Number Of Land Cover Classes Research Articles

Evaluation of the Spatial Representativeness of In Situ SIF Observations for the Validation of Medium-Resolution Satellite SIF Products

The upcoming Fluorescence Explorer (FLEX) mission will provide sun-induced fluorescence (SIF) products at unprecedented spatial resolution. Thus, accurate calibration and validation (cal/val) of these products are key to guarantee robust SIF estimates for the assessment and quantification of photosynthetic processes. In this study, we address one specific component of the uncertainty budget related to SIF retrieval: the spatial representativeness of in situ SIF observations compared to medium-resolution SIF products (e.g., 300 m pixel size). Here, we propose an approach to evaluate an optimal sampling strategy to characterise the spatial representativeness of in situ SIF observations based on high-spatial-resolution SIF data. This approach was applied for demonstration purposes to two agricultural areas that have been extensively characterized with a HyPlant airborne imaging spectrometer in recent years. First, we determined the spatial representativeness of an increasing number of sampling points with respect to a reference area (either monocultural crop fields or hypothetical FLEX pixels characterised by different land cover types). Then, we compared different sampling approaches to determine which strategy provided the most representative reference data for a given area. Results show that between 3 and 13.5 sampling points are needed to characterise the average SIF value of both monocultural fields and hypothetical FLEX pixels of the agricultural areas considered in this study. The number of sampling points tends to increase with the standard deviation of SIF of the reference area, as well as with the number of land cover classes in a FLEX pixel, even if the increase is not always statistically significant. This study contributes to guiding cal/val activities for the upcoming FLEX mission, providing useful insights for the selection of the validation site network and particularly for the definition of the best sampling scheme for each site.

A fusion based land cover classification model using remote sensed images

Classification of land cover from remote sensed image is quite challenging task. Since the satellite images preserve spatial and spectral information, thus it is essential to identify the land cover classes and classify them to generate the thematic map. The remote sensed images and thus produced thematic maps are useful for extracting the esteemed information in diagnosing, supervising, and management of earth’s surface. In this paper, a multiclass land cover classification model is proposed that comprise of pre-processing method, a multiclass classifier and performance evaluation strategy. The land cover-based satellite images are applied to this model to generate a land cover map labelled with seven land cover classes. The morphological opening, closing, and a fusion technique are involved in pre-processing stage to extract the spatial information as well as reduce the incurred noise from the input image. Then a supervised classification methodology is introduced to classify the image into 7 number of land cover classes based on the spectral values of each pixel of the image. The overall achievement of the proposed model is compared with some existing multiclass supervised and unsupervised classification techniques such as Naïve Bayes classifier (NBC), Decision tree (DT), K-nearest neighbour (KNN), Convolution Neural Network (CNN).

Probabilistic unsupervised classification for large-scale analysis of spectral imaging data

Land cover classification of remote sensing data is a fundamental tool to study changes in the environment such as deforestation or wildfires. A current challenge is to quantify land cover changes with real-time, large-scale data from modern hyper- or multispectral sensors. A range of methods are available for this task, several of them being based on the k-means classification method which is efficient when classes of land cover are well separated. Here a new algorithm, called probabilistic k-means, is presented to solve some of the limitations of the standard k-means. It is shown that the new algorithm performs better than the standard k-means when the data are noisy. If the number of land cover classes is unknown, an entropy-based criterion can be used to select the best number of classes. The proposed new algorithm is implemented in a combination of R and C computer codes which is particularly efficient with large data sets: a whole image with more than 3 million pixels and covering more than 10,000 km2 can be analysed in a few minutes. Four applications with hyperspectral and multispectral data are presented. For the data sets with ground truth data, the overall accuracy of the probabilistic k-means was substantially improved compared to the standard k-means. One of these data sets includes more than 120 million pixels, demonstrating the scalability of the proposed approach. These developments open new perspectives for the large scale analysis of remote sensing data. All computer code are available in an open-source package called sentinel.

