, , ,,

1. Key Laboratory of State Forestry Administration on Soil and Water Conservation & Ecological Restoration of Loess Plateau, Xi’an 710082, China; 2. Shaanxi Academy of Forestry, Xi’an 710082, China

1 Introduction

Mu Us sandy land is one of the four major sandy lands in China, and it is located in the border area of Shanxi, Shaanxi, Inner Mongolia and Ningxia along the Great Wall in the north of Yulin City, Shaanxi Province. It is characterized by strong transition and instability in climate, soil, vegetation and other factors due to the large amount of wind and sand, and the scarcity of precipitation[1-3]. At the same time, it is located in the hinterland of Shenfu coalfield, one of the eight major coalfields in China. It is a typical area where the energy-rich area and the fragile ecological area interact with each other. Especially with the rapid development of coal mining and development in recent years, the phenomenon of "secondary desertification" of sandy soil caused by strong human disturbance is becoming more and more obvious. Up to now, the area of coal mined-out area and subsidence area is more than 1 000 km2, in which the area of obvious subsiding fracture is about 130 km2, and the area of activated fixed and semi-fixed sand dunes is more than 300 km2[4-7]. This results in the distortion of sand texture, the crisscross of cracks, the further loosening of soil structure, the leakage of groundwater and the death of vegetation due to the lack of water, which seriously restricts the further economic and social development of the area. Previous studies on soil moisture distribution, dynamic change and spatial structure under different conditions in Mu Us sandy land have been done in detail[8-17], but few studies have been done on spatial variation of soil moisture caused by coal mining subsidence in sandy area. In order to investigate the influence of coal mining subsidence on soil moisture in Mu Us sandy land, the variation of soil moisture and spatial differentiation in different position of sand dunes after coal mining subsidence were systematically studied, by using the method of classical statistics and geostatistics through field observation, so as to provide scientific basis for ecological restoration and land reclamation in mining area.

2 Overviewofstudyarea

The study area is located in Daliuta mining area connecting the north of Yulin City, Shaanxi Province and the southeast of Ordos City, Inner Mongolia. The structure belongs to the Ordos syncline of the North China platform, and the main strata are Jurassic, Cretaceous mudstone, sedimentary sandstone and Quaternary aeolian sand[17]. The landform types are mostly fixed and semi-fixed star-moon sand dunes with relative height of 5-10 m. The main types of vegetation are artificially planted Salix, poplar as well asHedysarummongolicum,HedysarumscopariumandCaraganamicrophyllaaerially seeded in the 1980s. The coverage of vegetation is about 20%-30%. In terms of climatic conditions, the main features are scarce precipitation, sufficient light, strong evaporation, dry surface, frequent strong winds, so it is prone to wind erosion and sandstorm, resulting in land desertification.

At present, after many years of coal mining, a concentrated hollowed-out subsidence area is about 5 km2. The collapse fissures in the area interlaced and spread, the surface subsidence was obvious, and the width of the fissures varied from 3 to 60 cm. The height of the fissures ranged from 5 to 50 cm, and the surface breakage rate ranged from 20% to 40%, which resulted in the death of a large amount of artificial vegetation and the activation of some sand dunes[12]. After two years of coal mining subsidence, the coarsening of sandy soil is serious, and the sand is mainly medium grain size sand and coarse sand, accounting for more than 60% of the total composition. The bulk density of sandy soil decreased to be 1.5-1.6 g/cm3[2-5].

3 Researchmethods

3.1SamplingandobservationmethodsFrom March 2015 to October 2016, three sand dunes with similar trend, geomorphological type and vegetation coverage (assuming no other differences) were selected in the two-year coal mining subsidence area and non-subsidence area (control) of Daliuta mine area. In the typical positions of interdune, lower windward slope, middle windward slope, dune top,etc., through the field linear sampling method, the 1 m×1 m sample plot was adopted along the middle base line and both sides of dune, and the point was selected with the distance of 10 cm. TSV-V soil moisture instrument was used to observe the change of soil moisture in the range of 0-100 cm. It was repeated three times in parallel at each point, and the total number of sample points was 360.

