ISSN 1794-6190 e-ISSN 2339-3459 https://doi.org/10.15446/esrj.v29n1.118386 EARTH SCIENCES RESEARCH JOURNAL Earth Sci. Res. J. Vol. 29, No. 1 (March, 2025): 45 - 53 G EO M O R PH O LO G Y Transitional relationships among typical loess geomorphologic types: Loess Plateau of China Shifang Zhang1, Shangmin Zhao2* 1 School of Traffic Engineering, Shanxi Vocational University of Engineering Science and Technology, Jinzhong 030619, China; 2 Department of Surveying and Mapping, College of Geological and Surveying Engineering, Taiyuan University of Technology, Taiyuan, 030024, China * Email: zhaoshangmin@tyut.edu.cn How to cite item: Zhang, S., & Zhao, S. (2025). Transitional relationships among typical loess geomorphologic types: Loess Plateau of China. Earth Sciences Research Journal, 29(1), 45-53. https://doi.org/10.15446/esrj. v29n1.118386 Record Manuscript received: 17/11/2024 Accepted for publication: 10/02/2025 ABST R AC T Taking the distribution area of typical loess geomorphologic types in the Loess Plateau of China as the study area, this research aims to explore the transitional relationships among typical loess geomorphologic types based on their cons- tructed digital index systems using the data sources of digital geomorphologic database, Landsat TM/ETM+ images and SRTM1 DEM data. Based on Landsat TM/ETM+ images, the digital geomorphologic database was firstly checked and improved using visual interpretation combined with expert knowledge methods, so as to acquire the precise spatial distribution of the typical loess geomorphologic types. Then, the digital index systems for the typical loess geomor- phologic types were constructed, which included topographic indexes, quantitative indexes from topographic profile method and remote sensing indexes. Finally, the transitional relationships among the typical loess geomorphologic types were explored based on the constructed digital index systems and the spatial distribution pattern on the remote sensing images. The research results show that loess tableland may transform to loess ridge or loess knoll, and loess ridge may transform to loess knoll. As to morphologic types, loess tableland may transform to loess residual tableland, and then to loess beam tableland; loess wide acclivitous ridge may transform to loess narrow acclivitous ridge, and then to loess knoll ridge. For the valley shape types of the loess ridge and loess knoll, the shallow and high valley may transform to deep and high valley, and then to shallow and low valley. This research is meaningful in the fields such as digital mapping of loess geomorphology, water and soil loss, soil erosion and digital topographic analysis. Keywords: transitional relationship; digital index systems; typical loess geomorphologic type; digital geomorphologic database; Landsat TM/ETM+ images; SRTM1 DEM data; Loess Plateau Palabras clave: Relación transicional; sistemas de indexación digital; tipos geomorfológicos de loess; bases de datos geomorfológicas digitales; imágenes Landsat TM/ETM+; información SRTM1 DEM; Meseta de Loess Relaciones de transición entre tipos geomorfológicos típicos de loess: meseta de loess de China Este trabajo toma como objeto de estudio el área de distribución de los tipos geomorfológicos de loess típicos en la Meseta de Loess de China, con el objetivo de explorar las relaciones de transición entre los tipos geomorfológicos de loess típicos en función de sus sistemas de índices digitales a través de bases de datos geomorfológicos digitales, imágenes Landsat TM/ ETM+ y datos DEM SRTM1. Con base en las imágenes Landsat TM/ETM+, la base de datos geomorfológica digital fue verificada y mejorada a través de interpretación visual combinada con métodos de expertos para adquirir la distribución espacial precisa de los tipos geomorfológicos típicos de loess. Luego se construyeron los sistemas de indexación digital para esta clase de loess, los cuales incluyen índices topográficos, índices cuantitativos tomados de métodos de perfil topo- gráfico e índices remotos. Finalmente, se exploraron las relaciones transicionales entre los tipos gemorfológicos de loess con base en los sistemas de indexación digital construidos y los patrones de distribución espacial en las imágenes de de- tección remota. Los resultados de la investigación muestran que los loess tipo meseta pueden transformarse en loess tipo crestas o montículos, y las crestas pueden transformarse en loess tipo montículos. En cuanto a los tipos morfológicos, los loess tipo meseta pueden transformarse en meseta residual y luego en meseta viga; los loess tipo cresta ancha e inclinada pueden transformarse en cresta estrecha e inclinada y luego en cresta de montículo. Para los tipos de forma de valle de la cresta loess y montículo loess, el valle superficial y alto puede transformarse en valle profundo y alto, y luego en valle superficial y bajo. Esta investigación es significativa en campos como el mapeo digital de la geomorfología del loess, la pérdida de agua y suelo, la erosión del suelo y el análisis topográfico digital. RESU M EN https://doi.org/10.15446/esrj.v29n1.118386 mailto:zhaoshangmin@tyut.edu.cn https://doi.org/10.15446/esrj.v29n1.118386 https://doi.org/10.15446/esrj.v29n1.118386 46 Shifang Zhang, Shangmin Zhao 1. Introduction The popularization and advances of Geographic Information System (GIS) and remote sensing technologies have mean that they were widely used in geomorphologic researches (Remondo & Oguchi, 2009). Based on the digital elevation model (DEM), data and digital topographic analysis, spatial distribution of volcanic landforms, complexity and distribution pattern of natural landform, distribution characteristics of Aeolian sand dune and other geomorphologic- related fields have been extensively studied (Werner, 1999; Kocurek & Ewing, 2005; Bishop, 2007a, 2007b, 2009; Bullard et al., 2011). Meanwhile, topographic indexes, such as slope, fluvial knick zone and shape, are extracted from DEM data, whose spatial distribution and characteristics are analyzed specifically (Saito & Oguchi, 2005; Hashimoto et al., 2008; Lin et al., 2009; Hayakawab & Oguchi, 2009). Furthermore, based on DEM data and remote sensing images, geomorphologic