Next Article in Journal
Spatiotemporal Dynamic Evolution of PM2.5 Exposure from Land Use Changes: A Case Study of Gansu Province, China
Previous Article in Journal
Spatial Representation of Soil Erosion and Vegetation Affected by a Forest Fire in the Sierra de Francia (Spain) Using RUSLE and NDVI
Previous Article in Special Issue
Land Use and Land Cover Changes: A Case Study in Nigeria
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Land
Volume 14
Issue 4
10.3390/land14040794
land-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

The Potential of the Copernicus Product “Imperviousness Classified Change” to Assess Soil Sealing in Agricultural Areas in Poland and Norway

by
Wendy Fjellstad
1,*,
Agata Hościło
2,
Svein Olav Krøgli
1,
Jonathan Rizzi
1 and
Milena Chmielewska
3
1
Department of Landscape Monitoring, Norwegian Institute of Bioeconomy Research (NIBIO), P.O. Box 115, 1431 Ås, Norway
2
National Centre for Emissions Management, Institute of Environmental Protection—National Research Institute (IOŚ-PIB), 32 Slowicza St., 02-170 Warsaw, Poland
3
Institute of Geodesy and Cartography (IGiK), 27 Modzelewskiego Street, 02-679 Warsaw, Poland
*
Author to whom correspondence should be addressed.
Land 2025, 14(4), 794; https://doi.org/10.3390/land14040794
Submission received: 22 February 2025 / Revised: 17 March 2025 / Accepted: 3 April 2025 / Published: 7 April 2025
(This article belongs to the Special Issue Advances in Land Use and Land Cover Mapping (Second Edition))
Browse Figure
Versions Notes

Abstract

:
Many countries have goals to reduce soil sealing of agricultural land to preserve food production capacity. To monitor progress, reliable data are needed to quantify soil sealing and changes over time. We examined the potential of the Imperviousness Classified Change (IMCC) 2015–2018 product provided by the Copernicus Land Monitoring Service (CLMS) to assess soil sealing in agricultural areas in Poland and Norway. We found very high overall accuracy due to the dominance of the area with no change. When we focused on areas classified as change, we found low user accuracy, with over-estimation of soil sealing. The producer accuracy was generally much higher, meaning that real cases of soil sealing were captured. This is better than under-estimation of soil sealing because it highlights areas where sealing may have occurred, allowing the user to carry out further control of this much smaller area, without having to assess the great expanse of unchanged area. We concluded that the datasets provide useful information for Europe. They are standardized and comparable across countries, which can enable comparison of the effects of policies intended to prevent soil sealing. Some distinctions between classes are not reliable, but the general information about increase or decrease is useful.

1. Introduction

Soil sealing is the transformation of natural or agricultural land into impervious surfaces that prevent water infiltration into the soil [1]. Impervious surfaces include asphalt, concrete, and synthetic surfaces and are associated with land uses such as roads, pavements, car parks, buildings, and certain leisure areas. Soil sealing affects many ecosystem services, including water absorption, filtering, storage, temperature regulation, provision of habitats, and our capacity to grow food [2].
Soil sealing is an important part of land take. Land take is defined as “the transformation of agricultural, natural and semi-natural spaces into urban and other artificial uses” [3]. Whilst soil sealing refers specifically to the area that becomes impervious, land take is a broader term that captures all the land taken over by the new land use. For example, when agricultural land is converted to a residential area, some of the land will be covered by impervious buildings and roads (soil sealing), whilst other areas may become gardens, lawns, road verges, etc. These unsealed surfaces between the buildings and roads have also been taken out of agricultural production. Hence, the land take is greater than the area of soil sealing.
In this paper, we focus on land take and soil sealing of agricultural land. The irreversible loss of productive land is critical in relation to Sustainable Development Goal 2, to end hunger and achieve food security [4]. The global population is increasing and changing diets are leading to a higher demand for agricultural land per capita, whilst formerly productive areas are becoming degraded due to climate change and urbanization [5]. Many countries have goals and strategies to reduce the loss of agricultural land and preserve food production capacity [6,7,8]. To control whether these goals are fulfilled, it is important to be able to measure and monitor the loss of agricultural land.
For decades, efforts have been made to develop indicators and maps to report and assess the extent of soil sealing [9]. Sample-based surveys have been developed that can provide indicators of soil sealing, e.g., [10,11,12] and other methods have been devised using national cadastral data, including buildings and transport infrastructure [13,14]. To provide harmonized data across multiple countries, Gardi et al. [15] attempted to assess soil sealing using CORINE Land Cover maps. Their calculations for nineteen EU Member States for the period 1990 to 2006, suggested that the potential agricultural production capacity was reduced by more than 6.2 million tonnes of wheat due to the loss of agricultural land [15]. Although the method of calculation was imprecise, this study highlighted the magnitude of the issue. However, to assess and measure the loss of agricultural land, reliable geospatial datasets are essential.
The increase in easily and freely accessible remotely sensed data, and the rapid development of remote sensing techniques, have led to the development of a range of methods to map impervious surfaces and measure the processes of soil sealing and land take. The main methods are based on the analysis of spectral signatures and the calculation of vegetation indices [16,17].
A review paper by Peroni et al. [18] highlighted that studies of imperviousness far out-weighed studies of soil sealing and land take, but that the latter had gradually increased since 2014. This may be related to trends in terminology, and a shift in focus from the physical phenomenon of imperviousness to the consequences of changes [18]. Many studies of land take are related to the expansion of urban and sub-urban land [19] whilst the land taken outside urban areas is neglected [14].
The Copernicus Land Monitoring Service (CLMS), which is a part of the European Copernicus Program, has developed products to quantify imperviousness, so-called High-Resolution Layer (HRL) Imperviousness. The Imperviousness products are raster datasets derived from Earth Observation. The first HRL datasets, for the reference years 2006, 2009, 2012 and 2015 had a spatial resolution of 20 × 20 m. From 2018, the resolution improved to 10 × 10 m due to the high-resolution Sentinel missions. HRL Imperviousness Density (HRL IMD) has already been used in a comparative study of 60 European cities to assess the effectiveness of greenbelts [20]. Orsi [21] used HRL IMD in 2006 and 2012 to calculate changes in imperviousness and to investigate how population density, centrality, and the contiguity of settlements influenced the expansion of soil sealed areas in Italy. Although the HRL Imperviousness datasets have been suggested as a common data source across Europe [22], there has been little verification of these products, as it is a time-consuming and labor-intensive process.
Strand [23] verified the HRL Imperviousness Density (IMD) 2018 data for Norway and concluded that HRL IMD under-estimated the total sealed area in Norway by approximately 33%. This study suggested that weaknesses associated with the earlier datasets [24,25,26] had not been resolved in the higher resolution 2018 dataset. Of particular concern was the apparent under-estimation of sealed surfaces in rural areas. Unfortunately, due to the very small proportion of agricultural land in Norway, the study by Strand [23] was unable to provide accurate statistics for soil sealing of agricultural land.
To make it easier for people to use HRL IMD data, CLMS has also released datasets where the change in imperviousness between two time points has been classified as so-called Imperviousness Classified Change (IMCC). This assumes that the HRL IMD data are accurate and reliable. Our aim was to examine the potential and reliability of the Copernicus HRL Imperviousness Classified Change (IMCC) 2015–2018 product to assess soil sealing in agricultural areas in Poland and Norway. By comparing verification results from two countries, using the same methods, we aimed to determine the accuracy of IMCC 2015–2018 and examine whether the accuracy was similar across agricultural landscapes with different patterns of land use and proportion of agricultural areas.