3.2GeostatisticalanalysisUsing the measured soil moisture, the optimal linear unbiased estimation was carried out by the interpolation method of Surfer 8.0 according to the Point Kriging model and Block Kngmg model. Kriging interpolation, also called spatial local estimation or spatial local interpolation, is a geostatistical method[1, 18]. The method is widely used in groundwater modeling, soil mapping, metal grade estimation of ore deposits and other fields. In this paper, the spatial variation of soil moisture after coal mining subsidence was systematically analyzed using this method.

4 Resultsandanalysis

4.1SpatialdistributioncharacteristicsoflongitudinalsoilmoistureindunesofsubsidenceareaThe microgeomorphological changes of the sand dunes in the non-subsidence area (control) accorded with the characteristics of the sand dunes in the Mu Us sandy area, and the spatial distribution characteristics of the longitudinal soil moisture in the sand dunes are shown in Table 1.

Table1Statisticalcharacteristicsoflongitudinalsoilmoistureindunesofnon-subsidencearea(control)

Soildepth∥cmMinimum∥%Maximum∥%Mean∥%StandarddeviationSkewnessKurtosisCoefficientofvariation0-104．585．885．230．560．00-1．470．1110-204．725．765．270．51-0．11-4．850．1020-304．185．724．950．800．01-5．550．1630-404．145．764．920．690．25-0．660．1440-504．305．344．770．450．54-0．550．0950-604．285．224．730．390．301．220．0860-704．465．485．060．48-0．66-2．080．0970-804．745．305．030．24-0．10-1．030．0580-904．344．784．600．19-0．85-0．170．0490-1003．544．764．250．51-1．041．880．12

The longitudinal (0-100 cm) soil moisture from the top of the dune to the interdune decreased with the increasing depth of the soil layer. The maximum value of soil moisture was 5.88% at the surface layer of 0-10 cm, then decreased to about 4.00% (50-60 cm), and gradually increased to over 4.90% at 70-80 cm level, while the soil moisture was reduced to 4.76% at 90-100 cm level, and the average soil moisture was reduced by 18.74% at the whole longitudinal soil level. But from the whole section, the variation degree of soil moisture in each layer was slightly different, the distribution of probability curve was normal distribution, the variation of standard deviation was between 0.19 and 0.80, the degree of positive and negative deviation was not more than 1, and the variation of coefficient of variation was between 0.04 and 0.16. This change was consistent with the results of spatial-temporal variation of soil moisture in Mu Us sandy land[8-14].

Two years after the coal mining subsidence, the soil moisture change was disorderly at the longitudinal level of 0-100 cm from the top of the dune to the interdune (Table 2). The average soil moisture change of the whole dune was about 4.01%, which was 18.03% lower than that of the non-subsidence area (control). From the view of the whole section, the variation degree of soil moisture in each layer was quite different, and its probability curve distribution was normal distribution. The variation of standard deviation was from 0.54 to 1.05, which was 52.08% higher than that of the non-subsidence area (control), and the probability of positive and negative deviation was more than 50%. The coefficient of variation was from 0.14 to 0.28, which was 80% higher than that of the non-subsidence area (control).

Table2Statisticalcharacteristicsoflongitudinalsoilmoistureindunesofsubsidencearea

Soildepth∥cmMinimum∥%Maximum∥%Mean∥%StandarddeviationSkewnessKurtosisCoefficientofvariation0-103．585．084．660．72-1．933．730．1510-203．565．084．380．66-0．41-1．430．1520-303．284．523．760．541．331．730．1430-403．184．423．870．54-0．59-1．030．1440-503．264．703．950．680．16-3．950．1750-602．584．983．611．010．921．300．2860-702．884．363．940．71-1．903．620．1870-803．104．884．070．76-0．550．010．1980-902．945．404．441．05-1．392．610．2490-1003．484．844．380．62-1．702．970．140-1003．184．834．100．73-0．610．960．18

4.2SpatialdistributioncharacteristicsofhorizontalsoilmoistureindunesofsubsidenceareaThe spatial distribution characteristics of the horizontal soil moisture in the dunes of non-subsidence area (control) are shown in Table 3. The horizontal (0-100 cm) soil moisture from the top of the dune to the interdune gradually increased on the whole, the maximum and minimum soil moisture on the top of the dune was 4.94% and 3.54%, respectively, while the maximum and minimum soil moisture in the interdune was 5.88% and 4.40%, respectively. The average soil moisture increased by 13.73% from the top of the dune to the interdune in the whole horizontal soil layer. But from the overall horizontal variation, the variation degree of soil moisture in each layer was slightly different, and its probability curve distribution was normal distribution. The variation of standard deviation was from 0.34 to 0.57, the degree of positive and negative deviation was not more than 1, and the coefficient of variation was from 0.07 to 0.11, which was consistent with the temporal and spatial variation of soil moisture in Mu Us sandy land[8-14].

