- Email: [email protected]
Geostatistical analysis of pedodiversity in Taihang Mountain region in North China
Geoderma 328 (2018) 91–99
Contents lists available at ScienceDirect
Geoderma journal homepage: www.elsevier.com/locate/geoderma
Geostatistical analysis of pedodiversity in Taihang Mountain region in North China Tonggang Fua, Lipu Hana, Hui Gaoa,b, Hongzhu Lianga,b, Jintong Liua,
T
⁎
a Key Laboratory of Agricultural Water Resources, Center for Agricultural Resources Research, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, Shijiazhuang 050022, China b University of Chinese Academy of Sciences, Beijing 100049, China
A R T I C LE I N FO
A B S T R A C T
Handling editor: A.B. McBratney
The protection of pedodiversity is very important for biodiversity maintenance and food production. However, where and how large should the protection focus on in a given region (especially in mountain regions with complex environment) remains unclear. In this study, geostatistical analysis was used to study pedodiversity indexes (Shannon's index, the maximum Shannon's index, Richness index and Evenness index) in Taihang Mountain region in North China. Then pedodiversity protection and buffer zones were designed. Moreover, the relationship between protection zone and human activities was determined. The results showed that average Shannon's index for the study area was 1.72, with two lower value zones appeared in the north and the south of Taihang Mountain region, respectively. This suggested that there was the need to enhance the protection of pedodiversity in the two regions. The ranges of the Richness index and the Shannon's index were respectively 57.16 km and 96.00 km, indicating that pedodiversity had spatial dependence within these distances. Therefore, protection zones with radius of 57.16 km and buffer zones with radius of 96 km was designed. For both the northern and southern protection zones, the percent area of population density lower than 1 person/km2 (35.27% for the north and 51.71% for the south) was much higher than the average value (27.91%) of Taihang Mountain region. Furthermore, the percent area of farmland (43.72% for the north and 40.48% for the south) were higher than the average value (35.92%). This demonstrated that human activity, especially farming, was a key consideration in the protection of pedodiversity in the study region. The results of the study constitute a significant contribution to the theory of pedodiversity protection and soil resources management.
Keywords: Soil diversity Semivariance Land use Population density Protection zone
1. Introduction Pedodiversity, also known as soil diversity, is the basis for biodiversity. In spite of the increasing interest in recent years, the study of pedodiversity lags far behind that of biodiversity (Guo et al., 2003; Minasny et al., 2010; Fajardo et al., 2017; Fu et al., 2018). The concept of pedodiversity was originally proposed in the 1990s (Mcbratney, 1992; Ibañez et al., 1995). Since then, it has gradually attracted global scientific attention and focus (Amundson et al., 2003; Toomanian et al., 2006; Shangguan et al., 2014). Pedodiversity study reached a peak level in the 2010s, with the publication of the book Pedodiversity in Springer (Ibáñez and Bockheim, 2013) and the follow-up special issue in Geomorphology (Volume 135, Issues 3–4). Irrespectively, the study of pedodiversity still remains much weaker than that of biodiversity. This underscores the need for more studies on pedodiversity. Existing studies on pedodiversity have focused mainly on four areas.
First, pedodiversity index; which come from the methods used to study biodiversity. This mainly includes the Shannon's index, Richness index, Evenness index and their derivatives (Ibañez et al., 1995; Tan et al., 2003; Danek et al., 2016). Second, object abundance models; which describe the distribution of objects abundance (Ibañez et al., 1998; Toomanian et al., 2006; Kooch et al., 2015). Four main types of models always used for this purpose include geometric model, logarithmic model, logarithmic normal model and broken stick model (Ibañez et al., 1995). Each of these models is related to a distribution pattern of soil individuals. Third, pedodiversity-area or richness-area relationship; which relationship is given by power or logarithmic function (Guo et al., 2003; Ibáñez et al., 2005; Ren and Zhang, 2015). Pedodiversityarea relationship is closely related to the object abundance models, with the power curve accompanied by lognormal and broken-stick models, and the logarithmic expressions by geometric and logarithmic series models (Saldaña and Ibáñez, 2007). Fourth, rare and endangered
⁎ Corresponding author at: Key Laboratory of Agricultural Water Resources, Center for Agricultural Resources Research, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, 286 Huaizhong Road, 050021 Shijiazhuang, China. E-mail address: [email protected] (J. Liu).
https://doi.org/10.1016/j.geoderma.2018.05.010 Received 26 November 2017; Received in revised form 28 March 2018; Accepted 6 May 2018 0016-7061/ © 2018 Elsevier B.V. All rights reserved.
Geoderma 328 (2018) 91–99
T. Fu et al.
annual temperature is 11.4 °C, with the highest in July and lowest in January. The average annual precipitation is 457 mm, with the highest in July and lowest in December. Average elevation range is 1000–1500 m, with the northwest higher than the southeast. The slopes are generally steep and often over 25°. The main land use types include farmland (35.92%), forestland (29.05%), grassland (26.44%) and others (8.59%). The vegetation is influenced by the elevation, slope gradient, slope aspect and the interaction effects of these factors (Zhang et al., 2006). Soils are always thin in Taihang Mountain, with an average thickness of 35 cm (Cao, 2011). The soils are mainly developed from limestone in the northern and southern regions, but from gneiss in the central region. Based on the first-level FAO soil classification (Harmonized World Soil Database), 26 soil types exist in Taihang Mountain. The most common are Cambisols and the most uncommon are Podzoluvisols. Detailed information on soil types is given by Fu et al. (2018).
soils studies; which is mainly about soils that are significantly influenced by humans (Amundson et al., 2003; Shangguan et al., 2014; Tennesen, 2014). In summary, most of the existing studies on pedodiversity are concerned mainly with basic theories of pedodiversity. Only the studies of endangered soils provide information on soil protection. With the existing soil databases, land planning can be used to protect pedodiversity and cultural heritage (Costantini et al., 2007; Ibáñez et al., 2008). However, exactly where and to what extent is considered appropriate for sustainable protection remain largely unclear. Geostatistics, which is based on the theory of regionalized variables, is a branch of applied statistics that focuses on the detection, modeling and estimation of spatial patterns (Rossi et al., 1992; Goovaerts, 2000; Webster and Oliver, 2007). The most used methods in geostatistical analysis are semivariogram and kriging. From semivariogram, nugget (C0), sill (C0 + C), range and other parameters can be derived. The nugget-to-sill ratio (C0/(C0 + C)) reflects the degree of spatial dependence. Based on Cambardella et al. (1994), spatial dependence is strong when C0/(C0 + C) ratio < 25%, moderate when between 25 and 75% and weak when > 75%. The range reflects the correlation length beyond which the sampling points are independent. Because of the ease and practicability of geostatistical method, it is widely used in the study of soil moisture, soil nutrient, soil texture and other soil property analysis (Duffera et al., 2007; Bourennane et al., 2014; Fu et al., 2016; Bogunovic et al., 2017), as well as the study of biodiversity (Gholami et al., 2017; Liparoto et al., 2017). Nevertheless, geostatistics is hardly tried in the study of pedodiversity. The use of geostatistics in pedodiversity can be useful in understanding the spatial structures of pedodiversity and can provide meaningful information on pedodiversity protection zones. The study of pedodiversity began in China in 2001 (Chen et al., 2001). Since then, it has gradually intensified in Shandong Province, Nanjing City, Henan Province and other areas in the country (Tan et al., 2003; Zhang et al., 2007; Ren and Zhang, 2017). Besides regional studies, there also are studies which focus on the whole China (Zhang and Gong, 2004; Shangguan et al., 2014). However, mountain areas, which have complex environmental conditions, are largely ignored in these studies. Taihang Mountain, which is in North China, is an important boundary region between the Loess Plateau and the North China Plain (NCP). The special location of Taihang Mountain makes it an important ecological area (Yang et al., 2003). However, this mountain region is threatened by serious ecological degradation, including soil depletion, natural hazards, overgrazing, etc. (Li et al., 2004). The loss of pedodiversity is also a very serious problem in Taihang Mountain. There are more rare-unique soils and endangered soils in this mountain area than in other areas of China (Shangguan et al., 2014); requiring considerable attention for pedodiversity assessment, protection and sustainability. The objectives of this paper were to: (1) obtain the geostatistical characteristics of pedodiversity indexes, (2) determine protection zones and buffer zones, and (3) analyze the relationship between protection zones and human activities in the Taihang Mountain area, north China.
2.2. Pedodiversity indexes Richness index (S) defines the number of soil types within a given region, which can be a country, a catchment or even a regular grid of in the area. Another important pedodiversity index is the Shannon's index. It is based on proportional abundance of objects and it is expressed as: i=n
H = − ∑ pi × ln pi
(1)
i=1 th
where pi is the proportion of individual number of the i object (area of ith soil type) to the total number of individuals (total area of the given region). The value H varies between 1.5 and 3.5 and rarely exceeds 4.5 (Margalef, 1972). When all the objects (soil types) are evenly distributed, the Shannon's index is then maximum and is expressed as: (2)
Hmax = lnS
Actually, the Shannon's index rarely reaches its maximum value and the proportion of H to Hmax is the evenness index (E), expressed as:
E=
H Hmax
(3)
The range of E is 0–1 and higher E values means more even distribution of soil types. 2.3. Geostatistical analysis Semivariance and spatial interpolation (Kriging interpolation) are the two wildly used methods in geostatistical analysis. The semivariance (γ(h)) of a regionalized variable is expressed as (Webster and Oliver, 2007; Fu et al., 2015):
γ(h) =
2. Materials and methods
1 2N (h)
N (h)
∑
[Z (x i ) − Z (x i + h)]2
i=1
(4)
In the formula, N(h) is the paired points at lag distance h; Z(i) is the value of the variable at point xi; and Z(xi + h) is the value at a point which is h meters away from point xi. The relationship between semivariance and lag distance is always simulated by Spherical model, Exponential model or Gaussian model as:
2.1. Site description Taihang Mountain is located in North China, extending in the southwest-northeast direction (Fig. 1). Taihang Mountain region extents across a total of 4 provinces (Beijing, Hebei, Shanxi and Henan) and 101 counties. It is an important transition zone between the Loess Plateau and NCP. Because of its special location and orogeny, the western and eastern parts of Taihang Mountain are very different in structural evolutions, sedimentation processes and tectonic settings (Wang and Li, 2008). Taihang Mountain is influenced by the East Asian Monsoon climate. The summer is warm and rainy, and the winter is cold and dry. Based on 10-year (2005–2014) meteorological data in the region, average
γ(h) =
⎧ C0 + C ⎨ ⎩ C0 + C
(
3h 2a
−
(
h3 a3
92
0≤h≤a
h≥a h
γ(h) = C0 + C 1 − e− a γ(h) = C0 + C ⎜⎛1 − e ⎝
)
)
(5) (6)
2
− h2 ⎞ a ⎟
⎠
(7)
Geoderma 328 (2018) 91–99
T. Fu et al.
Fig. 1. A map depicting the location of the study area in China (inset) and an expanded map of the study area (main plate).