2. Data and Methods

2.1. The IMCC Dataset

For both Poland and Norway, we used the Imperviousness Classified Change product showing changes from 2015 to 2018. The data were downloaded spring 2022 from the CLMS website (https://land.copernicus.eu/en, accessed on 1 April 2025). All associated documentation was accessible at the same site. The IMCC dataset is the difference between IMD 2015, available at 20 m spatial resolution and IMD 2018, available at 10 m spatial resolution [27,28]. The IMCC data comprises change classes (Table 1) for 20 × 20 m pixels.

2.2. Maps of Agricultural Land

Existing national land cover and land use digital maps were used to identify agricultural land. In both countries, these maps are updated intermittently and although yearly versions exist, they contain areas that may not have been updated for several years.
For Poland, we used the 2015 version of the Topographic Objects Database BDOT10K to prepare a mask for the agricultural land. The BDOT10K provides data over the entire country with a level of detail corresponding to the topographic maps at 1:10,000. It is derived and updated partially using manual interpretation of aerial orthophotos. Two land cover classes were selected: (i) permanent crops (code: PTUT), and (ii) grassy vegetation and arable crops (code: PTTR). In addition, the class ‘land use complexes’ (code: KU) was used to eliminate non-agricultural areas such as sports and recreational complexes.
For Norway, we selected agricultural land from the National Land Resource Map AR5 (scale 1:5000). We used the 2010 dataset, rather than 2015, to account for a time lag in the national data, where agricultural land can be re-classified several years before it is built on. We selected classes 21: fully cultivated, 22: surface cultivated, and 23: improved pasture.

2.3. Verification of Using Stratified Random Sampling

For each country, we selected all pixels of IMCC that intersected with agricultural areas. We treated the six IMCC classes (Table 1) as strata and randomly selected 200 pixels for each class. For class 12 (decreased IMD), only five pixels were available for Poland and nine pixels for Norway. Each pixel was checked manually using imagery available through Google Earth Pro for the years 2015 and 2018. If imagery was lacking or unclear, we used aerial orthophotos from national archives (made available by Norge Digitalt at Norgeibilder.no in Norway, and by the Head Office of Geodesy and Cartography in Poland). All five authors were involved in the verification process, and we had calibration meetings to ensure that our interpretations were as similar as possible. We used a conservative evaluation. If we could not see that IMCC was obviously wrong, then we recorded it as being correct. If the classification was considered incorrect, we recorded the correct class. Figure 1 presents examples from the verification process.

2.4. Analysis of IMCC Verification

The IMCC verification data allowed us to set up error matrices, showing sample counts of pixels in each IMCC mapped class against the real class, as observed from the remote sensing imagery. From the error matrices, we calculated the overall accuracy as well as the producer’s and user’s accuracies, and associated errors of omission and commission for each change class. The producer accuracy expresses how often real features on the ground are correctly shown on the map. The opposite of this is the omission error when real features on the ground are not in the correct class on the map. The user’s accuracy expresses how often the class on the map is actually present on the ground. The opposite of this is commission error, when the map shows the presence of a class that is not there on the ground.