objects are classified and identified from geomorphometric characteristics (Székely & Karátson, 2004; Saadat et al., 2008; Hampton & Cole, 2009; Dragut & Eisank, 2011, 2012). In soil erosion region, geomorphologic research is meaningful in erosion evaluation and in soil and land-use mapping (Remortel et al., 2004; Saadat et al., 2008). Soil erosion and water flow in a basin are modeled based on the topographic variants using DEM datasets (Zhu et al., 1997; Pennock & Corre, 2001). Landform automatic or semi-automatic classification is widely studied based on derived parameters (elevation, slope, curvature and so on) from DEM datasets (Macmillan et al., 2000, 2004; Shary et al., 2002; Ballantine et al., 2005). Loess Plateau is one of the most famous soil erosion regions in China. Heavy water and soil loss forms typical loess geomorphologic types in the world. Unique morphologic features, abundant natural resources, dense population and serious soil erosion problems make Loess Plateau an important region to conduct geomorphologic research (Tang et al., 2005), and most researches are focused on loess geomorphologic classification and identification based on slope spectrum index, which is a histogram statistics method based on slope variant (Li et al., 2007; Tang et al., 2008; Wang et al., 2008; Zhou et al., 2010). The adventure of the Chinese Geomorphologic database at 1:1 M scale provides spatial distribution of the geomorphologic types in vast region (Cheng et al., 2011a). Based on the database and Shuttle Radar Topography Mission at 3” spatial resolution (SRTM3 DEM) data, Zhao & Cheng (2014) attempted to build the transitional relationship among typical loess geomorphologic types using slope spectrum index in the whole Loess Plateau, which is helpful in understanding the relationship among the geomorphologic types. However, in the research by Zhao & Cheng (2014), the geomorphologic types are simple and coarse; meanwhile, the spatial distribution of the geomorphologic types is not accurate enough, which is at 1:1M scale; further, the criteria only chooses slope spectrum, so there are many other indexes to improve the achievements. Choosing a representative distribution area of typical loess geomorphologic types in the Loess Plateau as the study area, this research aims to construct the digital index system based on the improved geomorphologic database, so as to explore the transitional relationships among the typical loess geomorphologic types specifically. This research is meaningful in automatic digital mapping and classification, understanding the loess geomorphology, soil erosion, and water and soil loss and so on. 2. Study area and primary data sources 2.1 Study area Loess Plateau distributed in the northern and middle part of China, which is the most widespread distribution region of the loess geomorphology (Fig. 1). The typical loess geomorphology types are mainly loess tableland, loess ridge and loess knoll in the Loess Plateau (Zhao et al., 2017). The distribution of the typical loess geomorphology types is shown in Figure 1. Loess Plateau is a big region with widely distributed typical loess geomorphologic types. It is difficult to conduct deep research in this vast region, so we choose a test area which is representative in typical loess geomorphologic type distribution. The test area is the region in the black rectangle in Figure 1. The topographic characteristic of the test area is shown in Figure 2: Figure 1. Location and the typical loess geomorphology distribution in the loess plateau of China. Figure 2. Relief shader map in the test area. Based on the SRTM1 DEM data, the relief shader map is produced using DEM data in Global Mapper software. Figure. 2 shows that the typical loess geomorphologic types mainly distribute in the fragmental areas with abundant incised valleys, and the elevation mainly ranges from about 500 m to 2000 m. the fragmental topography is due to soil erosion and water and soil loss, which also result in fragile ecology and environment. The soil erosion and water and soil loss are the main causes to form the typical topography and loess geomorphologic types, so there are many researches about topography and geomorphology in this region using digital topographic analysis methods (Tang et al., 2005, 2008; Zhao & Cheng, 2014). Hence, it is meaningful to acquire the digital index system for the typical loess geomorphologic types in this area. 2.2 Primary data sources The primary data sources in this research are digital geomorphologic database, Landsat TM/ETM+ images and SRTM1 DEM data. 2.2.1 Digital geomorphologic database Digital geomorphologic database means Chinese 1:1,000,000 geomorphologic database, which is used to provide the spatial distribution of the typical geomorphologic types in the study area. The database is acquired using visual interpretation method at about 1:250,000 scales. The main data sources are Landsat TM/ETM+ images, SRTM3 DEM data, geologic data, published geomorphologic maps and so on (Cheng et al., 2011a). The classification method adopted morpho-genetic types, which can be divided 47Transitional relationships among typical loess geomorphologic types: Loess Plateau of China into seven layers, they are, basic morphology and altitude, genesis, sub-genesis, morphology, micro-morphology, slope and aspect, material and lithology (Cheng et al., 2011b). The digital database is seamless with consistent classification system, so it can be clipped for any regions with the same quality. In this research, the fundamental data for typical loess geomorphologic type distribution is acquired from this database. 2.2.2 Landsat TM/ETM+ images Remote sensing images are used to revise and improve the digital geomorphologic database data; meanwhile, the remote sensing characteristics for the typical loess geomorphologic types are acquired based on the remote sensing images; finally, the transitional relationship among the typical loess geomorphologic types can be explored by referencing the spatial distribution pattern of them in the remote sensing images. The main remote sensing images select the Landsat TM/ETM + images, which are global false colour composite images and downloaded from Chinese Scientific Data Base. The acquisition periods of the Landsat TM images are mainly about 1990; as to the Landsat ETM + images, about 2000. They are also the data sources in remote sensing visual interpretation process of the Chinese 1:1,000,000 geomorphologic database data. The composited bands of the remote sensing images are 7, 4 and 2. The combined effect of the images is similar to natural colour, which provides convenience for the visual interpretation and revision of the typical loess geomorphologic types. 