**1. Introduction**

Subsurface drip irrigation (SDI) has been largely applied in arid and semi-arid regions to supply water due to greater yield production and water-saving characteristics [1,2]. Nevertheless, a large number of wetting fronts are generated near emitters, producing ethylene and CO2, which are harmful for crop growth [3]. Aerated irrigation (AI), a modified irrigation technique that involves injecting air into soils based on SDI, has been extensively proven to improve soil aeration, thus increasing crop yields and fruit quality [4–6]. Even so, the effect of AI on soil environmental pollution is relatively sparse.

Soil respiration, originating primarily from heterotrophic respiration and autotrophic respiration [7,8], is a principal component in the global carbon cycle. A few studies have reported an increase of soil respiration under AI [6,9,10], while the cause of CO2 release needs to be further explored. Previously, studies on drivers of soil respiration have been largely conducted on soil water content, temperature, and the interaction of these two parameters [9,11–14]. For AI treatment, a close correlation between soil CO2 fluxes with soil water content and temperature has been confirmed [9,10]. Soil microbes and enzymes as biocatalysts for all biochemical reactions in the soil would decompose and

oxidize soil organic matter [15] and intrinsically a ffect heterotrophic respiration, while the e ffect of AI on soil microbes and enzymes has been less tested [16]. Additionally, the properties of root morphology (total length, surface area, and volume) not only determine the ability of water and nutrient absorption but also determine the intensity of autotrophic respiration. Studies of root morphology under AI have been conducted in multiple crops [1,17–19], but the e ffect of AI on roots of greenhouse vegetables is still scarce. In recent years, researchers began to focus on the e ffects of soil microorganisms and plant growth on soil respiration [14,20,21]. However, the relationship between soil respiration and biotic components (microbes and plants) under AI remains unknown. Hence, studies of soil physical and biotic properties under AI are of critical significance to improve our mechanistic understanding of processes that release CO2 to the atmosphere.

To better understand the mechanism of soil respiration change under di fferent irrigation levels with and without aeration, soil respiration from greenhouse tomato fields, as well as soil physical and biotic components (soil water-filled pore space, temperature, abundance of soil bacteria, fungi, and actinomycetes, soil microbial biomass carbon, soil cellulase and dehydrogenase activity, tomato root morphology, and plant dry biomass) were investigated in the present study. We hypothesized that irrigation in combination with AI would increase soil respiration, soil microbes, soil enzyme activity, and plant growth. We also hypothesized that soil respiration would be closely related to soil physical and biotic components. Our results were used to manage irrigation measures under AI for CO2 mitigation and to reveal the mechanism of soil respiration.

#### **2. Results and Discussion**

## *2.1. Environmental Variables*

#### 2.1.1. Soil Water-Filled Pore Space and Temperature

Soil water availability influences organic carbon decomposition, and soil temperature a ffects microbial growth and activity. They are considered as two major factors driving the variation of soil respiration [14].

A distinct seasonal di fference of soil water-filled pore space (WFPS) and soil temperature can be observed (Figure 1). A sharp increase of WFPS occurred before 35 days after transplanting (DAT), and a decrease pattern was shown between 35 to 53 DAT. WFPS presented a total increase then decrease trend from 53 to 98 DAT. There was an upward trend of WFPS since 98 DAT (Figure 1a–c). As for soil temperature, a total decreasing trend was found throughout the whole tomato growing period except for a general increase between 35 to 49 DAT, between 70 to 83 DAT, as well as between 133 to 141 DAT (Figure 1d–f), which coincided with the seasonal patterns of air temperature. WFPS and soil temperature under aeration and high irrigation level were higher than the control and low irrigation level most of the time, which were in accordance with the findings of a previous study [9]. However, analysis of variance indicated that the e ffects of irrigation, aeration, and their interaction on mean WFPS and soil temperature were not significant (Table 1, *p* > 0.05).



Note: ns—significance at *p* > 0.05.

**Figure 1.** Soil water-filled pore space (WFPS) (**<sup>a</sup>**,**b**,**<sup>c</sup>**) and soil temperature (**d**,**e**,**f**) with and without aeration under the irrigation level of 60%W (**<sup>a</sup>**,**d**), 80%W (**b**,**<sup>e</sup>**), and 100%W (**<sup>c</sup>**,**f**) (mean ± SD, n = 3). W refers to full irrigation.

