Soil CO₂ Uptake in Deserts and Its Implications to the Groundwater Environment

Abstract

Recent studies of soil carbon cycle in arid and semi-arid ecosystems demonstrated that there exists an abiotic CO₂ absorption by saline-alkali soils (A_a) at desert ecosystems and suggested potential contributions of CO₂ dissolution beneath deserts to the terrestrial ecosystems carbon balance. However, the overall importance of such soil CO₂ uptake is still undetermined and its implications to the groundwater environment remain unaddressed. In this manuscript, a simple method is proposed for the direct computation of A_a from the total soil CO₂ flux (F_a) as well as for the evaluation of A_a importance to F_a. An artificial soil-groundwater system was employed to investigate the implications to groundwater environment and it was found that soil CO₂ uptake in deserts can contribute a possible influence on the evolution of the groundwater environment, providing that the absorbed CO₂ largely remained in the soil-groundwater system.

1. Introduction

Global soil CO₂ releases account for 20%–38% of the annual input of CO₂ from terrestrial and marine sources to the atmosphere [1,2,3,4,5,6,7,8,9,10,11]. Therefore, soil CO₂ flux (F_a) is a significant determinant of the ecosystem carbon balance. Among the components of F_a, the biotic processes (including the root and microbial respiratory components) were thought to be significantly contributing to it [2,3,4,5,6], while the abiotic processes were always neglected [12,13,14,15,16,17,18,19,20]. However, a series of recent studies of deserts ecosystems have highlighted that the abiotic processes (including subterranean CO₂ dissolution and other abiotic carbon sink beneath deserts) must be taken into account in the budget of F_a [7,8,9,10].

Some recent studies attributed the abiotic processes to the effects of the soil water dynamics on subterranean carbon cycle and presented theoretical and experimental analyses based on the biogeochemical reactive transport modeling [21,22]. This not only developed the methodology of the partition of F_a into the biotic and abiotic components across some special environments, but also highlighted that the subterranean dissolution of CO₂ has been involved in the ecosystem carbon balance in deserts, and can even temporally dominate [7,8,9,10,11]. However, soil CO₂ dissolution and effusion, as abiotic processes in deserts, are different from the considered processes in these previous studies. Especially in some experimental contrast, only the top few cm of the soil was considered. It became much more impressive after the abiotic CO₂ absorption (A_a) by the soil collected from a saline desert was directly observed after a minimizing-disturbance sterilization treatment on the soil [8]. This demonstrated that there exists an abiotic CO₂ uptake by saline-alkali soils (A_a) at desert ecosystems and suggested a potential contribution of such soil CO₂ uptake to the ecosystem carbon balance in deserts [8,10].

There is strong evidence suggesting that A_a can contribute significantly to F_a and the net ecosystem carbon balance, but the overall importance of the soil CO₂ uptake to F_a is still undetermined. Presently, both the field observations and the laboratory experiments are still limited on site scales. Because of the lack of large-scale experiments, the magnitude of such CO₂ uptake is still a matter of controversy. Some recent publications have further presented the laboratory data on the soil abiotic CO₂ absorption and the field data on the net CO₂ exchange in desert ecosystems [23,24], whereas it is still imperative to have a simple method for a direct computation of the abiotic soil CO₂ uptake from the total soil CO₂ flux and for quantifying its overall importance to F_a.

The minimizing-disturbance sterilization treatment on soils has been widely recognized as an experimental method for separating A_a [10]. Since the significance of CO₂ dissolution and effusion as the abiotic processes has become much more impressive after the soil CO₂ uptake was directly observed after such a sterilization treatment on the soil [8,10], another subsequent unresolved issue is whether the beneath CO₂ dissolution has a non-negligible implication to the groundwater environment beneath deserts, providing that the dissolved and absorbed CO₂ largely remained in the soil-groundwater system. This certainly desires a further investigation but remains unaddressed.

This manuscript aims to design a simple method for a direct computation of the abiotic soil CO₂ uptake from the total soil CO₂ flux and proposed a simple approach to quantify its overall importance to F_a. After a special treatment, F_a is reconciled as a simple function of A_a, which is implemented in two different approaches. In the experimental approach, all A_a data are collected after the minimizing-disturbance sterilization treatment on the soils, while in the other semi-experimental approach, most of the A_a data are directly calculated from a simple regressive model. The potential implications of A_a and the beneath CO₂ dissolution to the groundwater environment are also preliminarily discussed.

