Effects of Artificial Green Land on Land–Atmosphere Interactions in the Taklamakan Desert

Abstract

Land–atmosphere interactions are influenced by the earth’s complex underlying subsurface, which in turn indirectly affects atmospheric motion and climate change. Human activities are increasingly exerting an influence on desert ecosystems, and artificial green land with clear functional orientation has been established in many desert areas. Consequently, the previously dominant, shifting, sand-covered, underlying surface in these desert regions is gradually transforming. This transformation has significant implications for the characteristics of land–atmosphere interactions, causing them to deviate from their original state. At present, existing studies still have not presented a systematic understanding of this change and have ignored the impact of human activities on land–atmosphere interactions in artificial green land. To address these research gaps, this study specifically targets artificial green land in the Tazhong region of Taklamakan Desert. We carried out observation experiments on land–atmosphere interactions in three different functional units from outside to inside: natural shifting sands, the shelter forest, and the living area. We also analyzed the differences and attribution of land–atmosphere interactions characteristics of different functional units. Compared with the natural shifting sands, the daily average maximum values of wind speed in the shelter forest decreased by 78%, and the daily average maximum air temperature and soil (0 cm) temperature decreased by 2.6 °C and 7 °C, respectively. Additionally, the soil moisture level was significantly increased throughout the green land due to the shelter forest. The surface albedo experienced a decrease, with an annual average of 0.21. Furthermore, the aerodynamic roughness and bulk transport coefficient increased by two orders of magnitude. The daily average maximum values of sensible heat flux and soil heat flux (G₀₅) decreased by 18.7% and 75%, respectively, and the daily average maximum value of latent heat flux increased by 70.3%. This effectively improved the microclimate environment of the green land. The living area was greatly reduced by the shelter forest coverage and influenced by the buildings. Consequently, the environmental improvement was not as large as it was inside the shelter forest. However, it still provided a good shelter for production and living in the desert area. Throughout the year, a total of 4.60 × 10⁵ t water was consumed through evapotranspiration in the artificial green land. The findings of this study have the potential to enhance our comprehension of land–atmosphere interactions in desert regions, thereby offering valuable insights for the establishment and effective management of artificial desert green lands.

1. Introduction

The intricate exchange of energy, matter, and momentum between the land and atmosphere profoundly affects global atmospheric circulation as well as the climate system. This dynamic process serves as a pivotal link in the formation of extreme weather events and climate change. This exchange can be influenced not only by atmospheric circulation and solar radiation, but also by the complex subsurface of the land [1]. Hence, gaining a deeper understanding of the characteristics of radiation flux and surface energy flux within diverse regions and underlying surfaces is crucial. This knowledge not only aids in comprehending the interplay between various underlying surface materials, energy, momentum, and atmosphere, but also holds great significance for enhancing the accuracy and predictive power of climate models for the underlying surface area. To date, many countries around the world have conducted extensive field observation experiments on land surface processes in diverse representative regions worldwide. These experiments have yielded remarkable data and results, advancing the research on land–atmosphere interactions.

Deserts are the largest terrestrial ecosystem on Earth, covering approximately 21% of the global land area. These areas possess distinct characteristics, including arid climates, a high surface albedo, a special sandy substratum, and frequent dusty weather. These characteristics contribute to the unique ways deserts respond to climate change (including warming, cooling, wetting, or drying) [2]. With the continuous attention and investment in various aspects of desert regions in recent years, research on desert land surface processes has been advancing. The surface energy exchange characteristics of desert subsurfaces are significantly different compared to those of other regions, where a sensible heat flux is the main form of energy exchange in desert regions and the main source of energy, driving local atmospheric circulation [3,4,5,6]. The surface albedo is an important parameter of land surface processes in desert areas and is strongly influenced by the solar altitude angle, precipitation, and subsurface properties. Meanwhile, due to the prolonged extreme aridity of desert subsurface, their characteristics remain relatively stable throughout the year. Therefore, in parametric studies, the solar altitude angle is one of the important covariates in the empirical equation of surface albedo [7,8,9]. Roughness is a physical parameter characterizing the smoothness of the surface and is closely related to the morphological distribution and spatial distribution of subsurface roughness elements as well as atmospheric factors. The surface dynamics roughness in desert areas generally has minor seasonal variation and is closely related to wind speed [10,11,12,13]. Physically, the surface thermodynamic roughness z_0h is the height of the surface thermal transport impedance, and the thermal transport added damping KB⁻¹ is usually used to describe z_0h. It has been shown that the thermal transport added damping KB⁻¹ is not a fixed value, and there are significant daily variations and negative values of KB⁻¹ for bare ground and desert areas of the Gobi [14,15]. The bulk transport coefficient reflects the ability of material and energy exchange between the land and atmosphere. Numerous studies have demonstrated that the variation in this coefficient mainly depends on the surface roughness and atmospheric stability. Additionally, the wind speed affects the bulk transport coefficient by influencing the atmospheric stability [16,17,18].