#### 2.1.2. Soil Microbe and Enzyme Activity

Heterotrophic respiration, as a primary contributor to the soil respiration, is impacted by soil carbon-use efficiency which varies based on soil microbial abundance and richness [22]. A previous study demonstrated that the abundance of bacteria (*cfu*b), fungi (*cfu*f), and actinomycetes (*cfu*a) are involved in the soil carbon cycle by decomposing organic matter, degrading cellulose, and forming antibiotic substances [16]. Soil microbial biomass carbon (MBC), which affects the transformation of all organic matter entering the soil, is the key and driving force of the nutrient and energy cycle in the whole ecosystem and is also an important source and reservoir of soil nutrient transformation. Soil cellulose activity (CA), which participates in the decomposition and release of CO2 from soil organic substances, is the main enzyme activity in soil carbon cycle. Soil dehydrogenase activity (DHA), which catalyzes dehydrogenation of organic substances and plays an intermediate role in hydrogen transformation and transfer, can be used as an indicator of the microbial redox system and is considered to be a global indicator of microbial metabolic activity in soil. However, soil biological activity can be limited by many factors, i.e., soil water and soil aeration conditions [23,24]. Very few pieces of literature were concerned with the soil microbes under AI [23,25], and the effect of AI on MBC, CA, and DHA have rarely been reported. Hence, study of the soil microbe and enzyme activity (*cfu*b, *cfu*f, *cfu*a, MBC, CA, and DHA) has grea<sup>t</sup> significance to reveal the mechanism of CO2 release under AI.

As seen in Figure 2, *cfu*b made up the majority of soil microbes, followed by *cfu*a and *cfu*f, which was generally supported by the results of Li et al. [16] and Zhu et al. [25]. Nevertheless, the microorganism abundance in the study of Zhu et al. [25] were greater than the values of the current research (Figure 2), which was influenced by higher soil temperature in their study (their study vs. our study = 18–32 ◦C vs. 9–29 ◦C). Furthermore, there were different results about the changing trends of soil microbes in the tomato growing period. Zhu et al. [25] pointed out that *cfu*b, *cfu*f, and *cfu*a integrally presented an increase pattern. Chen et al. [26] concluded that *cfu*f and *cfu*a showed an initial increase then decrease

trend, and peaks were observed on 50 d. In our study (Figure 2), the number of *cfu*b as a function of the days after transplanting was normally distributed, with the highest values observed on 98 DAT (Figure 2a,b). The number of *cfu*f peaked on 98 DAT, and the values during other periods were relatively stable (Figure 2c,d). The number of *cfu*a peaked on 35 DAT, but remained at a relatively constant level during other periods (Figure 2e,f). The differences of changing patterns could have resulted from the combined effects of the availability of different rhizosphere secretions and substrates, changes of soil moisture, temperature, and fertility, as well as plant growth. Soil microbial abundance (especially for *cfu*b and *cfu*f, Figure 2) peaked when soil hydrothermal conditions were good (Figure 1) and crops were growing vigorously on 98 DAT. Peaks of *cfu*a during the early tomato growing period (on 35 DAT) were probably ascribed to the highest WFPS (64.5%–67.7%) and greater soil temperature (23.1–24.7 ◦C), as well as greater soil substrates resulted from base fertilizer application [9]. Compared with the control, aeration under each irrigation level slightly increased mean values of *cfu*b, *cfu*f, and *cfu*a (Table 1, *p* > 0.05), with average increases of 4.6%, 5.5%, and 3.4%, respectively. Similar results were also reported by Li et al. [16], Du et al. [23], and Zhu et al. [25]. The increases of soil microbes under the aeration were likely due to the frequent alternation of soil dry and wet zones, thereby enhancing soil nutrient mineralization to improve microbial growth. Additionally, in line with previous researches [24,25], *cfu*b, *cfu*f, and *cfu*a in this study increased as irrigation amount increased (Figure 2), which was in order of 60% full irrigation (W) level without aeration (S) (W0.6S) < 80%W irrigation level without aeration (W0.8S) < 100%W irrigation level without aeration (W1.0S). The enhancement of soil microbes under aeration or high irrigation level was also probably ascribed to greater temperature (Figure 1d–f), which stimulated more microbial growth and activity [14].

**Figure 2.** The abundance of soil bacteria (**<sup>a</sup>**,**b**), fungi (**<sup>c</sup>**,**d**), and actinomycetes (**<sup>e</sup>**,**f**) with the irrigation level of 60%W, 80%W, and 100% W under the aeration (**<sup>a</sup>**,**c**,**<sup>e</sup>**) and control (**b**,**d**,**f**) (mean ± SD, n = 3).

