类别制图及其误差建模的概念框架(英文稿)

尽管离散目标和连续场的误差建模已得到了发展, 名义场却存在实质性的和多半悬而未决的概念问题。致力于为确定信息和不确定特性整合出一个概念框架。这个概念模型是基于判别空间而构建的; 后者是由面状类别时空表象的特质或驱动过程定义的。这个模型通过加入特定类的平均结构( 其可进行基于判别变量的回归分析) 的方式, 奠定类别制图一致性的基础, 并且使基于尺度的误差建模变得更为简便易行。这种误差建模可以有效地仿效观测者在类别、边界位置、多边形个数和边界网络拓扑特性等方面的差异。通过基于模拟数据的实验, 与基于指示克里格的随机仿真结果相对比, 肯定判别空间模型在确定平均面状类别( 反映判别变量的平均响应) 以及空间不确定性( 实为空间自相关的残差在地理空间的镜像) 的复现性或可重复性。     

Building spatial models within GIS through Geo-Algebra

This paper describes Geo-Algebra, a mathematical framework for supporting geo-computational modelling in conjunction with GIS-based spacial data manipulation capabilities. Geo-Algebra overcomes the discrepancy between spacial modelling and GIS in the modes of representation as well as in their underlying concepts of space by using a common representational framework for (1) mathematical models expressing spatial relationships and (2) data models of georeferenced information. Geographic models of spatial structure and static as well as dynamic spatial interaction are formulated consistently within Gco-Algebra through a limited set of generic operations on map layers which are used simultaneously for GIS dam manipulation and analysis. A significant advantage of Geo-Algebra over other approaches to integration, such as high-level computational languages, is the development of theoretical concepts of the most general kind which allows the derivation of general properties of these in a deductive manner. In particular, Geo-Algebra formalizes and extends the mathematical notion of Map into the novel concepts of relational and metarelational Maps. These extensions lead to the novel concepts of space bridging the absolute and relative, as well as static and dynamic views of space. Such theoretical concepts are also implemented in a dynamic simulation tool called Geocellular .     

Semantic Interoperability through the Definition of Conceptual Model Transformations

The integration of distributed geospatial data is a great issue in GI Science. Hereby, the collection of data from different systems and different structures as well as across country borders is a problem. We explain a model‐based four‐step approach to conceptual data modelling. Within this approach, we introduce a methodology to map data models on the conceptual level without touching data formats. First, we developed a conceptual mapping language called “UMLT” which is the core of our approach. Furthermore, we integrated UMLT into the Model‐Driven Architecture where a conceptual model mapping is referred to as a platform independent model mapping. Finally, the data format transformation is deduced from the model mapping automatically. This represents the implementation mapping from a platform independent model to a platform specific model. Together with transformation operations for attribute mapping, this approach is a means for effective and enhanced semantic interoperability.     

Quantifying uncertainty in remote sensing-based urban land-use mapping

Land-use/land-cover information constitutes an important component in the calibration of many urban growth models. Typically, the model building involves a process of historic calibration based on time series of land-use maps. Medium-resolution satellite imagery is an interesting source for obtaining data on land-use change, yet inferring information on the use of urbanised spaces from these images is a challenging task that is subject to different types of uncertainty. Quantifying and reducing the uncertainties in land-use mapping and land-use change model parameter assessment are therefore crucial to improve the reliability of urban growth models relying on these data. In this paper, a remote sensing-based land-use mapping approach is adopted, consisting of two stages: (i) estimating impervious surface cover at sub-pixel level through linear regression unmixing and (ii) inferring urban land use from urban form using metrics describing the spatial structure of the built-up area, together with address data. The focus lies on quantifying the uncertainty involved in this approach. Both stages of the land-use mapping process are subjected to Monte Carlo simulation to assess their relative contribution to and their combined impact on the uncertainty in the derived land-use maps. The robustness to uncertainty of the land-use mapping strategy is addressed by comparing the most likely land-use maps obtained from the simulation with the original land-use map, obtained without taking uncertainty into account. The approach was applied on the Brussels-Capital Region and the central part of the Flanders region (Belgium), covering the city of Antwerp, using a time series of SPOT data for 1996, 2005 and 2012. Although the most likely land-use map obtained from the simulation is very similar to the original land-use map – indicating absence of bias in the mapping process – it is shown that the errors related to the impervious surface sub-pixel fraction estimation have a strong impact on the land-use map's uncertainty. Hence, uncertainties observed in the derived land-use maps should be taken into account when using these maps as an input for modelling of urban growth.     