Radiometric normalization of Landsat thermal imagery for detection of tundra land cover changes: experience from West Siberia

ABSTRACT The potentiality and specificity of using multitemporal stacks of thermal infrared (TIR) satellite scenes for modelling changes in tundra land cover and predicting geohazards in areas under development are studied using the example of two areas in northern West Siberia. The study aims at developing and testing a method for relative radiometric normalization of Land Remote-Sensing Satellite (Landsat) TIR images applied to estimate changes in land surface temperature (LST) in the tundra with regard to the high-latitude climate conditions. The suggested normalization method (RNUC) differs from other similar approaches in the use of unsupervised classification to determine pseudo-invariant features where master and slave images are spectrally identical. The error in LST contrasts is minimized iteratively by choosing the optimal number of land-cover classes and linear regression coefficients by cross-validation. The normalization procedure ensures error reduction by a factor of two to four, which is evidence of its high performance. The results of processing and analysis of 19 pairs of Landsat thermal images at the sites of the Marre-Sale and Urengoy oil and gas-condensate fields using the RNUC method showed that the relative error of LST contrast normalization could be in the range of approximately 3% to 10%. Based on the time series of normalized values of LST and the normalized difference vegetation index (NDVI) of sites with pyrogenic damage in the area of the Urengoy oil and gas-condensate field, it is shown that the process of stabilization of the tundra cover is characterized by a negative correlation of these parameters and lasts 25 to 28 years. A positive correlation between LST and NDVI characterizes a different type of landscape transformation; in particular, it may be an increase in the thickness of secondary vegetation cover or a change in the hydrological conditions.

Adaptive MultiScale Segmentations for Hyperspectral Image Classification

The number of superpixels (i.e., segmentation scale) is crucial for spectral–spatial hyperspectral image (HSI) classification. Existing methods always set the segmentation scale through a manually experimental strategy, which is time-consuming and unsuitable for various complicated practical applications. The information fusion of complementary multiple scales is proven to be more effective than the single scale for HSI classification, but the scale level is still set manually. In this article, we propose a novel adaptive multiscale segmentations (AMSs) method that can automatically provide a set of suitable scales that are adapted to different hyperspectral data. Specifically, based on the assumption that the segmentation scale of HSI is related to the image complexity itself, the texture ratio and the number of land cover classes are used to examine a candidate scale pool. A good scale means that it contains a small spectral difference between pixels within the same superpixel (intrasuperpixel discrimination index) and a large discrepancy between neighboring superpixels (intersuperpixel discrimination index). Thus, an intra–interscale discrimination index is defined and applied to depict the characteristics of the scale. Then, the scale with the best intra–inter discrimination index, which usually has satisfactory performance, is treated as the initially selected scale. The remaining suitable scales are iteratively compared with the selected ones and then added to the target scale pool, until the newly added scale can no longer provide significantly complementary information. Extensive experimental results on three HSI data sets have demonstrated the effectiveness of the proposed AMS when compared with state-of-the-art methods.

Investigation on Land Cover Mapping Capability of Maximum Likelihood Classifier: A Case Study on North Canara, India

The success of a large number of real-world applications depends on the efficiency with which land use land cover classes are extracted from Remotely Sensed (RS) imagery. This study aims at analysing the performance of Maximum Likelihood Classifier (MLC) over multispectral remote sensor data characterized by the overlapping of spectral classes and presence of mixed pixels, investigating the impact of prior class probabilities on the overall classification performance, the impact of the number of land cover classes on classification accuracy, and identifying the decision parameters for assigning a pixel to a land cover class. A quantitative analysis of decision parameters of MLC for pixel assignment is presented for five randomly selected pixels in the study area. Landsat 8 multispectral data of North Canara District was collected from USGS website and is considered for the research. Seven land use land cover classes were identified over the study area. Results obtained indicate that use of prior probability estimates improves the classification accuracy. The study also presents the variation in classification accuracy for four different cases of total number of land cover classes. The conducted research is useful in understanding the performance of MLC under multiple scenarios.

Segmentation parameter selection for object-based land-cover mapping from ultra high resolution spectral and elevation data

ABSTRACTSegmentation is the primary task for image analysis in many practical applications, such as object-based image analysis. Segmentation algorithms need to have properly estimated parameters to provide efficient performance and reliable results. Due to the fact that some features have different shapes and spectral characteristics, it is hard to find the proper parameters for the whole image. In this article, we propose a new method for resolving this issue through the building of a hierarchy of segmentations, based on the number of land-cover classes in the image, namely segmentation scale space (SSS). Both spectral and elevation data are employed in order to enhance the SSS and to obtain a single segmentation for the image. The performance of the proposed algorithm is evaluated using two data sets, which consist of ultra-high resolution aerial images and elevation data with ground sampling distance of 5 and 9 cm, respectively. The experiments demonstrate the efficiency of enhanced segmentation with respect to over and under segmentation cases. Finally, the comparative analysis shows that the accuracy of the proposed method is superior to the classical methods.