*Catalysts* **2019**, *9*, 945

MBC generally exhibited an initial increase followed by a volatility within the range of 210.43 to 289.75 mg·kg−<sup>1</sup> throughout the whole tomato growing period (Figure 3a,b). Across all sampling periods, CA among treatments varied from 0.63 to 1.00 mg·kg−<sup>1</sup> and peaked on 35 DAT except for W0.8S treatment on 119 DAT (Figure 3c,d). Contrary to the changing rule of CA, DHA generally increased throughout the tomato growing period (Figure 3e,f). The changing patterns of soil enzyme activities were primarily because soil enzymes were correlated with the growth stages, soil texture, soil water content, soil temperature, air availability, and other factors [16]. Compared with the control, mean MBC, CA, and DHA under the aeration were slightly greater (Table 1, *p* > 0.05). As noted by Li et al. [16], soil enzymes are secreted by crop roots and rhizosphere microorganisms, as well as the decomposition of plant residues and microbial cells. Under the aeration, enhanced tomato root (Figure 4) and increased soil microbes (Figure 2) could immobilize and release nutrients into the soil and ameliorate soil fertility [23,27], which ultimately improved the CA and DHA (Figure 3). Additionally, soil water availability affects substrate availability, O2 concentrations, osmotic potential, gas diffusion, and cellular metabolism [24,28], thus impacting soil microbes. Difference in mean MBC, CA, and DHA values among treatments in this experiment was not significant (Table 1, *p* > 0.05).

**Figure 3.** Soil microbial biomass carbon (**<sup>a</sup>**,**b**), soil cellulase activity (**<sup>c</sup>**,**d**), and soil dehydrogenase activity (**<sup>e</sup>**,**f**) with the irrigation level of 60%W, 80%W, and 100%W under the aeration (**<sup>a</sup>**,**c**,**<sup>e</sup>**) and control (**b**,**d**,**f**) (mean ± SD, n = 3).

**Figure 4.** Total root length (**<sup>a</sup>**,**b**), surface area (**<sup>c</sup>**,**d**), and volume (**<sup>e</sup>**,**f**) with the irrigation level of 60%W, 80%W, and 100%W under the aeration (**<sup>a</sup>**,**c**,**<sup>e</sup>**) and control (**b**,**d**,**f**) (mean ± SD, n = 3).

In our study, the highest mean values of soil microbe and enzyme activity were obtained when 100%W was applied coupled with AI. This indicated that in a way the effect of irrigation on soil microbe and enzyme activity was enhanced under AI and that the soil biological environment was improved.

## 2.1.3. Root Morphology

The root system plays a decisive role in water and nutrient absorption. The size of crop roots also determines autotrophic respiration. Hence, studies on root morphology are of grea<sup>t</sup> practical significance to the study of plant growth and root respiration.

Aeration has been determined to increase root dry biomass and root morphology in cucumber [17,18], soybean [1], and even in the conventional staple grain crop [19]. However, there have been few studies regarding tomato root morphology under AI. Our results showed that total root length, surface area, and volume on 104 and 141 DAT were significantly greater than those on 42 and 68 DAT (Figure 4, *p* < 0.05). Compared with the control, total root length was significantly increased by 22.2% on average under aeration (Table 2, *p* < 0.05). Meanwhile, total root surface area and volume under the aeration was 6.6% and 6.7% higher than that of the control, respectively (*p* > 0.05). Li et al. [18] also showed that root morphology (root length, surface area, and volume) increased with increasing frequency of aeration. Root length of greenhouse muskmelon was 7076, 5839, 5207, and 3864 cm, and root surface area was 1217, 1023, 998, and 746 cm2, while root volume was 31.0, 26.1, 25.7, and 20.1 cm<sup>3</sup> for daily, 2-day, 4-day, and no aeration, respectively (*p* < 0.05) [18]. These increases of root morphology under the aeration were attributed to elongation, branching, and curving, influenced by the shape and dimensions of the wetted soil volume [18]. The injected air changed the soil structure owing to the shrinking and movement of soil particles, and it also pushed the water downwards [29]. All these characters in conjunction with higher soil moisture under aeration (Figure 1) were conducive to elongation of roots due to hydro-tropism. With respect to W1.0S, W0.6S significantly decreased the total root volume by 18.6% (*p* < 0.05), while the e ffects of other irrigation levels on root morphology were not significant (*p* > 0.05). Contrary to the results of the current study, Li et al. [18] stated that high irrigation levels has a negative e ffect on total root length, surface area, and volume with root length of 5981, 5364, and 5145 cm, surface area of 1114, 947, and 927 cm2, and volume of 30.8, 22.7, and 23.6 cm<sup>3</sup> for the 70%, 80%, and 90% of field capacity level, respectively. Xu et al. [30] demonstrated that root length and surface area presented an increasing then decreasing trend as soil changes from dry to moist. Di fferences among literature were likely due to di fferent hydrophily of crops controlled by the genes and tropic response to stimuli [18].

