Warming Enhances CO₂ Flux from Saline–Alkali Soils by Intensifying Moisture–Temperature Interactions in the Critical Zone

Abstract

Saline–alkali soils in arid regions are increasingly recognized as critical yet underrepresented components of the global carbon cycle. However, their CO₂ flux dynamics under warming remain poorly understood. In this study, we conducted controlled growth-chamber experiments using typical saline–alkali soils from the Taklamakan Desert, where temperature, soil moisture, and atmospheric CO₂ concentrations were systematically manipulated. We quantified how warming reshaped moisture–temperature interactions regulating soil CO₂ fluxes. The results revealed a pronounced diurnal variation pattern, characterized by daytime CO₂ release and nighttime uptake. Temperature was identified as the dominant driver (R² > 0.93, p < 0.001), whereas soil moisture primarily modulated flux intensity; at 0.8 cm³ cm⁻³, fluxes declined by up to 61% compared with the baseline. Warming enhanced the temperature–moisture synergy (−43%, p < 0.01) and simultaneously reduced baseline fluxes (−56%, p < 0.01). These shifts fundamentally altered the regulation of CO₂ flux dynamics. Our findings highlight the necessity of integrating salt dynamics and carbonate equilibria into multiphase reactive transport models to improve regional carbon sink assessments. Ultimately, this study refines estimates of the contribution of saline–alkali soils to the global “missing carbon sink” (~1.7 Pg C a⁻¹) and emphasizes their overlooked role in the Earth’s carbon budget under a warming climate.

1. Introduction

The carbon cycle is a fundamental component of Earth’s geochemical system, directly regulating climate and ecosystem stability [1,2]. Since the Industrial Revolution, anthropogenic activities have disrupted this cycle by releasing excessive greenhouse gases (e.g., CO₂ and CH₄) [3,4], intensifying the greenhouse effect and triggering global warming with more frequent extreme climate events. Rising temperatures accelerate soil organic matter decomposition, alter vegetation carbon sequestration efficiency, and weaken oceanic carbon sinks, thereby reinforcing positive climate feedback. Under the influence of global warming, the intensification of the regional hydrological cycle has produced nonlinear climatic responses. These include more frequent extreme precipitation events, which nonetheless fail to alleviate the desert’s inherently arid nature [5]. Long-term, intense evaporation promotes salt accumulation on the soil surface. These salts are transported and concentrated in depressions via surface and subsurface water flow, leading to the widespread distribution of saline soils across the Tarim River Basin [6,7,8]. This hypersaline environment further destabilizes the ecological balance, rendering the regional ecosystem increasingly fragile. Despite considerable progress, a persistent gap of ~1.7 Pg C a⁻¹ remains in the global carbon budget [9,10,11], commonly referred to as the “missing carbon sink.” Recent studies suggest that terrestrial carbon pools under warming conditions may account for part of this discrepancy [12,13,14]. Among these pools, saline–alkaline soils are increasingly recognized for their abiotic inorganic carbon sequestration mechanisms and sensitivity to moisture–temperature interaction changes, highlighting their potential role in resolving the missing sink.

Previous studies first identified a diurnal dissolution–precipitation cycle of inorganic carbon flux in the Taklamakan Desert, tightly coupled with temperature fluctuations [15,16,17]. Similar patterns have since been reported in the Chihuahuan, Gurbantunggut, and Negev deserts [18,19,20], indicating that this phenomenon is widespread across arid regions (Table 1). Carbon isotope tracing in Ningxia further confirmed that negative CO₂ fluxes are associated with carbonate dissolution driven by diurnal temperature variation [21,22]. Additional studies have demonstrated that soil temperature and moisture jointly regulate CO₂ solubility, subsurface partial pressure, and the balance between CO₂ uptake and release [18,23,24]. Radiocarbon (^14C) analyses in the Tarim Basin revealed that modern atmospheric CO₂ can migrate downward with soil solution and accumulate in saline basins, providing direct evidence for an inorganic carbon sink and supporting the hidden sink hypothesis [25]. Collectively, these findings establish carbonate reactions—including the dissolution of primary minerals, atmospheric CO₂ absorption, and pedogenic carbonate formation—as critical processes of inorganic carbon sequestration in saline–alkaline soils. However, substantial uncertainty remains regarding how natural drivers interact to control the magnitude and direction of CO₂ fluxes, particularly under a warming climate.