3. Results

Having selected all IMCC pixels containing agricultural land in Poland and Norway, we found that by far the most common class in the dataset was class 0, i.e., unchanged pixels with no imperviousness (IMD = 0) either in 2015 or 2018 (Table 2). This class was entirely dominant in both countries. The second most common class was class 10, i.e., pixels containing some imperviousness, but showing no change between 2015 and 2018. The least common class was class 12, with decreased imperviousness, with only five pixels available to check in the Polish dataset and nine pixels in the Norwegian dataset. If we considered just the pixels expressing a change in imperviousness (i.e., excluding classes 0 and 10), we found in both countries that 99.6% of pixels classed as change were gain (class 1 or 11) and 0.4% loss (class 2 or 12).
From our verification of 200 pixels per IMCC class, we calculated user and producer accuracies for Poland (Table 3) and for Norway (Table 4). If IMCC states that an area is class 0, this is generally correct. In Poland, 99.0% of pixels selected from class 0 were correct (Table 3) and in Norway, 83.5% were correct (Table 4). The greater commission error in Norway was mainly due to pixels with no change but with some degree of imperviousness, being classified as class 0 when they should have been classified as class 10.
The spatial dominance of class 0 (Table 2), combined with the very high user accuracy for this class, resulted in a very high overall accuracy when weighted by the area of the IMCC classes: 97.7% in Poland and 83.5% in Norway.
Having established that the vast majority of both study areas remained unchanged between 2015 and 2018, the interesting question was whether IMCC accurately identified the small proportion of areas where change did occur (classes 1, 11, 2, and 12). In Poland, user accuracy was highest for class 11 (58.5%), similar for class 1 (54%), but much lower for class 2 (24%) and class 12 (20%) (Table 3). In Norway, user accuracy was lowest for class 1 (19.5%), and over 60% for classes 11, 2, and 12. In both countries, our verification showed that a high proportion of change pixels should have been classified as no change. In Poland, the correct class was most often assessed as class 0 (unchanged and IMD = 0), whereas in Norway the correct class was most often assessed as class 10 (unchanged but with some IMD).
All these extra cases of unchanged pixels meant that the producer accuracy for class 0 was much lower than the user’s accuracy: 68.7% in Norway and just 37.8% in Poland. In Norway, the producer accuracy for class 10 was just 39.1%. However, in Poland, the producer accuracy for class 10 (71.4%) was higher than the user’s accuracy (52.5%).
Producer accuracies for changed areas were generally much higher than user accuracies, 100% in both countries for class 2, loss of cover, and over 90% in both countries for class 11, increased IMD. In addition, Poland had high producer accuracy for class 12 (100%), although it should be noted that this was just one sample pixel) and Norway for class 1 (86.7%). This means that, where real change occurred on the ground, it was generally captured by IMCC. However, the producer accuracy for class 1 in Poland (new cover, increasing from IMD = 0 in 2015) was lower (66.3%), with a quarter of cases misclassified to class 11 (Table 3). In Norway, producer accuracy was low for class 12 (35%), where 60% of real cases seen in the remote sensing imagery were inaccurately assigned to class 2. Hence, in each case of low producer accuracy for changed areas, the error was to a thematically closely related class and the error was a failure to identify zero IMD in 2015 in Poland, or in 2018 in Norway. If class 1 and 11 merged, and class 2 and 12 merged, i.e., simply recording any increase or decrease, respectively, then the producer accuracies would be over 94% in both cases in both countries.
The number of real instances of change observed in the remotely sensed images was very similar in Poland (334 pixels) and Norway (323 pixels), but the distribution between the different classes was quite different. In Poland, almost half the observed change pixels were class 1, 37% class 11, 14% class 2, and just 0.3% class 12. In Norway, the proportion of class 1, new cover, was much lower (14%) and class 11 was the biggest of the change classes (41%), closely matched by class 2 (39%), whilst class 12 comprised 6% of real change observed.
A commonly observed reason for over-estimation of impervious areas was related to the misclassification of bare soil or sand as impervious. Thus, locations were wrongly assigned to class 2 or 12 when bare soil in 2015 was covered by vegetation in 2018 (Figure 1b), to class 1 or 11 when vegetated areas in 2015 were bare soil in 2018, or to class 11 when bare soil was interpreted as impervious in both years (Figure 1c).

4. Discussion

4.1. The Accuracy of IMCC 2015–2018

The high degree of stability in landscapes can be a challenge when trying to assess the accuracy of detecting change, since any model suggesting no change will generally be correct [29]. Therefore, the high overall accuracy of IMCC 2015–2018 in both Poland and Norway, due to the dominance of class 0, is not particularly relevant to our quest to detect soil sealing. Rather, we want to know how accurate IMCC is at detecting the relatively small areas where changes in imperviousness occur. For these areas the user accuracy was much lower, indicating a high commission error, i.e., an over-estimation both of gain and loss of impervious surfaces when, in fact, no change had occurred. The most extreme case was for class 12 in Poland, with a user accuracy of just 20%. With only five pixels of class 12, this result must be considered anecdotal rather than statistically valid. However, the result for class 2 was almost as poor at 24%. The producer accuracy, on the other hand, was generally high for the change classes, i.e., a low omission error, meaning that real cases of soil sealing and loss of imperviousness were captured correctly.
The producer accuracy would be even higher if we ignored misclassifications between the closely related classes 1 and 11 (new cover and increased cover) and classes 2 and 12 (loss of cover and decreased cover). If there had been just one class for gain and one class for loss, this would bring the producer accuracy to 94% and higher for each of the two classes in both countries. In other words, the vast majority of real changes are detected by IMCC 2015–2018.

4.2. Differences in Observed Cases of Soil Sealing in Norway and Poland

The absolute number of real instances of class 1, new imperviousness, in the verification data, was very much higher in Poland (163 pixels) than in Norway (45 pixels). This difference in the degree of soil sealing may be linked to policies. Norway has an exceptionally small amount of agricultural land, just 3% of the land area or around one million hectares. In addition, the agricultural land is often located close to settlements. Policies have therefore been in place for many decades aiming to preserve this area for food production [30,31,32]. The Norwegian Constitution from 1814 (§110b) protects the productivity of the land and the Agricultural Act states that “cultivable land must not be disposed of in such a way as to render it unfit for agricultural production in the future” (Land Act 1995, no.23 §9). The Agricultural Soil Protection Strategy states that, by 2030, the use of agricultural soil for other purposes must not exceed 200 hectares per year [33].
In Poland, on the other hand, almost half the country is agricultural land, around 14.7 million hectares [34]. Hence, with a population of around 38 million and rapid development of infrastructure, there may have been less policy focus on this apparently plentiful resource. Related to this, policies to encourage the densification of cities and prevent urban sprawl may have been introduced earlier in Norway than in other countries and adhered to more strongly [35]. Spatial development trends in Eastern Europe, on the other hand, have previously been found to be rather poorly in line with EU strategies aiming to increase densification and reduce land take [36]. This may be due to a combination of historical factors, land use legacies, planning traditions and attitudes, and socio-economic factors [37,38].

4.3. Differences in Observed Cases of Imperviousness Loss in Norway and Poland

The number of observed cases of imperviousness loss was also very different between Poland and Norway. As expected, imperviousness loss was a much smaller class than imperviousness gain in both countries; however, amongst all the pixels used for the verification, there was a much greater incidence of loss in Norway (145 instances of observed class 2 or 12, compared with 49 in Poland). In both countries, the producer accuracy for class 2, loss of cover, was 100%, i.e., all observed cases of complete loss of impervious cover were correctly classified as IMCC class 2. With such a low general occurrence, it seems likely that the greater incidence of loss in Norway represents a real difference between the countries. This could be linked to stricter or faster enforcement of policies in Norway, e.g., ensuring that agricultural land is restored quickly after the development of projects such as road construction. Or there could be some forms of temporary imperviousness that are more common in Norway due to climatic differences and more marginal conditions for agriculture. For example, crops may be covered by protective plastic longer into the summer in Norway than in Poland, being captured as impervious on occasion but later recorded without the plastic. In 2018, the annual temperature for Europe was one of the three highest on record, and average temperatures for April and May were particularly high. Therefore, it is probable that crops may have developed faster in the spring than usual in Norway and the plastic removed earlier.