**Table 2.** The e ffects of irrigation, aeration, and their interaction on mean root morphology and dry biomass using a two-way ANOVA.


Note: ns, \*, and \*\*—significance at *p* > 0.05, *p* < 0.05, and *p* < 0.01, respectively.

## 2.1.4. Dry Biomass

Soil respiration was influenced by not only soil physical environment (i.e., soil temperature and moisture), but also plant growth [14]. Study of dry biomass throughout the whole tomato growing period was an e ffective way to analyze the changes of soil respiration, especially for the autotrophic component.

An increasing trend of dry biomass was observed throughout the whole tomato growing period, and dry biomass on 42, 68, 104, and 142 DAT showed a similar changing pattern among treatments (Figure 5). Taking dry biomass at harvest (142 DAT) as an example, dry biomass of tomato leaf, stem, fruit, and root under aeration were higher than the control (Figure 5). As reported previously [31], the average increases of each part were 17.8%, 17.7%, 17.8%, and 8.4%, respectively, and the e ffect of aeration on leaf, stem, and fruit was significant (Table 2, *p* < 0.05). These improvements of dry biomass were in agreemen<sup>t</sup> with the results of former research [4,32], which were beneficial from increased soil aeration and reduced phytohormones under AI [31]. Dry biomass of tomato leaf, stem, fruit, and root increased as irrigation amount increased, and the e ffect was significant on leaf, fruit, and root (Table 2, *p* < 0.05). As noted previously [31], dry weight of root, stem, leaf, and fruit under 100%W was increased by 22.2%, 19.3%, 22.5%, and 19.0%, and by 20.1%, 5.4%, 7.0%, and 12.1% than that under 60%W and 80%W treatment, respectively. Zhu et al. [4] demonstrated that with crop-pan coe fficient increasing from 0.6 to 1.0, dry biomass of root, stem, and leaf was increased by 24.0%, 17.2%, and 22.8%, respectively. The enhancement of dry biomass as irrigation amount increased was primarily ascribed to the greater canopy and leaf area index [4], as well as increased assimilation rate under high irrigation level [33].

**Figure 5.** Dry biomass of tomato fruit, leaf, stem, and root among treatments on 42, 68, 104, and 142 days after transplanting (DAT). The number 1, 2, 3, 4, 5, and 6 represented treatment of 60%W with aeration, 60%W without aeration, 80%W with aeration, 80%W without aeration, 100%W with aeration, and 100%W without aeration, respectively.

## *2.2. Soil Respiration*

As presented in Figure 6, soil respiration showed fluctuated patterns during the whole tomato growing period, which varied from 139.19 to 748.64 mg·m<sup>−</sup>2·h−<sup>1</sup> among treatments. Ranges of soil respiration in the present study was similar to the results of Hou et al. [10] but was higher than the research of the same tomato cultivations [9]. Differences might be the results of different irrigation amount and weather condition based on the year of cultivation. The changing patterns of soil respiration could be explained mostly by the abiotic and biotic factors (Figures 1–5). The lowest values on 9 DAT were mainly due to lower soil microbes (especially for *cfu*b and *cfu*f, Figure 2) and undeveloped tomato roots (Figures 4 and 5) at the onset of transplantation. As days after transplanting increased, *cfu*b and DHA increased gradually (Figures 2 and 3), and the root growth enhanced slightly (Figures 4 and 5), inducing larger emissions on 83 DAT. Relatively lower WFPS and obvious increases of soil temperature on 49 DAT (Figure 1) resulted in the peaks of soil respiration under W0.6O, W0.8S, and W1.0S treatment. Higher soil respiration on 62 DAT was attributed to increased WFPS, resulting in peaks under W0.8O and W1.0O treatment. Lower soil respiration on 98 and 133 DAT was primarily ascribed to a sharp decline of WFPS (Figure 1). An increasing trend of soil respiration was detected since 133 DAT, which was probably due to the increase of WFPS and soil temperature.

**Figure 6.** Soil respiration with and without aeration under the irrigation level of 60%W (**a**), 80%W (**b**), and 100%W (**c**) (mean ± SD, n = 3).