2. Materials and Methods

2.1. Data Sources

Sampling data of soil CO₂ fluxes were collected from Xie et al. [8]. The measurements were conducted in the Gurbantunggut Desert, which is located at hinterland of the Eurasian Continent (87°56′ E, 44°17′ N; elevation: 461 m), with an arid, windy climate (sunshine: 3079 h; precipitation: 144.7 mm; evapotranspiration: 2020 mm; radiation: 5439 MJ/m²; velocity: 2.6 m/s). Note that two LI-8100s (LI-COR, Lincoln, Nebraska, NE, USA) have been employed for the valid comparison between the flux data measured at the sterilized and unsterilized samples. It should be noted that the soil was smashed, roots-sieved, air-dried and the many sinks and sources of CO₂ in the subsurface environment have been excluded. After sterilization, the net soil CO₂ fluxes were negative and the CO₂ exchange was abiotic. In order to highlight the soil CO₂ uptake, such abiotic exchange is hence defined as “soil abiotic CO₂ absorption” throughout this manuscript (Figure 1). Consequently, the variations of F_a were largely determined by A_a and the soil microbial respiration (R_m).

With the purpose of a preliminary discussion on the potential implications of the beneath CO₂ dissolution to the groundwater environment, an artificial soil-groundwater system was employed and the groundwater were sampled with six replications (Figure 2). The groundwater pH and the concentration of dissolved inorganic carbon (DIC) in the artificial soil-groundwater system were measured, where the concentration of DIC (symbolized as (DIC) was defined as the sum of the concentration of the dissolved carbonate ion and bicarbonate ion.

2.2. Modeling Approach

Neglecting A_a, it is well-known that F_a can be formulated as an exponential function of the soil surface temperature or the ambient air temperature [19]:

$${\mathrm{F}}_{\mathrm{a}}={\mathrm{F}}_{\mathrm{a}20}\times {{\mathrm{Q}}_{10}}^{（{\mathrm{T}}_{\mathrm{S}}-20）/10}$$

(1)

where F_a20 is measured F_a at 20 °C, T_S is the soil temperature (here it is measured at 0–10 cm), and Q₁₀ is the relative change in F_a with 10 °C increases.

R_m is formulated as in [20]:

$${\mathrm{R}}_{\mathrm{m}}={\mathrm{R}}_{\mathrm{m}20}\cdot {{\mathrm{Q}}_{10}}^{({\mathrm{T}}_{\mathrm{S}}-20)/10}$$

(2)

where R_m20 is the measured value of R_m at 20 °C.

Taking into account A_a, F_a is reconciled as,

$${\mathrm{F}}_{\mathrm{a}}={\mathrm{R}}_{\mathrm{m}}+{\mathrm{A}}_{\mathrm{a}}$$

(3)

The experimental method separates A_a by the sterilization treatment on the soils and hence R_m20 can be estimated as follows:

$${\mathrm{R}}_{\mathrm{m}20}={\mathrm{F}}_{\mathrm{a}20}-{\mathrm{A}}_{\mathrm{a}20}$$

(4)

where A_a20 are measured A_a at 20 °C.

Substituting Equation (2) into Equation (4), R_m is reconciled as:

$${\mathrm{R}}_{\mathrm{m}}=({\mathrm{F}}_{\mathrm{a}20}-{\mathrm{A}}_{20})\cdot {{\mathrm{Q}}_{10}}^{({\mathrm{T}}_{\mathrm{S}}-20)/10}$$

(5)

Correspondingly, F_a can be finally reconciled as:

$${\mathrm{F}}_{\mathrm{a}}=({\mathrm{F}}_{\mathrm{a}20}-{\mathrm{A}}_{20})\cdot {{\mathrm{Q}}_{10}}^{({\mathrm{T}}_{\mathrm{S}}-20)/10}+{\mathrm{A}}_{\mathrm{a}}$$

(6)

Unfortunately, extensive time is required during the minimizing-disturbance sterilization treatment on the soil. At large scales, it is necessary to couple with other methodologies. Development of a semi-experimental method for the rough partition of A_a is thus necessary.

Similar to semi-experimental approaches for the partition of the biotic components of F_a [12], the relationship between A_a and F_a could be assumed to be represented by a simple linear regressive model and written as:

$${\mathrm{A}}_{\mathrm{a}}={\mathsf{\beta }}_0{\mathrm{F}}_{\mathrm{a}}+{\mathsf{\beta }}_1$$

(7)

where β₀ and β₁ are determined by the contributions of A_a to F_a.