Simulated climate-warming conditions further complicate these dynamics by exerting both positive and negative effects on carbon cycling [26]. Elevated temperatures stimulate microbial activity and accelerate SOC mineralization, potentially transforming saline–alkaline soils from sinks into sources of CO₂ [3,27]. Concurrently, warming intensifies evapotranspiration, lowers soil moisture, alters soil structure, and constrains plant root development and vegetation carbon input [28]. These changes indirectly affect microbial processes and organic carbon turnover [29]. Conversely, warming can extend growing seasons, enhance productivity, and increase carbon input from salt-tolerant vegetation, partially offsetting losses [30]. Thus, warming exerts a dual influence: reinforcing some sequestration pathways while weakening others [31]. In arid Northwest China, where warming has progressed at three times the global average, precipitation has increased, but potential evapotranspiration has risen even faster. The net effect remains uncertain: wetter conditions may promote vegetation cover and carbon uptake, while extreme events such as heatwaves, floods, and salt crust accumulation destabilize ecosystems and weaken sink strength [11,32]. Given that China has ~99 million hectares of saline–alkaline soils, clarifying these mechanisms is crucial for assessing their carbon sequestration potential and contribution to the missing carbon sink [2,12,33].

To address these uncertainties, this study investigates the response of saline–alkali soil CO₂ fluxes to simulated climate-warming conditions, using controlled chamber experiments with soils collected from the Tarim River Basin. By manipulating temperature, moisture gradients, and atmospheric CO₂ concentrations, we systematically quantify how warming reshapes moisture–temperature interactions regulating CO₂ fluxes. Specifically, we test the hypothesis that warming enhances moisture–temperature interactions while simultaneously reducing baseline fluxes. Building on this framework, the objectives of this study are to (1) quantify moisture–temperature interaction effects on CO₂ flux, (2) assess warming’s regulatory role (+2 °C, ~650 ppm CO₂), and (3) develop estimation schemes for saline–alkaline soils under different scenarios. These objectives provide a clear pathway to evaluate the role of saline–alkaline soils in the global carbon budget and refine estimates of their contribution to the missing carbon sink.

Table 1. Summary of CO₂ flux ranges and key drivers in various arid and semi-arid regions.

Location	Flux Range (μmol·s−1·m−2)	Key Driver	Citation
Chihuahuan Desert	20–154.7	Soil Moisture, Temperature, Precipitation	[24]
Mojave Desert	0.39–1.49	Soil Temperature, Soil Moisture	[18]
Sonoran Desert	110–476.6	Soil Temperature, Moisture, Vegetation	[34]
Taklamakan Desert	−0.56 to 1.24	Inorganic Carbon, Soil Moisture	[15]
Gurbantunggut Desert	−0.28 to 1.2	Carbonates, Moisture	[19,20]

Table 1. Summary of CO₂ flux ranges and key drivers in various arid and semi-arid regions.

Location	Flux Range (μmol·s−1·m−2)	Key Driver	Citation
Chihuahuan Desert	20–154.7	Soil Moisture, Temperature, Precipitation	[24]
Mojave Desert	0.39–1.49	Soil Temperature, Soil Moisture	[18]
Sonoran Desert	110–476.6	Soil Temperature, Moisture, Vegetation	[34]
Taklamakan Desert	−0.56 to 1.24	Inorganic Carbon, Soil Moisture	[15]
Gurbantunggut Desert	−0.28 to 1.2	Carbonates, Moisture	[19,20]