In recent years, the pursuit of economic growth has intensified the competition between human societies, natural resources, and the environment. As a result, there has been a noticeable trend of reclaiming and utilizing previously unproductive land in desert areas. For example, activities like oil and gas extraction face numerous challenges such as extreme temperatures, sandstorms, and drought. To safeguard these operations, a large number of sand barriers have been artificially set up in desert areas, and shelter forest belts have been planted using artificial drip irrigation. These interventions aim to develop functional green spaces that provide both external protection and support internal production and living conditions. The original single surface with continuous shifting sands as the main cover type gradually changes, which leads to the variation in the original land–atmosphere interactions characteristics. How does the establishment of an artificial shelter forest affect the characteristics of land–atmosphere interactions with shifting sands cover as the subsurface, and how does it change when human activities are superimposed? Some scholars have paid attention to this issue. Aili et al. [19] noted that the relative humidity and latent heat flux of the near ground layer in artificial green land increased significantly under the influence of the desert shelter forest belt. This effect resulted in reduced air and soil temperatures, as well as a decrease in the sensible heat flux. Consequently, the hydrothermal conditions of the original natural shifting sands land were altered. The ground surface is affected by vegetation cover, which increases the wetness of the ground, reduces the reflectivity of the ground surface, increases the absorption of surface radiation, and changes the original surface radiation balance. Compared with the natural shifting sands outside the shelter forest, artificial green land increases the surface roughness, reduces the wind speed at the surface, and promotes water and soil conservation during the growing season. These functions play a positive role in preventing and controlling sand movement, as well as improving the microclimate environment [20,21]. In addition, the establishment of artificial green land has led to significant changes in the microbial structure of desert soils, while having an impact on the carbon and nitrogen cycles [22,23]. However, in previous analytical studies, researchers have tended to simplify the whole green land as consisting solely of plants, while ignoring the influence of human activities on artificial green land. As production operations continue to expand in desert areas, the emergence of a large number of artificial functional green land regions, combined with increasing population density and diversified production activities, is having a profound impact on the deserts and surrounding regions. Therefore, the anthropogenic influence on land–atmosphere interactions can no longer be overlooked or simplified.

In this study, we use the Taklamakan Desert hinterland oil production artificial green land region as a research example. We conduct simultaneous observations of land–atmosphere interactions fluxes in the peripheral shifting sand coverage area and the representative areas of different substrates inside the artificial green land. Based on the comparative analysis of microclimate differences across three functional units (the natural shifting sands, shelter forest, and living area), we reveal the characteristics of land surface processes and key land surface parameters and surface fluxes of each functional unit. The differences in land surface processes, key land surface parameters, and surface fluxes of each functional unit are revealed, and the reasons for the differences are discussed. The findings of this study are expected to enhance the understanding of the variability mechanisms of land–atmosphere interactions in the desert after anthropogenic modification and to provide reference and guidance for the establishment and scientific management of desert artificial green land and further development and utilization.

2. Materials and Methods

2.1. Site Description

The Taklamakan Desert, situated in the central part of the Tarim Basin in southern Xinjiang, spans a total area of 33.76 × 10⁴ km². Shifting sands cover approximately 70% of the desert, making it the second largest shifting desert globally. The National Observation and Research Station of Desert Meteorology, Taklimakan Desert of Xinjiang (38°58′51″ N, 83°38′28″ E, 1099 m above sea level), is located in the Taklamakan Desert hinterland in the Tazhong region. It holds the distinction of being the only desert observation station worldwide that extends over 200 km into the hinterland of a shifting desert. This station is widely regarded as the most representative and comprehensive observation experiment station for studying the Taklamakan Desert. It has a warm temperate arid desert climate, with an average annual temperature of 12.1 °C in the desert hinterland, with the highest recorded temperature reaching 40.0–46.0 °C and the lowest temperature reaching from −28.0 to −32.6 °C. Precipitation mainly occurs between May and August, with an annual average precipitation of 25.9 mm and an annual average potential evaporation of 3812.3 mm. The annual average wind speed is 2.3 m·s⁻¹, with easterly winds prevailing throughout the year. In the desert, tall composite longitudinal sand monopolies and intermonopoly lands are distributed among each other. The desert soil particles are light in texture, very fine in particle size, highly mobile, and made of very fine sand. This provides the ideal environmental conditions for the occurrence of sand and dust storms. According to the statistics, the study area has an annual average of 11 days of windy weather, more than 157 days of floating dust and sandy weather, and an average of 16 days of dust storm weather [24,25,26]. The harsh environmental conditions have resulted in almost no natural flora and fauna being distributed throughout the area. Since the 1990s, in order to support Tarim oil exploitation, artificial greenery has been established and maintained in Tazhong, located in the hinterland of the Taklamakan Desert. This has been achieved using artificial drip irrigation between two large sand dunes running northeast–southwest. The greenery consists of Tamarix ramosissima, Haloxylon ammodendron, and Calligonum arborescens Litv, and Phragmites communis. Over time, an area of 3.6 km² has been transformed into an artificial green land. It serves a distinct purpose of providing external protection and facilitating internal production and living. The vegetation within this area reached an average height of 1.5 m.