Under this assumption, R_m is reconciled as:

$${\mathrm{R}}_{\mathrm{m}}=(1-{\mathsf{\beta }}_0){\mathrm{F}}_{\mathrm{a}}-{\mathsf{\beta }}_1$$

(8)

and correspondingly

$${\mathrm{R}}_{\mathrm{m}10}=(1-{\mathsf{\beta }}_0){\mathrm{F}}_{\mathrm{a}10}-{\mathsf{\beta }}_1$$

(9)

By Equation (2), the Q₁₀ value for R_m can be estimated by

$${\mathrm{R}}_{\mathrm{m}}=[(1-{\mathsf{\beta }}_0){\mathrm{F}}_{\mathrm{a}10}-{\mathsf{\beta }}_1]\cdot {{\mathrm{Q}}_{10}}^{({\mathrm{T}}_{\mathrm{S}}-10)/10}$$

(10)

Combining Equations (8) and (10), the soil CO₂ flux could be finally reconciled as

$${\mathrm{F}}_{\mathrm{a}}=[{\mathrm{F}}_{\mathrm{a}10}-\frac{{\mathsf{\beta }}_1}{1-{\mathsf{\beta }}_0}]\cdot {{\mathrm{Q}}_{10}}^{({\mathrm{T}}_{\mathrm{S}}-10)/10}+\frac{{\mathsf{\beta }}_1}{1-{\mathsf{\beta }}_0}$$

(11)

The parameters β₀ and β₁ can be estimated by Equation (7) after a few experiments and a large magnitude of time will be saved in the partition of A_a.

Regarding the first approach, the relationship between the groundwater pH and (DIC) are analyzed by utilizing an exponentially regressive model:

$$\mathrm{p}\mathrm{H}={\mathsf{\lambda }}_0{\mathrm{e}}^{{\mathsf{\lambda }}_1(\mathrm{D}\mathrm{I}\mathrm{C})}$$

(12)

where λ₀ and λ₁ are determined by the contributions of (DIC) to the groundwater pH.

3. Results and Discussion

3.1. Separated Biotic and Abiotic Components of Soil CO₂ Fluxes

The aforementioned parameters of β₀ and β₁ are estimated using the measured values of A_a and F_a on two clear days in two typical desert ecosystems, saline desert and oasis farmland (Figure 3).

The estimated parameters of β₀ and β₁ are then applied to the semi-experimental separation of A_a at the corresponding ecosystems in a drought period of one growing season in 2006, taking into account their difference for different ecosystems (the estimated β₀ and β₁ from saline desert are also applied to its neighbor abandoned farmland, as provided in

Table 1. Suitable parameters for the isolation of A_a.

Study Sites	Parameters Applied in Semi-Experimental Method
Saline desert	β0 = 0.5284	β1 = −0.8975
Abandoned farmland	β0 = 0.5284	β1 = −0.8975
Oasis farmland	β0 = 0.4266	β1 = −0.8330

).

It is obvious that A_a separated in both the experimental and semi-experimental approaches acts as a CO₂ sink. A_a separated into two approaches varies in a similar pattern during the considered drought period. Through comparison with the total soil CO₂ flux F_a, it was found that most F_a data are positive. This implies that the other flux component of F_a, the microbial soil respiration R_m, plays a dominant role within this period (Figure 4)

Considering that β₀ and β₁ are determined by the contributions of A_a to F_a, they might be very different in various ecosystems (A_a is not very different, but F_a can be very different). Choosing soil sites for the few sampled experiments should be done cautiously.

The values of Q₁₀ for the microbial soil respiration R_m might also be different in various ecosystems. From the visualization of the separated R_m in considered drought period by a continuous curve, the gradient of R_m in some sites and sampling days might not be evident. For to this reason, only the separated results of R_m on nine sampled days are presented here and three kinds of ecosystems are selected for the comparison, the saline desert, abandoned farmland and oasis farmland. It could also facilitate the valid and explicit comparison between the separated results of R_m from the experimental method and the semi-experimental method. Obviously, the separated R_m in the semi-experimental and experimental approaches show similar patterns. The diurnal dynamics on nine sampled days also reveal a duplicated variability (Figure 5 and Figure 6).

The separated results of R_m in the former two ecosystems are almost the same (

Table 2. Q₁₀ values for isolated R_m by two methods.

Study Sites	Semi-Experimental Method	Experimental Method
Saline desert	0.53200	0.58711
Abandoned farmland	0.81835	0.77012
Oasis farmland	1.21009	1.19123

), which are all different with the oasis farmland. This might be the reason that the soil microbial activity in the saline desert is very different to that in oasis farmland.

The separated A_a from the semi-experimental approach represents about 80% of that from the experimental approach, which indicates a reliable cross validation (Figure 7). In fact, the experimental method should still be used with much caution since it indicates a large magnitude of carbon sink in the most barren ecosystems [10,14]. Thus, the cross validations between these two methods are necessary.