2. Materials and Methods

2.1. Site Description

The Taklimakan Desert (Figure 1a, 75° E–90° E, 35° N–42° N), located within the Tarim Basin of Northwest China, is the largest desert in China and the world’s second-largest shifting sand desert. It is characterized by an extremely arid desert climate. The mean annual precipitation is less than 50 mm, with pronounced spatiotemporal variability; temporally, rainfall is concentrated between May and August, while spatially, it decreases from north to south and from west to east. The lowest annual precipitation on record—only 25.9 mm—was observed at the Tazhong Meteorological Station in the desert interior. The annual potential evapotranspiration ranges from 1500 to 3815 mm (with an extreme of 3812.3 mm recorded at Tazhong), equivalent to approximately 150 times the annual precipitation [35,36,37]. The mean annual temperature is about 10 °C, but extremes range from −32.6 °C to 46 °C. Persistent winds (mean annual speed: 1.5–3.0 m·s⁻¹), combined with abundant sand sources, drive frequent dust events, establishing the Taklimakan as one of Asia’s primary dust source regions.

2.2. Soil Characterization

To better characterize the saline–alkali soils used in this study, a series of physicochemical analyses was conducted prior to the chamber experiments. Soil electrical conductivity (EC) was determined from 1:5 soil-to-water extracts, with measurements taken at depths of 3 cm and 10 cm. Calcium ion (Ca²⁺) concentrations were measured using acidimetry (HCl neutralization). Soil organic carbon (SOC) was quantified using the Walkley–Black dichromate oxidation method. Soil texture (sand, silt, and clay fractions) was determined using the hydrometer method after pretreatment to remove organic matter and carbonates.

Summary statistics for these parameters are presented in

Table 2. Physicochemical characteristics of surface saline–alkali soils (0–20 cm) used in the controlled chamber experiments. Values are expressed as mean ± standard deviation (SD), with minimum and maximum ranges shown. Parameters include soil electrical conductivity (EC, measured at 3 cm and 10 cm depths), total carbon, SOC, pH, Ca²⁺, Mg²⁺, and total salts.

Parameter	Mean	SD	Min	Max
Total Carbon (g·kg−1)	1.27	0.41	0.50	2.00
SOC (g·kg−1)	1.92	0.73	1.03	4.94
pH (1:2.5)	8.84	0.12	8.69	9.24
Ca2+ (mg·g−1)	3.91	0.62	2.55	6.66
Mg2+ (mg·g−1)	2.09	1.16	0.17	5.15
Total Salts (mg·g−1)	105.80	53.08	12.33	359.70
EC (3 cm depth, dS·m−1)	0.003	0.003	0.001	0.018
EC (10 cm depth, dS·m−1)	0.253	0.059	0.051	0.363

. The soils were moderately alkaline (mean pH 8.84), with relatively low SOC (mean 1.92 g·kg⁻¹) and notable calcium content (mean Ca²⁺ concentration 3.91 mg·g⁻¹). Electrical conductivity values reflected substantial salinity heterogeneity, ranging from 0.01 to 0.018 dS·m⁻¹ at 3 cm depth to 0.056 to 0.20 dS·m⁻¹ at 10 cm depth. The total salt content was also high (mean 105.80 mg·g⁻¹), with considerable variability among samples.

The 0–20 cm soil layer was selected for sampling as it represents the surface zone most directly influenced by atmospheric exchange and moisture–temperature interactions. This depth includes the zone of active carbonate dissolution–precipitation, salt accumulation, and microbial activity, and is therefore widely recognized as representative of surface CO₂ flux dynamics. Its strong sensitivity to diurnal variations in temperature and soil moisture further justifies its use in assessing flux responses under warming conditions.

2.3. Experimental Design

Prior to measurements, the LI-8100A automated soil CO₂ flux system (LI-8100A with LI-8150 multiplexer; LI-COR, Lincoln, NE, USA; accuracy ±1.5%) was calibrated. Calibration was verified by sealing soil collars with plastic film to block soil–atmosphere exchange. One-day monitoring showed fluxes near zero (±0.02 μmol·s⁻¹·m⁻²), confirming stable baseline performance under desert conditions.

