G. M. Smith

The UK Land Cover Map 2000: Construction of a parcel-based vector Map from satellite images

Abstract Land Cover Map 2000 (LCM2000) is a thematic classification of satellite image data covering the entire United Kingdom. The map updates and substantially upgrades the Land Cover Map of Great Britain (LCMGB), made in 1990–92. This paper outlines the character of the map through a description of its specification, production and outputs. The paper is aimed at users of LCM2000 and derived data who need to understand more of the map and its characteristics. The paper also outlines plans for making data available to researchers and applied users. The most important development in LCM2000 was the spectral segmentation of images to generate vector land parcels. Land cover was then identified by the spectral classification of the image data in these parcels. Classification used specially developed procedures which exploited known spatial, spectral and contextual characteristics of land cover. The resultant GIS incorporates, within its vector structure, detailed attribute data which record parcel-based land cover, plus information on class probabilities, data on within-parcel heterogeneity, information on landscape structure and context, cover information from LCMGB, together with a record of each parcels processing history.

The characterisation and measurement of land cover change through remote sensing: problems in operational applications?

Abstract This paper examines issues surrounding the detection, measurement and characterisation of landscape changes by remote sensing and other means. First it compares the remotely sensed Land Cover Map of Great Britain (LCMGB), 1990, with the UK Land Cover Map 2000 (LCM2000). Then it considers the more general circumstances where users compare any two thematic maps incorporating independent inaccuracies, to record changes. The paper concludes that the measurement of small to medium scale changes over large areas requires levels of precision in mapping which are near impossible to achieve with satellite image classifications alone; indeed they may be difficult to achieve in any form of survey unless it is tailor-made and rigorously applied specifically to the recording of real changes. The paper argues that users of all such data products therefore need to treat with extreme caution their maps of differences and their conclusions about change. It suggests that users need also to develop intelligent approaches, which draw upon a broader knowledge of the directions, patterns and scale of the changes to be recorded, in order to refine their assessments.

UK Land Cover Map Production Through the Generalisation of OS MasterMap

Abstract Earth Observation (EO) data is seen as a major source of information to characterise the Earths surface, but is conventionally analysed using pixel-based approaches that do not incorporate the concept of landscape features or real-world objects. The UK land cover maps to date have been developed in an attempt to exploit landscape features to improve the quality and accuracy of their derived products. For Land Cover Map 2007 (LCM2007) landscape features will be derived from a generalised version of OS MasterMap to capture the required real-world objects. This paper describes the generalisation process that aligns the scale of the landscape features with the information content of high spatial resolution EO data as the first step in the production of LCM2007.

Land cover heterogeneity in Great Britain as identified in Land Cover Map 2000

Land cover is an important component of landscape character. The spatial configuration and heterogeneity of land cover can influence species distributions and patterns of biodiversity. Sustainable countryside planning and land management policy in Great Britain require the mapping of land cover heterogeneity at a national scale. Here, we map four measures of land cover heterogeneity across Great Britain using Land Cover Map 2000. We calculate land cover richness, diversity, evenness and similarity within and between 1 × 1 km grid cells of the British National Grid. From this we are able to identify assemblages of land cover types that are associated with high or low landscape heterogeneity, and where they occur geographically.

Airborne remote sensing for monitoring the impact of coastal zone management

Managed realignment of the coastline is being trialled in the UK as a means of coping with elevated sea levels on sedimentary coasts. In this process, current sea defences are breached allowing the tide to flood previously reclaimed land up to a new defence line. This helps to reduce the cost of sea defences by decreasing their height requirements and by increasing the amount of intertidal area over which wave energy is absorbed. Managed realignment also promotes the creation or extension of important coastal habitats such as saltmarshes which are being squeezed against sea defences by rising sea levels.

Contextual Analyses of Remotely Sensed Images for the Operational Classification of Land Cover in United Kingdom

Th e use of remote sensing and image analysis for operational mapping at national scales requires a range of tools to help achieve the necessary accuracy and consistency despite variations in image quality, image acquisition dates, seasonal changes in vegetation, and a range of other confounding factors which distort or alter spectral signatures of the target classes. Contextual analyses can help to generate a coherent map structure, correct errors and extend the thematic scope of classifi cations. Th is is especially important when producers are asked to meet the wide-ranging needs of multiple end-users.

Potential of hyperspectral imaging to assess the stability of mudflat surfaces by mapping sediment characteristics

This work assessed the suitability of hyperspectral data for estimating mudflat surface characteristics related to stability. Due to the inaccessibility of intertidal areas, precise ground-based measurements of mudflat stability are difficult to conduct. Remote sensing can provide full spatial coverage and non-intrusive measurement. As stability changes on mudflats are linked to subtle differences in mudflat surface characteristics, they can potentially be mapped by hyperspectral data. Hyperspectral images were collected along with near contemporary ground measurements. An unsupervised classification gave a map which confirmed that a channel bar was mainly sand whereas soft mud dominated an adjacent embayment. Multiple regression analysis was used to relate surface characteristics to hyperspectral data to construct regression equations. Erosion shear stress was estimated directly from the hyperspectral data and also by a relationship with the surface characteristics. The results of the thematic class map matched well with the known situation at the site during image acquisition. The maps of surface characteristics highlighted the additional information that can be extracted from hyperspectral data. Stability maps, based on the erosion shear stress, can be used as a basis for predicting the likely future behaviour in this dynamic environment and will be of use for coastal zone management.