where γ(h) is the semivariance; h is the lag distance; C0 is the nugget; and C is the sill. The degree of spatial dependence of a variable is reflected by the ratio of C0 to C0 + C. When the ratio is < 25%, the spatial dependence is strong. Then when the ratio is between 25 and 75%, the spatial dependence is moderate. Yet when it is higher than 75%, the spatial dependence is weak (Cambardella et al., 1994). Semivariance can be different in different directions (Journel and Huijbregts, 1978). The semivariance of regional variable in a given direction can be calculated by:
γ(h, φ) =
1 2N (h, φ)
values were assigned to the center point of each grid, which computation was completed in ArcMap 9.3 (Environmental Systems Research Institute, Inc. (ESRI)). Then pedodiversity and geographical information of the points were imported to GS+ (Geostatistics for the Environmental Sciences, version 9.0) to calculate the semivariance, nugget, sill and range. The most suitable semivariance model was also simulated in GS+ software. In addition, anisotropy ratio between the 30° direction (parallel to the trend of the mountain) and the 120° direction (perpendicular to the trend of the mountain) was calculated. The spatial distribution map of the pedodiversity indexes and pedodiversity protection and buffer zones were drawn in ArcMap 9.3. The data source of elevation, land used type, and population density were introduced in Fu et al. (2018).
N (h, φ)
∑ i=1
[Z (x i ) − Z (x i + h)]2
(8)
Here, γ(h,φ) is the semivariance in φ direction; and N(h,φ) is the number of paired points with lag distance h in φ direction. Considering different directions, the anisotropy ratio is given as:
K(h) =
γ (h, φ1 ) γ (h, φ2 )
3. Results 3.1. Classical statistical characteristics of pedodiversity indexes
(9) The range of H value in Taihang Mountain region was 1.01–2.19 with mean and CV (Coefficient of Variation) of respectively 1.72 and 0.17, indicating a moderate variation (0.1 < CV < 1). The range of Hmax was 2.48–3.14 with mean and CV of respectively 2.87 and 0.05, showing a weak variation (CV < 0.1). The CV value of Hmax was the smallest among the four investigated indexes. The lowest S value was 12, highest 23, mean 17.75 and the CV 0.13; suggesting a moderate variation. The range of E was 0.37–0.78, with a mean value of 0.60 and CV value of 0.15; also indicating a moderate variation (Table 1). All the four indexes had normal distributions. Significant Kolmogorov-Smirnov test values were respectively 0.06, 0.19, 1.20 and 0.08, with a normal distribution for values higher than 0.05.
In the above formula, γ(h,φ1) is the semivariance in φ1 direction; and γ(h,φ2) is the semivariance in φ2 direction. 2.4. Data analysis The soil taxa data for Taihang Mountain were clipped from the Harmonized World Soil Database (HWSD) which describes global soil taxa. HWSD is a comprehensive database that combines vast volumes of regional and national updates of soil information with information already contained within the 1:5,000,000 scale FAO-UNESCO Digital Soil Map of the World. The work was done by the Food and Agriculture Organization of the United Nations (FAO) and the International Institute for Applied Systems Analysis (IIASA). The database is a raster with a resolution of about 1 km (30 arc sec by 30 arc sec). Every pixel has an attribute that noted the soil type. A bigger grid layer of 31.6 km × 31.6 km scale (1000 km2) was overlaid on the soil taxa map. Then the pedodiversity indexes (S, H, Hmax and E) were calculated for the individual 1000 km2 grids. The calculated pedodiversity index
3.2. Geostatistical characteristics of pedodiversity indexes Based on geostatistical analysis, the different indexes had different suitable models. The best-fit models were the Spherical model for the H index and the Gaussian model for and the Hmax and S indexes. However, 93
Geoderma 328 (2018) 91–99
T. Fu et al.
spatial dependence. The range of E was the largest (174.3 km) and that of Hmax (56.12 km) was the smallest. The ranges of S were similar with that of Hmax, with a value of 57.16 km (Table 2).
Table 1 Classical statistical analysis of pedodiversity indexes for Taihang Mountain region in North China. Minimum
Shannon's index Maximum Shannon's index Richness index Evenness index
Maximum
Mean
Standard deviation
CV
3.3. Anisotropy of pedodiversity indexes
1.01 2.48
2.19 3.14
1.72 2.87
0.29 0.13
0.17 0.05
12.00 0.37
23.00 0.78
17.75 0.60
2.35 0.09
0.13 0.15
The semivariance of the pedodiversity indexes were different in the two investigated directions (the 30° and 120° directions). Anisotropic ratio of H and E increased first and then decreased with increasing lag distance (Fig. 2a and d). It was higher than 1 when the lag distance was about 100–300 km, indicating that H and E variance in the 30° direction was lower than that in the 120° direction. The highest anisotropic ratio was 2.67, which was in the 232.56 km lag distance. It suggested that the difference in variance between the two directions was the highest for this lag distance. For Hmax and S indexes, anisotropic ratio gradually increased with increasing separation distance. It was < 1 when the separation distance was < 133 km, implying that the variation in the 120° direction was higher than that in the 30° direction (Fig. 2b and c). Moreover, the difference decreased with increasing separation distance. When the separation distance was higher than 133 km, the variance in the 30° direction exceeded that in the 120° direction.
Note: CV is coefficient of variation. Table 2 Geostatistical analysis of pedodiversity indexes for Taihang Mountain region in North China.
Shannon's index Maximum Shannon's index Richness index Evenness index
Best-fit model
C0
C0 + C
C0/(C0 + C)
Range (km)
R2
Spherical model Gaussian model
0.003
0.086
0.03
96.00
0.46
0.002
0.018
0.11
56.12
0.30
Gaussian model Exponential model
0.630
5.568
0.11
57.16
0.33
3.4. Spatial distribution of pedodiversity indexes
0.001
0.009
0.11
174.3
0.54
The spatial distribution of E was similar to that of H, both apparently fluctuating from north to south. There were two low value zones separately in the northern and southern regions of Taihang Mountain (Fig. 3a and d). The spatial distributions of Hmax and S were also similar, with higher zones occurring mainly in the north and lower zones in the south (Fig. 3b and c).
Note: R2 is Coefficient of determination; Co is nugget; Co + C is sill; and Co/ (Co + C) is nugget-sill ratio.
the best-fit model for the E index was the Exponential model. All the indexes had nugget effect, with nugget values of 0.003 (H index), 0.002 (Hmax index), 0.630 (S index) and 0.009 (E index). The sill values of the H, Hmax, S and E indexes were respectively 0.086, 0.018, 5.568 and 0.009. The nugget/sill values for Hmax, S and E were the same (all equal to 0.11), suggesting a moderate spatial dependence. However, the nugget/sill value for H was much smaller (0.03), indicating a strong
3.5. Spatial distribution of protection and buffer zones The H and E indexes each had two lower zones in the northern and southern parts of the study area. This suggested that protection zones and buffer zones should be designed around these area (Fig. 4a). The radii of the protection zone and the buffer zone were determined
Fig. 2. The anisotropy of pedodiversity indexes between 30° and 120° directions. a is Shannon's index, b is maximum Shannon's index, c is Richness index and d is Evenness index. 94
Geoderma 328 (2018) 91–99
T. Fu et al.
Fig. 3. Maps of Ordinary Kriging for spatial distributions of H (Shannon's index), Hmax (the maximum Shannon's index), S (Richness index) and E (Evenness index).
geostatistical analysis has already been used in the study of biodiversity, which is much similar to pedodiversity, and the semivariances of abundance, richness, Shannon's index, etc. were calculated (Gholami et al., 2017; Liparoto et al., 2017). Therefore, geostatistical methods were used in this study to analyze pedodiversity in the Taihang Mountain region. Pedodiversity indexes had a strong spatial dependence in Taihang Mountain region. The degrees of spatial dependence of soil properties (such as soil thickness, soil moisture, soil nutrients, soil texture etc.) were always different for different regions or scales (Duffera et al., 2007; Bourennane et al., 2014; Yang et al., 2016). The strong spatial dependence of pedodiversity in the study area was attributed to the fact that pedodiversity is less sensitive to the influencing factors than other soil properties. For example, the spatial structure of soil moisture is easily influenced by precipitation, irrigation or other processes (Yang et al., 2016); soil bulk density and porosity change after plowing or vehicle compaction (Fu et al., 2015). These processes hardly affect
respectively by the ranges of the richness index (57.16 km) and Shannon's index (96.00 km). Then the protection zones covered mainly 10 counties (Fig. 4b) which needed monitoring for pedodiversity protection. One of the two buffer zones covered 6 counties and the other 17 counties. The areas and counties under the two protection and buffer zones are shown in Table 3, which provide useful information for pedodiversity protection. 4. Discussions 4.1. Geostatistical method for pedodiversity analysis Soil properties, such as soil moisture, soil nutrient and soil texture, are regionalized variables and can be analyzed by geostatistical method (Duffera et al., 2007; Bourennane et al., 2014; Fu et al., 2016; Bogunovic et al., 2017). Pedodiversity could be considered as individual soils attribute (Saldaña and Ibáñez, 2007). Moreover, 95
Geoderma 328 (2018) 91–99
T. Fu et al.
Fig. 4. Maps of protection zones and buffer zones in Taihang Mountain region. a is the distribution of Shannon's index, b is the distribution of counties, c is the distribution of population density, and d is the distribution of land use types.
