Complex Geographical Phenomena Research Articles

A New Framework for Integrating DNN-Based Geographic Simulation Models within GISystems

As a crucial spatial decision support tool, Geographic Information Systems (GISystems) are widely used in fields such as digital watersheds, resource management, environmental assessment, and regional governance, with their core strength lying in the integration of geographic simulation models from various disciplines, enabling the analysis of complex geographical phenomena and the resolution of comprehensive spatial problems. With the rapid advancement of artificial intelligence, deep neural network-based geographic simulation models (DNN-GSMs) have increasingly replaced traditional models, offering significant advantages in simulation accuracy and inference speed, and have become indispensable components in GISystems. However, existing integration methods do not adequately account for the specific characteristics of DNN-GSMs, such as their formats and input/output data types. To address this gap, we propose a novel tight integration framework for DNN-GSMs, comprising four key interfaces: the data representation interface, the model representation interface, the data conversion interface, and the model application interface. These interfaces are designed to describe spatial data, the simulation model, the adaptation between spatial data and the model, and the model’s application process within the GISystem, respectively. To validate the proposed method, we construct a spatial morphology simulation model based on CNN-LSTM, integrate it into a GISystem using the proposed interfaces, and conduct a series of predictive experiments on island morphology evolution. The results demonstrate the effectiveness of the proposed integration framework for DNN-GSMs.

A neural network model to optimize the measure of spatial proximity in geographically weighted regression approach: a case study on house price in Wuhan

The estimation of spatial heterogeneity within real estate markets holds significant importance in house price modelling. However, employing a single or straightforward distance to measure spatial proximity is probably insufficient in complex urban areas, thereby resulting in an inadequate modelling of spatial heterogeneity. To address this issue, this paper incorporates multiple distance measures within a neural network framework to achieve an optimized measure of spatial proximity (OSP). Consequently, a geographically neural network weighted regression model with optimized measure of spatial proximity (osp-GNNWR) is devised for the purpose of spatially heterogeneous modeling. Trained as a unified model, osp-GNNWR obviates the need for separate pretraining of OSP. This enables OSP to delineate the modeled spatial process through a post hoc calculated value. Through simulation experiments and a real-world case study on house prices, the proposed model reaches more accurate descriptions of diverse spatial processes and exhibits better overall performance. The interpretable results of the case study in Wuhan demonstrate the efficacy of the osp-GNNWR model in addressing spatial heterogeneity within real estate markets, suggesting its potential for modelling and predicting complex geographical phenomena.

A Containerized Service-Based Integration Framework for Heterogeneous-Geospatial-Analysis Models

The integration of geospatial-analysis models is crucial for simulating complex geographic processes and phenomena. However, compared to non-geospatial models and traditional geospatial models, geospatial-analysis models face more challenges owing to extensive geographic data processing and complex computations involved. One core issue is how to eliminate model heterogeneity to facilitate model combination and capability integration. In this study, we propose a containerized service-based integration framework named GeoCSIF, specifically designed for heterogeneous-geospatial-analysis models. Firstly, by designing the model-servicized structure, we shield the heterogeneity of model structures so that different types of geospatial-analysis models can be effectively described and integrated based on standardized constraints. Then, to tackle the heterogeneity in model dependencies, we devise a prioritization-based orchestration method, facilitating optimized combinations of large-scale geospatial-analysis models. Lastly, considering the heterogeneity in execution modes, we design a heuristic scheduling method that establishes optimal mappings between models and underlying computational resources, enhancing both model stability and service performance. To validate the effectiveness and progressiveness of GeoCSIF, a prototype system was developed, and its integration process for flood disaster models was compared with mainstream methods. Experimental results indicate that GeoCSIF possesses superior performance in model management and service efficiency.

