measured in a single month of a season was owed to the effect of monsoon (June–September) and postmonsoon (October–November) rain which was the common phenomenon of this region which was also observed in the nearby grassland (23.3% to 59.6% and 15.6% to 42.2%) of the same region [34]. The nature of unpredictable and fluctuated climate of this region with having instant and short period of seasonal rain is the reason for lower and higher SWC before and after rain events. The instant events of rainfall in the forest might cause increasing the surface soil water level and the effect could not reach up to the deep soil tree rooting system due to the thick layer of the litter fall to vary the SR and the ST which indicated that increased precipitation may not always respond immediately to the increase in SR. Similar to our results, no SWC effects on variations of SR due to fluctuated climate were observed in the temperate forests in China [42]. The SWC effect on SR and its variations are not simply understood due to the intricacy of soil on climate and its limitations in the temperate regions under the influence of monsoon climate with exceptionally getting short period of saturation and drought of these regions [14, 47]. However, the SWC effect of SR is the effective measure to evaluate the SR in the semiarid ecosystems [8, 48] and tropical region [22, 43] where precipitation enhance the most accepted driver of the ecosystem function. The knowledge of soil-water effect on SR and the breach between them could be filled with the extended research for years with continuous measurements.

The effect of solar radiation/sunlight is the effective measure to increasing respiration rate of the plants importantly aboveground vegetation by enhancing the temperature and photosynthesis equally in high altitude regions [45]. In this study, the variations of the PPFD of the forest floor were different among time and date and even between the chambers, and that was caused due to the close canopy cover of the trees in forest. The canopy allows partial penetration of the sunlight that depends on the position and angle of the sun, and the forest floor receives sunlight in patches. The undefined scattered effect of PPFD on SR of this study established (Figures 3 a3, b3) in two years also owed the unpredictable seasonal and diurnal variations from clear and sunny to the cloudy and rainy weather within a single day faced during measurements, and particularly that phenomena was very common in the area. However, the open grassland of the same region has observed the visible linear relation with the defined PPFD effect on SR variation at 37.6% in October. The minimum (2.7 μ mol·m−2·s−1 in 2016 and 1.3 μ mol·m−2·s−1 in 2017) and maximum (96.2 μ·mol·m−2·s−1 in 2016 and 761.0 μ·mol·m−2·s−1 in 2017) PPFD values of the PPFD of this study were comparably lower than the PPFD (51.68 and 1526.0 μ·mol·m−2·s−1) obtained in the nearby grassland [34]. However, not only the ecologically adopted environmental parameters, but the land use is also considered as one of the major factors affecting soil carbon emissions via SR [49]. As compared to the grassland, the forest floor received only about 50% of the total PPFD to utilize. In two years of this study observed that most of the PPFD values were within 100 μ·mol·m−2·s−1 in the autumn season, and this revealed that only that much of the photosynthetic rays could be able to utilize the autotrophic and heterotrophic activities within the forest.

The seasonal variations of SR and its corresponding climatic factors are very common in the temperate forests of the Asian monsoon climate [14]. The temporal and diurnal variations in SR of this natural R. arboreum forest were very noticeable (Figure 5). The maximum values of SR (306.3 mg CO2 m−2·h−1 in 2016 and 348.9 mg CO2 m−2·h−1 in 2017) of this forest were recorded during the afternoon from 11 : 00 am to 1 : 00 pm and the evening from 3 : 00 pm to 5 : 00 pm that was much consistent to the nearby temperate grassland of the region [34]. However, the values of maximum SR of this forest were much less than (963.42 mg CO2 m−2·h−1 in October 2015 and 1132.6 mg CO2 m−2·h−1 in April 2016) that of the grassland site. This explained that the temperate R. arboreum forest could store the maximum amount of C in the soil by the storage of soil C with limiting C emission through SR and become a significant sink of atmospheric CO2. As compared to the values of the maximum SR, the minimum SR values between the years (201.6 mg CO2 m−2·h−1 in 2016 and 232.5 mg CO2 m−2·h−1 in 2017) of this forest were not much lowered than (45.3 mg CO2 m−2·h−1 in October and 172.5 mg CO2 m−2·h−1 in April) the minimum SR of the grassland. The minimum SR of this R. arboreum forest was higher than the nearby grassland which showed that the temperate forest possesses less difference between the maximum and minimum SR than the grassland. This proved that the forest of the temperate region is less sensitive to the increasing temperature than the grassland; hence, forest deserves more sink capacity of the atmospheric carbon than the open grassland which is considered as the major part for the land-use change conversion [14, 23]. Compared to our observations, the daily mean SR of this forest varied among the days of measurements in 2016 (278.2, 242.0, and 289.4 mg CO2 m−2·h−1) and 2017 (332.1, 301.6, 273.3, and 273.5 mg CO2 m−2·h−1) (Table 2) that were consistent to the diurnal variations of SR observed previously (517.0, 430.5, 123.4, and 357.0 mg CO2 m−2·h−1) in the nearby temperate open grassland of ACA during October [34]. The diurnal trend of the ST concerning the SR of the forest (Figure 5, Table 2) was also comparable to the nearby temperate grassland where the lowest SR and ST were detected during the morning, whereas they were very near in the afternoon and evening Similarly, the difference (1.1–1.7°C) between the maximum and minimum ST of the forest was very less as compared (9.2–20.3°C) to the grassland. This showed higher temperature sensitivity of the SR in this R. arboreum forest which might be vulnerable to the warming climate.

