分布式水文模型的并行计算研究进展

大流域、高分辨率、多过程耦合的分布式水文模拟计算量巨大,传统串行计算技术不能满足其对计算能力的需求,因此需要借助于并行计算的支持。本文首先从空间、时间和子过程三个角度对分布式水文模型的可并行性进行了分析,指出空间分解的方式是分布式水文模型并行计算的首选方式,并从空间分解的角度对水文子过程计算方法和分布式水文模型进行了分类。然后对分布式水文模型的并行计算研究现状进行了总结。其中,在空间分解方式的并行计算方面,现有研究大多以子流域作为并行计算的基本调度单元;在时间角度的并行计算方面,有学者对时空域双重离散的并行计算方法进行了初步研究。最后,从并行算法设计、流域系统综合模拟的并行计算框架和支持并行计算的高性能数据读写方法3个方面讨论了当前存在的关键问题和未来的发展方向。     

A machine learning approach for predicting computational intensity and domain decomposition in parallel geoprocessing

ABSTRACT

High performance computing is required for fast geoprocessing of geospatial big data. Using spatial domains to represent computational intensity (CIT) and domain decomposition for parallelism are prominent strategies when designing parallel geoprocessing applications. Traditional domain decomposition is limited in evaluating the computational intensity, which often results in load imbalance and poor parallel performance. From the data science perspective, machine learning from Artificial Intelligence (AI) shows promise for better CIT evaluation. This paper proposes a machine learning approach for predicting computational intensity, followed by an optimized domain decomposition, which divides the spatial domain into balanced subdivisions based on the predicted CIT to achieve better parallel performance. The approach provides a reference framework on how various machine learning methods including feature selection and model training can be used in predicting computational intensity and optimizing parallel geoprocessing against different cases. Some comparative experiments between the approach and traditional methods were performed using the two cases, DEM generation from point clouds and spatial intersection on vector data. The results not only demonstrate the advantage of the approach, but also provide hints on how traditional GIS computation can be improved by the AI machine learning.     

Strategies for real-time spatial analysis using massively parallel SIMD computers: an application to urban traffic flow analysis

Abstract

The current research focuses upon the development of a methodology for undertaking real-time spatial analysis in a supercomputing environment, specifically using massively parallel SIMD computers. Several approaches that can be used to explore the parallelization characteristics of spatial problems are introduced. Within the focus of a methodology directed toward spatial data parallelism, strategies based on both location-based data decomposition and object-based data decomposition are proposed and a programming logic for spatial operations at local, neighborhood and global levels is also recommended. An empirical study of real-time traffic flow analysis shows the utility of the suggested approach for a complex, spatial analysis situation. The empirical example demonstrates that the proposed methodology, especially when combined with appropriate programming strategies, is preferable in situations where critical, real-time, spatial analysis computations are required. The implementation of this example in a parallel environment also points out some interesting theoretical questions with respect to the theoretical basis underlying the analysis of large networks.     

Soil-landscape modelling and spatial prediction of soil attributes

Abstract

Explicit and quantitative models for the spatial prediction of soil and landscape attributes are required for environmental modelling and management. In this study, advances in the spatial representation of hydrological and geomorphological processes using terrain analysis techniques are integrated with the development of a field sampling and soil-landscape model building strategy. Statistical models are developed using relationships between terrain attributes (plan curvature, compound topographic index, upslope mean plan curvature) and soil attributes (A horizon depth, Solum depth, E horizon presence/absence) in an area with uniform geology and geomorphic history. These techniques seem to provide appropriate methodologies for spatial prediction and understanding soil landscape processes.     

Coordinating hazardous waste management activities using geographical information systems

Abstract

This paper describes a framework for the role of geographical information systems (GIS) in the monitoring and management of hazardous waste sites. Compilation of required information, incorporation of existing strategies for waste monitoring, analysis of these data in a GIS environment and the integration of computerized models for transport processes are discussed. Examples for the analysis of spatial data using techniques of cartographic overlay and the implementation of geo-statistical methods on monitoring data are provided from work in progress by the authors. These examples are set in the context of developing a fully integrated monitoring and management system utilizing GIS technology.     