2.2.3 SRTM1 DEM data SRTM1 DEM data is downloaded from the Earth Explorer of the USGS, USA. It has spatial resolution of 1” and continuous surface. Through downloading, merging, projection and clipping, the SRTM1 DEM data in the study area is shown in Figure 2. The SRTM1 DEM data not only provides topographic characteristics for the study area, but the elevation index and other indexes computed based on the DEM data. It is an important data sources in constructing the index system for the typical loess geomorphologic types. 3. Methodology The quality and scale of the digital geomorphologic database are improved through different methods. Based on the improved digital geomorphologic database, the spatial distribution of the typical loess geomorphology is acquired and illustrated accordingly. 3.1 Improvement methods to the digital geomorphologic database The digital geomorphologic database adopted in this research is interpreted at about 1:250,000 scales. The spatial resolution of the images is about 30m, which provides feasibility to improve the database at higher scales. So the database is revised using the Landsat TM/ETM + images at about 1:50,000 scales. The improvement methods to the digital geomorphologic database include improving the boundary of the units, revising the mis-classifying units and choosing the typical units for the typical loess geomorphologic types. 3.1.1 Improvement to the boundary At higher scale, the original boundary shows coarse and low quality. All the boundaries of the typical loess geomorphologic types are checked, and the boundary improvement is conducted especially for loess tableland. The change of the boundary can be shown in Figure 3: Figure 3 shows through boundary improvement, the boundary of the loess tableland becomes more accurate than before. The boundary represents the spatial distribution of the typical loess geomorphology, whose quality affects the construction of the digital index system. (a) (b) Figure 3. Boundary improvement to the units (a. before improvement; b. after improvement). 3.1.2 Revising the mis-classifying units The digital geomorphologic database covers the whole China territory. It is not easy to guarantee the attributes of all the units are correct. So the attribute of every unit is checked carefully to guarantee the quality of the data. To the mis-classifying units, their attributed would be changed to right if the correct attributes can be determined; otherwise would be deleted on the occasion that the proper attributes are not clear or difficult to determine. The distribution of the typical loess geomorphologic types needs not to cover the entire study area. 3.1.3 Choosing the typical units The revised digital geomorphologic database is checked firstly and the typical loess geomorphologic types are distilled. Then, all the units for every loess geomorphologic type are compared and judged. The typical units are chosen to represent the spatial distribution of the typical loess geomorphologic types, which can guarantee the quality of the constructed digital index system. 3.2 Spatial distribution of the typical loess geomorphology Through improvement, the spatial distribution of the typical loess geomorphologic types acquired (Figure 4). The classification method and distribution status for the typical loess geomorphologic types in the study area presented in Table 1, which also gives the codes for the typical loess geomorphologic types. The meanings of the codes are consistent and will be omitted in the following text. 48 Shifang Zhang, Shangmin Zhao Figure 4. Typical loess geomorphologic distribution type in the test area. Table 1 Classification and distribution for the typical loess geomorphologic types Typical loess geomorphology Morphology Valley shape Type Code Patch number Area(km2) Depth Height Loess tableland Loess tableland 11 29 254.9 Residual Loess residual tableland 12 8 1504.9 Beam Loess beam tableland 13 25 6172.0 Loess type Narrow acclivitous Shallow Low Loess narrow acclivitous ridge with shallow and low valley 21 13 1408.0 Narrow acclivitous Shallow High Loess narrow acclivitous ridge with shallow and high valley 22 7 1303.0 Narrow acclivitous Deep High Loess narrow acclivitous ridge with deep and high valley 23 7 838.7 Wide acclivitous Shallow Low Loess wide acclivitous ridge with shallow and low valley 24 2 62.6 Wide acclivitous Shallow High Loess wide acclivitous ridge with shallow and high valley 25 6 537.3 Wide acclivitous Deep High Loess wide acclivitous ridge with deep and high valley 26 3 298.1 Knoll type Shallow Low Loess knoll ridge with shallow and low valley 27 12 2262.3 Knoll type Shallow High Loess knoll ridge with shallow and high valley 28 47 11872.3 Knoll type Deep High Loess knoll ridge with deep and high valley 29 8 3093.3 Loess knoll Shallow Low loess knoll with shallow and low valley 31 10 1751.3 Shallow High loess knoll with shallow and high valley 32 18 7176.9 Deep High loess knoll with deep and high valley 33 7 5314.6 49Transitional relationships among typical loess geomorphologic types: Loess Plateau of China Figure 4 and Table 1 show that the loess tableland mainly distributes in middle and southern parts; the loess knoll, middle and northern parts; the loess ridge, northern and western parts. The loess tableland is divided by morphology; the loess knoll, valley shape; the loess ridge, both the morphology and valley, which has the most sub-types. The types are systematic, but their distribution status has large difference. For example, the loess wide acclivitous ridge with shallow and low valley only has two units with the area of 62.6 km2; whereas the loess knoll ridge with shallow and high valley has 47 units with the area of 11872.3 km2. The difference is due to the geomorphologic distribution characteristics in the study area. 4. Results Digital index system is firstly constructed based on the digital geomorphologic database, and then the transitional relationships are achieved for the typical loess geomorphologic types. 