Table3Statisticalcharacteristicsofhorizontalsoilmoistureindunesofnon-subsidencearea(control)

TypicalpositionsMinimum∥%Maximum∥%Mean∥%StandarddeviationSkewnessKurtosisCoefficientofvariationDunetop3．544．944．390．42-0．590．770．09Middlewindwardslope4．345．484．960．34-0．25-0．310．07Lowerwindwardslope4．285．765．180．50-0．55-0．820．10Interdune4．405．884．970．570．71-1．400．11Dunes4．145．524．880．46-0．17-0．440．09

After two years of coal mining subsidence, the soil moisture change was very great in the range of 0-100 cm from the top of the dune to the interdune. The maximum and minimum soil moisture on the top of the dune was 5.40% and 4.36%, respectively, while the maximum and minimum soil moisture in interdune was 4.90% and 3.48%, respectively (Table 4). The soil moisture on the top of the dune was lower than that in the interdune. The studies of Sun Jianhuaetal.[10], Lu Yizhongetal.[11], Feng Weietal.[13]showed that the soil moisture content at the top of the dune was obviously lower than in the interdune, which indicated that the soil moisture change after coal mining subsidence was affected by the collapse-fissure and surface breakage, and no longer followed the law of soil moisture change of the dune in Mu Us sandy land. Similarly, from the statistical characteristics of the horizontal variation of soil moisture, the variation of soil moisture from the top of the slope to the interdune was quite different. The variation of standard deviation was from 0.34 to 0.66, which was 10.87% higher than that of the non-subsidence area (control), the degree of positive and negative deviation increased in different degrees, and the coefficient of variation was from 0.07 to 0.15, which was 44.44% higher than that of the non-subsidence area (control).

Table4Statisticalcharacteristicsofhorizontalsoilmoistureofsanddunesinsubsidencearea

TypicalpositionsMinimum∥%Maximum∥%Mean∥%StandarddeviationSkewnessKurtosisCoefficientofvariationDunetop4．365．404．720．340．860．200．07Middlewindwardslope3．245．084．360．66-0．85)-0．210．15Lowerwindwardslope2．584．163．290．460．430．060．14Interdune3．484．904．040．570．65-1．450．14Dunes3．424．894．100．510．27-0．350．13

4.3SpatialvariabilityofsoilmoistureinsubsidenceareaAfter coal mining subsidence, the micro-landform of sand dune in sand area changed greatly, ground collapse and fissures interlaced and spread, and the surface subsidence was obvious. The width of the fissures was 3-60 cm, the height of the staggered dune was about 5-50 cm, and the degree of the surface breakage was about 20%-40% or above. Many scholars studied the physical and chemical properties of sandy soil in the mining subsidence area of Shenfu coalfield, and the results[3-7, 19]showed that the vibration and distortion during coal mining subsidence led to the decrease of sandy soil bulk density, the increase of porosity, the increase of saturated hydraulic conductivity, and the increase of soil moisture loss. The study of Zhao Guopingetal.[12]showed that after two years of coal mining subsidence, the groundwater was funneled, causing vegetation degradation or death, and the collapse-fissure could lead to coarsening of surface particles. These factors might cause the spatial distribution and variation of sandy soil moisture (Fig.1).

Fig.1Spatialvariationofsoilmoistureintypicalpositionsofdunesinsubsidenceareaandnon-subsidencearea(control)