Web-GIS based Dashboard for Real-Time Data Visualization & Analysis using Open Source Technologies

Real-time visualization is the requirement for immediacy of decision making, which tends to be role-based. Using maps to visualize data can enable quicker interpretation of complex geographical phenomena, identify patterns, and aid in planning, resource allocations for policy and decision making. In present study, an interactive Web GIS Dashboard is developed with the objectives to display the work progress of Department of Soil & Water Conservation. The data includes activities and schemes undergoing in the department which was validated and geo-tagged with district & block boundary. For real-time data visualization, the graphs for different year, activities and schemes are developed for number of beneficiaries and area benefitted in ha. Various filters i.e. Year, Scheme and District are provided for viewing map. Different levels of User Authentication are provided for uploading new data and updating data

Extracting spatial effects from machine learning model using local interpretation method: An example of SHAP and XGBoost

Machine learning and artificial intelligence (ML/AI), previously considered black box approaches, are becoming more interpretable, as a result of the recent advances in eXplainable AI (XAI). In particular, local interpretation methods such as SHAP (SHapley Additive exPlanations) offer the opportunity to flexibly model, interpret and visualise complex geographical phenomena and processes. In this paper, we use SHAP to interpret XGBoost (eXtreme Gradient Boosting) as an example to demonstrate how to extract spatial effects from machine learning models. We conduct simulation experiments that compare SHAP-explained XGBoost to Spatial Lag Model (SLM) and Multi-scale Geographically Weighted Regression (MGWR) at the parameter level. Results show that XGBoost estimates similar spatial effects as those in SLM and MGWR models. An empirical example of Chicago ride-hailing modelling is presented to demonstrate the utility of SHAP with real datasets. Examples and evidence in this paper suggest that locally interpreted machine learning models are good alternatives to spatial statistical models and perform better when complex spatial and non-spatial effects (e.g. non-linearities, interactions) co-exist and are unknown.

Analysis and Research on the Temporal and Spatial Changes of Land Use in Zhaozhou County Based on Geo-information TUPU

Geo-information maps use simple language and graphics to express complex geographical phenomena, and provide an effective means for spatial analysis of land use. This article uses Landsat 8 OLI in Zhaozhou County in 2013 and 2020 as the main data source, with the support of RS and GIS technology to process the land use data in Zhaozhou County, conduct spatial analysis and explore its temporal and spatial changes. The results show that: from the current land use map, Zhaozhou County is dominated by cultivated land, from the analysis of the land use change TUPU, we can see that grassland and unused land are increasing, and the mainly transfer map is from cultivated land to unused land and cultivated land →grassland with a transfer area of 354.22km2; the largest new grassland area was from 2013 to 2010, mainly from the transfer of cultivated land, forestland and construction land. Overall, the land use pattern tended to be stable from 2013 to 2020.

Two Sets of Simple Formulae to Estimating Fractal Dimension of Irregular Boundaries

Irregular boundary lines can be characterized by fractal dimension, which provides important information for spatial analysis of complex geographical phenomena such as cities. However, it is difficult to calculate fractal dimension of boundaries systematically when image data are limited. An approximation estimation formula of boundary dimension based on square is widely applied in urban and ecological studies. But the boundary dimension is sometimes overestimated. This paper is devoted to developing a series of practicable formulae for boundary dimension estimation using ideas from fractals. A number of regular figures are employed as reference shapes, from which the corresponding geometric measure relations are constructed; from these measure relations, two sets of fractal dimension estimation formulae are derived for describing fractal-like boundaries. Correspondingly, a group of shape indexes can be defined. A finding is that different formulae have different merits and spheres of application, and the second set of boundary dimensions is a function of the shape indexes. Under condition of data shortage, these formulae can be utilized to estimate boundary dimension values rapidly. Moreover, the relationships between boundary dimension and shape indexes are instructive to understand the association and differences between characteristic scales and scaling. The formulae may be useful for the prefractal studies in geography, geomorphology, ecology, landscape science, and especially, urban science.

Geo-informatic spectrum analysis of land use change in the Manas River Basin, China during 1975-2015.