Rank-Based Methods for Selection of Landscape Metrics for Land Cover Pattern Change Detection

Often landscape metrics are not thoroughly evaluated with respect to remote sensing data characteristics, such as their behavior in relation to variation in spatial and temporal resolution, number of land cover classes or dominant land cover categories. In such circumstances, it may be difficult to ascertain whether a change in a metric is due to landscape pattern change or due to the inherent variability in multi-temporal data. This study builds on this important consideration and proposes a rank-based metric selection process through computation of four difference-based indices (β, γ, ξ and θ) using a Max–Min/Max normalization approach. Land cover classification was carried out for two contrasting provinces, the Liverpool Range (LR) and Liverpool Plains (LP), of the Brigalow Belt South Bioregion (BBSB) of NSW, Australia. Landsat images, Multi Spectral Scanner (MSS) of 1972–1973 and TM of 1987–1988, 1993–1994, 1999–2000 and 2009–2010 were classified using object-based image analysis methods. A total of 30 landscape metrics were computed and their sensitivities towards variation in spatial and temporal resolutions, number of land cover classes and dominant land cover categories were evaluated by computing a score based on Max–Min/Max normalization. The landscape metrics selected on the basis of the proposed methods (Diversity index (MSIDI), Area weighted mean patch fractal dimension (SHAPE_AM), Mean core area (CORE_MN), Total edge (TE), No. of patches (NP), Contagion index (CONTAG), Mean nearest neighbor index (ENN_MN) and Mean patch fractal dimension (FRAC_MN)) were successful and effective in identifying changes over five different change periods. Major changes in land cover pattern after 1993 were observed, and though the trends were similar in both cases, the LP region became more fragmented than the LR. The proposed method was straightforward to apply, and can deal with multiple metrics when selection of an appropriate set can become difficult.

A New Self-Training-Based Unsupervised Satellite Image Classification Technique Using Cluster Ensemble Strategy

This letter addresses the problem of unsupervised land-cover classification of remotely sensed multispectral satellite images from the perspective of cluster ensembles and self-learning. The cluster ensembles combine multiple data partitions generated by different clustering algorithms into a single robust solution. A cluster-ensemble-based method is proposed here for the initialization of the unsupervised iterative expectation-maximization (EM) algorithm which eventually produces a better approximation of the cluster parameters considering a certain statistical model is followed to fit the data. The method assumes that the number of land-cover classes is known. A novel method for generating a consistent labeling scheme for each clustering of the consensus is introduced for cluster ensembles. A maximum likelihood classifier is henceforth trained on the updated parameter set obtained from the EM step and is further used to classify the rest of the image pixels. The self-learning classifier, although trained without any external supervision, reduces the effect of data overlapping from different clusters which otherwise a single clustering algorithm fails to identify. The clustering performance of the proposed method on a medium resolution and a very high spatial resolution image have effectively outperformed the results of the individual clustering of the ensemble.

A Novel Graph Based Fuzzy Clustering Technique For Unsupervised Classification Of Remote Sensing Images

Abstract. This paper addresses the problem of unsupervised land-cover classification of multi-spectral remotely sensed images in the context of self-learning by exploring different graph based clustering techniques hierarchically. The only assumption used here is that the number of land-cover classes is known a priori. Object based image analysis paradigm which processes a given image at different levels, has emerged as a popular alternative to the pixel based approaches for remote sensing image segmentation considering the high spatial resolution of the images. A graph based fuzzy clustering technique is proposed here to obtain a better merging of an initially oversegmented image in the spectral domain compared to conventional clustering techniques. Instead of using Euclidean distance measure, the cumulative graph edge weight is used to find the distance between a pair of points to better cope with the topology of the feature space. In order to handle uncertainty in assigning class labels to pixels, which is not always a crisp allocation for remote sensing data, fuzzy set theoretic technique is incorporated to the graph based clustering. Minimum Spanning Tree (MST) based clustering technique is used to over-segment the image at the first level. Furthermore, considering that the spectral signature of different land-cover classes may overlap significantly, a self-learning based Maximum Likelihood (ML) classifier coupled with the Expectation Maximization (EM) based iterative unsupervised parameter retraining scheme is used to generate the final land-cover classification map. Results on two medium resolution images establish the superior performance of the proposed technique in comparison to the traditional fuzzy c-means clustering technique.

