The study was carried out in Taihuyuan Town, Hangzhou City, Zhejiang Province, China (30°19′ N, 119°35′ E). It has a subtropical climate. The mean annual temperature is 15.8°C. The rainfall is 1600 mm. The monthly mean temperature and precipitation during the experiment are shown in Fig. 1. The attitude of the experiment site is about 100–150 m.

The Chinese hickory stands were established about 35 years before the experiment. At the time of the study, the stocking density was 300 trees ha⁻¹ with 70% coverage. The understorey vegetation was mainly Poa annua, Ceratium glomeratum, Aster ageratoides, Ehmannia chingii, Semiaguilegia adoxoides, and Rumex japonicas, with a mean height of 0.1 m and 30% coverage. The soil type was characterized as Ferralsols [24]. The important properties of the soil (0–0.2 m) are as follows: bulk density, 1.20 g cm⁻³; pH (H₂O), 4.69; organic C, 17.19 g kg⁻¹; and total N, 1.06 g kg⁻¹.

The study was established in a Chinese hickory stand, which had not received any specific previous management except for routine harvest. Sixteen 20 m × 20 m plots were randomly selected and subjected to four treatments with four replications: control (CK, no fertilizer), inorganic fertilizer (IF), organic fertilizer (OF), and equal parts organic and inorganic N fertilizers (OIF). Except for the control plots, all the other treatments received the same N rates (Table 1). The rate of N, P, and K application was 33.8, 10.8, and 22.4 kg ha⁻¹, respectively, according to the locally recommended doses for Chinese hickory (Table 1) [25]. The organic fertilizer was commercially produced in Deqing City, Zhejiang Province, China. The contents of C, N, P, and K in organic fertilizer were 351, 30, 7.9, and 22.0 g kg⁻¹, respectively. All urea, superphosphate, potassium chloride, and organic fertilizers were sprayed onto the soil surface and manually cultivated into 0–0.2 m soil in late May.

Soil CO₂ efflux was measured using the closed chamber method and gas chromatography techniques between July 2011 to June 2012. The chamber base (0.3 m × 0.3 m × 0.1 m) was installed and the tank was full of distilled water before gas sampling. Four samples were collected with 60 mL plastic syringes first, then stored in the airbags and determined by a chromatograph (Shimadzu, GC-2014, Japan) [26]. The sampling interval is 10 min. The gases were homogenized by pumping the gases inside three times with the sampling syringe before sampling [27]. All gas sampling was conducted at a frequency of approximately once a month, between 9:00 and 11:00 in the morning, as the CO₂ efflux rate in this period was approximate to the average value [26].

At the same time of gas sampling, four topsoil (0–0.2 m) was sampled from each plot for the determination of soil moisture content and DOC. At the last collection of gas samples, the composite soil samples were gathered to analyse MBC and MBN.

Soil temperature at a depth of 0.05 m was detected with a thermometer. Soil bulk density was determined by collecting soil samples with a bulk density corer of 200 cm³. Soil moisture content was determined using the drying method in the oven at 105°C for 24 h. Soil organic C was determined with the wet combustion method. Total N was determined by the Kjeldahl method and pH value was analysed with a pH meter at the soil to extractant ratio of 1:5 (w:v).

Soil DOC was extracted from 10 g moist soil by adding 20 mL of distilled water. After shaking at 200 rpm for 30 min and centrifuging for 10 min at 20,000 rpm, then filtering through 0.45 μm filterable membrane, the organic carbon in the filtrate was determined by an automated TOC-TN analyser [28].

Soil MBC and MBN were analyzed using the fumigation–extraction method [29,30]. Both the fumigated and non-fumigated soil samples were extracted with K₂SO₄ solutions of 0.5 mol L⁻¹ with a ratio of extractant to the soil of 5:1 (v:w). The resulting extracts were analyzed for organic carbon and nitrogen by an automated TOC-TN analyzer. The concentrations of C and N were calculated using the difference of the extracts of non-fumigated and those of the chloroform-fumigated, as described by Zhang et al. [27]. The extraction efficiency factor applied was 0.45 for C and 0.54 for N, respectively [29,30].

The effects of fertilization on CO₂ and environmental factors were examined by using the one-way ANOVA and LSD test. The data were log-transformed in case of heterogeneity of variance before the ANOVA analysis. A linear regression model was adopted to analyze the relationship between soil CO₂ and soil DOC, or moisture content. An exponential equation was adopted to build the relationship between soil CO₂ and soil temperature. A logarithmically transformed multiple regression was conducted to analyse the relationship between soil CO₂ efflux and environmental factors. The soil CO₂ efflux and the cumulative CO₂ emissions were calculated by the equation stated by Liu et al. [26].