Geo-informatic spectrum analysis method was used to understand the complex geogra-phical phenomena concisely in graphic language. It is important for the integration research on spatial pattern and temporal process of land use change over multiple temporal and space scales. Based on remote sensing images in 1975, 1990, 1995, 2000, 2005, 2010 and 2015, we built the geo-spectrum of land use/cover change (LUCC) and quantitatively analyzed both the process and rend of LUCC in Manas River Basin. Results showed that the range of land use change in this basin was gradually decreasing, and land use structure tended to be simplistic, then tended to stable in later stage during 1975-1990. At the same period, the area of grassland increased greatly, which was mainly derived from unused land. The area of cultivated land expanded from 1990 to 2015, which mainly converted from unused land, forestland, and grassland. The geo-spectrum of land use change model showed that prophase change type, anaphase change type and continuous change type accounted for 1.3% of the total basin area. The overall characteristics of geo-spectrum unit were dominated by unused land converted to cultivated land, forestland and grassland. The land use change process was relatively stable. The Manas River Basin was covered 61.2% by region with comparatively stable spatial pattern. The most frequent way of land use change was increasingly expanding of cultivated land, which was displayed as: Liumaowan Bay Town of Shawan County and Liuhudi Town of Manas County were the center, then expanded to the two sides and spread to the Gurbantunggut Desert. The construct geo-spectrum of LUCC enriched the LUCC spatial-temporal analysis method system and provided an effective approach for the future research of LUCC under multiple temporal and spatial scales.

Geographical transition of urban areas. When regeneration not necessarily means gentrification. A case study from Bologna (Italy)

Some cities or neighborhoods with a specific socio-economic vocation suffer the phenomenon of change and abandonment, which generates socio-functional and economic redefinitions. This has also happened to the city of Bologna which, over the last 150 years, has experienced a rapid process of economic development, moving from a primary sector economy to a economy based on the industry, to then end up with the most advanced specializations in the tertiary sector. Thus, industries and factories (which had attracted a substantial national workforce) were closed. This has produced the formation of numerous «urban voids», which have turned into abandoned and degraded areas, such as in the Bolognina neighborhood. The consequence was initially the abandonment by the resident population: dwellings remained empty were subsequently occupied by new arrived inhabitants. At first this transition generated contrasts between the pre-existing citizens and the new ones. For this reason, the municipal administration has initiated a series of interventions, aimed on the one hand at the regeneration of the neighborhood in its ethno-socio-economic peculiarities, on the other at the refunctionalization of abandoned areas, to the point of generating a partial process of gentrification. With reference to the territorial Municipality data and according to some published essays and field researches through the Bolognina neighborhood (site inspections and conversation with residents), this article confirms the idea that urban gentrification, urban transition, and urban regeneration are territorial processes in opposition to each other. Each of them can have characteristics and features that are appearently typical of the others, but which, taking shape in a particular territory, are complex geographical phenomena.

Integrating Cellular Automata with Unsupervised Deep-Learning Algorithms: A Case Study of Urban-Sprawl Simulation in the Jingjintang Urban Agglomeration, China

An effective simulation of the urban sprawl in an urban agglomeration is conducive to making regional policies. Previous studies verified the effectiveness of the cellular-automata (CA) model in simulating urban sprawl, and emphasized that the definition of transition rules is the key to the construction of the CA model. However, existing simulation models based on CA are limited in defining complex transition rules. The aim of this study was to investigate the capability of two unsupervised deep-learning algorithms (deep-belief networks, DBN) and stacked denoising autoencoders (SDA) to define transition rules in order to obtain more accurate simulated results. Choosing the Beijing–Tianjin–Tangshan urban agglomeration as the study area, two proposed models (DBN–CA and SDA–CA) were implemented in this area for simulating its urban sprawl during 2000–2010. Additionally, two traditional machine-learning-based CA models were built for comparative experiments. The implementation results demonstrated that integrating CA with unsupervised deep-learning algorithms is more suitable and accurate than traditional machine-learning algorithms on both the cell level and pattern level. Meanwhile, compared with the DBN–CA, the SDA–CA model had better accuracy in both aspects. Therefore, the unsupervised deep-learning-based CA model, especially SDA–CA, is a novel approach for simulating urban sprawl and also potentially for other complex geographical phenomena.