methods based on regular grids in this study had the same trend with that calculated by classical approach based on political boundary (Fu et al., 2018). This also indicates that geostatistic analysis is appropriate in the study of pedodiversity in Taihang Mountain.
pedodiversity in the short run. Therefore, the spatial dependence of pedodiversity can be strong. Pedodiversity indexes (e.g., richness index and Shannon's index) were derived from biodiversity analysis methods (Ibañez et al., 1995), but biodiversity is more sensitive to the influencing factors. For example, the diversity of neuston community showed a moderate to week spatial dependence (Liparoto et al., 2017). The spatial distribution of the four considered pedodiversity indexes were similar (Fig. 4). There were two lower zones, with one in the northern and the other in the southern part of Taihang Mountain. The spatial distribution of pedodiversity indexes calculated by geostatistical
4.2. Anisotropy of pedodiversity and its influencing factors Anisotropy has always existed in soils; saying that soils vary in different directions (Crawford and Hergert, 1997). Anisotropy of soil properties, such as soil physical properties, soil nutrient properties and 96
Geoderma 328 (2018) 91–99
T. Fu et al.
(2015). The trend for Taihang Mountain may cause the different variations of soil-forming factors in the two directions. As a result, anisotropy of pedodiversity could be caused by the differences in variation of the soilforming factors. Most of the soil-forming factors considered in this study (e.g., elevation, precipitation and temperature) had higher variations in the 120° direction than in the 30° direction (Fig. 5). This is due to the trend in elevation, which had a clear gradient from northwest to southeast. Most of the soil-forming factors had a closed correlation with elevation. Therefore, the natural factors had a clear anisotropy. However, the variances of pedodiversity indexes in the 120° direction were higher than those in the 30° direction for separation distances smaller than 100–150 km but lower for separation distances larger than 150 km (Fig. 2), which was different for the soil-forming factors. This could be due to the fact that many influencing factors were not considered. For example, farming and population density have been noted as very important factors influencing pedodiversity (Fu et al., 2015). These factors have poor spatial structure and always distributed randomly. This indicates that when the lag distance is < 100–150 km, controlling factors can be natural. But when the distance is longer than 100–150 km, the anthropogenic factors become more important.
Table 3 Counties in the protection and buffer zones in Taihang Mountain region in North China. Position
Zone
Area (km2)
Main county
North
Pz Bz
7476 17,961
South
Pz
10,264
Bz
27,710
Yuxian1, Yangyuan,Guangling Zhoulu, Laishui, Laiyuan, Lingqiu, Hunyuan, Yixian Jincheng, Zezhou, Lingchuan, Changzhi, Zhangzi, Qinshui, Yangcheng Mengzhou, Qinyang, Bo'ai, Xiuwu, Haozuo, Huixian, Huguan, Pingshun, Lucheng, Changzhishi, Tunliu, Anze, Fushan, Yicheng, jiangxiang, Yuanqu, Jiyuan.
Note: Pz is protection zone and Bz is buffer zone.
soil microorganisms, has been extensively studied (Fu et al., 2010; Fu et al., 2015; Wu et al., 2016). Pedodiversity, as aforementioned, is a form of soil property and is anisotropic (Fig. 2). The reason for anisotropy in the study area could be due to the southwest-northeast orientation of Taihang Mountain (Fig. 1). The distance from west to east is about 100–150 km, within which point pairs were equal for the two directions. With higher distances, however, point pairs hardly existed in the 120° direction, and the variation can be smaller. A similar anisotropic form of soil physical properties was also noted by Fu et al.
Fig. 5. The anisotropy of soil-forming factors: a is elevation, b is slope gradient, c is precipitation, d is temperature, e is population density and f is percent farmland. 97
Geoderma 328 (2018) 91–99
T. Fu et al.
Table 4 Population density in protection and buffer zones in Taihang Mountain region in North China. Position
Zone
<1 (person/km2)
P<1 (%)
1–25 (person/km2)
P1–25 (%)
26–100 (person/km2)
P26–100 (%)
> 100 (person/km2)
P > 100 (%)
North
Pz Bz Pz Bz
1787 6258 1815 5404 38,222
23.90 34.84 17.68 19.50 27.91
951 2399 1045 2589 15,151
12.72 13.36 10.18 9.34 11.06
2101 3957 2128 5389 25,670
28.10 22.03 20.73 19.45 18.74
2637 5347 5276 14,328 57,903
35.27 29.77 51.40 51.71 42.28
South Total
Note: Pz is protection zone, Bz is buffer zone, P < 1 is the percentage of the population density < 1 person/km2, P1–25 is the percentage of the population density between 1 person/km2 to 25 persons/km2, P26–100 is the percentage of population density between 26 persons/km2 to 100 persons/km2, and P > 100 is the percentage of population density higher than 100 persons/km2. Table 5 Land use types in the protection and buffer zones. Position
Zone
Fa (km2)
PFa (%)
Fo (km2)
PFo (%)
Gr (km2)
PGr (%)
Wa (km2)
PWa (%)
Co (km2)
PCo (%)
Un (km2)
PUn (%)
North
Pz Bz Pz Bz
3269 5776 4156 11429 49167
43.72 32.16 40.48 41.24 35.92
1692 5893 4110 10100 39773
22.64 32.81 40.04 36.45 29.05
2124 5415 1288 3879 36188
28.41 30.15 12.55 14.00 26.44
77 162 46 319 1725
1.03 0.90 0.45 1.15 1.26
305 702 661 1966 9831
4.07 3.91 6.44 7.09 7.18
10 13 4 17 210
0.13 0.07 0.04 0.06 0.15
South Total
Note: Pz is protection zone, Bz is buffer zone, Fa is farmland, PFa is the percentage of farmland, Fo is forestland, PFo is the percentage of forestland, Gr is grassland, PGr is the percentage of grassland, Wa is water land, PWa is the percentage of water land, Co is construction land, PCo is the percentage of construction land, Un is undeveloped land, and PUn is the percentage of undeveloped land.
4.3. The design of protection and buffer zones
4.4. Correlation between delineated protection area and human activities
The concept of protection and buffer zones first appeared in the 1970s in Man and Biosphere (MAB) and in Biosphere Reserves (BRs) (Martino, 2001). More recently, however, the terms have been used mainly in the study of national park, which include three rings. From the inner to the outside, the rings are protection zone (core zone), buffer zone and transmission zone (Mannetti et al., 2015; Squeo et al., 2016). In this paper, the terms protection zone and buffer zone were used in the context of pedodiversity and with different meanings as used in the context of national park. Protection zone in this study denotes areas with low pedodiversity (Fig. 4) and therefore requiring protection by government and residents. Buffer zone denotes areas where pedodiversity is closely related with protection zones. No further damage is allowed in these zones and the existing conditions should be maintained. Studies have also designed protection zones or buffer zones in protection of wetlands, forestlands or in basic biodiversity protection (Semlitsch and Bodie, 2003; Kueffer and Kaiser-Bunbury, 2014; Robalino et al., 2015). Numerous methods are available for delineating protection zone and buffer zone boundaries, including statistical habitat model and the least cost distance method (Li et al., 1999; Li and Liu, 2006). In this study, geostatistical method was used to delineate protection zone and buffer zone boundaries because the range of an index represents the length of spatial correlation (Fu et al., 2015; Liparoto et al., 2017). The radii of the protection and buffer zones were determined by the ranges of S and H indexes. The S index was the most sensitive index and directly influenced by human disturbance. Amundson et al. (2003) noted that 31 soils were essentially “extinct” due to farming, urbanization and other activities. Therefore, the correlation length of S could reflect protection zone. The H index considers both species and evenness (Ibañez et al., 1995) and is therefore less sensitive than S and its correlation length could reflect buffer zone. Based on these features, protection and buffer zone boundaries were delineated (Fig. 4), which was critical for soil protection in the study area.
In the protection zones, the P < 1 value (percent area with population density < 1 person/km2) was lower than that for the whole Taihang Mountain area (Table 4). Apparently, pedodiversity was negatively related with population density. Moreover, studies show that human activities can limit pedodiversity (Lo Papa et al., 2011; Tennesen, 2014). However, the P > 100 value (percent area with population density higher than 100 people/km2) was higher in the southern protection zone, but lower in the northern protection zone than the whole study area (Table 4). This was attributed to the fact that industries or tourist centers always have high population density, but relatively little effect on pedodiversity (Fu et al., 2018). In both of the two protection zones, percent farmland was higher than other land use types. Moreover, it was higher than average percent farmland for the whole Taihang Mountain (Table 5). This was due to the fact that farming always influences pedodiversity (Costantini and Dazzi, 2013; Tennesen, 2014). Study also show that farming has a significant negative effect on pedodiversity in Taihang Mountain (Fu et al., 2018). Besides farmland, forestland also affects pedodiversity and forest ecosystems have a greater deal of soil variability than non-forest ecosystems (Kooch et al., 2015). In this study, however, it was interesting to know that the southern protection zone had a much higher percent forestland (40.04%) than the average value (29.05%) of the Taihang Mountain region (Table 5). This was probably because the southern protection zone had lower elevation range, which has a significantly positive effect on pedodiversity (Fu et al., 2018). In the buffer zone, the effect of farming was apparently lower than that in the protection zone because the northern buffer zone had a lower percent farmland than the whole Taihang Mountain region (Table 5). This again demonstrated the need for more attention on protection zone.
5. Conclusions Spatial distribution of pedodiversity index clearly showed two lower patches in Taihang Mountain, highlighting the need for much closer monitoring for pedodiversity protection. Geostatistical analysis showed a clear spatial structure of pedodiversity index in the region. Therefore, 98
Geoderma 328 (2018) 91–99
T. Fu et al.
based on spatial correlation distance (range), boundaries for the protection and buffer zones were delineated. The location of the protection zones was closely related to farmland, suggesting that farming significantly influenced pedodiversity in the mountain region. Therefore, there was the need to restrict serious disturbances by farming in the protection zones in order to protect pedodiversity. The results of the study showed the spatial characteristics of pedodiversity and provided the critical basis for the delineation of pedodiversity protection zones. This was critical for pedodiversity protection and soil resource management in the mountain region with a complex terrain.