4.4. Differences in Unchanged Areas in Norway and Poland

Having discussed the differences between Poland and Norway in the change classes, we can return to the “unchanged” area, since this also differed between countries. In Poland, if we look across all strata we find 524 pixels that were observed to be class 0 (no change, zero imperviousness), compared with just 147 pixels observed to be class 10 (no change, but with a degree of imperviousness). Inaccurately classified pixels in Poland should generally have been classified as class 0.
In Norway, on the other hand, 243 pixels were observed to be class 0, and 443 were class 10. Inaccurately classified pixels in Norway should generally have been classified as class 10. Since the stratification procedure was the same in both countries, the different balance between class 0 and class 10 appears to indicate that some degree of imperviousness is more common in the agricultural landscape of Norway than in Poland. Probably this is related to the much greater total agricultural area in Poland and the presence of large, continuous agricultural fields, especially in the western and northern voivodeships [39]. In Norway, on the other hand, agriculture is small-scale, scattered all over the country, and often close to settlements [40]. Hence, there is a smaller proportion of 20 × 20 m pixels that completely lack any impervious area.

4.5. Potential Use of IMCC 2015–2018 and Future Prospects

IMCC 2015–2018 cannot provide reliable and accurate estimates of soil sealing, either in Poland or Norway. Nevertheless, the fact that IMCC over-estimates soil sealing, whilst having a low omission error, is a much better result than if IMCC had under-estimated soil sealing. It suggests that the dataset could play a role in helping to detect soil sealing, by highlighting areas where soil sealing may have occurred, allowing the user to carry out further control of this much smaller area, without having to assess the great expanse of unchanged area. This could be very time-saving for detecting areas of soil sealing.
Our finding that IMCC 2015–2018 over-estimated soil sealing, contrasts somewhat with the verification of HRL IMD 2018 carried out over Norway by Strand [23]. That study found that the total sealed area in Norway was under-estimated. We do not have a similar verification of HRL IMD 2015. Theoretically, if the under-estimation were even greater in 2015, that could explain the over-estimation of soil sealing in IMCC 2015–2018. However, another plausible explanation is that Strand used a random sample over the whole of Norway, stratified by the degree of imperviousness, whilst our study focused exclusively on the agricultural landscape (pixels containing agricultural land). Strand noted that one source of error was that roads and buildings beneath overhanging trees and in shade were often not classed as impervious. Since 33% of Norway is forest, this would have been a much more common cause of error in that sample than in our sample of agricultural landscapes.
Although IMCC 2015–2018 was rather poor at distinguishing between new imperviousness and increased degree of impervious, there is a strong policy relevance for keeping these classes separate. In many countries, including Poland and Norway, policies are in place to encourage densification of existing built-up areas rather than building on farmland [35,36,41,42]. Therefore, the occurrence of new cover, class 1, is a particularly interesting indicator. Further, it is important to remember that the IMD 2015 product is of 20 × 20 m resolution and is being compared with IMD 2018 10 × 10 m resolution. We are optimistic that the next round of classification of changes IMCC 2018–2021, when both datasets are at 10 × 10 m resolution, will produce better results. We therefore support keeping these separate classes, despite their shortcomings in the 2015–2018 dataset.
One source of error observed both in Poland and Norway was the misclassification of bare soil or sand as an impervious area. This source of error has been reported in previous verification work of soil sealing datasets [23,25,26]. Non-vegetated agricultural fields are not supposed to be included as imperviousness [28]; however, IMCC is based primarily on the analysis of the Normalized Difference Vegetation Index (NDVI), so it is not surprising that fields with no or very poor crop growth could be included by mistake. Considering the dry weather of 2018, which could have led to more bare soil than usual, this error could be a contributor to the over-estimation of soil sealing.
Even if the IMD datasets captured impervious areas perfectly, and IMCC gave a true reflection of the process of soil sealing, it is important to remember that the primary use of this information is related to soil infiltration and not to monitoring land take of agricultural land. The conversion of land to impervious surfaces is generally accompanied by additional land being taken out of agricultural production. For example, Figure 1h shows a typical case where houses are built on agricultural land and a large part of the pixel becomes covered by impervious surfaces. The garden area is still permeable so the IMD of the pixel is less than 100%, yet the land taken from agriculture is 100%. In addition, some urban or artificial uses that take land from agriculture do not result in soil sealing, such as sports grounds or unpaved roads and car parking areas. Therefore, land take is usually greater than the area of soil sealing. At the same time, although we selected pixels that contained agricultural land, it is not given that the soil sealing occurred on the agricultural land. It could have occurred elsewhere within the pixel. Changes in the degree of imperviousness in our dataset are thus a reflection of soil sealing in the agricultural landscape. From this perspective, the results could over-estimate land take of agricultural land. Again, these issues become less of a problem with improved resolution since the pixels will contain less of a mix of different land uses.
Cysek-Pawlak et al. [43] assessed CLMS products for use in spatial planning. They suggested that national data often have better thematic and geometric accuracy, but that CLMS data can fill gaps for specific topics missing in national mapping programs. In Poland, a “biologically active area”, which is to some extent the opposite of imperviousness, is an obligatory urban indicator in local spatial development plans. However, there are no systems in place to monitor soil sealing. Therefore, this indicator is usually calculated using existing national databases, which are not always up-to-date and precise, or by interpretation of aerial orthophotos, which is a time-consuming process. Thus, the IMD datasets could be an alternative source of information on soil sealing in Poland.
In Norway, a system based on reports from municipalities has been developed to monitor soil sealing. However, this system has been found to be inadequate. This is partly due to the time lag between the municipality giving permission for construction work and the actual implementation of the plans, which could take years or even be canceled. In addition, the municipal reporting system did not include construction by the agricultural sector, such as the building of new modern animal housing. An important source of soil sealing was thus missing from the statistics. To remedy this, a new monitoring method was recently created based on a combination of data sources, including the Land Resource Map (AR5) and the national building database, using a buffer around the buildings to represent land taken outside the building extent [13]. This methodology also has shortcomings and misses important sources of soil sealing such as conversion to sports facilities or car parks.
Thus, IMD and IMCC can contribute to filling the gaps in the national monitoring data by providing a consistent dataset of a specific indicator tailored to assess soil sealing, with the benefits of remotely sensed data to cover large areas and provide timely updates. Currently, IMCC is not up to date, with IMCC 2015–2018 being the newest data available at the start of 2025. Therefore, shortening the timeline for production should be a priority.
The finer resolution of future IMD and IMCC 2018–2021 should also allow the detection of smaller patches and linear features such as small roads. This could be an advantage but may also bring extra challenges if many small temporary features are detected that are not relevant in the long term, like vehicles or silage bales. These may be filtered out by size or by checking that they are present in all three years. Finally, IMD and IMCC provide standardized, comparable data across Europe, which may be useful for comparing the end results of policies and regulations and systems of control and enforcement in different countries.