Mining spatiotemporal association patterns from complex geographic phenomena

ABSTRACTSpatiotemporal association pattern mining can discover interesting interdependent relationships among various types of geospatial data. However, existing mining methods for spatiotemporal association patterns usually model geographic phenomena as simple spatiotemporal point events. Therefore, they cannot be applied to complex geographic phenomena, which continuously change their properties, shapes or locations, such as storms and air pollution. The most salient feature of such complex geographic phenomena is the geographic dynamic. To fully reveal dynamic characteristics of complex geographic phenomena and discover their associated factors, this research proposes a novel complex event-based spatiotemporal association pattern mining framework. First, a complex geographic event was hierarchically modeled and represented by a new data structure named directed spatiotemporal routes. Then, sequence mining technique was applied to discover the spatiotemporal spread pattern of the complex geographic events. An adaptive spatiotemporal episode pattern mining algorithm was proposed to discover the candidate driving factors for the occurrence of complex geographic events. Finally, the proposed approach was evaluated by analyzing the air pollution in the region of Beijing-Tianjin-Hebei. The experimental results showed that the proposed approach can well address the geographic dynamic of complex geographic phenomena, such as the spatial spreading pattern and spatiotemporal interaction with candidate driving factors.

Hyperparameter optimization of neural network-driven spatial models accelerated using cyber-enabled high-performance computing

ABSTRACTArtificial neural networks (ANNs) have been extensively used for the spatially explicit modeling of complex geographic phenomena. However, because of the complexity of the computational process, there has been an inadequate investigation on the parameter configuration of neural networks. Most studies in the literature from GIScience rely on a trial-and-error approach to select the parameter setting for ANN-driven spatial models. Hyperparameter optimization provides support for selecting the optimal architectures of ANNs. Thus, in this study, we develop an automated hyperparameter selection approach to identify optimal neural networks for spatial modeling. Further, the use of hyperparameter optimization is challenging because hyperparameter space is often large and the associated computational demand is heavy. Therefore, we utilize high-performance computing to accelerate the model selection process. Furthermore, we involve spatial statistics approaches to improve the efficiency of hyperparameter optimization. The spatial model used in our case study is a land price evaluation model in Mecklenburg County, North Carolina, USA. Our results demonstrate that the automated selection approach improves the model-level performance compared with linear regression, and the high-performance computing and spatial statistics approaches are of great help for accelerating and enhancing the selection of optimal neural networks for spatial modeling.

Computational steering for geosimulations

Geosimulations using computer simulation models provide researchers an effective way to study complex geographic phenomena and their outcomes. These simulations allow for scenario based exploration by capturing spatial and temporal relationships between various geographic processes in a region. However, current approaches to geosimulation limit manipulating model input and exploring alternative scenarios by controlling the simulation model at runtime. This paper proposes a computational steering system for geosimulation models and presents a prototype, tFUTURES , developed for the FUTURES Urban Growth Model (UGM). By allowing tangible inputs and implementing mechanisms to control model execution, this system solves the problem of lack of user-interactivity experienced at runtime. We develop a web interface and leverage the WMS, WFS-t and WPS OGC services to help visualize, modify and execute geosimulations. We define new steering controls within this interface and implement application checkpointing , allowing a user to provide new steering input and execute steering actions that can pause, advance or rollback a geosimulation and display the model outcomes in near real-time.