2.2. Experimental Design

The establishment of a shelter forest affects the characteristics of land–atmosphere exchange, which were previously dominated by continuous shifting sands. In addition, as production operations expand, population density increases, and production activities diversify, the contribution of anthropogenic activities to greenfield fluxes can no longer be overlooked or simplified. Therefore, by the end of 2017, we utilized unmanned aerial vehicle aerial photography to precisely categorize the green land into distinct functional areas based on the subsurface characteristics. We deployed a terrestrial air flux observation test system in three key locations: the central living area within the artificial green land in Tazhong, the representative area of the surrounding shelter forest, and the natural flowing sand surface located 2.2 km west of the artificial green land (Figure 1).

Table 1. Observation instrumentation and parameters.

Observation System	Observation Variables		Sensors	Installation Instructions	Observation Frequency
IRGASON	Three-dimensional wind speed and direction, Ta, H, LE, CO2 Flux, u*		IRGASON (Campbell Scientific Inc./Logan, UT, USA)	The equipment is installed at 3 m above the surface of the shifting sand land and shelter forestland. In addition, an equipment set is installed at 10 m on the 80 m meteorological tower located in the living area.	10 Hz
Surface solar radiation monitoring	Total solar radiation		CNR4 (Kipp & Zonen B.V. /Delft, The Netherlands)	The equipment is installed at 1.5 m above the surface of three sites.	1 Hz
	Atmospheric longwave radiation
	Ground-reflected radiation
	Ground longwave radiation
Soil detection instruments		Soil temperature	109 (Campbell Scientific Inc. /USA)	The equipment is installed at three stations at a depth of 0, 5, 10, 20, and 40 cm.	1 Hz
		Soil moisture	93640 Hydra (stevens Inc. /Teutopolis, IL, USA)	The equipment is installed at three stations at a depth of 5, 10, 20, and 40 cm.
		Soil heat flux	HFP01SC (Hukseflux Inc. /Delft, The Netherlands)	The equipment is installed at three stations at a depth of 5 and 10 cm.

u_*: represented as frictional wind speed.

presents the observation instrumentation and parameters. Each observation system is composed of an open-circuit eddy correlation, four-component radiometer, multilayered soil temperature and humidity and heat flux detection system. To ensure the accuracy of data acquisition and representativeness of the observed area, each observation system underwent calibration, and the thorough analysis of the source area was conducted prior to installation.

Synchronized observations from the above three stations from January to December 2018 were selected to first compare and analyze the microclimate differences between different functional units at the same temporal resolution. On this basis, we analyzed the key physical parameters of land–atmosphere interactions (including the surface albedo, aerodynamic roughness, thermodynamic roughness, and bulk transport coefficient). We also examined the variations in land–atmosphere energy distribution across each functional unit. Through this comprehensive investigation, we were able to identify the primary factors that contribute to the disparities between each unit, thereby revealing the extent of anthropogenic modification of the desert substratum. We then analyzed the influence of artificial green land with clear functional direction on the land–atmosphere interactions characteristics of the surrounding environment.

2.3. Data Processing and Methods

2.3.1. Quality Control of Conventional Meteorological Observations Data

Conventional meteorological observations are subjected to threshold testing (RT) and climate threshold testing (CRT) in order to ensure that the data are within the theoretical range and in accordance with the local climatic characteristics before they are used [27]. Secondly, we excluded anomalies that exhibited excessively large increases or decreases, as well as those that showed no change within a specific time period. Furthermore, adhering to the principles of meteorology, we examined closely related meteorological elements to determine whether they followed a specific pattern of change or not [27]. In this paper, we will utilize the aforementioned methods and refer to the quality control means of the meteorological observation data in the hinterland of Taklamakan Desert by Liu et al. [28]. This will allow us to ensure the quality of the routine meteorological observation data obtained from three distinct power units.

2.3.2. Micrometeorological Conditions

We used SPSS 27.0.1 software for analysis of variance (Anova) to investigate the differences in wind speed, air temperature, soil temperature (0 cm), and soil moisture (5 cm) among three different functional units. Based on the given significance level (p > 0.05 is not significant, 0.01 < p < 0.05 is significant, p < 0.01 is extremely significant, and p > 0.05 holds no statistical significance), we then determined the significance of the differences in micro meteorological elements among the three different functional units.

2.3.3. Surface Albedo

(1) The surface albedo (α) reflects the ability of the surface to reflect solar radiation [29]. It can be obtained from the shortwave radiation measured using the Radiation Observation System with the following equation:

$$\alpha =\frac{S_u}{S_d}$$

(1)

where S_u is the reflected shortwave radiation from the surface (unit: W·m⁻²), and S_d is the downwards shortwave radiation (unit: W·m⁻²). Considering that the lower solar altitude angle at sunrise and sunset affects the calculation of surface albedo, the observed data of shortwave radiation 30 min after sunrise and 30 min before sunset are ignored in the calculation of surface albedo, and exclude observations during rainfall periods. In addition, to avoid greater errors caused by nonclear sky periods when analyzing surface albedo changes at different solar altitude angles, the clear sky index (k_t) was used to distinguish between different sky conditions, i.e., k_t ≥ 0.6 for clear sky, 0.6 > k_t ≥ 0.3 for cloudy sky, and k_t ≤ 0.3 for cloudy sky [9,30], and to exclude non-clear sky values of surface albedo during the time period. The clear sky index was calculated using the following formula [31]:

$$k_t=\frac{S_d}{S_e}$$

(2)

$$S_e=S_{sc}[1+0.033\mathit{cos}\left(360t_d/365\right)]\mathit{sin}\beta$$

(3)

where S_e is the astronomical radiation; S_sc is the solar constant (1367 W·m⁻²); t_d is the day sequence number; and β is the solar altitude angle.