European soils: natural and cultural heritage. The soils of tomorrow. Adv. Geoecol. 39, 37–59. Journel, A.G., Huijbregts, C.J., 1978. Mining geostatistics. London Academic press. Kooch, Y., Hosseini, S.M., Scharenbroch, B.C., et al., 2015. Pedodiversity in the Caspian forests of Iran. Geoderma Reg. 5, 4–14. Kueffer, C., Kaiser-Bunbury, C.N., 2014. Reconciling conflicting perspectives for biodiversity conservation in the Anthropocene. Front. Ecol. Environ. 12 (2), 131–137. Li, J., Liu, X., 2006. Research of the nature reserve zonation based on the least, cost distance model. J. Nat. Resour. Policy Res. 21 (2), 217–224. Li, W., Wang, Z., Ma, Z., et al., 1999. Designing the core zone in a biosphere reserve based on suitable habitats: Yancheng biosphere reserve and the red crowned crane (Grus japonensis). Biol. Conserv. 90 (3), 167–173. Li, Y., Xu, H., Wang, Z., 2004. A talk on ecological rehabilitation of Taihang mountainous area in Hebei (in Chinese with English abstract). In: Soil And Water Conservation In China. 11. pp. 10–12. Liparoto, A., Mancinelli, G., Belmonte, G., 2017. Spatial variation in biodiversity patterns of neuston in the Western Mediterranean and Southern Adriatic seas. J. Sea Res. 129, 12–21 (Supplement C). Lo Papa, G., Palermo, V., Dazzi, C., 2011. Is land-use change a cause of loss of pedodiversity? The case of the Mazzarrone study area, Sicily. Geomorphology 135 (3–4), 332–342. Mannetti, L., Zeller, U., Esler, K., 2015. Evaluating Social-Ecological Aspects of Buffer Zones at the Borders of Etosha National Park, Namibia. USDA Forest Service Proceedings RMRS-P-74pp. 129–132. Margalef, R., 1972. Homage to Evelyn Hutchinson, or why there is an upper limit to diversity. Trans. Connecticut Acad. Arts Sci. 44, 211–235. Martino, D., 2001. Buffer zones around protected areas: a brief literature review. Electron. Green J. 1 (15). Mcbratney, A.B., 1992. On variation, uncertainty and informatics in environmental soil management. Soil Res. 30 (6), 913–935. Minasny, B., McBratney, A.B., Hartemink, A.E., 2010. Global pedodiversity, taxonomic distance, and the World Reference Base. Geoderma 155 (3–4), 132–139. Ren, Y., Zhang, X., 2015. Study on pedodiversity: status quo and future challenges (in Chinese with English abstract). Acta Pedol. Sin. 52 (1), 9–18. Ren, Y.Y., Zhang, X., 2017. Pedodiversity patterns based on terrain in Henan Province (in Chinese with English abstract). Chin. J. Soil Sc. 48 (1), 22–31. Robalino, J., Sandoval, C., Barton, D.N., et al., 2015. Evaluating interactions of Forest conservation policies on avoided deforestation. PLoS One 10 (4), e0124910. Rossi, R.E., Mulla, D.J., Journel, A.G., et al., 1992. Geostatistical tools for modeling and interpreting ecological spatial dependence. Ecol. Monogr. 62 (2), 277–314. Saldaña, A., Ibáñez, J., 2007. Pedodiversity, connectance and spatial variability of soil properties, what is the relationship? Ecol. Model. 208 (2), 342–352. Semlitsch, R.D., Bodie, J.R., 2003. Biological Criteria for Buffer Zones around Wetlands and Riparian Habitats for Amphibians and Reptiles. Conserv. Biol. 17 (5), 1219–1228. Shangguan, W., Gong, P., Liang, L., et al., 2014. Soil diversity as affected by land use in China: consequences for soil protection. Sci. World J. 2014. Squeo, F.A., Loayza, A.P., López, R.P., et al., 2016. Vegetation of Bosque Fray Jorge National Park and its surrounding matrix in the Coastal Desert of North-central Chile. J. Arid Environ. 126, 12–22 (Supplement C). Tan, M., Zhang, X., Chen, J., et al., 2003. Pedodiversity: a case study based on 1:1 million scale SOTER of Shandong Province, China. Pedosphere 13 (3), 219–226. Tennesen, M., 2014. Rare Earth. Science 346 (6210), 692–695. Toomanian, N., Jalalian, A., Khademi, H., et al., 2006. Pedodiversity and pedogenesis in Zayandeh-rud Valley, Central Iran. Geomorphology 81 (3–4), 376–393. Wang, Y., Li, H., 2008. Initial formation and Mesozoic tectonic exhumation of an intracontinental tectonic belt of the northern part of the Taihang Mountain belt, eastern Asia. J. Geol. 116 (2), 155–172. Webster, R., Oliver, M.A., 2007. Geostatistics for Environmental Scientists. Wiley. Wu, Y., Li, Y., Fu, X., et al., 2016. Three-dimensional spatial variability in soil microorganisms of nitrification and denitrification at a row-transect scale in a tea field. Soil Biol. Biochem. 103 (Supplement C), 452–463. Yang, Y.H., Watanabe, M., Wang, Z.P., et al., 2003. Prediction of changes in soil moisture associated with climatic changes and their implications for vegetation changes: WAVES model simulation on Taihang Mountain, China. Clim. Chang. 57 (1–2), 163–183. Yang, J., Chen, H., Nie, Y., et al., 2016. Spatial variability of shallow soil moisture and its stable isotope values on a karst hillslope. Geoderma 264, 61–70. Zhang, X., Gong, Z., 2004. A pedodiversity pattern: taxonomically established soil orders in China. J. Geogr. Sci. 14 (1), 52–56. Zhang, J., Xi, Y., Li, J., 2006. The relationships between environment and plant communities in the middle part of Taihang Mountain Range, North China. Community Ecol. 7 (2), 155–163. Zhang, X., Chen, J., Tan, M., et al., 2007. Assessing the impact of urban sprawl on soil resources of Nanjing city using satellite images and digital soil databases. Catena 69 (1), 16–30.
Acknowledgements This work was supported by the National Basic Research Program (973 Program) of China (No. 2015CB452705). References Amundson, R., Guo, Y., Gong, P., 2003. Soil diversity and land use in the United States. Ecosystems 6 (5), 470–482. Bogunovic, I., Trevisani, S., Seput, M., et al., 2017. Short-range and regional spatial variability of soil chemical properties in an agro-ecosystem in eastern Croatia. Catena 154, 50–62. Bourennane, H., Salvador-Blanes, S., Couturier, A., et al., 2014. Geostatistical approach for identifying scale-specific correlations between soil thickness and topographic attributes. Geomorphology 220, 58–67. Cambardella, C.A., Moorman, T.B., Novak, J.M., et al., 1994. Field-scale variability of soil properties in Central Iowa soils. Soil Sci. 58, 1501–1511. Cao, J., 2011. Research on Water Movement and High-Efficient Utilization on Slope Characterized by a Soil-rock Binary Medium in Mountain Area of Taihang Mountain (in Chinese). Graduate University of Chinese Academy of Science. Chen, J., Zhang, X., Gong, Z., et al., 2001. Pedodiversity: a controversial concept (in Chinese with English abstract). Adv. Earth Science 16 (2), 189–193. Costantini, E.A., Dazzi, C., 2013. The Soils of Italy. Springer. Costantini, E.A.C., Malucelli, F., Brenna, S., et al., 2007. Using existing soil databases to consider paleosols in land planning: case study of the Lombardy region (northern Italy). Quat. Int. 162–163, 166–171. Crawford, C.A.G., Hergert, G.W., 1997. Incorporating spatial trends and anisotropy in geostatistical mapping of soil properties. Soil Sci. Soc. Am. J. 61 (1), 298–309. Danek, P., Samonil, P., Phillips, J.D., 2016. Geomorphic controls of soil spatial complexity in a primeval mountain forest in the Czech Republic. Geomorphology 273, 280–291. Duffera, M., White, J.G., Weisz, R., 2007. Spatial variability of Southeastern U.S. coastal plain soil physical properties: implications for site-specific management. Geoderma 137 (3–4), 327–339. Fajardo, M.P., McBratney, A.B., Minasny, B., 2017. Measuring functional pedodiversity using spectroscopic information. Catena 152, 103–114. Fu, W., Tunney, H., Zhang, C., 2010. Spatial variation of soil nutrients in a dairy farm and its implications for site-specific fertilizer application. Soil Tillage Res. 106 (2), 185–193. Fu, T., Chen, H., Zhang, W., et al., 2015. Spatial variability of surface soil saturated hydraulic conductivity in a small karst catchment of Southwest China. Environ. Earth Sci. 74 (3), 2381–2391. Fu, T., Chen, H., Fu, Z., et al., 2016. Surface soil water content and its controlling factors in a small karst catchment. Environ. Earth Sci. 75 (21), 1406. Fu, T., Han, L., Gao, H., et al., 2018. Pedodiversity and its controlling factors in mountain regions — a case study of Taihang Mountain, China. Geoderma 310, 230–237. Gholami, S., Sheikhmohamadi, B., Sayad, E., 2017. Spatial relationship between soil macrofauna biodiversity and trees in Zagros forests, Iran. Catena 159, 1–8 (Supplement C). Goovaerts, P., 2000. Geostatistical approaches for incorporating elevation into the spatial interpolation of rainfall. J. Hydrol. 228 (1), 113–129. Guo, Y., Gong, P., Amundson, R., 2003. Pedodiversity in the United States of America. Geoderma 117 (1), 99–115. Ibáñez, J.J., Bockheim, J.G., 2013. Pedodiversity. CRC Press. Ibañez, J.J., De-Albs, S., Bermúdez, F.F., et al., 1995. Pedodiversity: concepts and measures. Catena 24 (3), 215–232. Ibañez, J.J., De-Alba, S., Lobo, A., et al., 1998. Pedodiversity and global soil patterns at coarse scales (with discussion). Geoderma 83 (3), 171–192. Ibáñez, J.J., Caniego, J., San José, F., et al., 2005. Pedodiversity-area relationships for islands. Ecol. Model. 182 (3–4), 257–269. Ibáñez, J.-J., Sánchez-Díaz, J., Rodríguez-Rodríguez, A., et al., 2008. Preservation of
99
Contents lists available at ScienceDirect
Geoderma journal homepage: www.elsevier.com/locate/geoderma
Geostatistical analysis of pedodiversity in Taihang Mountain region in North China Tonggang Fua, Lipu Hana, Hui Gaoa,b, Hongzhu Lianga,b, Jintong Liua,
T
⁎
a Key Laboratory of Agricultural Water Resources, Center for Agricultural Resources Research, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, Shijiazhuang 050022, China b University of Chinese Academy of Sciences, Beijing 100049, China
A R T I C LE I N FO
A B S T R A C T
Handling editor: A.B. McBratney
The protection of pedodiversity is very important for biodiversity maintenance and food production. However, where and how large should the protection focus on in a given region (especially in mountain regions with complex environment) remains unclear. In this study, geostatistical analysis was used to study pedodiversity indexes (Shannon's index, the maximum Shannon's index, Richness index and Evenness index) in Taihang Mountain region in North China. Then pedodiversity protection and buffer zones were designed. Moreover, the relationship between protection zone and human activities was determined. The results showed that average Shannon's index for the study area was 1.72, with two lower value zones appeared in the north and the south of Taihang Mountain region, respectively. This suggested that there was the need to enhance the protection of pedodiversity in the two regions. The ranges of the Richness index and the Shannon's index were respectively 57.16 km and 96.00 km, indicating that pedodiversity had spatial dependence within these distances. Therefore, protection zones with radius of 57.16 km and buffer zones with radius of 96 km was designed. For both the northern and southern protection zones, the percent area of population density lower than 1 person/km2 (35.27% for the north and 51.71% for the south) was much higher than the average value (27.91%) of Taihang Mountain region. Furthermore, the percent area of farmland (43.72% for the north and 40.48% for the south) were higher than the average value (35.92%). This demonstrated that human activity, especially farming, was a key consideration in the protection of pedodiversity in the study region. The results of the study constitute a significant contribution to the theory of pedodiversity protection and soil resources management.