Under the condition of similar trend, geomorphological type and vegetation coverage between subsidence area and control sand dunes, the surface fissures and ground subsidence caused by coal mining subsidence were the key factors that affect the spatial variation of soil moisture. According to Fig.1, under the same conditions (assuming that there is no other variability), after two years of coal mining subsidence, the 0-100 cm soil moisture loss of the whole dune in the subsidence area was serious, and the sandy soil began to dry, a cumulative decrease of 30.96% and an average decrease of 15.87% in the soil moisture when compared with the control dune. In the range of 0-100 cm from the top of the dune to the interdune, the soil moisture content in the top layer (0-20 cm) changed little, about 5%. The soil moisture in the middle layer (30-70 cm) changed most obviously, and the black area appeared in Fig.1. Compared with the control sand dunes, the soil moisture decreased by 21.24%, and the decrease of soil moisture was more obvious in lower windward slope and interdune, decreasing by 40.77% and 23.38%, respectively. The soil moisture content in the lower layer (80-100 cm) decreased, but it was not obvious in general, which was 4.42% and 4.41%, respectively. The soil moisture content in the whole sand dunes no longer followed the law of non-subsidence area (control). This might be due to impact of collapse-fissure and surface damage two years after the coal mining subsidence, which directly or indirectly increased the contact area with the air, and under the action of wind, soil moisture evaporation accelerated, resulting in serious loss of soil moisture in the middle layer.

To sum up, after two years of coal mining subsidence, there were great changes in mean, standard deviation, skewness, kurtosis and coefficient of variation of the soil in different sections (0-100 cm) from the top of the dune to the interdune (Table 1, Table 2), and especially the variance change was the most obvious. The overall variance of the control dunes was only 0.26, while it was 0.56 in the subsidence area, which was 113.24% larger than that of the control sand dunes, and the dispersion degree of soil moisture was greatly increased, which showed that the spatial variation of soil moisture occurred after coal mining subsidence. At the same time, through the multi-factor analysis of variance of soil moisture in typical positions, it was found that the soil moisture at different levels on the top of control dunes and collapsed dunes with no significant changes, while there was a significant difference (α=0.05) in the soil moisture among middle windward slope, lower windward slope and interdune.

5 Conclusionsanddiscussions

(i) In the range of 0-100 cm from the top of the dune to the interdune, the longitudinal soil moisture in the typical positions of dunes in non-subsidence area (control), showed the trend of first decreasing and then increasing and then decreasing with the increase of soil depth, while the horizontal soil moisture in the typical positions of dunes in non-subsidence area (control), peaked in the interdune, followed by dune top and middle slope. This was consistent with the characteristics and variation of soil moisture in sand dunes of Mu Us sandy land studied by Sun Jianhuaetal.[10]and Lu Yizhongetal.[11]. After two years of coal mining subsidence, both the longitudinal distribution of soil moisture in different layers and the horizontal change of soil moisture in the typical position of dune, a disorderly change process appeared, which did not accord with the change characteristics of non-subsidence area, and the soil moisture loss was serious, which was about 10%-30% less than that of the control dune.

(ii) Typical statistical analysis showed that the variation degree of soil moisture in different layers from the top of dune to the interdune was slightly different in the dunes of non-subsidence area (control), and its probability curve distribution was normal distribution. The variation of standard deviation was from 0.19 to 0.80, the degree of positive and negative deviation was not more than 1, the coefficient of variation was from 0.04 to 0.16, which was consistent with the temporal and spatial variation of soil moisture in Mu Us sandy land[10-11,14]. The variation degree of soil moisture in different layers of sand dunes in subsidence area was different, the standard deviation was between 0.54 and 1.05, 52.08% higher than that in non-subsidence area (control), the probability of positive and negative deviation higher than 1 was more than 50%, and the coefficient of variation was 0.14-0.28, 80% higher than that in non-subsidence area (control).

(iii) Geostatistical analysis showed that under the same conditions (assuming that there is no other variability), after two years of coal mining subsidence, the soil moisture changes were basically divided into three layers: the soil moisture changes in the surface layer (0-20 cm) were small, about 5%; the soil moisture in the middle layer (30-70 cm) changed most obviously, and the soil moisture decreased by 21.24%; the soil moisture content in the lower layer (80-100 cm) decreased, but it was not obvious in general, which was 4.42% and 4.41%, respectively, and the soil moisture content in the whole sand dunes no longer followed the law of changes under non-disturbance. The soil moisture mean, standard deviation, skewness, kurtosis and coefficient of variation all changed greatly, especially the variance was the most obvious. The overall variance of the control sand dunes was only 0.26, while it was 0.56 in the subsidence area, which was increased by 113.24% compared with the control sand dunes. The dispersion degree of soil moisture was greatly increased, which directly led to the spatial variation of soil moisture in sandy soil.

[1] ZHAO GP, LI H, ZHANG HJ,etal. Fuzzy model assessment on over-ground environment development of Shenfu Mining Subsidence Area[J]. Journal of Xi’an Jiaotong University, 2012, 46(5): 137-142. (in Chinese).