Simulation Modelling Applied to Runoff Modelling Using MapScript

Dynamic spatial analysis addresses computational aspects of space–time processing. This paper describes the development of a spatial analysis tool and modelling framework that together offer a solution for simulating landscape processes. A better approach to integrating landscape spatial analysis with Geographical Information Systems is advocated in this paper. Enhancements include special spatial operators and map algebra language constructs to handle dispersal and advective flows over landscape surfaces. These functional components to landscape modelling are developed in a modular way and are linked together in a modelling framework that performs dynamic simulation. The concepts and modelling framework are demonstrated using a hydrological modelling example. The approach provides a modelling environment for scientists and land resource managers to write and to visualize spatial process models with ease.     

Using Pattern‐oriented Modeling (POM) to Cope with Uncertainty in Multi‐scale Agent‐based Models of Land Change

Local land‐use and ‐cover changes (LUCCs) are the result of both the decisions and actions of individual land‐users, and the larger global and regional economic, political, cultural, and environmental contexts in which land‐use systems are embedded. However, the dearth of detailed empirical data and knowledge of the influences of global/regional forces on local land‐use decisions is a substantial challenge to formulating multi‐scale agent‐based models (ABMs) of land change. Pattern‐oriented modeling (POM) is a means to cope with such process and parameter uncertainty, and to design process‐based land change models despite a lack of detailed process knowledge or empirical data. POM was applied to a simplified agent‐based model of LUCC to design and test model relationships linking global market influence to agents’ land‐use decisions within an example test site. Results demonstrated that evaluating alternative model parameterizations based on their ability to simultaneously reproduce target patterns led to more realistic land‐use outcomes. This framework is promising as an agent‐based virtual laboratory to test hypotheses of how and under what conditions driving forces of land change differ from a generalized model representation depending on the particular land‐use system and location.     

Revealing Uncertainties in Land Change Modeling Using Probabilities

Land change models are frequently used to analyze current land change processes and possible future developments. However, the outcome of such models is accompanied by uncertainties that have to be taken into account in order to address their reliability for science and decision‐making. While a range of approaches exist that quantify the disagreement of land change maps, the quantification of uncertainty remains a major challenge. The aim of this article is therefore to reveal uncertainties in land change modeling by developing two measures: quantity uncertainty and allocation uncertainty. We choose a Bayesian Belief Network modeling approach for deforestation in Brazil to develop and apply the two measures to the resulting probability surface. Quantity uncertainty describes the uncertainty about the correct number of cells in a land change map assigned to different land change categories and allocation uncertainty expresses the uncertainty about the correct spatial placement of a cell in the land change map. Thus, uncertainty can be quantified even in those cases where no reference data exist. Informing about uncertainty in probabilistic outcomes may be an important asset when land change projections are being used in science and decision‐making and moreover, they may also be further evaluated for other spatial applications.     

Primitives as building blocks for constructing land cover maps

Land cover maps play an integral role in environmental management. However, countries and institutes encounter many challenges with producing timely, efficient, and temporally harmonized updates to their land cover maps. To address these issues we present a modular Regional Land Cover Monitoring System (RLCMS) architecture that is easily customized to create land cover products using primitive map layers. Primitive map layers are a suite of biophysical and end member maps, with land cover primitives representing the raw information needed to make decisions in a dichotomous key for land cover classification. We present best practices to create and assemble primitives from optical satellite using computing technologies, decision tree logic and Monte Carlo simulations to integrate their uncertainties. The concept is presented in the context of a regional land cover map based on a shared regional typology with 18 land cover classes agreed on by stakeholders from Cambodia, Laos PDR, Myanmar, Thailand, and Vietnam. We created annual map and uncertainty layers for the period 2000–2017. We found an overall accuracy of 94% when taking uncertainties into account. RLCMS produces consistent time series products using free long term historical Landsat and MODIS data. The customizable architecture can include a variety of sensors and machine learning algorithms to create primitives and the best suited smoothing can be applied on a primitive level. The system is transferable to all regions around the globe because of its use of publicly available global data (Landsat and MODIS) and easily adaptable architecture that allows for the incorporation of a customizable assembly logic to map different land cover typologies based on the user's landscape monitoring objectives     

Classifying historical remotely sensed imagery using a tempo-spatial feature evolution (T-SFE) model