4.6. Limitations of the Study

Manual visual verification using remote sensing imagery inevitably opens possible differences in evaluations between interpreters. We tried to minimize this through calibration meetings and being conservative in our interpretation. This may have resulted in us over-estimating the accuracy of the dataset. Ideally, cross-verification should be carried out by multiple interpreters. However, we experienced that the classification was generally straight-forward and we did not identify any disagreements in our meetings, therefore we prioritized a larger sample size rather than double-checking the interpretation.
The Copernicus Land Monitoring Service products are continuously being improved as new satellite data become available. In our verification, we used IMCC 2015–2018 since this was the newest data available. However, this dataset has limitations because it compares 2015 data that had a resolution of 20 m, against 2018 data, with a resolution of 10 m. Hence, each 20 m pixel of IMCC 2015–2018 comprises four separate pixels in the 2018 data. The coarse resolution in 2015 means that some small sealed surfaces, like small buildings or parts of a road, were not detected. Their signature would be mixed up with the signatures of surrounding land use. When these become apparent in the finer resolution data, it may look like change, even though the buildings were there the whole time. This could also contribute to the over-estimation of soil sealing in IMCC 2015–2018. To address this complicated situation, CLMS made available a raster layer called the Degree of Imperviousness Change—Support Layer (IMCS) that displays decreased and increased imperviousness density due to technical change. The technical changes are artifacts introduced because of the two different levels of resolution. The user manual ([28], p. 34) explains how the layers can be used together to distinguish real change and technical change. We did not use the IMCS since we wanted to test how the primary, publicly available product performed in relation to real change on the ground. Although the expert products may be useful for research, we consider that municipal or regional planning offices are unlikely to compare multiple products. It is time-consuming to discover and understand the different types of information available. Also, downloading and working with these large datasets is time-consuming and requires computer capacity and GIS skills that may not be available for many potential users.

5. Conclusions

Imperviousness is an important issue for environmental management and planning and the preservation of food production capacity. Therefore, for many years, people have been trying to find methods to accurately detect imperviousness from satellite imagery. The HRL IMD datasets can provide useful information for Europe, with the advantage of being standardized and comparable across countries. IMCC provides easy-to-use information about changes in imperviousness, which can enable the assessment of the effects of policies such as those intended to prevent soil sealing of agricultural land. We found that IMCC 2015–2018 generally captured real cases of soil sealing correctly but that both gain and loss of impervious surfaces were over-estimated (high commission error). For applications of IMCC 2015–2018, we advise merging the closely related classes 1, new cover, and 11, increased cover, and the same for 2, loss of cover, and 12, decreased cover. These distinctions are not reliable, but the general information about increase or decrease had low omission errors. The transition to finer resolution (10 × 10 m) from 2018 represents a great improvement and will make the data more reliable and useful, also in distinguishing between new cover/increased cover and removal of imperviousness/decrease in cover.

Author Contributions

Conceptualization, W.F., A.H., S.O.K., J.R. and M.C.; Methodology, W.F., A.H., S.O.K., J.R. and M.C.; Validation, W.F., A.H., S.O.K., J.R. and M.C.; Formal analysis, W.F., A.H., S.O.K. and J.R.; Investigation, W.F., A.H., S.O.K., J.R. and M.C.; Resources, A.H.; Data curation, S.O.K., J.R. and M.C.; Writing—original draft, W.F. and S.O.K.; Writing—review & editing, W.F., A.H., S.O.K., J.R. and M.C.; Visualization, A.H. and S.O.K.; Supervision, A.H.; Project administration, A.H.; Funding acquisition, A.H. All authors have read and agreed to the published version of the manuscript.

Funding

The research leading to these results received funding from Norway Grants 2014–2021 via the Polish National Centre for Research and Development, grant no: NOR/POLNOR/InCoNaDa/0050/2019-00.

Data Availability Statement

The raw data supporting the conclusions of this article will be made available by the authors on request.