4.1 Digital index system for the typical loess geomorphologic types In this research, the digital index system for the typical loess geomorphologic types includes the topographic indexes, the quantitative indexes based on the topographic profile methods and the remote sensing indexes. 4.1.1 Topographic indexes The topographic indexes are computed based on the SRTM1 DEM data, which include elevation, relief, slope and curvature indexes. 4.1.1.1 Elevation index Elevation is a basic topographic index, which can be acquired directly from the SRTM1 DEM data. The spatial distribution of the elevation index can be seen in Figure 2. Through making a numerical statistics for the elevation index of every typical loess geomorphologic type, the results are in Table 2. Table 2. Numerical statistics for elevation index of typical loess geomorphologic types (m) Code Min Max Range Mean Std. 11 770 1414 644 1200.2 87.1 12 421 1408 987 936.7 128.4 13 516 1631 1115 1017.4 145.9 21 484 1509 1025 950.2 130.2 22 684 1684 1000 1182.0 155.4 23 609 1592 983 1117.2 202.8 24 587 1401 814 830.4 124.6 25 515 1345 830 839.7 145.3 26 779 1689 910 1430.0 126.5 27 784 1740 956 1181.7 175.9 28 532 1819 1287 1182.5 150.0 29 446 1598 1152 1153.8 115.3 31 840 1739 899 1226.5 180.3 32 610 1638 1028 1066.5 156.4 33 721 1553 832 1111.5 98.3 Table 2 shows that the typical loess geomorphology types mainly distribute from 400 m to 1800 m; the mean elevation value ranges from about 800 m to 1400 m. The standard deviation value increases from 11 to 13, but decreases from 31 to 33 evidently; as to the loess ridge type, it increases from 21 to 23, decreases from 27 to 29, but undulates from 24 to 26. The morphology type of the typical loess geomorphology types changes at every three types (Table 1), so we analyze the statistics characteristics of the topographic indexes at the group of every three types. 4.1.1.2 Relief index Other than elevation, the flat degree of the topography is a more important factor in determining the loess geomorphologic types. So the relief index is adopted here. Based on SRTM1 DEM data, the relief index is computed using neighborhood statistics under the spatial analyst tools in the ArcGIS software. Conducting a numerical statistics to the relief index, the results of the typical loess geomorphologic types are in Table 3. Table 3. Numerical statistics of relief index for typical loess geomorphologic types (m) Code Min Max Range Mean Std. 11 6 202 196 91.9 39.6 12 9 280 271 127.1 35.2 13 10 273 263 133.3 26.6 21 23 310 287 119.6 28.5 22 26 292 266 125.8 34.6 23 49 276 227 143.0 26.7 24 46 299 253 117.8 25.7 25 18 300 282 108.9 30.4 26 36 262 226 148.6 27.9 27 26 288 262 144.6 25.0 28 13 276 263 124.2 32.2 29 24 264 240 113.8 29.8 31 11 293 282 127.9 27.0 32 17 287 270 119.4 26.2 33 14 254 240 113.2 25.3 Table 3 shows that the relief range changes from about 200 m to 300 m. The characteristics of the mean value are similar, but not consistent with that of the standard deviation value of the elevation index. For example, from 11 to 13, the mean value increases, but the standard deviation value decreases. 4.1.1.3 Slope index Slope index can affect the rate of the movement down slope, which is an important index in reflecting the undulating degree of topography. It is applied in many studies, especially in the Loess Plateau region (Wolinsky & Pratson, 2005; Bue & Stepinski, 2006; Zhou et al., 2010). The slope index is computed based on SRTM1 DEM data using 3D Analyst Tools in ArcGIS software. For every typical loess geomorphologic types in the study area, the numerical statistics results of slope index are in Table 4. Table 4. Numerical statistics of slope index for typical loess geomorphologic types (°) Code Min Max Range Mean Std. 11 0.0 54.4 54.4 8.3 9.2 12 0.0 58.7 58.7 16.4 9.4 13 0.0 62.3 62.3 18.0 8.7 21 0.0 65.7 65.7 15.2 7.0 22 0.0 63.9 63.9 16.5 8.0 23 0.0 67.4 67.4 18.8 8.4 24 0.0 51.7 51.7 15.5 6.7 25 0.0 51.3 51.3 13.8 6.7 26 0.0 58.7 58.7 19.4 8.3 27 0.0 66.7 66.7 19.3 8.4 28 0.0 62.9 62.9 17.2 8.3 29 0.0 57.9 57.9 16.3 7.9 31 0.0 66.2 66.2 16.8 8.5 32 0.0 66.2 66.2 17.1 8.0 33 0.0 62.7 62.7 16.6 8.0 Table 4 shows that the typical loess geomorphologic types distribute in middle slope region, whose mean value ranges from about 10° to 20°. The standard deviation value mainly ranges 6° to 10°. The mean value increases from 11 to 13 and from 21 to 23, but decreases from 27 to 29, and undulates from 31 to 33 and 24 to 26. 50 Shifang Zhang, Shangmin Zhao 4.1.1.4 Curvature index Compared to slope index, the curvature index can affect the acceleration and deceleration of flow, so as to influence the erosion and deposition process (Moore et al., 1991; Zeverbergen & Thorne, 1987). Using 3D Analyst Tools in ArcGIS software, the curvature index is computed based on SRTM1 DEM data. Making a numerical statistics of curvature index for every typical loess geomorphologic types in the study area, the results are in Table 5. Table 5. Numerical statistics of curvature index for typical loess geomorphologic types (m-1) Code Min Max Range Mean Std. 11 -11.5 9.2 20.7 0.1 0.9 12 -25.5 18.6 44.1 0.0 1.4 13 -20.0 22.9 42.9 0.0 1.5 21 -22.9 20.1 43.0 0.0 1.2 22 -24.7 19.5 44.2 0.0 1.3 23 -15.8 17.3 33.1 0.0 1.4 24 -7.6 8.1 15.7 0.0 1.2 25 -10.3 10.2 20.5 0.0 1.1 26 -12.6 13.7 26.3 0.0 1.4 27 -19.8 21.4 41.3 0.0 1.4 28 -20.0 21.9 41.9 0.0 1.4 29 -17.8 18.3 36.0 0.0 1.3 31 -23.4 23.8 47.2 0.0 1.2 32 -26.2 20.3 46.6 0.0 1.4 33 -17.9 23.5 41.4 0.0 1.4 Table 5 show that the mean value of the curvature index is about 0.0 m-1; the standard deviation value mainly changes from 0.9 m-1 to 1.5 m-1. As to the range, it increases from 24 to 26, but undulating from other groups. 