As shown in Figure 1a, soils were collected from the Tarim River floodplain at the northeastern margin of the Taklamakan Desert (41°35′23″ N, 84°44′34″ E; Xinjiang, China). The surface is dominated by saline–alkali soils with a thick salt crust. Surface soils (0–20 cm, including crust) were excavated (123 kg total), homogenized, air-dried, and divided into 12 portions (60 × 45 × 20 cm boxes). Four were used for a pre-incubation homogeneity test, which showed stable fluxes across replicates (±0.02 μmol·s⁻¹·m⁻²).

Two climate-controlled chambers simulated (i) ambient conditions (400 ppm CO₂) and (ii) warming conditions (+2 °C, ~650 ppm CO₂; Figure 1b,c). The specific settings for air temperature and CO₂ concentration in the chambers at each experimental stage are provided in

Table 3. Settings of air temperature and CO₂ concentration in the controlled-environment chamber at each experimental stage.

Time		Synchrony Test	Ambient Environment	Climate Warming
Temperature (°C)	1:00	23.6	23.6	25.6
	4:00	22.4	22.4	24.4
	7:00	25.8	25.8	27.8
	10:00	32.5	32.5	34.5
	13:00	33.8	33.8	35.8
	16:00	34.7	34.7	36.7
	19:00	30.9	30.9	32.9
	22:00	27.4	27.4	29.4
CO2 concentration (ppm)		400	400	650

. Elevated CO₂ was applied alongside temperature to emulate coupled future scenarios. Soil collars were inserted 10 cm deep in incubation boxes, and after 24 h stabilization, the LI-8100A system was mounted sequentially across samples to capture multiple diurnal cycles.

The remaining eight samples were distributed equally between chambers (labeled C1–C4 and W1–W4,

Table 4. The basic properties and water addition of saline alkali soil in ordinary environments and simulated climate-warming conditions.

Stage	No.	Initial VWC3cm (m3 m−3)	Initial T0cm (°C)	Add Water (mL)
Ambient conditions	C1	0.02	24.53	0
	C2	0.11	24.60	500
	C3	0.27	24.79	1000
	C4	0.91	24.80	1500
simulated climate-warming conditions	W1	0.00	26.15	0
	W2	0.09	26.06	500
	W3	0.23	26.13	1000
	W4	0.83	26.01	1500

). Within each chamber, Samples 2–4 received 500, 1000, and 1500 mL of distilled water, respectively, while Sample 1 served as the dry control [38]. From 22 November to 7 December, CO₂ fluxes were recorded every 30 min. CO₂ flux measurements were conducted at half-hourly intervals using an automated system. This high temporal resolution was adopted to effectively capture the diurnal dynamics of flux rates in response to rapid changes in soil temperature and moisture. Each measurement cycle consisted of a 30 s pre-purge with ambient air to completely evacuate the chamber headspace, followed by a 2 min closure period for flux calculation. The 30 min interval between measurements ensured minimal disturbance to the in-chamber environment while providing a dense time series for analyzing temporal patterns. Simultaneously, surface temperature (0 cm) was measured using sensors (Model 109, Campbell Scientific, Logan, UT, USA), and soil moisture at 3 and 10 cm depths was monitored with probes (Hydra 93640, Stevens Inc., Portland, OR, USA).

These measurements enabled systematic analysis of how soil temperature and moisture regulate CO₂ fluxes in saline–alkali soils under simulated climate warming. We note that the short-term laboratory setting does not fully capture microbial or root-driven processes typical of field environments.

2.4. Data Processing

Data quality control and correction of aberrant values for CO₂ flux measurements were performed using the LI8100 File Viewer 2.0 software. All statistical analyses were conducted on the Origin 2021 and IBM SPSS Statistics 29.0 platforms, generating corresponding statistical graphs. Linear regression analysis was employed to comparatively analyze the quantitative relationships between soil temperature, moisture, and CO₂ flux under simulated climate-warming conditions.

