Automatic detection of fault-controlled rivers using spatial pattern matching

Highlights

• An automatic detection method for fault-controlled rivers was developed.

• ARG was used to model the river system's geographic scene and the spatial patterns of FCRs.

• A spatial pattern library was formed and can be modified manually.

• The algorithm of spatial pattern matching was used to detect the FCR patterns.

Abstract

The river system of a region records its tectonic evolution. This study expects to infer some potential faults by effectively quantifying river morphology and identifying rivers with specific morphological and structural characteristics. Thus, this study aims to develop a novel method for automatically detecting fault-controlled rivers (FCRs) from drainage maps using spatial pattern matching and to support the scientific inference of potential faults. The method involves (1) constructing a scene model for the entire drainage basin in the study area based on the attributed relational graph (ARG) model; (2) defining spatial patterns for four types of FCRs using the ARG model, including straight river reaches, right-angle river reaches, barb rivers, and contra-aperture rivers, and storing all river patterns in a spatial pattern library; and (3) detecting the four types of FCRs with a spatial pattern-matching algorithm and the FCR pattern library. Case studies demonstrated the basic effectiveness of this method for detecting FCRs in the study areas of Bomi County (Parong Tsangpo River Valley and Yigong Zangbo River Valley) and southeast of the Qinghai Tibet Plateau. The goals of this study are to quantify river shape better, support the preliminary detection of potential faults, and provide a useful reference for structure-based spatial queries in GIS. Additionally, more river pattern types with specific morphological and structural characteristics can be automatically detected through dynamic maintenance of the spatial pattern library.

Introduction

Rivers are sensitive to subtle tectonic changes. River terraces, alluvial fans, river valley forms, and river patterns are related to tectonic activity and history (Howard, 1967; Audley-charles et al., 1978; Cox, 1994; Holbrook and Schumm, 1999; Shahzad et al., 2010; Gasparini et al., 2016). The river system of a region, particularly a lowland river system, can record important information about tectonic deformation (Kirby et al., 2003). Correspondingly, the morphological and structural characteristics of river systems could be used to decipher potential tectonic controls in the region (Mahmood and Gloaguen, 2012; Tian and Zhan,2013; Luirei et al., 2015).

Recently, scholars have increasingly studied the relationship between rivers and tectonic systems, mainly focusing on two aspects: the longitudinal profiles and planform morphology of rivers. Longitudinal river profiles show tectonic activity through their overall morphology (steepness and concavity) and knickpoints (Seeber and Gornitz, 1983; Schumm et al., 2000; Wobus et al., 2006; Ambili and Narayana, 2014). Furthermore, steepness and concavity are often used as effective parameters to evaluate the influence of uplift tectonics on fluvial systems (Kirby et al., 2003; Mahmood and Gloaguen, 2012; Sharma and Sarma, 2017). Knickpoints provide a relative indication of the uplift rate and gradient change along the longitudinal river profiles (Hack, 1973).

Similarly, the planform morphology of a river can also be affected by tectonic activity of different degrees and scales (Hallet and Molnar, 2001). For a slight dip on the surface caused by tectonic activity, rivers usually undergo minor changes in sinuosity and channel width (Holbrook and Schumm, 1999). As a result, the sinuosity and channel width of the river increase with an increased valley gradient to maintain an equilibrium slope (Maynard, 2006; Gasparini et al., 2016). Meanwhile, the change in the surface gradient transforms the spatial structure of the river system. The regional tectonic activity commonly played an essential role in forming river captures and reversed rivers (Clark et al., 2004) and is one of the driving factors in channel migration. The large-scale lateral tilting of the drainage basin leads to overall bending in the same direction or even reorganizing the river system (Cox, 1994; Brookfield, 1998; Hallet and Molnar, 2001; Maroukian et al., 2008). Many researchers have recently evaluated tectonic activity using drainage networks and geomorphic indices, such as stream-length gradient, fractal dimension, basin asymmetry factor, and basin shape index (Mahmood and Gloaguen, 2012; Elio et al., 2015; Sharma and Sarma, 2017).