4.1.2 Quantitative indexes from digital topographic profile method The subtypes of the loess ridge and loess knoll are classified by their morphology and valley shape (Table 1), which are quantified tentatively by using the digital topographic profile method. To represent the morphologic change of the valley, the position of the topographic profile line is selected to perpendicular to the trend of the ridge (Cheng et al., 2013). Taking the type of loess narrow acclivitous ridge with deep and high valley as an example, the position of the topographic profile line is in Figure 5 in the remote sensing images. Figure 5. Position of the digital topographic profile line in the remote sensing images. In Figure 5, the yellow line is the topographic profile line. Based on SRTM1 DEM data, the topographic profile is drawn to quantify the valley change using ArcGIS software (Figure 6). Valley deep Valley height Horizontal distance (m) El ev at io n (m ) 1150 1100 1050 1000 950 900 0 1000 2000 3000 4000 5000 6000 7000 8000 SE Figure 6. Topographic profile line in loess narrow acclivitous ridge with deep and high valley. Figure 6 shows the valley height of about 940 m and valley deep of about 200 m. Through making a calculation to all the loess ridge and loess knoll types, the results are in Table 6. Table 6. Valley shape for loess ridge and loess knoll types (m) Type Valley height Valley depth Type Valley height Valley depth 21 725 120 27 1000 140 22 1080 170 28 1100 220 23 940 200 29 850 150 24 760 110 31 1020 150 25 850 105 32 960 120 26 1300 220 33 960 210 Table 6 presents the valley height and valley depth for the subtypes of loess ridge and loess knoll, which are generally consistent with the valley shape of the types, although some minor contradictions exist. The existence of the contradictions may be due to the valley shape is acquired from the distribution of the types, but they may have a little change in different regions. 4.1.3 Remote sensing indexes for typical loess geomorphologic types Remote sensing indexes include both the remote sensing characteristics and the spatial distribution pattern in the remote sensing images for the typical loess geomorphologic types. 4.1.3.1 Remote sensing characteristics Based on the spatial distribution of the typical loess geomorphologic types and the Landsat TM/ETM+ images, the remote sensing characteristics for the typical loess geomorphologic types are acquired. Remote sensing images are the fundamental data for acquiring the distribution status of typical loess geomorphologic types. So the remote sensing characteristics reflect the change and transition in colour and texture in the remote sensing images for the typical loess geomorphologic types. 4.1.3.2 Spatial distribution pattern in remote sensing images Other than the remote sensing characteristics, the spatial distribution of the typical loess geomorphologic types in the remote sensing images is helpful to understand the transition and change relationship among them. Figure 7 presents an example of this transition: from right to left, the typical loess geomorphologic types change from loess residual tableland to loess beam tableland, and then the loess knoll ridge with shallow and high valley. 51Transitional relationships among typical loess geomorphologic types: Loess Plateau of China Figure 7. Spatial distribution pattern of the typical loess geomorphologic type (a. loess residual tableland; b. loess beam tableland; c. loess knoll ridge with shallow and high valley). Through exploring and analyzing the spatial distribution pattern of the typical loess geomorphologic types in the remote sensing images, we attempted to construct the transitional relationship among them. 4.2 Transitional relationship exploration Referencing the previous research based on the slope spectrum (Zhao & Cheng, 2014), we explored to construct the transitional relationships among the typical loess geomorphologic types by using the digital index systems (Fig. 8). Figure 8 shows that the loess tableland may transform to loess ridge or to loess knoll, and loess ridge may transform to loess knoll. For the subtypes of the loess tableland, the loess tableland may transform to loess residual tableland, and then transform to loess beam tableland. For the morphology of loess ridge, the wide acclivitous types transform to narrow acclivitous, and then transform to knoll types. The valley shape transforming status is the same for both loess ridge and loess knoll: shallow and high valley transforms to deep and high valley, and then transform to shallow and low valley. 5. Discussions 5.1 Comparison with relevant researches This research is compared with international relevant researches, the relevant researches in the Loess Plateau and the transitional relationship exploration among loess geomorphologic types. 5.1.1Comparison with internal relevant researches International relevant researches mostly focus on morphologic analysis, pattern analysis, classification, reconstruction and identification for geomorphologic types (Werner, 1999; Kocurek & Ewing, 2005; Saito & Oguchi, 2005; Bishop, 2007a, 2007b, 2009; Hashimoto et al., 2008; Hampton & Cole, 2009; Bullard et al., 2011; Dragut & Eisank, 2012). Nevertheless, these researches mainly studied on sole geomorphologic type and the spatial distribution of the geomorphologic type was not acquired in advance, which seldom pay attentions to the relationships among geomorphologic types. In this research, the transitional relationships among typical loess geomorphology are constructed based on the digital index system. 