Subsequently, the differences in CO₂ flux between saline–alkaline soils at varying moisture gradients and those under completely dry conditions were quantified. Ultimately, a CO₂ flux estimation scheme for saline–alkaline soils, incorporating moisture–temperature interactions and regulatory factors, was developed. Based on our findings, we propose that the CO₂ flux (F) in saline–alkaline soils at any given moisture level consists of two components: (1) the component driven solely by soil temperature under completely dry conditions (fdry), and (2) the component jointly regulated by moisture–temperature factors (f(T₀cm, VWC₃cm)):

F = F_dry − F(T_0cm, VWC_3cm)

(1)

F(T_0cm, VWC_3cm) = a + bVWC_3cm + cT_0cm + dVWC_3cm·T_0cm

(2)

To isolate the moisture–temperature interactions component, f(T₀cm, VWC₃cm), we analyzed the difference in CO₂ flux between each moist sample and the dry baseline (C₁ − C₂, C₁ − C₃, C₁ − C₄; W₁ − W₂, W₁ − W₃, and W₁ − W₄) in relation to the corresponding soil temperature and moisture. A unified equation for f(T₀cm, VWC₃cm) was then established through nonlinear surface fitting.

3. Results

3.1. Fluctuation of CO₂ Flux at Different Experimental Stages

During the consistency test (Figure 2), CO₂ fluxes of the four soil samples showed a stable diurnal pattern, ranging from −0.56 to 1.24 μmol·s⁻¹·m⁻², confirming high reproducibility among replicates. After applying moisture gradients, fluxes under both ambient and warming conditions retained this diurnal pattern (Figure 3). Under ambient conditions, the amplitude ranged from −0.51 to 0.70 μmol·s⁻¹·m⁻², whereas under warming conditions it increased to −0.87 to 1.10 μmol·s⁻¹·m⁻². In the simulated climate warming environment, the CO₂ flux of dry saline–alkali soils fluctuated more strongly (−0.87 to 1.10 μmol·s⁻¹·m⁻²) compared with watered samples. The daily amplitude was greater under warming than under ambient conditions, while increasing soil moisture gradually suppressed flux amplitude. These results demonstrate that diurnal temperature variation primarily drives flux dynamics, with soil moisture acting as a modulator of flux intensity.

3.2. Warming Reshapes Moisture–Temperature Interactions in Saline–Alkali Soils

Analysis of soil surface temperature and CO₂ flux revealed a significant positive linear correlation. Figure 4a,b illustrates the relationships for samples 1–4, with the key statistical values for the CO₂ flux across these samples provided in

Table 5. Key statistical values (slopes, intercepts, and p-values) for CO₂ flux across samples.

Stage	No.	Slopes	Intercepts	R2	P
Ambient conditions	C1 = 0.098T0cm − 2.55	0.02	24.53	0	8.72
	C2 = 0.073T0cm − 1.90	0.11	24.60	500	8.76
	C3 = 0.075T0cm − 1.94	0.27	24.79	1000	8.88
	C4 = 0.039T0cm − 1.02	0.91	24.80	1500	8.96
simulated climate-warming conditions	W1 = 0.14T0cm − 4.01	0.00	26.15	0	8.70
	W2 = 0.11T0cm − 3.03	0.09	26.06	500	8.91
	W3 = 0.10T0cm − 2.89	0.23	26.13	1000	8.83
	W4 = 0.058T0cm − 1.61	0.83	26.01	1500	8.70

. Quantitatively, under ambient conditions (Figure 4a), the temperature sensitivity (regression slope) decreased from 0.098 for sample 1 (dry) to 0.039 for sample 4, representing a reduction of 60%. Under warming conditions (Figure 4b), sensitivity decreased from 0.14 to 0.058, a reduction of 59%. A direct comparison of the dry soils showed that warming significantly increased the baseline temperature sensitivity by 43% (from 0.098 to 0.14; p < 0.01, ANCOVA). Under ambient conditions, a significant negative linear relationship existed between daily-scale soil moisture and CO₂ flux (R² > 0.58, p < 0.001; Figure 5b). In contrast, no significant correlation between moisture and flux was observed at any timescale under warming conditions.