Land Cover Heterogeneity Effects on Sub-Pixel and Per-Pixel Classifications

Per-pixel and sub-pixel are two common classification methods in land cover studies. The characteristics of a landscape, particularly the land cover itself, can affect the accuracies of both methods. The objectives of this study were to: (1) compare the performance of sub-pixel vs. per-pixel classification methods for a broad heterogeneous region; and (2) analyze the impact of land cover heterogeneity (i.e., the number of land cover classes per pixel) on both classification methods. The results demonstrated that the accuracy of both per-pixel and sub-pixel classification methods were generally reduced by increasing land cover heterogeneity. Urban areas, for example, were found to have the lowest accuracy for the per-pixel method, because they had the highest heterogeneity. Conversely, rural areas dominated by cropland and grassland had low heterogeneity and high accuracy. When a sub-pixel method was used, the producer’s accuracy for artificial surfaces was increased by more than 20%. For all other land cover classes, sub-pixel and per-pixel classification methods performed similarly. Thus, the sub-pixel classification was only advantageous for heterogeneous urban landscapes. Both creators and users of land cover datasets should be aware of the inherent landscape heterogeneity and its potential effect on map accuracy.

Factors biasing the correlation structure of patch level landscape metrics

Landscape metrics are in varying correlations with each other. Several authors have revealed their correlation structure and determined sets of metrics which can be used in landscape analysis. We assumed that correlation structure is not stable and is biased by several factors, thus, selection based on the correlation can vary by case studies. In this study we dealt with 13 patch level landscape metrics using three landscape types, consisting of 9 subregions with 7 and 14 land cover classes, applying 5 different cell sizes. In each step of the analysis other factors that can bias the results were controlled, or the analyses were carried out separately. In accordance with our aims, we uncovered the factor structure of the metrics in different situations, with the parameters which might possibly bias the results. Results showed that cell size, landscape types and number of land cover classes had a greater or lesser effect on cross-correlations. However, the greatest effect was experienced when variables were changed slightly (i.e. two metrics were replaced with two new ones). A comparison of factor structure was conducted with the coefficient of congruence, rank order based on factor loadings, and biplots. According to our findings, congruence values are not reliable in all cases, while ranks and biplots were not sensitive to the changes in circumstances. Possible outcomes were tested with calculations of 3 test areas (a large landscape from NE-Hungary and two countries). Results can be relevant for landscape ecologists dealing with many variables and multivariate techniques.

Indicators of landscape diversity derived from remote sensing based land cover maps – spatial and thematic aspects

Landscape is heterogeneous part of the Earth surface, forming mosaic of various habitats organized at different scales and levels (Johnson et al. 1992). The landscape pattern has important impact on ecological processes; hence its analysis through quantitative measures is essential for environmental studies. There are many indicators characterizing spatial structure of landscape at different level of detail; they enable analysis of landscape fragmentation at patch level, through studies at habitat level up to complex analyses at landscape level. Seven indicators, which are related to various levels of detail, were selected at the presented work. The following indicators have been studied: Patch Density, Edge Density, Patch Richness, Simpson Diversity Index, Natural Patch Richness, Percentage of Natural Landscape, Mean Natural Patch Area (McGarigal & Marks 1995). First two indicators were used for analysis of landscape fragmentation at patch level, next two at land cover level, while the last three were applied for studies of natural and semi-natural classes at both levels. The studies were performed at six test areas located in different regions of Europe (France, Germany, Poland, Latvia, Spain and Italy), using two different land cover maps. First map was based on Very High Resolution (VHR) Kompsat satellite images (4 m spatial resolution); it included 8 land cover categories with 0.25 ha Minimum Mapping Unit (MMU). CORINE Land Cover (CLC) map 2006 (25 ha MMU) was the second map used for analyses. Number of land cover classes in case of CLC map varied from 9 for Poland till 14 for France. All above mentioned indicators were calculated for grids with 100, 200, 500 and 1000 meter cell size, corresponding to 1, 4, 25 and 100 ha, respectively. The obtained results reveal high usefulness of land cover maps based on VHR satellite images for analysis of landscape fragmentation, even for grids with 100 m cell size. It was found that at patch level these materials are superior to CLC classifications, irrespective of cell area. In case of land cover level VHR data are better while using 100 and 200 m grid cells, whereas for larger cell sizes – 500 and 1000 m – results are not so evident, depending on degree of landscape fragmentation and spatial structure characteristic for individual land cover classes.