5.1.2 Comparison with the relevant researches in the Loess Plateau In this research, the spatial distribution of typical loess geomorphologic types can be acquired from digital geomorphologic database at 1:1,000,000 scale; meanwhile, the transitional relationships among typical loess geomorphologic types were explored based on the constructed digital index system. The achievement of the relationship among different geomorphologic types is a meaningful research method compared with international relevant researches. Transitional relationships among typical loess geomorphologic types Loess tableand Loess ridge Narrow acclivitous Knoll type Loess ridge Loess wide acclivitous ridge with shallow and high valley Loess wide acclivitous ridge with deep and high valley Loess wide acclivitous ridge with deep and low valley Loess narrow acclivitous ridge with shallow and high valley Loess narrow acclivitous ridge with deep and high valley Loess Knoll ridge with shallow and high valley Loess Knoll ridge with shallow and low valley Loess Knoll ridge with shallow and high valley Loess Knoll ridge with deep and high valley Loess Knoll ridge with shallow and low valley Loess Knoll ridge with deep and high valley Loess narrow acclivitous ridge with deep and low valley Loess tableand Residual tableand Beam tableand Loess Knoll Figure 8. Transitional relationship among typical loess geomorphologic types. In the loess Plateau of China, loess geomorphology forms due to soil erosion and water and soil loss, so they have close relationship. Most researches based on DEM datasets attempted to classify the typical loess geomorphologic types and analyze their characteristics, especially for morphologic characteristics (Tang et al., 2005, 2008; Zhou et al., 2010). In this research, the classification and accurate distribution can be acquired from digital geomorphologic database, and our aim is to explore the transitional relationship among typical loess geomorphologic types based on the constructed digital index system. Compared with previous research in the Loess Plateau, this research has more close relationship to soil erosion problems, which is one of the most serious problems in the Loess Plateau (Ma, 1996; Liang et al., 2004; Sang et al., 2007). 5.1.3 Comparison with the transitional relationship exploration researches Compared with the previous research about the transitional relationship exploration among the typical loess geomorphology by Zhao & Cheng (2014), this research selects the typical test area in Loess Plateau, where distribute dense and typical loess geomorphologic types; in addition, the digital geomorphologic database is revised and improved using visual interpretation and remote sensing images, which guarantees the spatial distribution accuracy 52 Shifang Zhang, Shangmin Zhao of typical loess geomorphologic types; meanwhile, digital index system is construct, which provide more reasonable transitional relationships among typical loess geomorphologic types; furthermore, typical loess geomorphologic types have more subtypes and are subdivided by morphologic type and valley shape type, which makes this research much more meaningful compared with previous research. Overall, this research is a developed and improved version of the research by Zhao & Cheng (2014). 5.2 Prospects about this research Based on this research, the future researches can be conducted on the following three aspects: Firstly, choose more reasonable study area. The subtypes for the typical loess geomorphologic types are not complete in this research (Cheng et al., 2011b). For example, there is another valley shape type: deep and low valley, which has limited distribution and does not exist in the study area However, more geomorphologic types’ distribution means much more big study area, which will increase the workload and decrease the accuracy of the research (Cheng et al., 2011a), especially in digital geomorphologic database improvement and topographic profile line selection. So it is important to seek the balance between the size of the study area and the distribution of the typical geomorphologic types. Then, construct digital index system using more indexes. This research adopts topographic indexes, digital indexes from topographic profile method and remote sensing indexes. Other than the indexes in this research, there are many other indexes which are useful, such as slope angle and its other derivations, complexity, point pattern, spacing and so on (Werner, 1999; Székely & Karatson, 2004; Bishop, 2007a, 2007b; Bullard et al., 2011). Indexes are the basis for exploring the transitional relationship among the typical loess geomorphologic types, so using more indexes may help deeply understanding the relationship among geomorphologic types. Finally, besides the loess geomorphology types, digital index system and relationship exploration can also be studied for other geomorphologic types. Taking Aeolian sand dune for example, its point pattern, sand-dune pattern, morphometry and complexity are studied in previous researches (Werner, 1999; Kocurek & Ewing, 2005; Bishop, 2007b; Bullard et al., 2011). Digital index system construction and relationships exploration helps understanding the nature of the geomorphologic types much further. 6. Conclusions Based on 1:1,000,000 digital geomorphologic database in China, the spatial distribution of the typical loess geomorphologic types is revised and improved using remote sensing images in the test area of Loess Plateau; then, the digital index system is constructed for the typical loess geomorphologic types based on topographic analysis methods and remote sensing characteristics; finally, the transitional relationships among the typical loess geomorphologic types are explored based on the digital index system and previous research. Through this research, the following conclusions can be acquired: Through boundary improvement, revision to the mis-clssifying units and selection to the typical units, the improved digital geomorphologic database has the typical loess geomorphologic types including loess tableland, loess ridge and loess knoll. Due to the morphologic types, the loess tableland is divided into loess tableland, loess ridge and loess knoll; the loess ridge is divided into narrow acclivitous types, wide acclivitous types and knoll types. As to the valley shape, the loess ridge and loess knoll are divided into shallow and low valley, shallow and high valley and deep and high valley. Digital index system includes topographic indexes, quantitative indexes from topographic profile and remote sensing indexes. In the test area, the typical loess geomorphologic types mainly distributes in the elevation of about 1000 m, relief of about 250 m and middle slope region; as to the curvature index, it has mean value of 0.0 m-1 and standard deviation values from 0.9 m-1 to 1.5 m-1. The topographic profile presents the valley height and valley depth for the typical