Acknowledgments

We would like to thank two anonymous reviewers for helpful comments on the manuscript.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Arnold, C.L.; Gibbons, C.J. Impervious Surface Coverage: The Emergence of a Key Environmental Indicator. J. Am. Plan. Assoc. 1996, 62, 243–258. [Google Scholar] [CrossRef]
  2. Jónsson, J.Ö.G.; Davíðsdóttir, B. Classification and valuation of soil ecosystem services. Agric. Syst. 2016, 145, 24–38. [Google Scholar] [CrossRef]
  3. Colsaet, A.; Laurans, Y.; Levrel, H. What drives land take and urban land expansion? A systematic review. Land Use Policy 2018, 79, 339–349. [Google Scholar] [CrossRef]
  4. UN General Assembly. Transforming Our World: The 2030 Agenda for Sustainable Development. Resolution 70/1 Adopted by the General Assembly on 25 September 2015; United Nations: New York, NY, USA, 2015. [Google Scholar]
  5. Shukla, P.R.; Skeg, J.; Calvo Buendia, E.; Masson-Delmotte, V.; Pörtner, H.-O.; Roberts, D.C.; Zhai, P.; Slade, R.; Connors, S.; van Diemen, S.; et al. (Eds.) Climate Change and Land: An IPCC Special Report on Climate Change, Desertification, Land Degradation, Sustainable Land Management, Food Security, and Greenhouse Gas Fluxes in Terrestrial Ecosystems; IPCC: Geneva, Switzerland, 2019; Available online: https://www.ipcc.ch/srccl/ (accessed on 1 April 2025).
  6. European Commission; Directorate-General for Environment; Jobstmann, H.; Prokop, G.; Schönbauer, A. Overview of Best Practices for Limiting Soil Sealing or Mitigating Its Effects in EU-27: Final Report; Publications Office: Luxembourg, 2011; Available online: https://data.europa.eu/doi/10.2779/15146 (accessed on 1 April 2025).
  7. European Commission. Communication from the Commission to the European Parliament, the Council, the European Economic and Social Committee and the Committee of the Regions, EU Soil Strategy for 2030—Reaping the Benefits of Healthy Soils for People, Food, Nature and Climate, COM/2021/699 Final. 52021DC0699, 17 November 2021. Available online: https://eur-lex.europa.eu/legal-content/EN/ALL/?uri=CELEX:52021DC0699 (accessed on 1 April 2025).
  8. Heuser, D.I. Soil Governance in current European Union Law and in the European Green Deal. Soil Secur. 2022, 6, 100053. [Google Scholar] [CrossRef]
  9. Turner, S. Proceedings of the Technical Workshop on Indicators for Soil Sealing; Technical report 80/-1; European Environment Agency: Copenhagen, Denmark, 2002; Available online: https://www.eea.europa.eu/en/analysis/publications/technical_report_2002_80 (accessed on 1 April 2025).
  10. Dramstad, W.E.; Fjellstad, W.J.; Strand, G.H.; Mathiesen, H.F.; Engan, G.; Stokland, J.N. Development and implementation of the Norwegian monitoring programme for agricultural landscapes. J. Environ. Manag. 2002, 64, 49–63. [Google Scholar] [CrossRef]
  11. Ståhl, G.; Allard, A.; Esseen, P.-A.; Glimskär, A.; Ringvall, A.; Svensson, J.; Sundquist, S.; Christensen, P.; Torell, Å.G.; Högström, M.; et al. National Inventory of Landscapes in Sweden (NILS)—Scope, design, and experiences from establishing a multiscale biodiversity monitoring system. Environ. Monit. Assess. 2011, 173, 579–595. [Google Scholar] [CrossRef]
  12. Eurostat 2025. LUCAS Web Site. Available online: https://ec.europa.eu/eurostat/web/lucas (accessed on 13 March 2025).
  13. Fadnes, K.D.; Munsterhjelm, N. Registrert Nedbygd Jordbruksareal. Kartbasert Måling Basert på Registreringar i Perioden 2020–2021; NIBIO Rapport 8 (142); NIBIO: Ås, Norway, 2022; Available online: https://nibio.brage.unit.no/nibio-xmlui/handle/11250/3033086 (accessed on 1 April 2025).
  14. Moser, A.; van Vliet, J.; Wissen Hayek, U.; Grêt-Regamey, A. Analyzing the extent and use of impervious land in rural landscapes. Geogr. Sustain. 2024, 5, 625–636. [Google Scholar] [CrossRef]
  15. Gardi, C.; Panagos, P.; Van Liedekerke, M.; Bosco, C.; De Brogniez, D. Land take and food security: Assessment of land take on the agricultural production in Europe. J. Environ. Plan. Manag. 2015, 58, 898–912. [Google Scholar] [CrossRef]
  16. Weng, Q. Remote sensing of impervious surfaces in the urban areas: Requirements, methods, and trends. Remote Sens. Environ. 2012, 117, 34–49. [Google Scholar] [CrossRef]
  17. Wang, Y.; Li, M. Urban impervious surface detection from remote sensing images: A review of the methods and challenges. IEEE Geosci. Remote Sens. Mag. 2019, 7, 64–93. Available online: https://ieeexplore.ieee.org/document/8846614 (accessed on 28 March 2025).
  18. Peroni, F.; Pappalardo, S.E.; Facchinelli, F.; Crescini, E.; Munafò, M.; Hodgson, M.E.; De Marchi, M. How to Map Soil Sealing, Land Take and Impervious Surfaces? A Systematic Review. Environ. Res. Lett. 2022, 17, 053005. [Google Scholar] [CrossRef]
  19. Güneralp, B.; Reba, M.; Hales, B.U.; Wentz, E.A.; Seto, K.C. Trends in urban land expansion, density, and land transitions from 1970 to 2010: A global synthesis. Environ. Res. Lett. 2020, 15, 044015. [Google Scholar] [CrossRef]
  20. Pourtaherian, P.; Jaeger, J.A.G. How effective are greenbelts at mitigating urban sprawl? A comparative study of 60 European cities. Landsc. Urban Plan. 2022, 227, 104532. [Google Scholar] [CrossRef]
  21. Orsi, F. Exploring the role of compactness in path-dependent land-taking processes in Italy. GeoJournal 2023, 88, 69–87. [Google Scholar] [CrossRef]
  22. Vallecillo, S.; Maes, J.; Teller, A.; Babí Almenar, J.; Barredo, J.; Trombetti, M.; Abdul Malak, D.; Paracchini, M.; Carré, A.; Addamo, A.; et al. EU-Wide Methodology to Map and Assess Ecosystem Condition: Towards a Common Approach Consistent with a Global Statistical Standard; Joint Research Centre Science for Policy report, European Commission; Publications Office of the European Union: Luxembourg, 2022; Available online: https://data.europa.eu/doi/10.2760/13048 (accessed on 1 April 2025).