Soil CO₂ effluxes showed distinct dynamic changes with seasons (Fig. 2). Regardless of fertilization treatment, the effluxes were the highest in summer and the lowest in winter. Soil CO₂ effluxes ranged from 51.3 to 306.2, 96.5 to 576.0, 74.3 to 284.6, and 39.1 to 360.4 mg m⁻² h⁻¹ for CK, IF, OIF, and OF treatments, respectively. The mean annual soil CO₂ efflux in the IF treatment was 266.9 ± 8.4 mg m⁻² h⁻¹ (mean ± s.e.), significantly higher than that in the CK, OIF, and OF treatments (P < 0.05). The cumulative emission in the IF treatment was 18.6 ± 1.4 Mg ha⁻¹ year⁻¹, which was significantly higher than that in the CK, OIF, and OF treatments (P < 0.05). No significant difference was detected between CK, OIF, and OF treatments.

Soil temperature at 0.05 m depth followed a seasonal pattern similar to soil CO₂ efflux within the range of 0.30°C to 29.31°C (Fig. 2). The average values under CK, IF, OF, and OIF treatments were 16.17°C, 16.15°C, 16.16°C, and 16.19°C, respectively. A significant exponential correlation between soil CO₂ effluxes and soil temperature was found under all treatments (Table 2).

Soil moisture content ranged from 218 to 413 g kg⁻¹, with an average of 306 g kg⁻¹ during the experimental period (Fig. 2). No significant difference was detected between treatments. Correlation analysis showed that there was no significant relationship between soil CO₂ effluxes and soil moisture in all treatments (Table 2).

Concentrations of DOC under all treatments peaked in summer and were lowest between December to February, increasing after March (Fig. 2). Soil DOC concentrations ranged from 30.4 to 138.5, 24.2 to136.2, 39.0 to 145.4, and 40.6 to 142.0 mg kg⁻¹ in the CK, IF, OIF, and OF treatments, respectively. The DOC concentration was significantly lower under the IF treatment than that under the CK, OF, and OIF treatments (P < 0.05). A significant correlation was determined between DOC and soil CO₂ efflux in the CK treatment (P < 0.05), however, no significant correlation was observed in the three fertilization treatments (Table 2). A multiple regression analysis including soil temperature, soil moisture, and soil DOC accounted for 52%–89% of temporal variation in soil CO₂ effluxes, which was very close to the results from a regression equation including soil temperature alone (Tables 2 and 3).

MBC and MBN ranged from 296.4 to 319.6 mg kg⁻¹ and 34.6 to 74.2 mg kg⁻¹, respectively (Fig. 3). The OF and OIF treatments showed significant increases in MBC and MBN, compared to the control (Fig. 3). The ratio of MBC:MBN was 8.3 in the control, and decreased to 4.3–6.4 when organic fertilizer was applied. The application of inorganic fertilizer did not affect soil MBC, MBN, or the ratio of MBC:MBN (Fig. 3).

Information on soil respiration from Chinese hickory stands is scarce. Our results revealed that the annual soil CO₂ emissions ranged from 11.8 to 18.6 Mg ha⁻¹ year⁻¹ in different fertilizer treatments from Chinese hickory stands in subtropical China. On average, soils in the experimental field released CO₂ about 14.7 Mg ha⁻¹ year⁻¹ into the atmosphere, which was similar to that in Chinese fir forest (15.9 Mg ha⁻¹ year⁻¹) in subtropical China [31]. However, this value was much lower than that in other subtropical orchards, such as longan orchards (52.3 Mg ha⁻¹ year⁻¹) [32], citrus orchards (25.6 Mg ha⁻¹ year⁻¹) [33], and peach orchards (26.7 Mg ha⁻¹ year⁻¹) [34]. According to a review conducted by Song et al. [35], soil CO₂ emissions range from 15.9 to 68.2 Mg ha⁻¹ year⁻¹ in major forest ecosystems in subtropical China. These results demonstrate that the types of land use and tree species substantially influence soil CO₂ emissions in subtropical regions.

Soil temperature has widely been identified as the dominant factor leading to the temporal variation of soil CO₂ effluxes [36,37]. The increase of soil temperature will promote microbial growth and enzyme activity and thus soil respiration. There was a significant exponential correlation between soil CO₂ effluxes and soil temperature in this study, which was consistent with the results shown from other subtropical forest soils [7,33].