Large and growing archives of orbital imagery of the earth’s surface collected over the past 40 years provide an important resource for documenting past and current land cover and environmental changes. However uses of these data are limited by the lack of coincident ground information with which either to establish discrete land cover classes or to assess the accuracy of their identification. Herein is proposed an easy-to-use model, the Tempo-Spatial Feature Evolution (T-SFE) model, designed to improve land cover classification using historical remotely sensed data and ground cover maps obtained at later times. This model intersects (1) a map of spectral classes (S-classes) of an initial time derived from the standard unsupervised ISODATA classifier with (2) a reference map of ground cover types (G-types) of a subsequent time to generate (3) a target map of overlaid patches of S-classes and G-types. This model employs the rules of Count Majority Evaluation, and Subtotal Area Evaluation that are formulated on the basis of spatial feature evolution over time to quantify spatial evolutions between the S-classes and G-types on the target map. This model then applies these quantities to assign G-types to S-classes to classify the historical images. The model is illustrated with the classification of grassland vegetation types for a basin in Inner Mongolia using 1985 Landsat TM data and 2004 vegetation map. The classification accuracy was assessed through two tests: a small set of ground sampling data in 1985, and an extracted vegetation map from the national vegetation cover data (NVCD) over the study area in 1988. Our results show that a 1985 image classification was achieved using this method with an overall accuracy of 80.6%. However, the classification accuracy depends on a proper calibration of several parameters used in the model.     

An Assessment of Deforestation Models for Reducing Emissions from Deforestation and Forest Degradation (REDD)

With the increasing concerns in developing methodologies for Reducing Emissions from Deforestation and forest Degradation (REDD) projects, there is a need to understand the characteristics of existing Land‐Use/Cover Change (LUCC) modules. This research presents a modular framework for assessing predictive accuracy of business‐as‐usual deforestation in the future by comparing two existing approaches: GEOMOD Modeling (GM) and Land Change Modeler (LCM). The comparison uses data from a case study in Chiquitanía, Bolivia. Data from 1986 and 1994 are used to simulate land‐cover of 2000; the resulting maps are compared with an observed land‐cover map of 2000. GM and LCM simulate business‐as‐usual deforestations at the pixel level. The model structures of GM's linear extrapolation and LCM's Markov Chain are compared to review quantity of LUCC; and the model structures of GM's empirical frequency, LCM's logistic regression, and LCM's multilayer perceptron are compared to review (spatial) allocation of LUCC. Relative operating characteristics, figure of merit, and multiple resolution analysis are employed to assess predictive accuracy of multiple transition modeling. By design, GM lacks the potential to model multiple transitions, and the LCM's multilayer perceptron may produce different results for each simulation due to its stochastic element. Based on the model structure and predictive accuracy comparisons, the LCM seems more suitable than the GM for a REDD application. When a project is to employ a predictive method for its spatially explicit baseline setting, then it is highly recommended to use the proposed framework to assess accuracy of the baseline as part of a project design document.     

Developing spatially and thematically detailed backdated maps for land cover studies

ABSTRACT

Global or regional land cover change on a decadal time scale can be studied at a high level of detail using the availability of remote sensing data such as that provided by Landsat. However, there are three main technical challenges in this goal. First, the generation of land cover maps without reference data is problematic (backdating). Second, it is important to maintain high accuracies in land cover change map products, requiring a reasonably rich legend within each map. Third, a high level of automation is necessary to aid the management of large volumes of data. This paper describes a robust methodology for processing time series of satellite data over large spatial areas. The methodology includes a retrospective analysis used for the generation of training and test data for historical periods lacking reference information. This methodology was developed in the context of research on global change in the Iberian Peninsula. In this study we selected two scenes covering geographic regions that are representative of the Iberian Peninsula. For each scene, we present the results of two classifications (1985–1989 and 2000–2004 quinquennia), each with a legend of 13 categories. An overall accuracy of over 92% was obtained for all 4 maps.     

Categories are in flux,but their computational representations are fixed: That's a problem

As research advances, our conceptual understanding also changes. Computational approaches do little to recognize the evolution that occurs at the conceptual level during the research process. This can result in misunderstanding between knowledge producers and consumers and so inhibit the reusability of outcomes. In this article, we describe how changes at the conceptual level can be represented, along with related changes to data and methods, and how appropriate connections between these various artefacts can be maintained. To demonstrate these ideas, we show how categories used in remote sensing and land cover analysis change over time and how these changes are linked to various research activities. We present a new system (called AdvoCate) that augments typical GIS and remote sensing functionality with a conceptual model of categories that can undergo change, and that also captures the cause of conceptual change and its extent. We argue that concepts and categories should be represented explicitly and richly within GIS, because without this, we have a poor idea of what our modeled entities really mean, and by implication how they should be used appropriately. We demonstrate the usefulness of this deeper representation using examples of category evolution from a land cover mapping exercise.     