Keywords: Soil diversity Semivariance Land use Population density Protection zone
1. Introduction Pedodiversity, also known as soil diversity, is the basis for biodiversity. In spite of the increasing interest in recent years, the study of pedodiversity lags far behind that of biodiversity (Guo et al., 2003; Minasny et al., 2010; Fajardo et al., 2017; Fu et al., 2018). The concept of pedodiversity was originally proposed in the 1990s (Mcbratney, 1992; Ibañez et al., 1995). Since then, it has gradually attracted global scientific attention and focus (Amundson et al., 2003; Toomanian et al., 2006; Shangguan et al., 2014). Pedodiversity study reached a peak level in the 2010s, with the publication of the book Pedodiversity in Springer (Ibáñez and Bockheim, 2013) and the follow-up special issue in Geomorphology (Volume 135, Issues 3–4). Irrespectively, the study of pedodiversity still remains much weaker than that of biodiversity. This underscores the need for more studies on pedodiversity. Existing studies on pedodiversity have focused mainly on four areas.
First, pedodiversity index; which come from the methods used to study biodiversity. This mainly includes the Shannon's index, Richness index, Evenness index and their derivatives (Ibañez et al., 1995; Tan et al., 2003; Danek et al., 2016). Second, object abundance models; which describe the distribution of objects abundance (Ibañez et al., 1998; Toomanian et al., 2006; Kooch et al., 2015). Four main types of models always used for this purpose include geometric model, logarithmic model, logarithmic normal model and broken stick model (Ibañez et al., 1995). Each of these models is related to a distribution pattern of soil individuals. Third, pedodiversity-area or richness-area relationship; which relationship is given by power or logarithmic function (Guo et al., 2003; Ibáñez et al., 2005; Ren and Zhang, 2015). Pedodiversityarea relationship is closely related to the object abundance models, with the power curve accompanied by lognormal and broken-stick models, and the logarithmic expressions by geometric and logarithmic series models (Saldaña and Ibáñez, 2007). Fourth, rare and endangered
⁎ Corresponding author at: Key Laboratory of Agricultural Water Resources, Center for Agricultural Resources Research, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, 286 Huaizhong Road, 050021 Shijiazhuang, China. E-mail address: [email protected] (J. Liu).
https://doi.org/10.1016/j.geoderma.2018.05.010 Received 26 November 2017; Received in revised form 28 March 2018; Accepted 6 May 2018 0016-7061/ © 2018 Elsevier B.V. All rights reserved.
Geoderma 328 (2018) 91–99
T. Fu et al.
annual temperature is 11.4 °C, with the highest in July and lowest in January. The average annual precipitation is 457 mm, with the highest in July and lowest in December. Average elevation range is 1000–1500 m, with the northwest higher than the southeast. The slopes are generally steep and often over 25°. The main land use types include farmland (35.92%), forestland (29.05%), grassland (26.44%) and others (8.59%). The vegetation is influenced by the elevation, slope gradient, slope aspect and the interaction effects of these factors (Zhang et al., 2006). Soils are always thin in Taihang Mountain, with an average thickness of 35 cm (Cao, 2011). The soils are mainly developed from limestone in the northern and southern regions, but from gneiss in the central region. Based on the first-level FAO soil classification (Harmonized World Soil Database), 26 soil types exist in Taihang Mountain. The most common are Cambisols and the most uncommon are Podzoluvisols. Detailed information on soil types is given by Fu et al. (2018).
soils studies; which is mainly about soils that are significantly influenced by humans (Amundson et al., 2003; Shangguan et al., 2014; Tennesen, 2014). In summary, most of the existing studies on pedodiversity are concerned mainly with basic theories of pedodiversity. Only the studies of endangered soils provide information on soil protection. With the existing soil databases, land planning can be used to protect pedodiversity and cultural heritage (Costantini et al., 2007; Ibáñez et al., 2008). However, exactly where and to what extent is considered appropriate for sustainable protection remain largely unclear. Geostatistics, which is based on the theory of regionalized variables, is a branch of applied statistics that focuses on the detection, modeling and estimation of spatial patterns (Rossi et al., 1992; Goovaerts, 2000; Webster and Oliver, 2007). The most used methods in geostatistical analysis are semivariogram and kriging. From semivariogram, nugget (C0), sill (C0 + C), range and other parameters can be derived. The nugget-to-sill ratio (C0/(C0 + C)) reflects the degree of spatial dependence. Based on Cambardella et al. (1994), spatial dependence is strong when C0/(C0 + C) ratio < 25%, moderate when between 25 and 75% and weak when > 75%. The range reflects the correlation length beyond which the sampling points are independent. Because of the ease and practicability of geostatistical method, it is widely used in the study of soil moisture, soil nutrient, soil texture and other soil property analysis (Duffera et al., 2007; Bourennane et al., 2014; Fu et al., 2016; Bogunovic et al., 2017), as well as the study of biodiversity (Gholami et al., 2017; Liparoto et al., 2017). Nevertheless, geostatistics is hardly tried in the study of pedodiversity. The use of geostatistics in pedodiversity can be useful in understanding the spatial structures of pedodiversity and can provide meaningful information on pedodiversity protection zones. The study of pedodiversity began in China in 2001 (Chen et al., 2001). Since then, it has gradually intensified in Shandong Province, Nanjing City, Henan Province and other areas in the country (Tan et al., 2003; Zhang et al., 2007; Ren and Zhang, 2017). Besides regional studies, there also are studies which focus on the whole China (Zhang and Gong, 2004; Shangguan et al., 2014). However, mountain areas, which have complex environmental conditions, are largely ignored in these studies. Taihang Mountain, which is in North China, is an important boundary region between the Loess Plateau and the North China Plain (NCP). The special location of Taihang Mountain makes it an important ecological area (Yang et al., 2003). However, this mountain region is threatened by serious ecological degradation, including soil depletion, natural hazards, overgrazing, etc. (Li et al., 2004). The loss of pedodiversity is also a very serious problem in Taihang Mountain. There are more rare-unique soils and endangered soils in this mountain area than in other areas of China (Shangguan et al., 2014); requiring considerable attention for pedodiversity assessment, protection and sustainability. The objectives of this paper were to: (1) obtain the geostatistical characteristics of pedodiversity indexes, (2) determine protection zones and buffer zones, and (3) analyze the relationship between protection zones and human activities in the Taihang Mountain area, north China.
2.2. Pedodiversity indexes Richness index (S) defines the number of soil types within a given region, which can be a country, a catchment or even a regular grid of in the area. Another important pedodiversity index is the Shannon's index. It is based on proportional abundance of objects and it is expressed as: i=n
H = − ∑ pi × ln pi
(1)
i=1 th
where pi is the proportion of individual number of the i object (area of ith soil type) to the total number of individuals (total area of the given region). The value H varies between 1.5 and 3.5 and rarely exceeds 4.5 (Margalef, 1972). When all the objects (soil types) are evenly distributed, the Shannon's index is then maximum and is expressed as: (2)
Hmax = lnS
Actually, the Shannon's index rarely reaches its maximum value and the proportion of H to Hmax is the evenness index (E), expressed as:
E=
H Hmax
(3)
The range of E is 0–1 and higher E values means more even distribution of soil types. 2.3. Geostatistical analysis Semivariance and spatial interpolation (Kriging interpolation) are the two wildly used methods in geostatistical analysis. The semivariance (γ(h)) of a regionalized variable is expressed as (Webster and Oliver, 2007; Fu et al., 2015):
γ(h) =
2. Materials and methods
1 2N (h)
N (h)
∑
[Z (x i ) − Z (x i + h)]2
i=1
(4)
In the formula, N(h) is the paired points at lag distance h; Z(i) is the value of the variable at point xi; and Z(xi + h) is the value at a point which is h meters away from point xi. The relationship between semivariance and lag distance is always simulated by Spherical model, Exponential model or Gaussian model as:
2.1. Site description Taihang Mountain is located in North China, extending in the southwest-northeast direction (Fig. 1). Taihang Mountain region extents across a total of 4 provinces (Beijing, Hebei, Shanxi and Henan) and 101 counties. It is an important transition zone between the Loess Plateau and NCP. Because of its special location and orogeny, the western and eastern parts of Taihang Mountain are very different in structural evolutions, sedimentation processes and tectonic settings (Wang and Li, 2008). Taihang Mountain is influenced by the East Asian Monsoon climate. The summer is warm and rainy, and the winter is cold and dry. Based on 10-year (2005–2014) meteorological data in the region, average
γ(h) =
⎧ C0 + C ⎨ ⎩ C0 + C
(
3h 2a
−
(
h3 a3
92
0≤h≤a
h≥a h
γ(h) = C0 + C 1 − e− a γ(h) = C0 + C ⎜⎛1 − e ⎝
)
)
(5) (6)
2
− h2 ⎞ a ⎟
⎠
(7)
Geoderma 328 (2018) 91–99
T. Fu et al.
Fig. 1. A map depicting the location of the study area in China (inset) and an expanded map of the study area (main plate).