Improvement and evaluation of simulated global biogenic soil NO emissions in an AC-GCM

Abstract. Biogenic NO emissions from soils (SNOx) play important direct and indirect roles in tropospheric chemistry. The most widely applied algorithm to calculate SNOx in global models was published 15 years ago by Yienger and Levy (1995), and was based on very few measurements. Since then, numerous new measurements have been published, which we used to build up a compilation of world wide field measurements covering the period from 1978 to 2010. Recently, several satellite-based top-down approaches, which recalculated the different sources of NOx (fossil fuel, biomass burning, soil and lightning), have shown an underestimation of SNOx by the algorithm of Yienger and Levy (1995). Nevertheless, to our knowledge no general improvements of this algorithm, besides suggested scalings of the total source magnitude, have yet been published. Here we present major improvements to the algorithm, which should help to optimize the representation of SNOx in atmospheric-chemistry global climate models, without modifying the underlying principals or mathematical equations. The changes include: (1) using a new landcover map, with twice the number of landcover classes, and using annually varying fertilizer application rates; (2) adopting a fraction of 1.0 % for the applied fertilizer lost as NO, based on our compilation of measurements; (3) using the volumetric soil moisture to distinguish between the wet and dry states; and (4) adjusting the emission factors to reproduce the measured emissions in our compilation (based on either their geometric or arithmetic mean values). These steps lead to increased global annual SNOx, and our total above canopy SNOx source of 8.6 Tg yr−1 (using the geometric mean) ends up being close to one of the satellite-based top-down approaches (8.9 Tg yr−1). The above canopy SNOx source using the arithmetic mean is 27.6 Tg yr−1, which is higher than all previous estimates, but compares better with a regional top-down study in eastern China. This suggests that both top-down and bottom-up approaches will be needed in future attempts to provide a better calculation of SNOx.

Classification of Landsat Thematic Mapper imagery for land cover using neural networks

Landsat Thematic Mapper (TM) imagery can be used to classify different land cover types based on reflectance and emittance characteristics in seven wavelength bands. Various methods, including NDVI and other simple mathematical transformations, can be used to show strong variations in band intensity ratios from different surfaces. However, the number of land cover classes used is commonly low, preventing a detailed mapping of the region of interest. A neural network trained with the backpropagation method should be able to improve on these simple mathematical calculations by developing complex functions which allow recognition of different land cover or land use types. Landsat imagery of Aberdeen and the surrounding area was used to develop a land cover map highlighting areas of residential, commercial and industrial land use, along with various natural and semi‐natural land cover classes. Confusion between specific classes is highlighted by the use of a Kohonen self‐organizing map to categorize the Landsat multispectral imagery, resulting in a description of the land cover categories that can actually be distinguished from one another using Landsat TM imagery.

Mapping tropical dry forest habitats integrating Landsat NDVI, Ikonos imagery, and topographic information in the Caribbean Island of Mona

Assessing the status of tropical dry forest habitats using remote sensing technologies is one of the research priorities for Neotropical forests. We developed a simple method for mapping vegetation and habitats in a tropical dry forest reserve, Mona Island, Puerto Rico, by integrating the Normalized Difference Vegetation Index (NDVI) from Landsat, topographic information, and high-resolution Ikonos imagery. The method was practical for identifying vegetation types in areas with a great variety of plant communities and complex relief, and can be adapted to other dry forest habitats of the Caribbean Islands. NDVI was useful for identifying the distribution of forests, woodlands, and shrubland, providing a natural representation of the vegetation patterns on the island. The use of Ikonos imagery allowed increasing the number of land cover classes. As a result, sixteen land-cover types were mapped over the 5500 ha area, with a kappa coefficient of accuracy equal to 79%. This map is a central piece for modeling vertebrate species distribution and biodiversity patterns by the Puerto Rico Gap Analysis Project, and it is of great value for assisting research and management actions in the island.