Unambiguous Reconstruction and High-Resolution Imaging for Multiple-Channel SAR and Airborne Experiment Results

Azimuth ambiguity occurs in synthetic aperture radar (SAR) systems due to the well-known constraint of minimum antenna area, particularly at high resolutions and wide swaths. A space time domain method can be utilized to remove this ambiguity if the multiple-channel data are available. In this letter, a modified approach is presented to determine the filter weight vectors. This approach was successfully applied to the real data, which were collected by an experimental airborne multiple-channel SAR system. The channel imbalance and the error in antenna phase center position are analyzed in detail.     

Creating a conceptual framework to improve the re‐usability of open geographic data in cities

Open data has a profound effect on the working environment within which information is created and shared at all levels. At the local government level, open data initiatives have resulted in higher transparency in policy, a greater engagement between decision‐makers and citizens, and have changed the culture about how data analysis and evidence are used to support local governance. This article, based on data collected through an on‐line survey, participatory workshops with data user communities in four cities (in Colombia and Spain), and interviews with Valencia good‐government office, identifies four elements for a conceptual framework to improve the re‐usability of open geographic data in cities. The essential elements defined in this research are the definition of data user communities and their needs, the creation of the community of reuse, user‐focused metadata, and reuse‐focused legal terms. The definition of these indicators provides a framework for authorities to re‐shape their current open data strategy to include data user requirements. At the end of this article, a roadmap for future research and implementation is presented, considering some reflections on the conceptual framework.     

Estimating the uncertainty of land-cover extrapolations while constructing a raster map from tabular data

This paper presents novel techniques to estimate the uncertainty in extrapolations of spatially-explicit land-change simulation models. We illustrate the concept by mapping a historic landscape based on: 1) tabular data concerning the quantity in each land cover category at a distant point in time at the stratum level, 2) empirical maps from more recent points in time at the grid cell level, and 3) a simulation model that extrapolates land-cover change at the grid cell level. This paper focuses on the method to show uncertainty explicitly in the map of the simulated landscape at the distant point in time. The method requires that validation of the land-cover change model be quantified at the grid-cell level by Kappa for location (Klocation). The validation statistic is used to estimate the certainty in the extrapolation to a point in time where an empirical map does not exist. As an example, we reconstruct the 1951 landscape of the Ipswich River Watershed in Massachusetts, USA. The technique creates a map of 1951 simulated forest with an overall estimated accuracy of 0.91, with an estimated users accuracy ranging from 0.95 to 0.84. We anticipate that this method will become popular, because tabular information concerning land cover at coarse stratum-level scales is abundant, while digital maps of the specific location of land cover are needed at a finer spatial resolution. The method is a key to link non-spatial models with spatially-explicit models.     

Impacts of Spatial Partitioning in Hydroecological Models: Predicting Grassland Productivity with RHESSys

Environmental models constructed with a spatial domain require choices about the representation of space. Decisions in the adaptation of a spatial data model can have significant consequences on the ability to predict environmental function as a result of changes to levels of aggregation of input parameters and scaling issues in the processes being modelled. In some cases, it is possible to construct a systematic framework to evaluate the uncertainty in predictions using different spatial models; in other cases, the realm of possibilities plus the complexity of the environmental model in question may inhibit numeric uncertainty estimates. We demonstrate a range of potential spatial data models to parameterize a landscape‐level hydroecological model (RHESSys). The effects of data model choice are illustrated, both in terms of input parameter distributions and resulting ecophysiological predictions. Predicted productivity varied widely, as a function of both the number of modelling units, and of arbitrary decisions such as the origin of a raster grid. It is therefore important to use as much information about the modelled environment as possible. Combinations of adaptive methods to evaluate distributions of input data, plus knowledge of dominant controls of ecosystem processes, can help evaluate potential representations. In this case, variance‐based delineation of vegetation patches is shown to improve the ability to intelligently choose a patch distribution that minimizes the number of patches, while maintaining a degree of aggregation that does not overly bias the predictions.