units of all the types. The remote sensing indexes include the remote sensing characteristics of all the types and the spatial distribution pattern in the remote sensing images, which is helpful to construct the transitional relationships among typical loess geomorphologic types. The loess tableland may transform to both loess ridge and loess knoll, and the loess ridge may transform to loess knoll. As to the morphologic types, loess tableland may transform to loess residual tableland, and then to the loess beam tableland; the loess wide acclivitous ridge transform to loess narrow acclivitous ridge, and then to loess knoll ridge. For the valley shape types of the loess ridge and loess knoll, the shallow and high valley may transform to deep and high valley, and finally transform to shallow and low valley. Acknowledgments This research is under the auspices of the National Natural Science Foundation of China (42271432, 41631179), the Foundation of Shanxi Vocational University of Engineering Science and Technology (KJ202426), the National Key Research and Development Program (2017YFB0503603), and the Natural Science Foundation of Shanxi Province (201901D111098). We would like to express our sincere gratitude to the editor and anonymous reviewers. References Ballantine, J. C., Okin, G. S., Prentiss, D. E., & Roberts, D. A. (2005). Mapping North African landforms using continental scale unmixing of MODIS imagery. Remote Sensing of Environment, 97(4), 470-483. https://doi. org/10.1016/j.rse.2005.04.023 Bishop, M. A. (2007a). Point pattern analysis of eruption points for the Mount Gambier Volcanic sub-Province: A geographical approach to the understanding of volcano distribution. Area, 39(2), 230–241. DOI: https://doi.org/10.1111/j.1475-4762.2007.00729.x Bishop, M. A. (2007). Point pattern analysis of north polar crescentic dunes, Mars: A geography of dune self-organization. Icarus, 191(1), 151-157. https://doi.org/10.1016/j.icarus.2007.04.027 Bishop, M. A. (2009). A generic classification for the morphological and spatial complexity of volcanic (and other) landforms. Geomorphology, 111(1- 2), 104-109. https://doi.org/10.1016/j.geomorph.2008.10.020 Bue, B., & Stepinski, T. (2006). Automated classification of landforms on Mars. Computers & Geosciences, 32(5), 604-614. https://doi.org/10.1016/j. cageo.2005.09.004 Bullard, J. E., White, K., & Livingstone, I. (2011). Morphometric analysis of aeolian bedforms in the Namib Sand Sea using ASTER data. Earth Surface Processes and Landforms, 36(11), 1534-1549. https://doi. org/10.1002/esp.2189 Cheng, W. M., Zhao, S. M., Zhou, C. H., & Chen, X. (2013). Topographic characteristics for the geomorphologic Zones in the northwestern edge of the Qinghai-Tibet Plateau. Journal of Mountain Science, 10(6), 1039- 1049. https://doi.org/10.1007/s11629-013-2582-z Cheng, W., Zhou, C., Chai, H., Zhao, S., Liu , H., & Zhou, Z. (2011a). Research and compilation of the Geomorphologic Atlas of the People’s Republic of China (1:1,000,000). Journal of Geographical Sciences, 21, 89-100. https://doi.org/10.1007/s11442-011-0831-z Cheng, W., Zhou, C., Li, B., Shen, Y., & Zhang, B. (2011b). Structure and contents of layered classification system of digital geomorphology for China. Journal of Geographical Sciences, 21, 771-790. https://doi.org/10.1007/ s11442-011-0879-9 Drăguţ, L., & Eisank, C. (2011). Object representations at multiple scales from digital elevation models. Geomorphology, 129(3-4), 183-189. https://doi. org/10.1016/j.geomorph.2011.03.003 Drăguţ, L., & Eisank, C. (2012). Automated object-based classification of topography from SRTM data. Geomorphology, 141-142, 21-33. https:// doi.org/10.1016/j.geomorph.2011.12.001 Hampton, S., & Cole, J. (2009). Lyttelton Volcano, Banks Peninsula, New Zealand: Primary volcanic landforms and eruptive centre identification. Geomorphology, 104(3-4), 284-298. https://doi.org/10.1016/j. geomorph.2008.09.005 Hashimoto, A., Oguchi, T., Hayakawa, Y., Lin, Z., Saito, K., & Wasklewicz, T. A. (2008). GIS analysis of depositional slope change at alluvial-fan toes in https://doi.org/10.1016/j.rse.2005.04.023 https://doi.org/10.1016/j.rse.2005.04.023 https://doi.org/10.1111/j.1475-4762.2007.00729.x https://doi.org/10.1016/j.icarus.2007.04.027 https://doi.org/10.1016/j.geomorph.2008.10.020 https://doi.org/10.1016/j.cageo.2005.09.004 https://doi.org/10.1016/j.cageo.2005.09.004 https://doi.org/10.1002/esp.2189 https://doi.org/10.1002/esp.2189 https://doi.org/10.1007/s11629-013-2582-z https://doi.org/10.1007/s11442-011-0831-z https://doi.org/10.1007/s11442-011-0879-9 https://doi.org/10.1007/s11442-011-0879-9 https://doi.org/10.1016/j.geomorph.2011.03.003 https://doi.org/10.1016/j.geomorph.2011.03.003 https://doi.org/10.1016/j.geomorph.2011.12.001 https://doi.org/10.1016/j.geomorph.2011.12.001 https://doi.org/10.1016/j.geomorph.2008.09.005 https://doi.org/10.1016/j.geomorph.2008.09.005 53Transitional relationships among typical loess geomorphologic types: Loess Plateau of China Japan and the American Southwest. Geomorphology, 100(1-2), 120-130. https://doi.org/10.1016/j.geomorph.2007.10.027 Hayakawa, Y. S., & Oguchi, T. (2009). GIS analysis of fluvial knickzone distribution in Japanese mountain watersheds. Geomorphology, 111(1- 2), 27-37. https://doi.org/10.1016/j.geomorph.2007.11.016 Kocurek, G., & Ewing, R. C. (2005). Aeolian dune field self-organization – implications for the formation of simple versus complex dune-field patterns. Geomorphology, 72(1-4), 94-105. https://doi.org/10.1016/j. geomorph.2005.05.005 Li, F. Y., Tang, G. A., Jia, Y. N., & Cao, Z. D. (2007). Scale Effect and Spatial Distribution of Slope Spectrum’s Information Entropy. Journal of Geo-information Science, 9(4), 13-18. DOI: 10.3969/j.issn.1560- 8999.2007.04.003 Liang, G. L., Chen, H., Cai, Q. G., & Hu, W. S. (2004). Research Progress of Modern Topographic Evolvement and Landform Erosion in Loess Plateau. Research of Soil and Water Conservation, 11(4): 131-137. DOI: 10.3969/j.issn.1005-3409.2004.04.031 Lin, Z., Oguchi, T., Chen, Y. G., & Saito, K. (2009). Constant-slope alluvial fans and source basins in Taiwan. Geology, 37(9), 787–790. DOI: 10.1130/ G25675A.1 Ma, N. X. (1996). Relationship between loess geomorphic evolution and soil erosion. Bulletin of Soil and Water Conservation, 16(2), 6-10.  