A critical evaluation of 3D spatial information models for managing legal arrangements of multi-owned developments in Victoria,Australia

Three-dimensional (3D) spatial information models are increasingly being adopted to help communicate the spatial dimensions of complex built environments. Land administration practices in multi-owned developments include the subdivision, registration and management of legal interests associated with private, communal and public properties, which are often located along the vertical dimension. The spatial structure of each legal interest is normally composed of invisible spaces, defined as the inside and outside of multi-owned developments, as well as physically built structures. Additionally, a wide variety of legal boundary types mark out the spatial limits of the individual parts of each legal interest. These legal boundaries are typically delineated by either relying on fixed surveying measurements or referencing physically existent objects. This article provides a critical assessment of 3D spatial information models in terms of their capabilities for modelling legal interests and legal boundaries defined in the Victorian jurisdiction. We classify these models into three categories: purely legal, purely physical and integrated. This assessment provides the basis for developing a new 3D spatial information model, which would subsequently support a pathway towards realising the Victorian land administration system in a 3D digital environment.     

Identification of field attribute error under different models of spatial variation

Abstract

Recent developments in theory and computer software mean that it is now relatively straightforward to evaluate how attribute errors are propagated through quantitative spatial models in GIS. A major problem, however, is to estimate the errors associated with the inputs to these spatial models. A first approach is to use the root mean square error, but in many cases it is better to estimate the errors from the degree of spatial variation and the method used for mapping. It is essential to decide at an early stage whether one should use a discrete model of spatial variation (DMSV—homogeneous areas, abrupt boundaries), a continuous model (CMSV—a continuously varying regionalized variable field) or a mixture of both (MMSV—mixed model of spatial variation). Maps of predictions and prediction error standard deviations are different in all three cases, and it is crucial for error estimation which model of spatial variation is used. The choice of model has been insufficiently studied in depth, but can be based on prior information about the kinds of spatial processes and patterns that are present, or on validation results. When undetermined it is sensible to adopt the MMSV in order to bypass the rigidity of the DMSV and CMSV. These issues are explored and illustrated using data on the mean highest groundwater level in a polder area in the Netherlands.     

Developing a parallel computational implementation of AMOEBA

As geospatial researchers' access to high-performance computing clusters continues to increase alongside the availability of high-resolution spatial data, it is imperative that techniques are devised to exploit these clusters' ability to quickly process and analyze large amounts of information. This research concentrates on the parallel computation of A Multidirectional Optimal Ecotope-Based Algorithm (AMOEBA). AMOEBA is used to derive spatial weight matrices for spatial autoregressive models and as a method for identifying irregularly shaped spatial clusters. While improvements have been made to the original ‘exhaustive’ algorithm, the resulting ‘constructive’ algorithm can still take a significant amount of time to complete with large datasets. This article outlines a parallel implementation of AMOEBA (the P-AMOEBA) written in Java utilizing the message passing library MPJ Express. In order to account for differing types of spatial grid data, two decomposition methods are developed and tested. The benefits of using the new parallel algorithm are demonstrated on an example dataset. Results show that different decompositions of spatial data affect the computational load balance across multiple processors and that the parallel version of AMOEBA achieves substantially faster runtimes than those reported in related publications.     

NEAT approach for testing and validation of geospatial network agent-based model processes: case study of influenza spread

ABSTRACT

Agent-based models (ABM) are used to represent a variety of complex systems by simulating the local interactions between system components from which observable spatial patterns at the system-level emerge. Thus, the degree to which these interactions are represented correctly must be evaluated. Networks can be used to discretely represent and quantify interactions between system components and the emergent system structure. Therefore, the main objective of this study is to develop and implement a novel validation approach called the NEtworks for ABM Testing (NEAT) that integrates geographic information science, ABM approaches, and spatial network representations to simulate complex systems as measurable and dynamic spatial networks. The simulated spatial network structures are measured using graph theory and compared with empirical regularities of observed real networks. The approach is implemented to validate a theoretical ABM representing the spread of influenza in the City of Vancouver, Canada. Results demonstrate that the NEAT approach can validate whether the internal model processes are represented realistically, thus better enabling the use of ABMs in decision-making processes.     