The diurnal fluctuations of the SWC of the forest were very low (Table 2) and there were not much difference of the maximum (36.7% on 16th October, 2016 and 36.9% on 27th October, 2017) and minimum (31.8% on 16th October, 2016 and 31.4% on 30th October, 2017) soil water values (Figure 6) while averaging all the point; however, the rain was frequent and the weather was unpredictable. The coverage of litter on forest floor in October is high due to the seasonal litter fall preventing the soil surface from getting wet with a lowest SWC value and the SWC value was higher where the litter coverage was less. The maximum SWC values were measured during the evening and minimum were during the afternoon because, in most of the days in our measurements, the short and frequent rain events were detected between the late afternoon and early evening measurements. The occurrence of both varying precipitation with the lower and higher rate than the usual has significant impacts on SR with the association of ecosystem components in the seasonally dry tropical forest [22]. However, the tropical forests are more sensitive to the changing rainfall than that of the forests in temperate region where the temperature becomes the most prominent in controlling the SR [20, 49].

Canopy coverage plays a major role to prevent the light from the sun to the undervegetation and suffers from inadequate light intensity influencing their adaptability and optimum growth by the limitations of nutrient-use efficiency. The research revealed that the light intensity (PPFD) enhanced the soil-water use efficiency by increasing nutrient uptake significantly [50] and hence the net ecosystem production [51]. In this study, the PPFD values randomly fluctuated throughout the day and only partial light could reach the forest floor due to the close canopy cover of the R. arboreum forest structure. The diurnal variations of the PPFD within our entire study period were not comparable to the days of the same season (Table 2) and even between the years (Figure 7). The higher PPFD recorded during the afternoon in the sunny weather of this forest was better compared to the grassland of the region [34], however the forest is different from the grassland under the coverage of the leafy canopy with differences in exposure to the sunlight. The higher PPFD in the afternoon and evening than in the morning during clear sunny weather of this forest much followed the trend of higher SR during the afternoon and evening than in the morning showed that the PPFD could be the major variable of SR when the forest is not dense and has open canopy. This could be better understood while the measurements of SR and PPFD were carried out in selected sunny weather days in the different seasons and days of a year.

` The interannual variations in SR are much common in the forests owing to the common soil factors altered each year known by temperature and precipitation [52]. In this study, interannual variation of SR was observed; however, the variation between the years (269.9 mg CO2 m−2·h−1 in 2016 and 295.1 mg CO2 m−2·h−1 in 2017) was most likely comparable (Figure 8.) The SR rate in 2016 was comparatively lower than the year 2017; however, they were nearly corresponding to each other. The two years average seasonal mean SR (282.5 mg CO2 m−2·h−1) of this temperate R. arboreum forest was within the range (90–551 mg CO2 m−2·h−1 and 98–305 mg CO2 m−2·h−1) of daily mean SR during the autumn season in cool-temperate Mongolian Quercus forests in Korea [14]. Due to the complexity of the components and spatially heterogeneous nature of the forest’s soil, it is much difficult to understand the environmental variability [49], and uniquely coupled with other biotic processes, the most common might be the soil properties in generating a broad spectrum of CO2 emission rates in the temperate forests [20]. The prior observations by Cui et al. [43] and Kochiieru et al. [53] suggested that it is not surprising that SR rates vary considerably in different seasons, years, and in the different forests, eco-zones, and even in the different soils and land use types.

The interannual variations of ST observed (10.2°C in 2016 and 8.1°C in 2017) in this study was similar to those seen for SR and it was surprised to record a low SR rate in 2016 at the time of high ST and the higher SR rate in 2017 at low ST. The cause might be the early measurements in 2016, i.e., mid of October and in the year 2017, i.e., late October. The temperature decreases day-by-day with an increase in the autumn season. Slightly higher SWC was detected in 2016 (34.3%) than in 2017 (33.9%) owed to the nearby summer and postmonsoon rain. Thus, higher SWC might play the role to suppress the SR in 2016 even though the ST was high. The decrease in increasing SR rate was detected with the increase in SWC beyond its limit in temperate forests in different soils and land use types [53]. The accountability of the soil-water effect on SR in the temperate region was less [34] to overcome the temperature effect; however, it was detected while comparing in two years separately. As the variety of biotic and abiotic factors regulates the rate of SR [54], the optimum soil moisture (30–40%, Figures 4 a2, b2) with maximal water-holding capacity allows in achieving the maximal respiration rate [20], and the moisture content in soil may modify the temperature sensitivity of soil microbial respiration, and thus activate the temperature sensitivity of the SR [55]. The previous study also revealed that when both ST and SWC are not at their extremes, these two factors interactively influence the SR and integrated can account for most of its variability and more during active growing season than the dormant seasons [56].