MacMillan, R., Pettapiece, W., Nolan, S., & Goddard, T. (2000). A generic procedure for automatically segmenting landforms into landform elements using DEMs, heuristic rules and fuzzy logic. Fuzzy Sets and Systems, 113(1), 81-109. https://doi.org/10.1016/S0165-0114(99)00014- 7 MacMillan, R., Jones, R., & McNabb, D. H. (2004). Defining a hierarchy of spatial entities for environmental analysis and modeling using digital elevation models (DEMs). Computers, Environment and Urban Systems, 28(3), 175-200. https://doi.org/10.1016/S0198-9715(03)00019-X Moore, I. D., Grayson, R. B., & Ladson, A. R. (1991). Digital terrain modelling: A review of hydrological, geomorphological, and biological applications. Hydrological Processes, 5(1), 3-30. https://doi.org/10.1002/ hyp.3360050103 Pennock, D., & Corre, M. (2001). Development and application of landform segmentation procedures. Soil and Tillage Research, 58(3-4), 151-162. https://doi.org/10.1016/S0167-1987(00)00165-3 Remondo, J., & Oguchi, T. (2009). GIS and SDA applications in geomorphology. Geomorphology, 111(1-2), 1-3. https://doi.org/10.1016/j. geomorph.2009.04.015 Remortel, R. V., Maichle, R., & Hickey, R. (2004). Computing the LS factor for the Revised Universal Soil Loss Equation through array-based slope processing of digital elevation data using a C++ executable. Computers & Geosciences, 30(9-10), 1043-1053. https://doi.org/10.1016/j. cageo.2004.08.001 Saadat, H., Bonnell, R., Sharifi, F., Mehuys, G., Namdar, M., & Ale-Ebrahim, S. (2008). Landform classification from a digital elevation model and satellite imagery. Geomorphology, 100(3-4), 453-464. https://doi. org/10.1016/j.geomorph.2008.01.011 Saito, K., & Oguchi, T. (2005). Slope of alluvial fans in humid regions of Japan, Taiwan and the Philippines. Geomorphology, 70(1-2), 147-162. https:// doi.org/10.1016/j.geomorph.2005.04.006 Sang, G. S., Chen, X., Chen, X., & Che, Z. L. (2007). Formation model and geomorphic evolution of loess hilly landforms. Arid Land Geography, 30(3), 375–380. DOI: 10.1016/S1872-5791(07)60044-X Shary, P. A., Sharaya, L. S., & Mitusov, A. V. (2002). Fundamental quantitative methods of land surface analysis. Geoderma, 107(1-2), 1-32. https://doi. org/10.1016/S0016-7061(01)00136-7 Székely, B., & Karátson, D. (2004). DEM-based morphometry as a tool for reconstructing primary volcanic landforms: Examples from the Börzsöny Mountains, Hungary. Geomorphology, 63(1-2), 25-37. https:// doi.org/10.1016/j.geomorph.2004.03.008 Tang, G., Ge, S., Li, F., & Zhou, J. (2005). Review of Digital Elevation Model (DEM) Based Research on China Loess Plateau. Journal of Mountain Science, 2(3), 265-270. https://doi.org/10.1007/BF02973200 Tang, G., Li, F., Liu, X., Long, Y., & Yang, X. (2008). Research on the slope spectrum of the Loess Plateau. Science in China Series E: Technological Sciences, 51, 175-185. https://doi.org/10.1007/s11431-008-5002-9 Wang, C., Tang, G. A., Li, F. Y., Zhu, X. J., & Jia, Y. N. (2008). The Uncertainty of Slope Spectrum Derived from Grid Digital Elevation Model. Geo-Information Science, 10, 539–544. DOI: 10.3969/j.issn.1560- 8999.2008.04.019 Werner, B. T. (1999). Complexity in Natural Landform Patterns. Science, 284(5411), 102-104. https://doi.org/10.1126/science.284.5411.102 Wolinsky, M. A., & Pratson, L. F. (2005). Constraints on landscape evolution from slope histograms. Geology, 3(6), 477-480. https://doi.org/10.1130/ G21296.1 Zevenbergen, L. W., & Thorne, C. R. (1987). Quantitative analysis of land surface topography. Earth Surface Processes and Landforms, 12(1), 47-56. https://doi.org/10.1002/esp.3290120107 Zhao, S., & Cheng, W. (2014). Transitional relation exploration for typical loess geomorphologic types based on slope spectrum characteristics. Earth Surface Dynamics, 2, 433–441. https://doi.org/10.5194/ esurf-2-433-2014 Zhao, S., Cheng, W., Zhou, C., Liu, H., Su, Q., Zhang, S., He, W., Wang, L., & Wu, W. (2017). Using MLR to model the vertical error distribution of ASTER GDEMV2 data based on ICEsat/GLA14 data in the Loess Plateau of China. Zeitschrift für Geomorphologie, Supplementary Issues, 61(2), 9-26. DOI: 10.1127/ZFG_SUPPL/2016/0325 Zhou, Y., Tang, G. A., Yang, X., Xiao, C. C., Zhang, Y., & Luo, M. L. (2010). Positive and negative terrains on northern Shaanxi Loess Plateau. Journal of Geographical Sciences, 20(1), 64-76. https://doi.org/10.1007/ s11442-010-0064-6 Zhu, X., Band, L., Vertessy, R., & Dutton, B. (1997). Derivation of Soil Properties Using a Soil Land Inference Model (SoLIM). Soil Science Society of America Journal, 61(2), 523-533. https://doi.org/10.2136/ sssaj1997.03615995006100020022x https://doi.org/10.1016/j.geomorph.2007.10.027 https://doi.org/10.1016/j.geomorph.2007.11.016 https://doi.org/10.1016/j.geomorph.2005.05.005 https://doi.org/10.1016/j.geomorph.2005.05.005 https://doi.org/10.1016/S0165-0114(99)00014-7 https://doi.org/10.1016/S0165-0114(99)00014-7 https://doi.org/10.1016/S0198-9715(03)00019-X https://doi.org/10.1002/hyp.3360050103 https://doi.org/10.1002/hyp.3360050103 https://doi.org/10.1016/S0167-1987(00)00165-3 https://doi.org/10.1016/j.geomorph.2009.04.015 https://doi.org/10.1016/j.geomorph.2009.04.015 https://doi.org/10.1016/j.cageo.2004.08.001 https://doi.org/10.1016/j.cageo.2004.08.001 https://doi.org/10.1016/j.geomorph.2008.01.011 https://doi.org/10.1016/j.geomorph.2008.01.011 https://doi.org/10.1016/j.geomorph.2005.04.006 https://doi.org/10.1016/j.geomorph.2005.04.006 https://doi.org/10.1016/S0016-7061(01)00136-7 https://doi.org/10.1016/S0016-7061(01)00136-7 https://doi.org/10.1016/j.geomorph.2004.03.008 https://doi.org/10.1016/j.geomorph.2004.03.008 https://doi.org/10.1007/BF02973200 https://doi.org/10.1007/s11431-008-5002-9 https://doi.org/10.1126/science.284.5411.102 https://doi.org/10.1130/G21296.1 https://doi.org/10.1130/G21296.1 https://doi.org/10.1002/esp.3290120107 https://doi.org/10.5194/esurf-2-433-2014 https://doi.org/10.5194/esurf-2-433-2014 https://doi.org/10.1007/s11442-010-0064-6 https://doi.org/10.1007/s11442-010-0064-6 https://doi.org/10.2136/sssaj1997.03615995006100020022x https://doi.org/10.2136/sssaj1997.03615995006100020022x