  23. Strand, G.-H. Accuracy of the Copernicus High-Resolution Layer Imperviousness Density (HRL IMD) Assessed by Point Sampling within Pixels. Remote Sens. 2022, 14, 3589. [Google Scholar] [CrossRef]
  24. Hurbanek, P.; Atkinson, P.M.; Pazur, R.; Rosina, K.; Chockalingam, J. Accuracy of Built-up Area Mapping in Europe at Varying Scales and Thresholds. In Proceedings of the Accuracy 2010, Ninth International Symposium on Spatial Accuracy Assessment in Natural Resources and Environmental Sciences, Leicester, UK, 20–23 July 2010; University of Leicester: Leicester, UK, 2010; Volume 385388. Available online: https://gsweb4.umd.edu/sites/default/files/lcluc_documents/pazur_lcluc_8-2010_poster_0.pdf (accessed on 1 April 2025).
  25. Krówczyńska, M.; Soszyńska, A.; Pabjanek, P.; Wilk, E.; Hurbanek, P.; Rosina, K. Accuracy of the Soil Sealing Enhancement Product for Poland. Quaest. Geogr. 2016, 35, 89–95. [Google Scholar] [CrossRef]
  26. Pabjanek, P.; Krówczyńska, M.; Wilk, E.; Miecznikowski, M. An accuracy assessment of European Soil Sealing Dataset (SSL2009): Stara Miłosna area, Poland—A case study. Misc. Geogr. 2017, 20, 59–63. [Google Scholar] [CrossRef]
  27. EEA. HRL Copernicus Land Monitoring Service—High Resolution Layer Imperviousness: Product Specifications Document; European Environment Agency: Copenhagen, Denmark, 2018; Available online: https://land.copernicus.eu/en/technical-library/hrl-imperviousness-technical-document-prod-2015/@@download/file (accessed on 1 April 2025).
  28. Copernicus; EEA. Copernicus Land Monitoring Service: High Resolution Land Cover Characteristics; Lot1: Imperviousness 2018, Imperviousness Change 2015–2018 and Built-Up 2018; User Manual; Document Version 2.2; Copernicus Land Monitoring Service (CLMS); European Environment Agency (EEA), GeoVille GmbH: Innsbruck, Austria, 2020; Available online: https://land.copernicus.eu/en/technical-library/hrl-imperviousness-2018-user-manual/@@download/file (accessed on 1 April 2025).
  29. Pontius, R.G.; Shusas, E.; McEachern, M. Detecting important categorical land changes while accounting for persistence. Agric. Ecosyst. Environ. 2004, 101, 251–268. [Google Scholar] [CrossRef]
  30. Forbord, M.; Bjørkhaug, H.; Burton, R.J.F. Drivers of change in Norwegian agricultural land control and the emergence of rental farming. J. Rural. Stud. 2014, 33, 9–19. [Google Scholar] [CrossRef]
  31. Vinge, H. Food Security, Food Sovereignty, and the Nation-State: Historicizing Norwegian Farmland Policy. In Food Sovereignty in International Context: Discourse, Politics and Practice in Place; Trauger, A., Ed.; Routledge: New York, NY, USA, 2015; pp. 87–105, Chapter 6. [Google Scholar] [CrossRef]
  32. Vinge, H.; Sørensen, S.O. From Agri-Culture to Agri-Nature: New Alliances for Farmland Preservation in Norway. In Finance or Food? Hilde, B., Philip, M., Bruce, M., Eds.; University of Toronto Press: Toronto, ON, USA, 2019; pp. 225–242, Chapter 11. [Google Scholar] [CrossRef]
  33. Norwegian Ministry of Finance. Prop. 121 S, 2022–2023. Updated Agricultural Soil Protection Strategy. In Proposition to the Storting: Changes to the State Budget 2023 Under the Ministry of Agriculture and Food; Attachment 9; Norwegian Ministry of Finance: Oslo, Norway, 2023. (In Norwegian) [Google Scholar]
  34. Statistics Poland, Statistical Yearbook of Agriculture 2024, Warsaw 2024. Available online: https://stat.gov.pl/en/topics/statistical-yearbooks/statistical-yearbooks/statistical-yearbook-of-agriculture-2024,6,19.html (accessed on 1 April 2025).
  35. Tiitu, M.; Naess, P.; Ristimäki, M. The urban density in two Nordic capitals—Comparing the development of Oslo and Helsinki metropolitan regions. Eur. Plan. Stud. 2021, 29, 1092–1112. [Google Scholar] [CrossRef]
  36. Cortinovis, C.; Haase, D.; Zanon, B.; Geneletti, D. Is urban spatial development on the right track? Comparing strategies and trends in the European Union. Landsc. Urban Plan. 2019, 181, 22–37. [Google Scholar] [CrossRef]
  37. Stanilov, K. (Ed.) The Post-Socialist City: URBAN Form and Space Transformations in Central and Eastern Europe After Socialism; Springer: Dordrecht, The Netherlands, 2007; Volume 92. [Google Scholar] [CrossRef]
  38. Marks-Bielska, R. Factors shaping the agricultural land market in Poland. Land Use Policy 2013, 30, 791–799. [Google Scholar] [CrossRef]
  39. Badach, E.; Szewczyk, J.; Lisek, S.; Bożek, J. Size Structure Transformation of Polish Agricultural Farms in 2010–2020 by Typological Groups of Voivodeships. Agriculture 2023, 13, 1789. [Google Scholar] [CrossRef]
  40. Krøgli, S.O.; Aune-Lundberg, L.; Dramstad, W.E. Presence of agriculture in photos of Norwegian landscapes uploaded to Flickr. Nor. Geogr. Tidsskr. Nor. J. Geogr. 2023, 77, 243–254. [Google Scholar] [CrossRef]
  41. UN-Habitat. A New Strategy of Sustainable Neighbourhood Planning: Five Principles. Urban Planning Discussion Note 3. The United Nations Human Settlements Programme (UN-Habitat). 2015. Available online: https://unhabitat.org/five-principles-of-neighbourhood-design (accessed on 1 April 2025).
  42. Pelczynski, J.; Tomkowicz, B. Densification of cities as a method of sustainable development. IOP Conf. Ser. Earth Environ. Sci. 2019, 362, 012106. [Google Scholar] [CrossRef]
  43. Cysek-Pawlak, M.; Misiak, J.; Hościło, A.; Strand, G.-H.; Eiter, S. Spatial planning needs towards Copernicus Land Monitoring Services: Case studies from Poland and Norway. Eur. Spat. Res. Policy 2023, 30, 235–255. [Google Scholar] [CrossRef]
Land 14 00794 g001