where γ(h) is the semivariance; h is the lag distance; C0 is the nugget; and C is the sill. The degree of spatial dependence of a variable is reflected by the ratio of C0 to C0 + C. When the ratio is < 25%, the spatial dependence is strong. Then when the ratio is between 25 and 75%, the spatial dependence is moderate. Yet when it is higher than 75%, the spatial dependence is weak (Cambardella et al., 1994). Semivariance can be different in different directions (Journel and Huijbregts, 1978). The semivariance of regional variable in a given direction can be calculated by:
γ(h, φ) =
1 2N (h, φ)
values were assigned to the center point of each grid, which computation was completed in ArcMap 9.3 (Environmental Systems Research Institute, Inc. (ESRI)). Then pedodiversity and geographical information of the points were imported to GS+ (Geostatistics for the Environmental Sciences, version 9.0) to calculate the semivariance, nugget, sill and range. The most suitable semivariance model was also simulated in GS+ software. In addition, anisotropy ratio between the 30° direction (parallel to the trend of the mountain) and the 120° direction (perpendicular to the trend of the mountain) was calculated. The spatial distribution map of the pedodiversity indexes and pedodiversity protection and buffer zones were drawn in ArcMap 9.3. The data source of elevation, land used type, and population density were introduced in Fu et al. (2018).
N (h, φ)
∑ i=1
[Z (x i ) − Z (x i + h)]2
(8)
Here, γ(h,φ) is the semivariance in φ direction; and N(h,φ) is the number of paired points with lag distance h in φ direction. Considering different directions, the anisotropy ratio is given as:
K(h) =
γ (h, φ1 ) γ (h, φ2 )
3. Results 3.1. Classical statistical characteristics of pedodiversity indexes
(9) The range of H value in Taihang Mountain region was 1.01–2.19 with mean and CV (Coefficient of Variation) of respectively 1.72 and 0.17, indicating a moderate variation (0.1 < CV < 1). The range of Hmax was 2.48–3.14 with mean and CV of respectively 2.87 and 0.05, showing a weak variation (CV < 0.1). The CV value of Hmax was the smallest among the four investigated indexes. The lowest S value was 12, highest 23, mean 17.75 and the CV 0.13; suggesting a moderate variation. The range of E was 0.37–0.78, with a mean value of 0.60 and CV value of 0.15; also indicating a moderate variation (Table 1). All the four indexes had normal distributions. Significant Kolmogorov-Smirnov test values were respectively 0.06, 0.19, 1.20 and 0.08, with a normal distribution for values higher than 0.05.
In the above formula, γ(h,φ1) is the semivariance in φ1 direction; and γ(h,φ2) is the semivariance in φ2 direction. 2.4. Data analysis The soil taxa data for Taihang Mountain were clipped from the Harmonized World Soil Database (HWSD) which describes global soil taxa. HWSD is a comprehensive database that combines vast volumes of regional and national updates of soil information with information already contained within the 1:5,000,000 scale FAO-UNESCO Digital Soil Map of the World. The work was done by the Food and Agriculture Organization of the United Nations (FAO) and the International Institute for Applied Systems Analysis (IIASA). The database is a raster with a resolution of about 1 km (30 arc sec by 30 arc sec). Every pixel has an attribute that noted the soil type. A bigger grid layer of 31.6 km × 31.6 km scale (1000 km2) was overlaid on the soil taxa map. Then the pedodiversity indexes (S, H, Hmax and E) were calculated for the individual 1000 km2 grids. The calculated pedodiversity index
3.2. Geostatistical characteristics of pedodiversity indexes Based on geostatistical analysis, the different indexes had different suitable models. The best-fit models were the Spherical model for the H index and the Gaussian model for and the Hmax and S indexes. However, 93
Geoderma 328 (2018) 91–99
T. Fu et al.
spatial dependence. The range of E was the largest (174.3 km) and that of Hmax (56.12 km) was the smallest. The ranges of S were similar with that of Hmax, with a value of 57.16 km (Table 2).
Table 1 Classical statistical analysis of pedodiversity indexes for Taihang Mountain region in North China. Minimum
Shannon's index Maximum Shannon's index Richness index Evenness index
Maximum
Mean
Standard deviation
CV
3.3. Anisotropy of pedodiversity indexes
1.01 2.48
2.19 3.14
1.72 2.87
0.29 0.13
0.17 0.05
12.00 0.37
23.00 0.78
17.75 0.60
2.35 0.09
0.13 0.15
The semivariance of the pedodiversity indexes were different in the two investigated directions (the 30° and 120° directions). Anisotropic ratio of H and E increased first and then decreased with increasing lag distance (Fig. 2a and d). It was higher than 1 when the lag distance was about 100–300 km, indicating that H and E variance in the 30° direction was lower than that in the 120° direction. The highest anisotropic ratio was 2.67, which was in the 232.56 km lag distance. It suggested that the difference in variance between the two directions was the highest for this lag distance. For Hmax and S indexes, anisotropic ratio gradually increased with increasing separation distance. It was < 1 when the separation distance was < 133 km, implying that the variation in the 120° direction was higher than that in the 30° direction (Fig. 2b and c). Moreover, the difference decreased with increasing separation distance. When the separation distance was higher than 133 km, the variance in the 30° direction exceeded that in the 120° direction.
Note: CV is coefficient of variation. Table 2 Geostatistical analysis of pedodiversity indexes for Taihang Mountain region in North China.
Shannon's index Maximum Shannon's index Richness index Evenness index
Best-fit model
C0
C0 + C
C0/(C0 + C)
Range (km)
R2
Spherical model Gaussian model
0.003
0.086
0.03
96.00
0.46
0.002
0.018
0.11
56.12
0.30
Gaussian model Exponential model
0.630
5.568
0.11
57.16
0.33
3.4. Spatial distribution of pedodiversity indexes
0.001
0.009
0.11
174.3
0.54
The spatial distribution of E was similar to that of H, both apparently fluctuating from north to south. There were two low value zones separately in the northern and southern regions of Taihang Mountain (Fig. 3a and d). The spatial distributions of Hmax and S were also similar, with higher zones occurring mainly in the north and lower zones in the south (Fig. 3b and c).
Note: R2 is Coefficient of determination; Co is nugget; Co + C is sill; and Co/ (Co + C) is nugget-sill ratio.
the best-fit model for the E index was the Exponential model. All the indexes had nugget effect, with nugget values of 0.003 (H index), 0.002 (Hmax index), 0.630 (S index) and 0.009 (E index). The sill values of the H, Hmax, S and E indexes were respectively 0.086, 0.018, 5.568 and 0.009. The nugget/sill values for Hmax, S and E were the same (all equal to 0.11), suggesting a moderate spatial dependence. However, the nugget/sill value for H was much smaller (0.03), indicating a strong
3.5. Spatial distribution of protection and buffer zones The H and E indexes each had two lower zones in the northern and southern parts of the study area. This suggested that protection zones and buffer zones should be designed around these area (Fig. 4a). The radii of the protection zone and the buffer zone were determined
Fig. 2. The anisotropy of pedodiversity indexes between 30° and 120° directions. a is Shannon's index, b is maximum Shannon's index, c is Richness index and d is Evenness index. 94
Geoderma 328 (2018) 91–99
T. Fu et al.
Fig. 3. Maps of Ordinary Kriging for spatial distributions of H (Shannon's index), Hmax (the maximum Shannon's index), S (Richness index) and E (Evenness index).
geostatistical analysis has already been used in the study of biodiversity, which is much similar to pedodiversity, and the semivariances of abundance, richness, Shannon's index, etc. were calculated (Gholami et al., 2017; Liparoto et al., 2017). Therefore, geostatistical methods were used in this study to analyze pedodiversity in the Taihang Mountain region. Pedodiversity indexes had a strong spatial dependence in Taihang Mountain region. The degrees of spatial dependence of soil properties (such as soil thickness, soil moisture, soil nutrients, soil texture etc.) were always different for different regions or scales (Duffera et al., 2007; Bourennane et al., 2014; Yang et al., 2016). The strong spatial dependence of pedodiversity in the study area was attributed to the fact that pedodiversity is less sensitive to the influencing factors than other soil properties. For example, the spatial structure of soil moisture is easily influenced by precipitation, irrigation or other processes (Yang et al., 2016); soil bulk density and porosity change after plowing or vehicle compaction (Fu et al., 2015). These processes hardly affect
respectively by the ranges of the richness index (57.16 km) and Shannon's index (96.00 km). Then the protection zones covered mainly 10 counties (Fig. 4b) which needed monitoring for pedodiversity protection. One of the two buffer zones covered 6 counties and the other 17 counties. The areas and counties under the two protection and buffer zones are shown in Table 3, which provide useful information for pedodiversity protection. 4. Discussions 4.1. Geostatistical method for pedodiversity analysis Soil properties, such as soil moisture, soil nutrient and soil texture, are regionalized variables and can be analyzed by geostatistical method (Duffera et al., 2007; Bourennane et al., 2014; Fu et al., 2016; Bogunovic et al., 2017). Pedodiversity could be considered as individual soils attribute (Saldaña and Ibáñez, 2007). Moreover, 95
Geoderma 328 (2018) 91–99
T. Fu et al.
Fig. 4. Maps of protection zones and buffer zones in Taihang Mountain region. a is the distribution of Shannon's index, b is the distribution of counties, c is the distribution of population density, and d is the distribution of land use types.