Where Non-academic Geographers are Employed

In traditional, constrained spatial interaction models, the number of predicted movers leaving origins and entering destinations is constrained to match exactly the observed number. In relaxed models, these constraints are allowed to vary over a range of values in order to provide greater flexibility in calibration. This paper identifies a new, seven-member family of relaxed spatial interaction models, based upon the generalization of the constraint sets used in model derivation. Three categories are suggested, including single and doubly relaxed models, cost-relaxed models, and totally relaxed models. This paper introduces these relaxed models as entropy-maximizing ones, proposes a terminology for them, and describes empirical situations in which they are useful.     

Topographic radiation balance models: Sensitivity and application in periglacial geomorphology

Direct solar radiation integrated over one year is a function of latitude and time of year, and topographic slope , aspect and shadowing control the local distribution. Recently, several spatial models have been developed which estimate the radiation balance based on digital elevation models, taking into account aspect, slope and shadowing effects. For the periglacial realm, these models are integrated both in models estimating possible occurrence of mountain permafrost and in studies of active layer dynamics. In this article our aim is to assess and discuss sensitivity and validations of the radiation balance model SRAD, in comparison with two other topographic-based radiation models. The study site and field data are from Finse, Southern Norway.     

Handling Geographic Information. Report of the Committee of Enquiry Chaired by Lord Chorley,Department of the Environment. (London: HMSO, 1987.) [Pp. 208. ] Price £l14-95. ISBN Oil 752015 2.

ABSTRACT

A wide range of environmental process simulations would benefit from the development of GIS able to cope with the additional dimensions of vertical space and time, and having extended spatial modelling facilities. The results of a project in the first of these areas, namely on the development of techniques for handling four-dimensional (4-D) data, are described. The key topics of data models, visualization and interpolation have been studied. Problems include the large size of some 4-D data sets, the sparseness of sampling in some dimensions compared with others, and the need to combine data sets from different sensors which may be of different dimensionalities and scales. The gridded volume data common to most environmental models are stored in a 4-D bintree form in which all dimensions are treated identically. Simple 4-D objects may also be stored, and links are provided between volume and object databases. The techniques have been implemented within a computational testbed allowing parallel processing to cope with large data sizes.     

The potential of GIS modelling of gravity flows and slope instabilities

Abstract

The component of gravitational acceleration parallel to the slope of the local surface partly determines the state of slope stability and the kinematics of flow under gravity on that slope. Geographical information systems based on digital elevation models offer the potential to be able to map this variable and permit the modelling of a variety of stability criteria and surface processes including landslides, rock avalanches, pyroclastic flows and lava flows. Three types of models and the basic map operations required to run them are discussed. The models are as follows: (i) sites of potential shallow slope failure (e.g. landslides), (ii) maps of flow deposition based on energy balance calculations (e.g. rock avalanches) and (iii) finite difference, initial value type simulations of dynamic flow (e.g. lava flows). The potential value of these models to hazard assessment is great but their application in specific cases must be assessed with reference to the accuracy of the digital elevation model used.     

Design of a spatial data structure using the relational normal forms

Abstract

In previous work, a relational data structure aimed at the exchange of spatial data between systems was developed. As this data structure was relational it was of first normal form, but compliance with the higher normal forms was not investigated. Recently, a new procedural method for composing fully normalized data structures from the basic data fields has been developed by H. C. Smith, as an alternative to the process of non-loss decomposition which is difficult to understand. Smith's method has been applied to data fields required to store points, lines and polygons in a chain-node spatial data model. When geographic domain, coverage layer and map are also considered, the procedure naturally leads to a catalogue model, needed for the exchange of spatial data. Although the method produces a fully normalized data structure, it is not as easy to identify which normal forms are responsible for the ultimate arrangement of the data fields into relations, but the benefits of these criteria for data base development also apply to spatial data structures and related ancillary data.     