Previous research has shown a good correlation between soil respiration and soil temperature, oxygen concentration, and air-filled porosity [34]. Nevertheless, the correlation between soil respiration and soil microbe and enzyme activity, as well as plant growth under the aeration and irrigation treatments has not ye<sup>t</sup> been well studied. In our study, regression analysis (linear, polynomial, and exponential) between soil respiration and WFPS was conducted, and a significant polynomial function was observed (Figure 7a,b, *p* < 0.05), similar to previous studies [11,12]. Further analysis found that a polynomial correlation was detected between soil respiration and WFPS when WFPS was below 60% (*p* < 0.01), while a linear positive correlation was observed when WFPS was above 60% (Figure 7c,d, *p* = 0.245 and 0.001 for the aeration and control, respectively). Moreover, there were significant negative correlations between soil respiration and *cfu*b, as well as between soil respiration and *cfu*f (Table 3, *p* < 0.01), which was different from the result of Zhu et al. [25] where soil respiration showed strong positive correlations with *cfu*b, *cfu*f, and *cfu*a. The reason for the inconsistent conclusions was probably due to the different growing seasons. Zhu et al. [25] conducted the experiment in the spring–summer period where the weather was gradually raised, while the present experiment was finished in the autumn–winter period where the weather was gradually reduced. Different variation of soil temperature would lead to different changing rules of soil respiration, microbial activity, and water content. In the present study, the interactive effect of WFPS, *cfu*b, and *cfu*f on soil respiration was extremely significant (Table 3, *p* < 0.01), which collectively accounted for 70.2% and 61.6% of changes in soil respiration under aeration and control, respectively. Unfortunately, correlations between soil respiration and other soil physical and biotic components (soil temperature, *cfu*a, MBC, CA, DHA, tomato root morphology, and plant dry biomass) were not significant (*p* > 0.05, data not shown), which required further study.

**Figure 7.** Correlation between soil respiration with soil water-filled pore space (WFPS) for the all WFPS data (**<sup>a</sup>**,**b**) and data with WFPS piecewise analysis (**<sup>c</sup>**,**d**). (**a**)(**c**) and (**b**)(**d**) represented correlation under the aeration and control, respectively.

**Table 3.** Relationships of soil respiration (Rs) with soil water-filled pore space (WFPS), the abundance of soil bacteria (*cfu*b) and fungi (*cfu*f) under the aeration and control treatments.


Similar to previous results [6,9,10,34], soil respiration under the aeration in our study was typically and on average 6.2% greater but no significant different to that under the control according to ANOVA (Figure 6, Table 1, *p* > 0.05). Chen et al. [6] found that soil respiration increased by 42%–100% for oxygenation compared to control. Hou et al. [10] stated that aeration increased soil CO2 emissions by 11.8% (*p* = 0.394) compared to the control. Zhu et al. [34] revealed that mean soil respiration under the aeration was 22.5% higher than the control. Potential reasons explaining the enhancement of soil respiration under aeration include: (1) aeration increased soil microbes (Figure 2,3) and root growth (Figures 4 and 5), which essentially controlled heterotrophic respiration and autotrophic respiration [22,35]; (2) greater CA and DHA under AI (Figure 3), which were involved in the decomposition and release of CO2 from soil organic substances, in turn promoted soil respiration; and (3) as a result of the enhanced aboveground dry biomass under AI, the increased demand for nutrients stimulated belowground C allocation and root growth (Figures 4 and 5), which increased substrates to soil organisms and stimulated organic matter turnover [36,37], leading to higher biomass and/or activity that might stimulate the decomposition of soil organic matter [38,39]. All of these factors increased soil respiration conclusively. Consistent with previous research [9,40], soil respiration increased in the order of W0.6S < W0.8S < W1.0S (Figure 6), which resulted from increased soil microbial biomass (Figure 2), soil enzyme activity (Figure 3), root biomass (Figure 5), mineralization and decomposition rate of soil organic matter, as well as the diffusion rate of gases in soil pores [41]. Soil respiration under W1.0S was 16.0% and 13.9% higher than that under W0.6S and W0.8S, respectively. Nevertheless, the effect of irrigation on soil respiration was not significant (Table 1, *p* > 0.05).

Although this paper analyzed the response of soil respiration to soil physical and biotic variables, we do not know the proportion of root or microbial respiration to soil respiration as no measurement was made in this study, which was the deficiency of this study, and needs to be further carried out in future experiments.

#### **3. Materials and Methods**