3.3. Estimation of CO₂ Flux Estimation Scheme for Saline Alkali Soil

The consistent reduction in temperature sensitivity under moisture influence—evidenced by a 59–60% decrease in slope across moisture gradients—indicates a dominance of physical buffering mechanisms, including increased heat capacity, reduced gas diffusion, and dilution effects, in regulating CO₂ flux. This behavior contrasts with biotically driven processes that typically display optimal moisture thresholds.

A complete CO₂ flux estimation scheme was developed by integrating fdry and f(T₀cm, VWC₃cm) (

Table 6. Significant estimation schemes (p < 0.001) for CO₂ fluxes in saline–alkali soils under ambient and simulated climate-warming conditions.

Stage	Number	Equation	R2	RMSE	AIC
Ambient conditions	C1	0.098 T0cm − 2.55	0.93
	C(VWC3cm, T0cm)	−0.23 − 1.70 VWC3cm + 0.009 T0cm + 0.065 VWC3cm · T0cm	0.94
	FC = C1 − C(VWC3cm, T0cm)	−2.32 + 1.70 VWC3cm + 0.084 T0cm − 0.065 VWC3cm · T0cm	0.594	0.186	−1930.514
Simulated climate-warming conditions	W1	0.14 T0cm − 4.01	0.88
	W(VWC3cm, T0cm)	−0.39 − 2.63 VWC3cm + 0.012 T0cm + 0.091 VWC3cm · T0cm	0.87
	FW = W1 − W(VWC3cm, T0cm)	−3.62 + 2.63 VWC3cm + 0.128 T0cm − 0.091 VWC3cm · T0cm	0.610	0.269	1503.849

). The model achieved high accuracy under ambient conditions (R² > 0.94, p < 0.001). Under warming conditions, however, explanatory power decreased (R² > 0.87, p < 0.001), accompanied by greater data dispersion, especially at higher moisture levels (Figure 6b,d). All regulatory coefficients (for T₀cm, VWC₃cm, and their interaction) increased significantly under warming.

Model performance was evaluated using R², RMSE, and AIC with a 70%/30% training/testing split. The Ambient model explained 59.4% of the variance in CO₂ flux (R² = 0.594, RMSE = 0.186, AIC = −1930.514), while the Warming model showed a slightly improved fit but higher prediction error and complexity (R² = 0.610, RMSE = 0.269, AIC = −1503.849).

4. Discussion

Unlike the well-studied biotic fluxes of forests and grasslands, this study focuses on the abiotic carbonate-driven fluxes in saline–alkali soils, which are a critical but underrepresented component of the global carbon cycle. Across two climate-controlled chambers, saline–alkali soils exhibited a robust diurnal, bidirectional CO₂ flux tightly synchronized with surface soil temperature (T₀cm); daytime warming enhanced efflux, whereas nocturnal cooling enhanced uptake (Figure 2 and Figure 3). During the homogeneity test, fluxes ranged from −0.56 to 1.24 μmol·s⁻¹·m⁻², providing a consistent baseline for subsequent treatments. Along the full moisture gradient (VWC ≈ 0.02–0.91 m³·m⁻³), increasing soil moisture consistently dampened diurnal amplitude—by up to ~61%—and reduced temperature sensitivity, with slopes declining from 0.098 to 0.039 under ambient conditions and from 0.14 to 0.058 under warming (Figure 4 and Figure 5). At the daily scale, moisture and flux were significantly negatively correlated under ambient conditions (R² > 0.58, p < 0.001), but this relationship disappeared under warming, suggesting a reorganization of moisture–temperature interactions. Simulated warming (+2 °C, 650 ppm CO₂) further elevated temperature sensitivity across all moisture levels and expanded the diurnal amplitude by ~37–75%. Based on these patterns, we developed a two-term estimation scheme consisting of a dry, temperature-only component and a joint moisture–temperature interactions component f(T₀cm, VWC₃cm). The scheme performed well under ambient conditions (R² ≈ 0.94) but less so under warming (R² ≈ 0.87), with systematic coefficient shifts (intercept from −2.32 to −3.62; moisture coefficient from +1.70 to +2.63; temperature coefficient from +0.084 to +0.128; interaction term from −0.065 to −0.091), pointing to the involvement of additional physicochemical processes under warmer conditions (Table 5).