No significant correlation was found between soil CO₂ effluxes and soil moisture in all treatments. This was in accordance with the results reported for Chinese chestnut plantations [27] and Chinese fir and Moso bamboo forests in subtropical condition [35]. This is probably because the soils are seldom affected by frequent prolonged drought in the subtropical areas. Thus, soil moisture might not limit the activities of microbes [33].

In this study, DOC showed a similar seasonal pattern to soil CO₂ effluxes; however, soil DOC and CO₂ efflux was positively correlated only in the CK treatment. There was no significant relationship between DOC and soil CO₂ effluxes following fertilizer application. Soil disturbance caused by fertilizer application may weaken the relationship between DOC and soil CO₂ efflux, as reported in other studies [26,38].

Due to the simultaneous changes of soil temperature and soil moisture, the impact of soil moisture on soil CO₂ effluxes may be masked by soil temperature [39,40]. A multiple regression equation including several environmental factors showed different results when compared with an equation with only one factor included [41,42]. The results in the current study indicated that the multiple regression equation accounted for 52%–89% of the soil CO₂ effluxes, which was almost the same as the conclusion of the regression model using only soil temperature (Tables 2 and 3). This suggests that soil temperature was the main factor affecting soil respiration in Chinese hickory stands.

Organic amendments can promote soil microbial population increase and thereby enhance soil MBC and MBN contents due to the increased organic matter substrates and the increased root biomass [43–46]. Soil MBC and MBN were increased by 3.9%–4.2%, and 37%–112%, respectively, in the OF and OIF treatments compared with the CK treatment. In contrast, inorganic fertilizer tended to decrease soil MBC and MBN, although this was not statistically significant [47]. The ratio of MBC:MBN was between 4.3 and 8.6 in the present study, which is comparable to the result (4.7 to 8.9) of eroded red soil by artificial revegetation in subtropical China reported by Xu et al. [28]. The application of organic fertilizer alone or combined with inorganic fertilizer significantly decreased the ratio of MBC:MBN, compared with the CK treatment [47,48]. In general, fungi have a higher MBC:MBN ratio than bacteria [49,50]. Therefore, the decrease in the MBC:MBN ratio suggests that the dominant soil microbes change from fungi to bacteria groups [51]. This is supported by Joergensen et al. who reported that farmyard fertilizer application decreased the ratio of fungal groups to bacterial groups [52]. The increases in soil FB ratio might promote the conversion of more refractory C to CO₂ [53–55]. This may explain the decrease of soil CO₂ emissions in OF and OIF treatments of the current study partly.

The nutrients from organic fertilizer are expected to affect soil CO₂ efflux in the soil by increasing the availability of C substrates and enhancing microbial activity. At present, few datasets can be used to describe the impact of organic fertilizer application (either alone or in combination with inorganic fertilizer) on soil CO₂ efflux from forest soil [14]. In this study, no significant difference in soil CO₂ emissions was detected between the OF, OIF, and CK treatments. Microbial activity and the corresponding decomposition rate of organic amendments are usually related to the ratio of C/N [56]. Microbes need to immobilize more inorganic N when they decompose organic matter with a C/N ratio exceeding 25, the threshold for mineralization of organic N [57–60]. Although the C/N ratio of organic fertilizer applied in this experiment was only 12, there was a significant increase of MBN, and reduction of MBC:MBN under OF and OIF treatments compared with those under the CK treatment (Fig. 3). This suggests that N immobilization occurs during the decomposition of organic fertilizer. Thus, the stimulating effects of organic fertilizer application on soil CO₂ efflux might be partially masked by microbial N immobilization, but further long-term studies are required to verify this potential suggestion. In addition, the relatively low application rate of organic fertilizer (34 kg ha⁻¹ of N) may contribute to the response of soil CO₂ efflux in the present study, because higher rates of organic manure (60 kg ha⁻¹ of N) have been shown to increase soil CO₂ emission in tropical forest [61].

The application of inorganic fertilizer increased soil CO₂ efflux compared with the CK treatment, which is identical with the study of van Miegroet et al. [62]. With the application of inorganic fertilizer, soil microbial activity and demand for substrate was increased, which could lead to the enhancement of soil CO₂ efflux and reduction of soil C content [63]. Dissolved organic C is considered to be easily utilized by microorganisms to promote CO₂ emissions [27,34]. The highest cumulative soil CO₂ emission and the lowest DOC concentration of IF among all the treatments in this study further confirm that the application of inorganic fertilizer increases the utilization of easily degradable organic carbon for microorganisms. Additionally, the N-limited experimental stand was partially accounting for the increase of soil CO₂ efflux [64,65]. Soil CO₂ efflux decreased with N additions in N-rich soils and increased in N-limited forests [66,67], which are similar to our experimental stands.