Sensitivity of landscape metrics to classification scheme

Landscape metrics are a standard tool in the study and monitoring of landscape pattern and change, but their statistical properties and behaviour across a range of classification schemes and landscapes, as well as their sensitivity to changing landscape patterns, are still not fully understood. We therefore investigated the sensitivity of 24 metrics to a number of land cover classes for three Arizona landscapes with different spatial patterns. To do so, we applied unsupervised classification of remotely sensed data with two different nominal spatial resolutions to generate maps containing 2–35 classes. We calculated metric values for these thematic maps and classified the metrics into six groups using principal components analysis. For each group, the nature and sensitivity of responses to differences in resolution, landscape pattern, and classification detail were assessed. Our results indicated that many metrics behaved predictably with increasing classification detail, increasing or decreasing at rates that were often relatively similar and independent to sensor and landscape pattern. At lower class numbers, metrics were most sensitive to increasing classification detail, and the effects of classification scheme were most erratic and sensitive to resolution and underlying landscape pattern. Overall, this study provides a descriptive overview of the sensitivity of common metrics to changes in classification scheme, as well as a first attempt to draw some generalizations about the importance of classification scheme in conjunction with resolution effects.

Incorporating domain knowledge and spatial relationships into land cover classifications: a rule‐based approach

For some tropical regions, remote sensing of land cover yields unacceptable results, particularly as the number of land cover classes increases. This research explores the utility of incorporating domain knowledge and multiple algorithms into land cover classifications via a rule‐based algorithm for a series of satellite images. The proposed technique integrates the fundamental, knowledge‐based interpretation elements of remote sensing without sacrificing the ease and consistency of automated, algorithm‐based processing. Compared with results from a traditional maximum likelihood algorithm, classification accuracy was improved substantially for each of the six land cover classes and all three years in the image series. Use of domain knowledge proved effective in accurately classifying problematic tropical land covers, such as tropical deciduous forest and seasonal wetlands. Results also suggest that ancillary data may be most useful in the classification of historic images, where the greatest improvement was observed relative to results from maximum likelihood. The cost of incorporating contextual knowledge and extensive spatial data sets may be justified, since results from the proposed technique suggest a considerable improvement in accuracy may be achieved.

PLANT SPECIES INVASIONS ALONG THE LATITUDINAL GRADIENT IN THE UNITED STATES

It has been long established that the richness of vascular plant species and many animal taxa decreases with increasing latitude, a pattern that very generally follows declines in actual and potential evapotranspiration, solar radiation, temperature, and thus, total productivity. Using county-level data on vascular plants from the United States (3000 counties in the conterminous 48 states), we used the Akaike Information Criterion (AIC) to evaluate competing models predicting native and nonnative plant species density (number of species per square kilometer in a county) from various combinations of biotic variables (e.g., native bird species density, vegetation carbon, normalized difference vegetation index), environmental/topographic variables (elevation, variation in elevation, the number of land cover classes in the county, radiation, mean precipitation, actual evapotranspiration, and potential evapotranspiration), and human variables (human population density, cropland, and percentage of disturbed lands in a county). We found no evidence of a latitudinal gradient for the density of native plant species and a significant, slightly positive latitudinal gradient for the density of nonnative plant species. We found stronger evidence of a significant, positive productivity gradient (vegetation carbon) for the density of native plant species and nonnative plant species. We found much stronger significant relationships when biotic, environmental/topographic, and human variables were used to predict native plant species density and nonnative plant species density. Biotic variables generally had far greater influence in multivariate models than human or environmental/topographic variables. Later, we found that the best, single, positive predictor of the density of nonnative plant species in a county was the density of native plant species in a county. While further study is needed, it may be that, while humans facilitate the initial establishment invasions of nonnative plant species, the spread and subsequent distributions of nonnative species are controlled largely by biotic and environmental factors.

IRS-1C LISS III land cover maps at different spatial resolutions used in real-time accidental air pollution deposition modelling

Remote-sensing scenes can provide information highly useful for real-time modelling of the atmospheric deposition in the case of an accidental nuclear release. Accurate prediction of spatial patterns in the deposition is vital to establish countermeasures to avoid unnecessary exposure of the public. This is demonstrated by a hypothesised case for a nuclear power plant in Lithuania. From classification analysis a number of land cover classes are mapped in an IRS-1C LISS III scene. The roughness of each land cover class is assigned and used as input to a real-time dispersion model. Calculation results with input of a high-resolution roughness map provide the most realistic deposition pattern. A scale analysis of degrading the land cover map into lower resolutions is performed. The atmospheric deposition model results based on lower resolution roughness maps are compared and evaluated by the so-called Dose Isopleth Correlation Area (DICA) method. The recommendation is that a 1-km resolution is adequate for real-time deposition modelling. Furthermore, the land cover mapping should be performed with care. Temporal variations in roughness at a seasonal time scale should be considered.