Placing these results in a broader context, CO₂ fluxes vary considerably across ecosystems due to differences in soil properties, vegetation cover, and climatic conditions. The flux range observed here was lower than those of forests (3.13–3.45 μmol·s⁻¹·m⁻²) [39,40], grasslands (1.06–1.59 μmol·s⁻¹·m⁻²) [33,41,42], and croplands (1.30–4.32 μmol·s⁻¹·m⁻²) [43,44], but higher than that of desert mobile sands (−0.10 to 0.40 μmol·s⁻¹·m⁻²) [45]. Although smaller in magnitude, saline–alkali soils exhibited a distinct diurnal bidirectional flux pattern, similar to reports from other arid deserts such as the Taklamakan and Gurbantunggut, where fluxes are governed largely by inorganic carbonate processes. Recent work by Schlesinger (2017) emphasized the importance of abiotic sinks in arid desert ecosystems, challenging previous notions of them being a “dead pool” in the global carbon cycle [34,46]. This finding is consistent with observations from the Mojave and Chihuahuan deserts, where inorganic carbon fluxes are primarily influenced by carbonate dissolution processes [18,24,34]. In contrast, forests, grasslands, and croplands—dominated by microbial and root respiration—show consistently positive fluxes, with soil moisture effects typically displaying a nonlinear “low–optimum–high” threshold response. In saline–alkali soils, however, moisture exerted a monotonic damping effect, reducing both diurnal amplitude and temperature sensitivity [46,47]. Recent studies further corroborate these findings, showing that in arid regions, inorganic carbon dynamics are more sensitive to abiotic factors than to biological processes [20,48]. This divergence highlights the fundamentally different mechanisms regulating fluxes in abiotic versus biotic systems, further supporting the hypothesis that saline–alkali soils act as significant but understudied carbon sinks. These findings align with global studies that have confirmed SIC as a significant carbon sink in arid regions [7,25]. Our study contributes to this growing body of work by demonstrating the importance of abiotic processes in regulating SIC dynamics, with implications for climate change mitigation strategies in dryland ecosystems [49].

Mechanistically, the observed patterns can be explained by the carbonate equilibrium system [50]:

CaCO₃ + CO₂ + H₂O ⇌ Ca²⁺ + 2HCO₃⁻

(3)

The equilibrium constant for this reaction (K_eq) is inherently temperature dependent, as described by the van’t Hoff equation:

ln(K_eq) = −ΔH/(R × T) + ΔS/R

(4)

where ΔH is the enthalpy change, R is the gas constant, T is the absolute temperature, and ΔS is the entropy change. This reversible reaction links gaseous CO₂, dissolved inorganic carbon (DIC), and solid carbonate minerals. Temperature thus acts as the dominant driver: daytime warming decreases Keq, reducing the solubility of CO₂ and CaCO₃, shifting the equilibrium leftward, and promoting carbonate precipitation and CO₂ efflux. Conversely, nocturnal cooling increases Keq, enhances solubility, shifts the equilibrium rightward, and favors dissolution with CO₂ uptake. This thermodynamic dependence directly explains the diurnal bidirectional fluxes observed in our study.

Moisture exerts a complementary influence by regulating flux amplitude through several pathways [48,51,52]. The progress of the reaction is conditioned by the availability of water and dissolved CO₂. The solubility of CO₂ in water ([CO₂(aq)]) is itself a function of temperature and ionic strength (I), expressed as follows:

[CO₂(aq)] = K_H × P_CO₂ × exp(k × I),

(5)

where K_H is Henry’s constant (temperature dependent), P_CO₂ is the partial pressure of CO₂, and k is a solute-specific parameter. Higher water content thickens pore-water films and restricts gas diffusion, increases soil heat capacity and buffers diurnal temperature fluctuations, and alters ionic strength and pH, thereby modifying carbonate equilibria. Collectively, these effects explain the monotonic damping of flux amplitude and temperature sensitivity across the VWC gradient [53]. Warming further reshapes this system. Elevated temperature and CO₂ not only increase temperature sensitivity but also amplify the temperature–moisture interaction. Enhanced evaporation raises ionic strength, which reduces CO₂ solubility via the Setschenow salting-out effect (ks), while surface salt crusts may physically hinder diffusion and alter subsurface equilibria [54]. These coupled processes explain the systematic coefficient shifts in our estimation model.