(2) In addition, we used the Pearson correlation coefficient to test the relationship between the surface albedo of three different functional units and solar altitude angle, soil moisture (5 cm), and vegetation coverage. The formula is as follows:

$$R_{XY}=\frac{{{\displaystyle \sum }}_{i=1}^n\left(X_i-\overline{X}\right)\left(Y_i-\overline{Y}\right)}{\sqrt{{{\displaystyle \sum }}_{i=1}^n{\left(X_i-\overline{Y}\right)}^2{{\displaystyle \sum }}_{i=1}^n{\left(Y_i-\overline{Y}\right)}^2}}$$

(4)

In the formula, n is the total number; R_XY is the correlation coefficient between variables X_i and Y_i (i = 1, 2,..., n); X_i and Y_i are the i-th values of each variable, respectively.

2.3.4. Roughness

(1) Aerodynamic roughness

According to the Monin–Obukhov similarity theory, the aerodynamic roughness z_0m is formulated as follows [13]:

$$ln{z}_{0m}=\ln \left(z-d\right)-\frac{ku}{u_{*}}-{\phi }_m\left[\frac{z-d}{L}\right]$$

(5)

where φ_m is the stability correction function for kinetic energy, which was calculated using Equation (5) for neutral atmospheric stability conditions with a zero value, unstable conditions, and stable conditions [13]:

$$\left\{\begin{array}{ll}{\phi }_m=2\mathit{ln}[\left(1+x)/2\right]+\mathit{ln}[\left(1+x^2\right)/2]-2{\mathit{tan}}^{-1}x+\pi /2& \frac{z-d}{L}<0\\ {\phi }_m=-5.5\left[\frac{z-d}{L}\right],& \frac{z-d}{L}\ge 0\end{array}\right.$$

(6)

where x = [(1 − 12(z − d)/L]^1/4, L is the Monin–Obukhov rough length, d is the zero-plane displacement height (m), z is the observed height (m), u is the average wind speed at z (m) (m·s⁻¹), u_* is the friction velocity (m·s⁻¹), and k is the Karman constant, which takes the value of 0.4 in this paper. Data for wind speeds less than 1 m·s⁻¹, friction roughness less than 0.01 m·s⁻¹, and rainfall periods were excluded [32].

(2) Thermodynamic roughness

Thermodynamic roughness (z_0h) is the height at which the air temperature equals the surface temperature when the near-surface meteorological conditions satisfy the Monin–Obukhov similarity theory and is usually described by KB⁻¹ using the additional damping of heat transport, which is calculated as follows [33]:

$$K{B}^{-1}=\frac{{\rho }_a{\mathrm{C}}_{\rho }\left(T_s-T_a\right)}{H}k{u}_{*}-\ln \left(\frac{z-d}{z_{0m}}\right)+{\phi }_h\left(\frac{z-d}{L}\right)$$

(7)

where the values of φ_h for the unstable and stable conditions are obtained using Equation (7), respectively:

$$\left\{\begin{array}{ll}{\phi }_h=2\mathit{ln}[\left(1+x^2\right)/2]& \frac{z-d}{L}<0\\ {\phi }_h=-\frac{5\left(z-d\right)}{L}& \frac{z-d}{L}\ge 0\end{array}\right.$$

(8)

where x = [1 − 16((z − d)/L)]^1/4, ρ_a and C_ρ are the air density and air specific heat at constant pressure, respectively; T_a and T_s are the air level temperature and surface temperature, respectively; H is the sensible heat flux; and H, C_ρ, ρ_a, T_a, and T_s can be obtained from the data information observed by the vortex-related system.

2.3.5. Bulk Transport Coefficient

The bulk transport coefficient includes the momentum transport coefficient (C_d), heat transport coefficient (C_h) and water vapor transport coefficient (C_e). In this paper, we study the momentum transport coefficient (C_d) and heat transport coefficient (C_h). C_d and C_h characterize the dynamic and thermal effects of turbulence, respectively, and are physical quantities that measure the strength of turbulence, which are calculated as follows [16]:

$$C_d={(u_{*}/u)}^{2 }$$

(9)

$$C_h=\frac{H}{\rho {\mathrm{C}}_{\rho }\left(T_s-T_a\right)}$$

(10)

2.3.6. Heat Flux and Latent Heat Flux

The 10 Hz observations collected from the vorticity correlation system of the three subsurface observing systems underwent several data processing steps. These steps involved the utilization of Eddy Pro 7.0.6 software to remove erroneous data points, as well as delay time correction, quadratic coordinate rotation, frequency response correction, ultrasonic false temperature correction, and density correction [34,35,36,37]. After a series of necessary processes, the flux values of sensible heat (H) and latent heat (LE) between land and air were obtained. In addition, data during precipitation, dust storms, or other high-impact weather periods were excluded to produce a more standardized 30 min observational experiment:

$$\mathrm{H}=\rho C_{\rho }\overline{W^{\prime }{T}^{\prime }}$$

(11)

$$\mathrm{LE}=L_v\overline{W^{\prime }{\rho }_v^{\prime }}$$

(12)

where ρ is the air density in kg·m⁻³, C_ρ is the air specific heat at constant pressure, w′ is the vertical wind speed pulsation value, T′ is the potential temperature pulsation value, ρv′ is the specific humidity pulsation value, and L_v is the latent heat of evaporation.