A bat-inspired approach to define transition rules for a cellular automaton model used to simulate urban expansion

ABSTRACTA new metaheuristic approach is presented to discover transition rules for a cellular automaton (CA) model using a novel bat movement algorithm (BA). CA is capable of simulating the evolution of complex geographical phenomena, and transition rules lie at the core of these models. An intelligence algorithm based on the echolocation behavior of bats is used to discover explicit transition rules for use in simulating urban expansion. CA transition rules are formed by links between attribute constraint items and classification items. The transition rules are derived using the BA to optimize the lower and upper threshold values for each attribute. The BA-CA model is then constructed for the simulation of urban expansion observed for Nanjing City, China. The total accuracy of newly formulated BA-CA model for this application is 86.9%, and the kappa coefficient is 0.736, which strongly suggest that the interactions of bats are effective in capturing the relationships between spatial variables and urban dynamics. It is further demonstrated that this bat-inspired BA-CA model performs better than the null model, the particle swarm optimization-based CA model (PSO-CA), and the ant colony optimization-based CA model (ACO-CA) using the same dataset. The model validation and comparison illustrate the novel capability of BA for discovering transition rules of CA during the simulation of urban expansion and potentially for other geographic phenomena.

Exploratory Method for Spatio-Temporal Feature Extraction and Clustering: An Integrated Multi-Scale Framework

This paper presents an integrated framework for exploratory multi-scale spatio-temporal feature extraction and clustering of spatio-temporal data. The framework combines the multi-scale spatio-temporal decomposition, feature identification, feature enhancing and clustering in a unified process. The original data are firstly reorganized as multi-signal time series, and then decomposed by the multi-signal wavelet. Exploratory data analysis methods, such as histograms, are used for feature identification and enhancing. The spatio-temporal evolution process of the multi-scale features can then be tracked by the feature clusters based on the data adaptive Fuzzy C-Means Cluster. The approach was tested with the global 0.25° satellite altimeter data over a period of 21 years from 1993 to 2013. The tracking of the multi-scale spatio-temporal evolution characteristics of the 1997–98 strong El Niño were used as validation. The results show that our method can clearly reveal and track the spatio-temporal distribution and evolution of complex geographical phenomena. Our approach is efficient for global scale data analysis, and can be used to explore the multi-scale pattern of spatio-temporal processes.

Simulating urban growth in a metropolitan area based on weighted urban flows by using web search engine

As a consequence of rapid and immoderate urbanization, simulating urban growth in metropolitan areas effectively becomes a crucial and yet difficult task. Cellular automata (CA) model is an attractive tool for understanding complex geographical phenomena. Although intercity urban flows, the key factors in metropolitan development, have already been taken into consideration in CA models, there is still room for improvement because the influences of urban flows may not necessarily follow the distance decay relationship and may change over time. A feasible solution is to define the weights of intercity urban flows. Therefore, this study presents a novel method based on weighted urban flows (CAWeightedFlow) with the support of web search engine. The relatedness measured by the co-occurrences of the cities’ names (toponyms) on massive web pages can be deemed as the weights of intercity urban flows. After applying the weights, the gravitational field model is integrated with Logistic-CA to fulfill the modeling task. This method is employed to the urban growth simulation in the Pearl River Delta, one of the most urbanized metropolitan areas in China, from 2005 to 2008. The results indicate that our method outperforms traditional methods with respect to two measures of calibration goodness-of-fit. For example, CAWeightedFlow can yield the best value of ‘figure of merit’. Moreover, the proposed method can be further used to explore various development possibilities by simply changing the weights.

Deriving urban dynamic evolution rules from self-adaptive cellular automata with multi-temporal remote sensing images