Incorporation of multi-scale spatial autocorrelation in soil moisture–landscape modeling

Effects of spatial autocorrelation (SAC), or spatial structure, have often been neglected in the conventional models of pedogeomorphological processes. Based on soil, vegetation, and topographic data collected in a coastal dunefield in western Korea, this research developed three soil moisture–landscape models, each incorporating SAC at fine, broad, and multiple scales, respectively, into a non-spatial ordinary least squares (OLS) model. All of these spatially explicit models showed better performance than the OLS model, as consistently indicated by R², Akaike’s information criterion, and Moran’s I. In particular, the best model was proved to be the one using spatial eigenvector mapping, a technique that accounts for spatial structure at multiple scales simultaneously. After including SAC, predictor variables with greater inherent spatial structure underwent more reduction in their predictive power than those with less structure. This finding implies that the environmental variables pedogeomorphologists have perceived important in the conventional regression modeling may have a reduced predictive power in reality, in cases where they possess a significant amount of SAC. This research demonstrates that accounting for spatial structure not only helps to avoid the violation of statistical assumptions, but also allows a better understanding of dynamic soil hydrological processes occurring at different spatial scales.     

A Hierarchy of Spatial Marginality Through Spatial Filtering

Abstract

Taking a clue from classic time-series decomposition, this article demonstrates a spatial filtering and search technique that permits the partitioning of a measure of marginality, here measured as the percent of the population living at less than 50 percent of the poverty level, into macro, meso, and micro components. This approach supports theory that has argued for scale-specific explanations of spatial marginality. The technique also offers promise for many other types of investigations such as disease incidence, microclimate dynamics, and consumer preferences.     

An iconic query language for topological relationships in GIS

Abstract

The study of query languages for spatial databases is an active research area. This paper describes a new spatial query language that uses a visual grammar to express topological relationships. It is supplemented by text and icons to handle other spatial and non-spatial queries. A graphical user interface is also developed to provide an interactive environment for composing the iconic query command. To test the language, the interface is implemented on a SUN 4 Workstation and linked to Ingres, a relational DBMS. Preliminary tests show that the iconic query language is more convenient for expressing spatial concepts than conventional textual languages. This is due mainly to the two-dimensionality of iconic languages in contrast with the linear nature of conventional languages.     

Local interpolation using a distributed parallel supercomputer

Abstract

Large spatial interpolation problems present significant computational challenges even for the fastest workstations. In this paper we demonstrate how parallel processing can be used to reduce computation times to levels that are suitable for interactive interpolation analyses of large spatial databases. Though the approach developed in this paper can be used with a wide variety of interpolation algorithms, we specifically contrast the results obtained from a global ‘brute force’ inverse–distance weighted interpolation algorithm with those obtained using a much more efficient local approach. The parallel versions of both implementations are superior to their sequential counterparts. However, the local version of the parallel algorithm provides the best overall performance.     

Logical data modelling for geographical applications

Abstract.

A formal, yet practical, GeoRelational Data Model (GRDM) is presented for the logical database design phase of the development of spatial information systems. Geographic applications are viewed in the context of information systems development. The generic needs of modelling spatial data are analyzed; it is concluded that they are not served satisfactorily by existing data models, so specifications of modelling tools for spatial application design are given. GRDM provides a set of representational constructs (relations and layers for the logical schema; virtual layers, object classes and spatial constraints for the user views) on top of well-established models. It constitutes part of a full, easily automated application design methodology. Extensive examples demonstrate the relevance, and ease-of-use of the platform-independent GRDM.