3. Results and Discussion

3.1. Micrometeorological Conditions

The interannual variation in the meteorological parameters of natural shifting sands, the shelter forest, and the living area during the observation period is shown in Figure 2. The maximum daily average wind speeds of the three functional units were 14 m·s⁻¹, 2.9 m·s⁻¹, and 5.3 m·s⁻¹, respectively. The natural shifting sands experienced unobstructed wind flow, resulting in consistently high and variable wind speeds throughout the year. The shelter forest was affected by the surface vegetation. As a result, the wind speed was significantly lower than that of the natural shifting sands. The difference between the two was highly significant, with a p-value of 0.009, which is less than the commonly accepted threshold of 0.01. Despite being situated in the center of the green land, the living area experiences different wind speeds compared to the natural shifting sands and the shelter forest. This is primarily due to the low vegetation cover and the limited number of buildings obstructing the area (living area). The wind speed in the living area is significantly lower than that of the shifting sands (p = 0.019 > 0.01), while it is higher than that of the shelter forest; although, this difference is not statistically significant (p = 0.061 > 0.05) (Figure 2a). The trends of air temperature and soil temperature (0 and 5 cm) of the three functional units were consistent, all showing high values in summer and low values in spring and winter. The air temperature and soil temperature (0 and 5 cm) in the shelter forest were slightly lower than those in the other two functional units throughout the year (

Table 2. Annual and summer average values of wind speed, air temperature, soil temperature (0 cm and 5 cm) and soil humidity (5 cm and 10 cm) for three different functional units.

	Wind Speed (m·s−1)		Air Temperature (°C)		Soil Temperature (0 cm) (°C)		Soil Temperature (5 cm) (°C)		Soil Moisture (5 cm) (m3/m3)		Soil Moisture (10 cm) (m3/m3)
	Annual	Summer	Annual	Summer	Annual	Summer	Annual	Summer	Annual	Summer	Annual	Summer
Shifting Sands	7.0	7.6	13.2	31.4	16.0	36.0	16.7	34.0	0.02	0.02	0.03	0.04
Shelter Forest	0.1	1.2	9.8	26.0	11.0	28.0	12.7	28.9	0.07	0.1	0.06	0.09
Living Area	2.0	2.5	12.5	28.9	15.0	33.1	15.6	32.5	0.01	0.01	0.03	0.03

) and were significantly different from those of the natural shifting sands (p = 0.03 < 0.05) and living area (p = 0.02 < 0.05). This is due to the fact that the shelter forest, under the influence of vegetation, reduces direct solar radiation to the surface of the unit, resulting in lower air and soil temperatures (0 and 5 cm) compared to those of the other two functional units (Figure 2b–d). In addition, due to the extreme aridity in Taklamakan Desert, except for the periods influenced by precipitation, the soil moisture in the natural shifting sands and living areas was low, and the difference was not obvious (p = 0.066 > 0.05). In contrast, the shelter forest area exhibits relatively high soil moisture levels at depths of 5 and 10 cm. This is primarily due to artificial drip irrigation activities carried out during the vegetation growing season (Figure 2e,f).

3.2. Surface Albedo

As shown in Figure 3, the daily variation in surface albedo in all three functional units showed a pattern of being high in the morning and late evening and low at midday, with seasonal fluctuations of being high in winter and low in summer. The surface albedo of the living area is similar to that of the natural shifting sands, and there was no difference between the two (p = 0.076 > 0.05), but it is significantly greater than that of the shelter forest. The annual mean values of surface albedo for the natural shifting sands, living area, and sheltered forest were 0.30, 0.31, and 0.21, respectively. In the context of the extreme aridity of the Taklamakan Desert, the albedo of the shifting sands and the living area, which are always bare, remains high. The shelter forest’s surface is covered with vegetation. Anthropogenic drip irrigation activities during the growing season increase the soil water content. As a result, the surface albedo is significantly lower compared to the natural shifting sands and living areas. This lower albedo shows some seasonal fluctuations. These findings align with the conclusions of Wang et al. (2006) and Webster Clare et al. (2018) [38,39]. When comparing the albedo values of various subsurface types, we found that alpine meadows have the highest annual mean albedo, followed by the deserts, semi-arid grasslands, and farmland (

Table 3. Monthly average variation in surface albedo in 3 different functional units and other areas [40,41].