Previous studies have mainly focused on the macroscopic characteristics of river systems and their basins, but few have concentrated on analyzing the morphological and structural characteristics at a local scale. In addition, structure-based spatial querying methods using raster images have been widely used, such as intelligent image retrieval (Chu et al., 1998; Tang et al., 2003; Mckay and Blain, 2014), spatial similarity assessment (Petrakis et al., 2002; Li and Fonseca, 2006; Nedas and Egenhofer, 2008; Zhang and Guilbert, 2013) and morphological feature recognition of river systems using artificial neural networks (Monica et al., 2015). These methods can perform simple structural recognition using raster images based on image segmentation, feature recognition, and object recognition. In these methods, the spatial information expressed in the image is primarily used. However, these methods lack sufficient consideration of the spatial relationship information between scene elements and the attribute information of scene elements. Therefore, these methods do not support accurate semantic constraints and complex structural definitions, and the recognized structure is relatively simple and rough.

Different fault-controlled rivers (FCRs) contain important local characteristics and fault-activity information. FCRs might be a good location for identifying cryptic faults using small-scale characteristics of the river system. Existing methods for fault screening in large drainage basins often require traditional geological field surveys, which are costly in terms of time, instrumentation, and logistically challenging. Additionally, the spatial relationship and contour of features in a digital map obtained directly from vector data may be more accurate, detailed, and can support accurate semantic constraints and complex structure definitions.

Therefore, using vector data and the method of spatial pattern matching may be more accurate for identifying objects with complex spatial structures. Compared with existing automatic methods, the method proposed in this study pays more attention to the local characteristics of river models, which are suitable for river systems with small-scale characteristics in maps with a scale greater than or equal to 1:250,000. Furthermore, using vector data makes it possible to identify objects with complex spatial structures more accurately. This method is mainly used to infer faults in areas where geological surveys are inadequate or difficult to conduct (such as uninhabited areas). Detecting fault-controlled rivers can also help researchers better understand the river systems' developmental mechanisms.

Although FCRs differ in morphology and spatial structure, the same type of river pattern has similar characteristics that can be abstracted as a spatial pattern. Spatial pattern libraries can be built based on definitions of different FCR patterns. Different FCR patterns can be detected in a watershed based on the spatial pattern-matching algorithm. The attributed relational graphs (ARG) model has been proven to effectively express geographic scenes by storing relationship information between spatial objects (Li et al., 2019). In ARG, objects are defined as graph nodes, relationships are defined as edges between the nodes, and both nodes and edges are labeled by attributes corresponding to the properties of objects and relationships respectively (Petrakis et al., 2002). The FCR pattern mainly defines the specific spatial relationship information between the elements of the FCR; therefore, it can be efficiently expressed by the ARG model. Similarly, the ARG model presents the study area's specific morphological and structural characteristics.

The fault is not the only factor controlling the river shape, but it is a relatively important factor (Gaudemer et al., 1989; Huang,1993; Ouchi, 2005; Tian and Zhan, 2013; Bufe et al., 2016; Sarma and Sharma, 2018; Duvall et al., 2020). There is no necessary relationship between faults and river shape; however, they are highly correlated. The geomorphologic characteristics of an area result from the combined action of internal and external geological forces on the crust. Fluvial landforms also follow this pattern. When a fault is just developing, under the influence of faulting, the plane morphology of related rivers usually changes to some extent and shows the morphological characteristics of a type of FCR pattern. As time progresses, if the fault is further developed, the planform morphological characteristics of the FCR pattern are further strengthened. In contrast, if the fault activity stops, the planform morphological characteristics of the FCR pattern will gradually weaken owing to the influence of external geological processes on the topography and landform. Inferring potential faults according to specific river morphologies has research significance and application value. First, it can provide support for analyzing the controlling factors of river morphology and predicting the changing trends in river morphology, which is essential for managing water resources, planning hydraulic projects, and preventing geological disasters (Kondrat'yev,1968; Zhang et al., 2008; Bruce, 2020); Second, it helps obtain fault information in hard-to-reach regions or areas that are no longer suitable for fault exploration by geological means (e.g., urban built-up areas).

The core of this work centers on using the four planform morphologies as a proxy for fault-impacted river reaches. Therefore, the reaches with these four planform morphologies identified by the spatial pattern-matching algorithm can be regarded as the reaches that faults may control. This method is not only helpful for detecting faults in tectonically developed regions but also has a unique advantage in identifying geographical scenes with complex spatial structures. The remainder of this paper is organized as follows: Section 2 presents the methodology, Section 3 presents the experimental results with case studies, Section 4 presents the discussion, and Section 5 provides the conclusion and recommendations for future work.