methods based on regular grids in this study had the same trend with that calculated by classical approach based on political boundary (Fu et al., 2018). This also indicates that geostatistic analysis is appropriate in the study of pedodiversity in Taihang Mountain.
pedodiversity in the short run. Therefore, the spatial dependence of pedodiversity can be strong. Pedodiversity indexes (e.g., richness index and Shannon's index) were derived from biodiversity analysis methods (Ibañez et al., 1995), but biodiversity is more sensitive to the influencing factors. For example, the diversity of neuston community showed a moderate to week spatial dependence (Liparoto et al., 2017). The spatial distribution of the four considered pedodiversity indexes were similar (Fig. 4). There were two lower zones, with one in the northern and the other in the southern part of Taihang Mountain. The spatial distribution of pedodiversity indexes calculated by geostatistical
4.2. Anisotropy of pedodiversity and its influencing factors Anisotropy has always existed in soils; saying that soils vary in different directions (Crawford and Hergert, 1997). Anisotropy of soil properties, such as soil physical properties, soil nutrient properties and 96
Geoderma 328 (2018) 91–99
T. Fu et al.
(2015). The trend for Taihang Mountain may cause the different variations of soil-forming factors in the two directions. As a result, anisotropy of pedodiversity could be caused by the differences in variation of the soilforming factors. Most of the soil-forming factors considered in this study (e.g., elevation, precipitation and temperature) had higher variations in the 120° direction than in the 30° direction (Fig. 5). This is due to the trend in elevation, which had a clear gradient from northwest to southeast. Most of the soil-forming factors had a closed correlation with elevation. Therefore, the natural factors had a clear anisotropy. However, the variances of pedodiversity indexes in the 120° direction were higher than those in the 30° direction for separation distances smaller than 100–150 km but lower for separation distances larger than 150 km (Fig. 2), which was different for the soil-forming factors. This could be due to the fact that many influencing factors were not considered. For example, farming and population density have been noted as very important factors influencing pedodiversity (Fu et al., 2015). These factors have poor spatial structure and always distributed randomly. This indicates that when the lag distance is < 100–150 km, controlling factors can be natural. But when the distance is longer than 100–150 km, the anthropogenic factors become more important.
Table 3 Counties in the protection and buffer zones in Taihang Mountain region in North China. Position
Zone
Area (km2)
Main county
North
Pz Bz
7476 17,961
South
Pz
10,264
Bz
27,710
Yuxian1, Yangyuan,Guangling Zhoulu, Laishui, Laiyuan, Lingqiu, Hunyuan, Yixian Jincheng, Zezhou, Lingchuan, Changzhi, Zhangzi, Qinshui, Yangcheng Mengzhou, Qinyang, Bo'ai, Xiuwu, Haozuo, Huixian, Huguan, Pingshun, Lucheng, Changzhishi, Tunliu, Anze, Fushan, Yicheng, jiangxiang, Yuanqu, Jiyuan.
Note: Pz is protection zone and Bz is buffer zone.
soil microorganisms, has been extensively studied (Fu et al., 2010; Fu et al., 2015; Wu et al., 2016). Pedodiversity, as aforementioned, is a form of soil property and is anisotropic (Fig. 2). The reason for anisotropy in the study area could be due to the southwest-northeast orientation of Taihang Mountain (Fig. 1). The distance from west to east is about 100–150 km, within which point pairs were equal for the two directions. With higher distances, however, point pairs hardly existed in the 120° direction, and the variation can be smaller. A similar anisotropic form of soil physical properties was also noted by Fu et al.
Fig. 5. The anisotropy of soil-forming factors: a is elevation, b is slope gradient, c is precipitation, d is temperature, e is population density and f is percent farmland. 97
Geoderma 328 (2018) 91–99
T. Fu et al.
Table 4 Population density in protection and buffer zones in Taihang Mountain region in North China. Position
Zone
<1 (person/km2)
P<1 (%)
1–25 (person/km2)
P1–25 (%)
26–100 (person/km2)
P26–100 (%)
> 100 (person/km2)
P > 100 (%)
North
Pz Bz Pz Bz
1787 6258 1815 5404 38,222
23.90 34.84 17.68 19.50 27.91
951 2399 1045 2589 15,151
12.72 13.36 10.18 9.34 11.06
2101 3957 2128 5389 25,670
28.10 22.03 20.73 19.45 18.74
2637 5347 5276 14,328 57,903
35.27 29.77 51.40 51.71 42.28
South Total
Note: Pz is protection zone, Bz is buffer zone, P < 1 is the percentage of the population density < 1 person/km2, P1–25 is the percentage of the population density between 1 person/km2 to 25 persons/km2, P26–100 is the percentage of population density between 26 persons/km2 to 100 persons/km2, and P > 100 is the percentage of population density higher than 100 persons/km2. Table 5 Land use types in the protection and buffer zones. Position
Zone
Fa (km2)
PFa (%)
Fo (km2)
PFo (%)
Gr (km2)
PGr (%)
Wa (km2)
PWa (%)
Co (km2)
PCo (%)
Un (km2)
PUn (%)
North
Pz Bz Pz Bz
3269 5776 4156 11429 49167
43.72 32.16 40.48 41.24 35.92
1692 5893 4110 10100 39773
22.64 32.81 40.04 36.45 29.05
2124 5415 1288 3879 36188
28.41 30.15 12.55 14.00 26.44
77 162 46 319 1725
1.03 0.90 0.45 1.15 1.26
305 702 661 1966 9831
4.07 3.91 6.44 7.09 7.18
10 13 4 17 210
0.13 0.07 0.04 0.06 0.15
South Total
Note: Pz is protection zone, Bz is buffer zone, Fa is farmland, PFa is the percentage of farmland, Fo is forestland, PFo is the percentage of forestland, Gr is grassland, PGr is the percentage of grassland, Wa is water land, PWa is the percentage of water land, Co is construction land, PCo is the percentage of construction land, Un is undeveloped land, and PUn is the percentage of undeveloped land.
4.3. The design of protection and buffer zones
4.4. Correlation between delineated protection area and human activities
The concept of protection and buffer zones first appeared in the 1970s in Man and Biosphere (MAB) and in Biosphere Reserves (BRs) (Martino, 2001). More recently, however, the terms have been used mainly in the study of national park, which include three rings. From the inner to the outside, the rings are protection zone (core zone), buffer zone and transmission zone (Mannetti et al., 2015; Squeo et al., 2016). In this paper, the terms protection zone and buffer zone were used in the context of pedodiversity and with different meanings as used in the context of national park. Protection zone in this study denotes areas with low pedodiversity (Fig. 4) and therefore requiring protection by government and residents. Buffer zone denotes areas where pedodiversity is closely related with protection zones. No further damage is allowed in these zones and the existing conditions should be maintained. Studies have also designed protection zones or buffer zones in protection of wetlands, forestlands or in basic biodiversity protection (Semlitsch and Bodie, 2003; Kueffer and Kaiser-Bunbury, 2014; Robalino et al., 2015). Numerous methods are available for delineating protection zone and buffer zone boundaries, including statistical habitat model and the least cost distance method (Li et al., 1999; Li and Liu, 2006). In this study, geostatistical method was used to delineate protection zone and buffer zone boundaries because the range of an index represents the length of spatial correlation (Fu et al., 2015; Liparoto et al., 2017). The radii of the protection and buffer zones were determined by the ranges of S and H indexes. The S index was the most sensitive index and directly influenced by human disturbance. Amundson et al. (2003) noted that 31 soils were essentially “extinct” due to farming, urbanization and other activities. Therefore, the correlation length of S could reflect protection zone. The H index considers both species and evenness (Ibañez et al., 1995) and is therefore less sensitive than S and its correlation length could reflect buffer zone. Based on these features, protection and buffer zone boundaries were delineated (Fig. 4), which was critical for soil protection in the study area.
In the protection zones, the P < 1 value (percent area with population density < 1 person/km2) was lower than that for the whole Taihang Mountain area (Table 4). Apparently, pedodiversity was negatively related with population density. Moreover, studies show that human activities can limit pedodiversity (Lo Papa et al., 2011; Tennesen, 2014). However, the P > 100 value (percent area with population density higher than 100 people/km2) was higher in the southern protection zone, but lower in the northern protection zone than the whole study area (Table 4). This was attributed to the fact that industries or tourist centers always have high population density, but relatively little effect on pedodiversity (Fu et al., 2018). In both of the two protection zones, percent farmland was higher than other land use types. Moreover, it was higher than average percent farmland for the whole Taihang Mountain (Table 5). This was due to the fact that farming always influences pedodiversity (Costantini and Dazzi, 2013; Tennesen, 2014). Study also show that farming has a significant negative effect on pedodiversity in Taihang Mountain (Fu et al., 2018). Besides farmland, forestland also affects pedodiversity and forest ecosystems have a greater deal of soil variability than non-forest ecosystems (Kooch et al., 2015). In this study, however, it was interesting to know that the southern protection zone had a much higher percent forestland (40.04%) than the average value (29.05%) of the Taihang Mountain region (Table 5). This was probably because the southern protection zone had lower elevation range, which has a significantly positive effect on pedodiversity (Fu et al., 2018). In the buffer zone, the effect of farming was apparently lower than that in the protection zone because the northern buffer zone had a lower percent farmland than the whole Taihang Mountain region (Table 5). This again demonstrated the need for more attention on protection zone.
5. Conclusions Spatial distribution of pedodiversity index clearly showed two lower patches in Taihang Mountain, highlighting the need for much closer monitoring for pedodiversity protection. Geostatistical analysis showed a clear spatial structure of pedodiversity index in the region. Therefore, 98
Geoderma 328 (2018) 91–99
T. Fu et al.
based on spatial correlation distance (range), boundaries for the protection and buffer zones were delineated. The location of the protection zones was closely related to farmland, suggesting that farming significantly influenced pedodiversity in the mountain region. Therefore, there was the need to restrict serious disturbances by farming in the protection zones in order to protect pedodiversity. The results of the study showed the spatial characteristics of pedodiversity and provided the critical basis for the delineation of pedodiversity protection zones. This was critical for pedodiversity protection and soil resource management in the mountain region with a complex terrain.