Month	Shifting Sands	Shelter Forest	Living Area	Zhangye Desert	Zhangye Oasis Farmland	Yakou Alpine Meadows
1	0.32	0.22	0.34	0.34	0.30	0.39
2	0.32	0.22	0.34	0.31	/	0.37
3	0.30	/	0.33	0.29	0.19	0.66
4	0.29	/	0.32	0.27	0.16	0.41
5	0.27	/	0.31	0.27	0.17	0.60
6	0.28	0.19	0.28	0.26	0.16	0.29
7	0.30	0.21	0.29	0.26	0.14	0.18
8	0.30	0.20	0.30	0.26	0.13	0.24
9	0.31	0.20	0.32	0.28	0.18	0.38
10	0.31	0.21	0.32	0.30	0.18	0.45
11	0.31	0.22	0.33	0.33	0.15	0.44
12	0.32	0.22	0.35	0.37	0.23	0.30
Average	0.30	0.21	0.31	0.30	0.19	0.39

) [40,41]. This is mainly because alpine meadows are affected by long-term snow cover, which leads to a higher overall surface albedo. In the study area, the latitude difference between the shifting sands, living areas, and Zhangye Desert is not significant, and the substrate types are similar. As a result, the albedo of these three areas is similar and demonstrates higher values. Our findings reveal that the albedo of the shelter forest in this study is comparable to that of farmland and semi-arid grassland. It is important to note that wheat and maize are cultivated in the spring and summer seasons in these areas. Additionally, the vegetation cover in the shelter forest is higher compared to the grassland, which leads to a relatively lower albedo.

Based on the correlation analysis, it can be seen that the albedo is not much affected by soil moisture and vegetation cover in desert areas, while solar altitude angle is the critical influence factor (

Table 4. Correlation between surface albedo and solar altitude angle, soil moisture (5 cm), and vegetation cover among the 3 functional units.

Function Unit	Solar Altitude Angle	Soil Moisture (5 cm)	Vegetation Cover
Shifting Sands	−0.63	−0.15	/
Shelter Forest	−0.48	−0.3	−0.047
Living Area	−0.60	−0.16	/

). From the relationship between the albedo and solar altitude angle for the three different functional units (Figure 3g–i), it can be seen that the albedo was higher after sunrise and on the eve of sunset when the solar altitude angle was low and was lower for the altitude angle at noon. In addition, it was found, by comparison, that the surface albedo of natural shifting sands and shelter forest gradually became flatter when the solar altitude angle was ≥15°. However, in the living area, the surface albedo variation tended to level off only when the solar altitude angle was ≥30°. The reason for this difference is that the living area is in the intermonopoly land between two composite longitudinal sand dunes of approximately 30 m in height, and there are some buildings distributed in the living area. This results in a significant difference in the relationship between surface albedo and solar altitude angle in the living area and the other two substrates.

3.3. Aerodynamic Roughness

Figure 4a–c shows the month-to-month characteristics of the aerodynamic roughness z_0m of the three different functional units, with the annual means z_0m of the natural shifting sands, shelter forest, and the living area being 5.7 × 10⁻³ m, 1.72 × 10⁻¹ m, and 5.7 × 10⁻² m, respectively. Overall, the shelter forest z_0m is two orders of magnitude larger than that of the natural shifting sands and one order of magnitude larger than that of the living area. The natural shifting sand subsurface is mainly covered by continuous flat shifting sands, with a relatively smooth surface, minimal z_0m and gentle variations, and no seasonal variation characteristics. The results of this unit are more similar to those measured in the Dunhuang Gobi and Heihe Desert, where the subsurface is homogeneous (

Table 5. Monthly average values of aerodynamic roughness z_0m × 10⁻¹ (m) for 3 different functional units.

Month	Shifting Sands	Shelter Forest	Living Area
1	0.054	1.49	0.49
2	0.056	1.50	0.50
3	0.056	1.57	0.51
4	0.054	1.59	0.52
5	0.057	1.76	0.55
6	0.058	2.08	0.61
7	0.058	2.17	0.66
8	0.059	2.07	0.65
9	0.057	1.78	0.64
10	0.058	1.58	0.53
11	0.053	1.47	0.53
12	0.056	1.46	0.53

) [10]. In contrast, the z_0m of shelter forest and living area was in the same order of magnitude as that of the forest [42] (pp. 160–161), farmland, and alpine sparse grasslands [43,44] (

Table 6. Aerodynamic roughness in 3 different functional units and other areas z_0m × 10⁻¹ (m) [10].

z0m × 10−1	Shifting Sands	Shelter Forest	Living Area	Dunhuang Gobi	Heihe Desert	Farmland	Alpine Sparse Grass
Average	0.057	1.72	0.57	0.019	0.045	3.02	0.29

) and showed some fluctuating trends throughout the year. Among them, the shelter forest has larger z_0m values and significant seasonal variations due to high vegetation cover and vigorous vegetation growth in summer, which increases the drag of plants on air currents [12,45]. The living area has a relatively high z_0m due to the fact that it is bare ground and is influenced by the surrounding greenery and artificial buildings.

As shown in Figure 4d–f, the z_0m of each functional unit fluctuates and changes under different wind conditions. The z_0m values of the shifting sands become larger in the due north and south-west directions, which is related to the multiple small dunes distributed in these two directions. The shelter forest and the living area are situated amidst large sandy ridges that run in the north-east and south-west directions. This positioning leads to lower z_0m values in both areas along this orientation. In this case, the data for the living area are more dispersed due to the surrounding buildings and vegetation.