The PPFD between the years were varied and it was lower (18.3 μ·mol·m−2·s−1) in 2016 than (41.4 μ·mol·m−2·s−1) in 2017 (Figure 8). The light penetration within the forest floor is determined by the canopy structure. The closed canopy cover of the forest structure was the reason for the low value of the PPFD and the cause for sparsely distributed understory vegetation of the forest floor. The seasonal two years average PPFD (29.9 μ·mol·m−2·s−1) of this R. arboreum forest was visibly much lower than (51.68 μ·mol·m−2·s−1) the minimum PPFD recorded in the nearby grassland [34]. However, the study in a similar temperate deciduous forest revealed that the temperature and light both led to the inhibition of leaf-level respiration and photosynthesis; as result carbon-use efficiency declined with increasing the leaf-level temperature and hence the basal respiration rate [57]. The overall net ecosystem production is controlled by the sunlight (PPFD), and the increasing rate of the PPFD accelerates the respiration of the vegetation only up to its maximum limit and beyond the limit, the respiration started to decline [51]. The PPFD of this forest could not reach to its maximum to the peak and decline in SR were not visible. Therefore, in this study, a higher seasonal SR was recorded in the year when the PPFD was recorded higher.

The interannual variations (8.9 g·d. w. m−2 in 2016 and 15.3 g·d. w. m−2 in 2017) of the litter biomass was well observed in this study and it was higher in the year 2017 when the SR, SWC, and PPFD was higher than in the year 2016 (Figure 8) except the ST which corresponds to the air temperature (7.1°C in 2016 and 5.6°C in 2017) of the forest. The SR visibly reduces (35%) with the reduction of litter input and induces (77%) with the addition of the litter in the forests [58]. The SR, like many other physiological processes of plants and microbes, are very sensitive to environmental changes; often interactively affected by multiple factors, although it is often difficult to separate their interactions. Thus, our initiative investigation of SR is the vital and adequate measure in the Rhododendron arboreum forest which further study needs to seek how carbon emission of this temperate forest responds to regional climatic warming and land use and integrates these feedbacks into global C budget and climate models.

5. Conclusions

This study elucidates SR in the natural R. arboreum forest located in the temperate region of Nepal and highlights the sensitivity of SR correlated with multiple ecological parameters known as climatic and biotic factors. In the study, ST was found the most visible factor to the variations of SR and within a short range (2–3°C) of the ST difference, the variations of SR showed significant exponential relation with the ST. This showed that the rising temperature is much vulnerable to the climatic change through the emission of the forest. The SWC effect on SR variation was found scattered and no clear relation between SR and SWC was detected which explained the effect of soil temperature on SR was more visible than the SWC and the combined effect of the soil temperature and soil moisture are effective to determine the SR variations. The effect of PPFD on SR variations was weak to represent but the seasonal SR of the forest was much consistent with the seasonal PPFD and litter input. Accountable temporal, diurnal, and interannual variations of SR with the modification of major climatic and biological factors of the forest were determined. The study revealed that the temperate forest could store the maximum amount of C in soil with limiting C emission through SR and could be the more sink of atmospheric CO2. Thus, the SR is very sensitive to environmental changes and interactively affected by multiple factors, although it is often difficult to separate their interactions. Our study concluded that temperate forests could store maximum soil carbon with limited emission through SR and become a larger sink of atmospheric carbon dioxide even though SR is very sensitive to environmental changes and interactively affected by multiple ecological factors. Thus, this result is an appreciable measure for the temperate forest to understand the regional carbon balance and suggested temperate forests are valued to incorporate them in evaluating global climate models and carbon budget.

Data Availability

The data used to support the findings of this study are available from the corresponding author upon request.

Additional Points

This paper has been published in a preprint by Research Square [59].

Conflicts of Interest

The authors declare that they have no conflicts of interest regarding the publication of this study.


This research was funded by the Nepal Academy of Science and Technology (NAST) for equipment, field visits, and facilities support. The authors would like to thank the Department of National Parks and Wildlife Conservation (DNPWC), Kathmandu and Annapurna Conservation Area Project (ACAP), Pokhara for granting research permission for this study. The authors are thankful to the people of Tadapani, ACA for their cooperation during the study.