Cellular automata (CA) have proven to be very effective for simulating and predicting the spatio-temporal evolution of complex geographical phenomena. Traditional methods generally pose problems in determining the structure and parameters of CA for a large, complex region or a long-term simulation. This study presents a self-adaptive CA model integrated with an artificial immune system to discover dynamic transition rules automatically. The model’s parameters are allowed to be self-modified with the application of multi-temporal remote sensing images: that is, the CA can adapt itself to the changed and complex environment. Therefore, urban dynamic evolution rules over time can be efficiently retrieved by using this integrated model. The proposed AIS-based CA model was then used to simulate the rural-urban land conversion of Guangzhou city, located in the core of China’s Pearl River Delta. The initial urban land was directly classified from TM satellite image in the year 1990. Urban land in the years 1995, 2000, 2005, 2009 and 2012 was correspondingly used as the observed data to calibrate the model’s parameters. With the quantitative index figure of merit (FoM) and pattern similarity, the comparison was further performed between the AIS-based model and a Logistic CA model. The results indicate that the AIS-based CA model can perform better and with higher precision in simulating urban evolution, and the simulated spatial pattern is closer to the actual development situation.

Geons – domain-specific regionalization of space

The design of methods and tools to build adequate representations of complex geographical phenomena in a way that spatial patterns are emphasized is one of the core objectives of GIScience. In this paper, we build on the concept of geons as a strategy to represent and analyze latent spatial phenomena across different geographical scales (local, national, regional) incorporating domain-specific expert knowledge. Focusing on two types, we illustrate and exemplify how geons are generated and explored. So-called composite geons represent functional land-use classes, required for regional planning purposes. They are created via class modeling to translate interpretation schemes from mapping keys. Integrated geons, on the other hand, address abstract, yet policy-relevant phenomena such as societal vulnerability to hazards. They are delineated by regionalizing continuous geospatial data sets representing relevant indicators in a multidimensional variable space. Using the geon approach, we create spatially exhaustive sets of units, scalable to the level of policy intervention, homogenous in their domain-specific response, and independent from any predefined boundaries. From a GIScience perspective, we discuss either type of geons in a semantic hierarchy of geographic information constructs. Despite its validity for decision-making and its transferability across scales and application fields, the delineation of geons requires further methodological research to assess their statistical and conceptual robustness.

A cellular automata model for simulating the evolution of positive–negative terrains in a small loess watershed

Cellular automata (CA) have been used increasingly to simulate complex geographical phenomena. This paper proposes a CA model for simulating the evolution of dynamic positive and negative (P–N) terrains in a small loess watershed. The CA model involves a large number of attributes, including the state of P–N terrains, distance to the shoulder-line, neighbourhood condition and topographic factors. Topographic factors include the slope gradient, aspect, slope length, slope variation, aspect variation, plan curvature, profile curvature, relief amplitude and flow accumulation. The CA model was applied to simulate the evolution of P–N terrains in an indoor, small loess watershed under artificial rainfall. The transition rules for CA were constructed automatically using a decision-tree algorithm. The derived transition rules are explicit for decision-makers and helpful for generating more reliable terrains. The simulation produces encouraging results in terms of numeric accuracy and spatial distribution, in agreement with natural P–N terrains. The iterative processes show that loess negative terrains continuously erode positive terrains. The development of a loess sinkhole near the centre gully head was reproduced as well, which shows the function of loess sinkholes in the formation of loess channel systems.

Exploiting empirical knowledge for automatic delineation of city centres from large-scale topographic databases

Current topographic databases rarely represent higher order geographic phenomena, such as city centres. However, such concepts are often referred to by humans and used in various forms of spatial analysis. Hence, the value and usability of topographic databases can greatly be improved by methods that automatically create such higher order phenomena through cartographic pattern recognition techniques, departing from the very detailed, geometry-oriented representations of topographic databases. As many higher order phenomena are only vaguely defined, this paper develops and evaluates a methodology to acquire definitional knowledge about geographic phenomena by participant experiments and use this knowledge to drive the cartographic pattern recognition process. The method is applied to acquire knowledge about British city centres and delineate referents of city centre from topographic data. City centres produced for ten British cities are compared to areas derived from alternative sources. F1-scores between 0.45 and 0.88 are achieved, suggesting that the delineation produced plausible city centre areas. The benefits of our work are better (and user-driven) descriptions of complex geographic phenomena that can form the basis for accurately enriching topographic databases with additional semantics, thus yielding added value for the data producer and the end user.