3.4. Thermodynamic Roughness

There are many factors that affect KB⁻¹, such as meteorological conditions, vegetation structures, surrounding obstacles, radiation, and soil surface resistance [44]. As shown in Figure 5, the daily variation in KB⁻¹ in natural shifting sands is parabolic; specifically, it is high at noon and low in the morning and evening, with negative values. This is consistent with the variation characteristics of bare ground studied by Verhoef et al. (1997) Yang et al. (2008) [14,15,46]. Comparatively, the daily variation in KB⁻¹ in the shelter forest was not significant, with KB⁻¹ being more dispersed at night and more concentrated during the day, with mean values of 6.2 and 6.7, respectively. This difference is due to the relatively high level of land–atmosphere heat transport on the exposed surface and the increased heat transport during the daytime than that at night. Therefore, the daily variation in KB⁻¹ in the shifting sands is obvious and negative in the morning and evening, indicating that the heat transport efficiency can exceed the momentum transport in the morning and evening. In contrast, due to the vegetation cover on the surface of the shelter forest, dense vegetation increases the impedance of energy transfer between the land and atmosphere, while weakening the difference of energy gradient between the land and atmosphere. Consequently, the daily change of KB⁻¹ in the shelter forest is less noticeable and more concentrated in the daytime compared to that of the shifting sands. This change characteristic was also found in the observation of the grassland [46]. For the living area, the subsurface is a mixture of bare ground, sparse vegetation, and artificial buildings. This functional unit is an area for human production and living, with frequent activities occurring, such as oil factories and vehicle transportation. This resulted in the KB⁻¹ of this subsurface being similar to that of shifting sands and showing a more pronounced daily trend [46].

3.5. Bulk Transport Coefficient

Figure 6 shows the month-to-month variation characteristics of the momentum exchange coefficient C_d and sensible heat exchange coefficient C_h for three different functional units. The annual average values of C_d for natural shifting sands, shelter forest, and the living area were 5.3 × 10⁻³, 1.9 × 10⁻¹, and 4.5 × 10⁻², respectively, and the annual average values of C_h were 3.6 × 10⁻³, 1.3 × 10⁻¹, and 1.7 × 10⁻², respectively. The main reason for this difference is that the exposed surface of natural shifting sands is dominated by sand material cover, the subbedding surface is relatively smooth, and the smooth surface reduces the traction and the transport rate through the surface, whereas shelter forests, due to vegetation, have a high surface roughness, wetter subsurface, high vegetation transpiration, and surface water vapor evaporation, and an intense exchange of water-heat fluxes between the ground and air [47]. The living area is peripheral to artificial greenery, but the observation site is exposed and surrounded by sparse vegetation and artificial buildings. Therefore, the C_d and C_h of the shelter forest are larger by a magnitude of one than those of the living area and larger by a magnitude of two than those of the natural shifting sands. In general, the bulk transport coefficient of the vegetated subsurface is larger than that of the non-vegetated subsurface; the bulk transport coefficient of the densely vegetated, and the tall subsurface is one order of magnitude larger than that of the sparsely vegetated and short subsurface, and the bulk transport coefficient of the smooth subsurface is the smallest [43]. In addition, the unique atmospheric instability characteristics of the Taklamakan Desert led to larger values of C_d and C_h in winter and smaller values in summer for the three different functional units (

Table 7. Comparison of monthly average values of C_d × 10⁻¹ and C_h × 10⁻¹ for the 3 different functional units.

	Cd					Ch
Month	Shifting Sands	Shelter Forest	Living Area	Month	Shifting Sands	Shelter Forest	Living Area
1	0.054	2.63	0.52	1	0.038	1.29	0.18
2	0.055	2.38	0.50	2	0.041	1.31	0.17
3	0.054	2.02	0.46	3	0.036	1.16	0.17
4	0.052	1.64	0.36	4	0.033	1.01	0.13
5	0.052	1.48	0.36	5	0.031	1.08	0.14
6	0.052	1.23	0.39	6	0.028	1.08	0.15
7	0.053	1.16	0.36	7	0.034	1.09	0.16
8	0.052	1.42	0.45	8	0.033	1.14	0.17
9	0.053	1.69	0.49	9	0.033	1.18	0.18
10	0.054	1.91	0.53	10	0.034	1.25	0.18
11	0.056	2.42	0.55	11	0.038	1.34	0.19
12	0.056	2.63	0.57	12	0.039	1.50	0.19

). This phenomenon is consistent with Liu et al.’s [13] previous research results on the Tazhong area of Taklamakan Desert, but contrary to the results of C_d and C_h from Huang et al.’s [48] and Feng et al.’s [16] research on degraded grassland and farmland underlying surface in Naiman (sand dune) and the Tongyu area of Inner Mongolia, respectively. This is mainly because the phenomenon in the Taklamakan Desert is quite different from other regions. In spring and summer, the desert experiences high wind speeds and frequent dusty weather. These conditions increase the intensity of atmospheric turbulence and reduce atmospheric instability, thus affecting the magnitude of C_d and C_h in spring and summer [13].