[2] ZHAO GP, BI YL, YANG W,etal. The impact of coal mining subsiding on particle size composition in Shenfu coalfield[J]. Journal of Desert Research, 2015, 26(6): 1461-1466. (in Chinese).

[3] WANG J. Influence of coal mining subsidence on physical and chemical properties of sandy soil in semi-arid area[D]. Hohhot：Inner Mongolia Agricultural University, 2007: 13-26. (in Chinese).

[4] ZHAO HM. Research of soil water distribution and dynamic characteristics under the coal mining condition[D]. Beijing: Chinese Academy of Geologecal Sciences, 2006. (in Chinese).

[5] ZHAO HM, ZHANG FW, SONG YX,etal. Spatial variation of soil moisture content in mining subsidence areas of Daliuta, Shenmu County, Shannxi Province[J]. Geo-information Science, 2010, 12(6): 753-760. (in Chinese).

[6] ZANG YT, WANG J, BAI T,etal. Study on environmental variability of soil moisture in windy desert area after mining subsiding[J]. Journal of Arid Land Resources and Environment, 2009, 23(6):151-156. (in Chinese).

[7] WEI JS, HE X, HU CY,etal. Influence of ground collapse caused by coal mining activities on the water characteristics of sandy soil in arid and semi-arid area[J]. Journal of Arid Land Resources and Environment, 2006, 20(5): 84-88. (in Chinese).

[8] LIU X, MAN XL. Soil hydro-physical properties of artificialPopulussimoniiCarr. forest soil in Wusu Desert[J]. Acta Pedologica Sinica, 2009, 46(2): 348-351. (in Chinese).

[9] YAO YF, MAN XL. Soil hydro-physical properties of artificialPopulussimoniiCarr. forest soil in Mao Wusu Desert [J]. Acta Pedologica Sinica, 2007, 21(1): 111-115. (in Chinese).

[10] SUN JH, LIU JJ, KANG BW,etal. Studies on the spatio-temporal variation regularity of soil water in Mu Us sandy land in north Shaanxi Province[J]. Agricultural Research in the Arid Areas, 2009,27(2):244-247. (in Chinese).

[11] LU YZ, HU L, LI BG. The spatio-temporal variability of soil water in sand dunes in Maowusu Desert[J]. Acta Pedologica Sinica, 2006, 43(1): 152-154. (in Chinese).

[12] CUI LQ, WU B, YANG WB,etal. The characteristics of soil water in different vegetation coverage on the southeastern margin Mu Us sandy land[J]. Journal of Arid Land Resources and Environment, 2010, 24(2): 177-182. (in Chinese).

[13] FENG W, YANG B, LI W,etal. Response of soil moisture on rainfall in theSalixpsammophilafixed dunes of Mu Us sand land[J]. Journal of Soil and Water Conservation, 2014, 28(5): 95-99. (in Chinese).

[14] SHANG AJ, BO YJ, AI HJ,etal. Research on temporal and spatial patterns and dynamic laws of soil moisture in Yulin Sand District[J]. Journal of Soil and Water Conservation, 2008, 22(4): 86-89. (in Chinese).

[15] LI XR, MA FY, XIAO HL,etal. Long-term effects of revegetation on soil water content of sand dunes in arid region of Northern China[J]. Journal of Arid Environments, 2004, 57(1): 1-16.

[16] RUZ MH, CATHERINE MF. Influence of high water levels on aeolian sand transport:upper beach/dune evolution on a macrotidal coast,Wissant Bay, northern France Marie[J]. Geomorphology, 2004, 60(1/2): 73-87.

[17] STOYAN H, DEPOLLI H, BOHM S,etal. Spatial heterogeneity of soil respiration and related properties at the plant scale[J]. Plant and Soil, 2000, 222(1/2): 203-214.

[18] BI HX, LI XY, LIU X,etal. Spatial heterogeneity of soil moisture using geological statistics method in the loess region of west Shanxi Province[J]. Journal of Beijing Forestry University, 2006, 28(5):59-66. (in Chinese).

[19] SONG YR, WANG JM, LI XF,etal. Simulation of moisture transfer law with different soil reconstruction models in coal mining subsided area with high ground-water level[J]. Journal of Soil and Water Conservation, 2016, 30(2): 143-154. (in Chinese).