European soils: natural and cultural heritage. The soils of tomorrow. Adv. Geoecol. 39, 37–59. Journel, A.G., Huijbregts, C.J., 1978. Mining geostatistics. London Academic press. Kooch, Y., Hosseini, S.M., Scharenbroch, B.C., et al., 2015. Pedodiversity in the Caspian forests of Iran. Geoderma Reg. 5, 4–14. Kueffer, C., Kaiser-Bunbury, C.N., 2014. Reconciling conflicting perspectives for biodiversity conservation in the Anthropocene. Front. Ecol. Environ. 12 (2), 131–137. Li, J., Liu, X., 2006. Research of the nature reserve zonation based on the least, cost distance model. J. Nat. Resour. Policy Res. 21 (2), 217–224. Li, W., Wang, Z., Ma, Z., et al., 1999. Designing the core zone in a biosphere reserve based on suitable habitats: Yancheng biosphere reserve and the red crowned crane (Grus japonensis). Biol. Conserv. 90 (3), 167–173. Li, Y., Xu, H., Wang, Z., 2004. A talk on ecological rehabilitation of Taihang mountainous area in Hebei (in Chinese with English abstract). In: Soil And Water Conservation In China. 11. pp. 10–12. Liparoto, A., Mancinelli, G., Belmonte, G., 2017. Spatial variation in biodiversity patterns of neuston in the Western Mediterranean and Southern Adriatic seas. J. Sea Res. 129, 12–21 (Supplement C). Lo Papa, G., Palermo, V., Dazzi, C., 2011. Is land-use change a cause of loss of pedodiversity? The case of the Mazzarrone study area, Sicily. Geomorphology 135 (3–4), 332–342. Mannetti, L., Zeller, U., Esler, K., 2015. Evaluating Social-Ecological Aspects of Buffer Zones at the Borders of Etosha National Park, Namibia. USDA Forest Service Proceedings RMRS-P-74pp. 129–132. Margalef, R., 1972. Homage to Evelyn Hutchinson, or why there is an upper limit to diversity. Trans. Connecticut Acad. Arts Sci. 44, 211–235. Martino, D., 2001. Buffer zones around protected areas: a brief literature review. Electron. Green J. 1 (15). Mcbratney, A.B., 1992. On variation, uncertainty and informatics in environmental soil management. Soil Res. 30 (6), 913–935. Minasny, B., McBratney, A.B., Hartemink, A.E., 2010. Global pedodiversity, taxonomic distance, and the World Reference Base. Geoderma 155 (3–4), 132–139. Ren, Y., Zhang, X., 2015. Study on pedodiversity: status quo and future challenges (in Chinese with English abstract). Acta Pedol. Sin. 52 (1), 9–18. Ren, Y.Y., Zhang, X., 2017. Pedodiversity patterns based on terrain in Henan Province (in Chinese with English abstract). Chin. J. Soil Sc. 48 (1), 22–31. Robalino, J., Sandoval, C., Barton, D.N., et al., 2015. Evaluating interactions of Forest conservation policies on avoided deforestation. PLoS One 10 (4), e0124910. Rossi, R.E., Mulla, D.J., Journel, A.G., et al., 1992. Geostatistical tools for modeling and interpreting ecological spatial dependence. Ecol. Monogr. 62 (2), 277–314. Saldaña, A., Ibáñez, J., 2007. Pedodiversity, connectance and spatial variability of soil properties, what is the relationship? Ecol. Model. 208 (2), 342–352. Semlitsch, R.D., Bodie, J.R., 2003. Biological Criteria for Buffer Zones around Wetlands and Riparian Habitats for Amphibians and Reptiles. Conserv. Biol. 17 (5), 1219–1228. Shangguan, W., Gong, P., Liang, L., et al., 2014. Soil diversity as affected by land use in China: consequences for soil protection. Sci. World J. 2014. Squeo, F.A., Loayza, A.P., López, R.P., et al., 2016. Vegetation of Bosque Fray Jorge National Park and its surrounding matrix in the Coastal Desert of North-central Chile. J. Arid Environ. 126, 12–22 (Supplement C). Tan, M., Zhang, X., Chen, J., et al., 2003. Pedodiversity: a case study based on 1:1 million scale SOTER of Shandong Province, China. Pedosphere 13 (3), 219–226. Tennesen, M., 2014. Rare Earth. Science 346 (6210), 692–695. Toomanian, N., Jalalian, A., Khademi, H., et al., 2006. Pedodiversity and pedogenesis in Zayandeh-rud Valley, Central Iran. Geomorphology 81 (3–4), 376–393. Wang, Y., Li, H., 2008. Initial formation and Mesozoic tectonic exhumation of an intracontinental tectonic belt of the northern part of the Taihang Mountain belt, eastern Asia. J. Geol. 116 (2), 155–172. Webster, R., Oliver, M.A., 2007. Geostatistics for Environmental Scientists. Wiley. Wu, Y., Li, Y., Fu, X., et al., 2016. Three-dimensional spatial variability in soil microorganisms of nitrification and denitrification at a row-transect scale in a tea field. Soil Biol. Biochem. 103 (Supplement C), 452–463. Yang, Y.H., Watanabe, M., Wang, Z.P., et al., 2003. Prediction of changes in soil moisture associated with climatic changes and their implications for vegetation changes: WAVES model simulation on Taihang Mountain, China. Clim. Chang. 57 (1–2), 163–183. Yang, J., Chen, H., Nie, Y., et al., 2016. Spatial variability of shallow soil moisture and its stable isotope values on a karst hillslope. Geoderma 264, 61–70. Zhang, X., Gong, Z., 2004. A pedodiversity pattern: taxonomically established soil orders in China. J. Geogr. Sci. 14 (1), 52–56. Zhang, J., Xi, Y., Li, J., 2006. The relationships between environment and plant communities in the middle part of Taihang Mountain Range, North China. Community Ecol. 7 (2), 155–163. Zhang, X., Chen, J., Tan, M., et al., 2007. Assessing the impact of urban sprawl on soil resources of Nanjing city using satellite images and digital soil databases. Catena 69 (1), 16–30.
Acknowledgements This work was supported by the National Basic Research Program (973 Program) of China (No. 2015CB452705). References Amundson, R., Guo, Y., Gong, P., 2003. Soil diversity and land use in the United States. Ecosystems 6 (5), 470–482. Bogunovic, I., Trevisani, S., Seput, M., et al., 2017. Short-range and regional spatial variability of soil chemical properties in an agro-ecosystem in eastern Croatia. Catena 154, 50–62. Bourennane, H., Salvador-Blanes, S., Couturier, A., et al., 2014. Geostatistical approach for identifying scale-specific correlations between soil thickness and topographic attributes. Geomorphology 220, 58–67. Cambardella, C.A., Moorman, T.B., Novak, J.M., et al., 1994. Field-scale variability of soil properties in Central Iowa soils. Soil Sci. 58, 1501–1511. Cao, J., 2011. Research on Water Movement and High-Efficient Utilization on Slope Characterized by a Soil-rock Binary Medium in Mountain Area of Taihang Mountain (in Chinese). Graduate University of Chinese Academy of Science. Chen, J., Zhang, X., Gong, Z., et al., 2001. Pedodiversity: a controversial concept (in Chinese with English abstract). Adv. Earth Science 16 (2), 189–193. Costantini, E.A., Dazzi, C., 2013. The Soils of Italy. Springer. Costantini, E.A.C., Malucelli, F., Brenna, S., et al., 2007. Using existing soil databases to consider paleosols in land planning: case study of the Lombardy region (northern Italy). Quat. Int. 162–163, 166–171. Crawford, C.A.G., Hergert, G.W., 1997. Incorporating spatial trends and anisotropy in geostatistical mapping of soil properties. Soil Sci. Soc. Am. J. 61 (1), 298–309. Danek, P., Samonil, P., Phillips, J.D., 2016. Geomorphic controls of soil spatial complexity in a primeval mountain forest in the Czech Republic. Geomorphology 273, 280–291. Duffera, M., White, J.G., Weisz, R., 2007. Spatial variability of Southeastern U.S. coastal plain soil physical properties: implications for site-specific management. Geoderma 137 (3–4), 327–339. Fajardo, M.P., McBratney, A.B., Minasny, B., 2017. Measuring functional pedodiversity using spectroscopic information. Catena 152, 103–114. Fu, W., Tunney, H., Zhang, C., 2010. Spatial variation of soil nutrients in a dairy farm and its implications for site-specific fertilizer application. Soil Tillage Res. 106 (2), 185–193. Fu, T., Chen, H., Zhang, W., et al., 2015. Spatial variability of surface soil saturated hydraulic conductivity in a small karst catchment of Southwest China. Environ. Earth Sci. 74 (3), 2381–2391. Fu, T., Chen, H., Fu, Z., et al., 2016. Surface soil water content and its controlling factors in a small karst catchment. Environ. Earth Sci. 75 (21), 1406. Fu, T., Han, L., Gao, H., et al., 2018. Pedodiversity and its controlling factors in mountain regions — a case study of Taihang Mountain, China. Geoderma 310, 230–237. Gholami, S., Sheikhmohamadi, B., Sayad, E., 2017. Spatial relationship between soil macrofauna biodiversity and trees in Zagros forests, Iran. Catena 159, 1–8 (Supplement C). Goovaerts, P., 2000. Geostatistical approaches for incorporating elevation into the spatial interpolation of rainfall. J. Hydrol. 228 (1), 113–129. Guo, Y., Gong, P., Amundson, R., 2003. Pedodiversity in the United States of America. Geoderma 117 (1), 99–115. Ibáñez, J.J., Bockheim, J.G., 2013. Pedodiversity. CRC Press. Ibañez, J.J., De-Albs, S., Bermúdez, F.F., et al., 1995. Pedodiversity: concepts and measures. Catena 24 (3), 215–232. Ibañez, J.J., De-Alba, S., Lobo, A., et al., 1998. Pedodiversity and global soil patterns at coarse scales (with discussion). Geoderma 83 (3), 171–192. Ibáñez, J.J., Caniego, J., San José, F., et al., 2005. Pedodiversity-area relationships for islands. Ecol. Model. 182 (3–4), 257–269. Ibáñez, J.-J., Sánchez-Díaz, J., Rodríguez-Rodríguez, A., et al., 2008. Preservation of
99