Nevertheless, some limitations must be acknowledged. This study relied on short-term, controlled chamber experiments designed to isolate moisture–temperature interaction effects, which by design excluded vegetation and intact microbial communities. As such, plant-mediated fluxes, microbial feedback, and the influence of natural soil structure and near-surface turbulence were not represented. The flux magnitudes and variability reported here should therefore be interpreted as process-level responses rather than ecosystem-level rates. To ensure ecological validity and transferability, long-term field validation is required, ideally through in situ warming experiments, automated chambers, and eddy-covariance observations. Where feasible, isotopic partitioning (e.g., δ¹³C) combined with measurements of salinity and ionic strength would help separate inorganic and organic contributions and test the robustness of the proposed moisture–temperature interactions scheme.

Taken together, our results show that saline–alkali soils exhibit robust diurnal bidirectionality and strong moisture–temperature interactions control of CO₂ exchange, with moisture exerting a monotonic damping effect and warming amplifying temperature sensitivity and the T×VWC interaction. These responses indicate that abiotic carbonate equilibria and transport processes must be explicitly represented in models; otherwise, land-sector carbon inventories will be biased. In particular, our warming experiments imply that the sink potential may decline by ~50–60%, which could exacerbate the “missing sink” gap in regional and global budgets. In the Paris Agreement context—where Parties set NDCs and report AFOLU/LULUCF fluxes—saline–alkali soils should not be treated as uniformly weak sources; rather, their source–sink balance depends on moisture–temperature interactions and salinity and may shift toward weaker sinks as warming enhances evaporation, ionic strength, and salt-crust formation. Although per-unit fluxes are lower than those of forests or croplands, the large areal extent of saline–alkali lands means their aggregate impact could be significant. Priorities, therefore, include field validation under ambient variability (in situ warming, eddy covariance, automated chambers with isotopic partitioning) and improved MRV that parameterizes carbonate equilibria (temperature, moisture, salinity/ionic strength, pH). Practically, Paris-aligned saline-land management—e.g., optimized irrigation/drainage, targeted salinity control (e.g., Ca²⁺ amendments where appropriate), and surface mulching or roughness management to limit midday efflux spikes—can help stabilize carbonate equilibria, avoid inadvertent CO₂ releases, tighten inventories, and reduce uncertainty for the Global Stocktake.

5. Conclusions

This study shows that saline–alkali soils in arid regions display strong diurnal bidirectionality of CO₂ fluxes, driven primarily by surface temperature, with soil moisture acting as a monotonic dampening factor. Warming (+2 °C, 650 ppm CO₂) amplified flux amplitude by 37–75%, increased temperature sensitivity slopes by43% (from 0.098 to 0.14), and reduced sink potential by ~50–60%, suggesting a possible shift from net sinks to sources and aggravating the “missing sink” gap in carbon budgets. Chamber experiments clarified moisture–temperature interactions but excluded vegetation and microbial processes, so the results represent process-level rather than ecosystem-scale fluxes, highlighting the need for long-term field validation with isotopic tracing, automated chambers, and eddy covariance. These findings underscore the importance of explicitly representing abiotic carbonate equilibria and moisture–temperature interactions in carbon models; future research should prioritize multiphase models that capture coupled processes and salinity effects. Embedding such schemes in Earth system models will refine land-sector inventories, improve global carbon projections, and support Paris Agreement–aligned saline-land management strategies, including irrigation and drainage optimization, targeted salinity control, and surface mulching in drylands.