Employing the estimated β₀ and β₁ in Equation (10) and fitted Q₁₀ provided in Table 2, the separated R_m by the semi-experimental method explain well the variability of separated R_m by the experimental method (Figure 8). Both semi-experimental and experimental methods reconcile F_a as a direct sum of A_a and R_m, rendering the cross validation of reconciled F_a using these two methods necessary. In Figure 9, it is also found that F_a reconciled by Equation (11) using the semi-experimental approach almost coincides with that reconciled by Equation (6) in the experimental approach. The aforementioned results of cross validation also show that the semi-experimental separations of R_m and A_a proposed in Section 2 are both feasible and reliable (

Table 3. Fitted relationships between isolated components by two methods.

Flux Components	Slope	RMSE
Alkaline absorption	0.80630	0.29089
Microbial respiration	0.71842	0.14753
Reconciled soil CO2 flux	0.91428	0.25061

).

3.2. Implications of Soil CO₂ Uptake to the Groundwater Environment

Although soil CO₂ uptake in deserts has been investigated in many previous studies, its potential influences on the groundwater environment remain unaddressed [25,26,27,28,29,30,31,32,33,34,35,36,37,38,39]. This is partially due to the lack of a simple method for the separation and quantification of such CO₂ uptake. This study presents a simple yet efficient method for partitioning CO₂ influxes to and efflux from bare saline-alkali soils, implying a time to consider the implications of such CO₂ uptake beneath deserts to the groundwater environment. An unresolved issue, where the absorbed CO₂ has gone, still remains. Most studies assumed that these absorbed CO₂ are partially dissolved in the soil-groundwater system. However, the magnitude of the beneath CO₂ dissolution remains undetermined [25,26,27,28,29,30].

Furthermore, the experimental partition method is still desired for further improvement. Based on some recent studies, there exists little difference between abiotic soil CO₂ influx and efflux when excluding the role of groundwater and soil is left to absorb/release CO₂ naturally [28]. The difference might be attributed to the extent of the sterilization treatment on the soil and the existence of some soil microbial respiration (R_m), or the interaction between the existing R_m and A_a. Future studies must keep a most cautious mind in extending the results from the laboratory experiments to field research. In reality, the conditions are much more complex and the additional confounding factors for many flux components exist and might interact with each other.

It has been widely recognized that the overall magnitude of CO₂ dissolution beneath deserts and its contributions to the ecosystem carbon balance can be huge [30,31,32,33,34,35,36,37,38,39], but its effect on the local environment are seldom reported. As a first attempt to analyze the implications of the soil CO₂ uptake to the groundwater environment, the present study simply assumes that the absorbed CO₂ has been largely dissolved in the soil-groundwater system beneath the deserts. Under this assumption, an artificial soil-groundwater system has been incubated for a long period in the laboratory. Collected data at the first stage confirm that the implications to the water environment should be taken into account (Figure 10).

Since the potential contributions of these abiotic components to the total soil CO₂ fluxes were non-negligible, the implication of the biotic CO₂ effluxes and abiotic CO₂ influxes to the soil-groundwater environment cycle is a subsequent problem. Despite the recent studies of soil carbon cycle in the desert ecosystems and the biogeochemical investigations in the arid ecosystems [40,41,42,43,44,45,46,47,48,49], the dynamics of beneath CO₂ dissolution [7,10] is still not well-understood. Furthermore, even the mechanisms of subterranean abiotic processes are still not wholly known. It is worth noting that the groundwater pH is essentially significant for the local water environment and has attracted a wide attention [50,51,52,53,54,55,56,57,58,59,60]. Attaining a reliable quantification for the contributions of soil CO₂ uptake to the groundwater environment remains one major challenge and hence should be a research priority.

Because the mechanisms for the abiotic CO₂ absorption and the overall magnitude of the soil CO₂ uptake are still not well-understood, there are still considerable uncertainties in more explicit analyses of the contributions of such soil CO₂ uptake to the groundwater environment [8,10,14]. Additionally, the data required by the present experimental method can only be collected from the top few cm of the soil. The whole atmospheric exchange of CO₂ requires some further investigations towards better and more suitable experimental and non-experimental methodologies.

4. Conclusions

This study presented a simple yet efficient method to work out the magnitude of soil abiotic CO₂ absorption in deserts from the data set of soil CO₂ fluxes, along with a first approach to illustrate its potential role in influencing environments in an artificial soil-groundwater system. The first approach highlighted significant events where an evident variation of the groundwater pH, provided that the absorbed CO₂ largely remained in the soil-groundwater system. Considering the CO₂ sink size across different environments remains unknown, there are still considerable uncertainties and further investigations are necessary.