3.6. Surface Flux

The daily and seasonal trends of the three different functional units H were consistent. As solar radiation heats the ground surface after sunrise, the ground mainly transports heat to the atmosphere with H, which reaches the daily peak at noon and decreases thereafter. After sunset, H becomes negative due to the cooling of the ground surface radiation. The variation in H throughout the seasons is primarily driven by solar radiation. Spring and summer are the peak seasons for H, while autumn and winter are not peak seasons. The natural shifting sands had the largest H value. This was due to their exposed and dry nature, particularly in spring and summer when high temperatures and strong radiation resulted in H values of 284.5 W·m⁻², on the other hand, the shelter forest received less solar radiation on the surface because of the lush vegetation, with a maximum H value of 228.8 W·m⁻². Therefore, the natural shifting sand’s H value was greater than that of the shelter forest, and the difference was significant in spring and summer. The living area was also affected by the peripheral artificial greenery, and H was also reduced. However, it was not affected as much as the shelter forest was (Figure 7a).

The study area is located in the desert hinterland and has an extremely arid climate, resulting in low values of LE when compared to H. It is worth noting that the increase in precipitation events between May and September led to a significant increase in LE in all three functional units. The average daily peaks of the natural shifting sand, shelter forest, and living areas were observed to increase during the period from May to September. Specifically, the peak values rose from 13.8 to 47.5 W·m⁻² for natural shifting sand, from 65.7 to 179.1 W·m⁻² for the shelter forest, and from 37.5 to 67.9 W·m⁻² for the living areas. Under the condition of equal precipitation, the shelter forest due to plant transpiration and artificial drip irrigation activities made the LE significantly higher than the natural shifting sands and living areas (Figure 7b).

The trend of soil heat flux (G₀₅) also had obvious unimodal daily and seasonal variation characteristics. Overall, G₀₅ in all three different functional units was significantly greater in summer than it was in winter. Comparing the three different functional units, the shelter forest, due to the protection of vegetation, has minimal fluctuation in the daily change in G₀₅, showing the trend of the lowest in the daytime and the highest in the nighttime. The living area is influenced by the peripheral green land and artificial buildings, and the fluctuation in G₀₅ change is the second highest. The natural shifting sands has low soil water content and large voids due to the bare surface receiving direct solar radiation. Therefore, the daily variation of G₀₅ fluctuated the most, showing a trend of it being the highest during the day and lowest at night (Figure 7c).

4. Conclusions

In this study, artificial greenery used for oil production in Tazhong in the hinterland of Taklamakan Desert was used as a research example. Based on the comparative analysis of the micrometeorological conditions among three functional units, namely, natural shifting sands, shelter forest and a living area, the study analyzed the characteristics of land surface processes and the differences in and causes of key land surface parameters and surface fluxes in each functional unit. The exposed surface and extremely arid nature of the natural shifting sands result in high wind speeds, high temperatures, extremely low-level soil moisture, and a high surface albedo in this functional unit. High radiation and high temperatures cause a high H and a very low LE, with occasional precipitation stimulating a sudden increase in LE. However, with the gradual increase in vegetation cover and the strengthening of transpiration of vegetation in the desert artificial green land, together with the superimposed influence of human production activities, the average daily maximum values of wind speed in the shelter forest decreased by 78%, and the daily average maximum air temperature and soil (0 cm) temperature decreased by 2.6 ℃ and 7 ℃, respectively. The degree of surface wetness increased and the surface albedo decreased with an annual average value of 0.21. This area has the best moisture conditions in the desert, and the average daily maximum values of H and G₀₅ decreased by 18.7% and 75%, respectively. For LE, artificial drip irrigation increased the average daily maximum value of LE by 70.3% under consistent precipitation conditions. In addition, the change in the nature of the subbedding surface led to the difference in surface roughness of the three functional units. This difference, in turn, caused a significant contrast in the aerodynamic roughness and bulk transport coefficient between the shelter forest and the bare natural shifting sands, and the seasonal variation was obvious. Moreover, the heat transport capacity of the shelter forest was relatively weak due to the influence of surface vegetation, and the KB⁻¹ daily variation was not obvious. In contrast, the living area inside the artificial green land was protected by the artificial green land and the surrounding buildings. Thus, the wind speed, temperature, and H were reduced, and the LE, aerodynamic roughness, and bulk transport coefficient increased, but not so substantially as in the shelter forest. It is worth mentioning that the mixed surface of the living area leads to the extremely pronounced and largest values of daily KB⁻¹ changes in this unit. Based on the evapotranspiration calculations, a minimum of 4.60 × 10⁵ t of groundwater per year should be withdrawn to maintain the current steady state of the artificial green land in the extremely arid Taklamakan Desert. In conclusion, the change in the nature of the subsurface under the influence of anthropogenic activities leads to completely different surface energy and turbulence transport capacities of the artificial green area than those of the natural shifting sand area. The establishment of artificial green land has changed the hydrothermal conditions of the flowing desert. The degree of vegetation cover has altered the original radiation balance of the surface, moderated the temperature changes, and reduced the degree of erosion due to sandy and windy weather. Ensuring the safety of production and life in the artificial green land during sand and high temperatures plays a positive role in improving the microclimate environment of the artificial green land. Based on the above research results, we will continue to conduct research on desert land surface processes using land surface model simulations in the future in order to deepen our understanding of the variations in land surface processes caused by changes in the underlying surface of deserts.