Figure 1. Pairwise examples of pixels (yellow box) from IMCC (20 × 20 m), against a background of remote sensing images from 2015 (left) and 2018 (right). Examples (a–d) are from Poland and (e–h) from Norway (Source of images for Poland: Google Earth Pro, for Norway: Norge Digitalt). (a) IMCC class 2 (loss of imperviousness), correct, impervious plastic-covered silage in 2015 removed in 2018. (b) IMCC class 2 (loss of imperviousness), should have been class 0 (no change, no imperviousness), but incorrectly classed as impervious in 2015. (c) IMCC class 11 (increased imperviousness), should have been class 0 (no change, no imperviousness). (d) IMCC class 11 (increased imperviousness), should have been class 10 (no change, with imperviousness). (e) IMCC class 0 (no change, no imperviousness), should have been class 10 (no change, with imperviousness) due to the existence of a road in 2015 and 2018. (f) IMCC class 1 (new imperviousness), should have been class 11 (increased imperviousness). (g) IMCC class 2 (loss of imperviousness), should have been class 10 (no change). (h) IMCC class 11 (increased imperviousness), correct.
Figure 1. Pairwise examples of pixels (yellow box) from IMCC (20 × 20 m), against a background of remote sensing images from 2015 (left) and 2018 (right). Examples (a–d) are from Poland and (e–h) from Norway (Source of images for Poland: Google Earth Pro, for Norway: Norge Digitalt). (a) IMCC class 2 (loss of imperviousness), correct, impervious plastic-covered silage in 2015 removed in 2018. (b) IMCC class 2 (loss of imperviousness), should have been class 0 (no change, no imperviousness), but incorrectly classed as impervious in 2015. (c) IMCC class 11 (increased imperviousness), should have been class 0 (no change, no imperviousness). (d) IMCC class 11 (increased imperviousness), should have been class 10 (no change, with imperviousness). (e) IMCC class 0 (no change, no imperviousness), should have been class 10 (no change, with imperviousness) due to the existence of a road in 2015 and 2018. (f) IMCC class 1 (new imperviousness), should have been class 11 (increased imperviousness). (g) IMCC class 2 (loss of imperviousness), should have been class 10 (no change). (h) IMCC class 11 (increased imperviousness), correct.
Land 14 00794 g001
Table 1. Definitions of classes in the Imperviousness Classified Change (IMCC) 2015–2018 layer. IMD stands for imperviousness density, or the degree of imperviousness.
Table 1. Definitions of classes in the Imperviousness Classified Change (IMCC) 2015–2018 layer. IMD stands for imperviousness density, or the degree of imperviousness.
Class CodeClass Description
0Unchanged, with zero imperviousness in both 2015 and 2018
10Unchanged degree of imperviousness (>0 IMD)
1New cover, with zero imperviousness in 2015, but some IMD in 2018
11increased IMD
2Loss of cover, from some imperviousness in 2015 to zero in 2018
12decreased IMD
Table 2. The percentage distribution of pixels containing agricultural land amongst the classes of Imperviousness Classified Change (IMCC) 2015–2018, in Poland and Norway.
Table 2. The percentage distribution of pixels containing agricultural land amongst the classes of Imperviousness Classified Change (IMCC) 2015–2018, in Poland and Norway.
% of Area in Class
IMCC ClassesPolandNorway
0: Unchanged, IMD = 097.2999.76
10: Unchanged, IMD > 02.580.17
1: New cover, from IMD = 0 in 20150.120.07
11: Increased IMD0.010.00111
2: Loss of cover, to IMD = 0 in 20180.000.00020
12: Decreased IMD0.00000220.00004
Total area (km2) of pixels containing agricultural land181,9125935
Table 3. Confusion matrix for Poland, showing the number of pixels in each class of Imperviousness Classified Change—IMCC 2015–2018 (rows) and their classification according to our verification against aerial photographs or very high-resolution satellite images (columns). The sample size was 200 pixels per class, except for class 12, where only 5 pixels were available. The diagonal (bold) are correct classifications. UA = user’s accuracy and PA = producer accuracy.
Table 3. Confusion matrix for Poland, showing the number of pixels in each class of Imperviousness Classified Change—IMCC 2015–2018 (rows) and their classification according to our verification against aerial photographs or very high-resolution satellite images (columns). The sample size was 200 pixels per class, except for class 12, where only 5 pixels were available. The diagonal (bold) are correct classifications. UA = user’s accuracy and PA = producer accuracy.
Correct Classification According to Remote Sensing
Imagery
Classes010111212No. of PixelsUACommission Error
IMCC 2015–20180: Unchanged, IMD = 0198 2 20099.01.0
10: Unchanged, IMD > 08810561 20052.547.5
1: New cover, from IMD = 0 in 201567221083 20054.046.0
11: Increased IMD271640117 20058.541.5
2: Loss of cover, to IMD = 0 in 201814236148 20024.076.0
12: Decreased IMD211 1520.080.0
No. of pixels5241471631224811005
PA37.871.466.395.9100.0100.0
Omission error62.228.633.74.10.00.0
Table 4. Confusion matrix for Norway, showing the number of pixels in each class of Imperviousness Classified Change IMCC 2015–2018 (rows) and their classification according to our verification against aerial photographs or very high-resolution satellite images (columns). The sample size was 200 pixels per class, except for class 12, where only 9 pixels were available. The diagonal (bold) are correct classifications. UA = user’s accuracy and PA = producer accuracy.
Table 4. Confusion matrix for Norway, showing the number of pixels in each class of Imperviousness Classified Change IMCC 2015–2018 (rows) and their classification according to our verification against aerial photographs or very high-resolution satellite images (columns). The sample size was 200 pixels per class, except for class 12, where only 9 pixels were available. The diagonal (bold) are correct classifications. UA = user’s accuracy and PA = producer accuracy.
Correct Classification According to Remote Sensing
Imagery
Classes010111212No. of PixelsUACommission Error
IMCC 2015–20180: Unchanged, IMD = 01673111 20083.516.5
10: Unchanged, IMD > 02217323 20086.513.5
1: New cover, from IMD = 0 in 201519135397 20019.580.5
11: Increased IMD7673122 120061.039.0
2: Loss of cover, to IMD = 0 in 20182835 1251220062.537.5
12: Decreased IMD 2 7977.822.2
No. of pixels24344345133125201009
PA68.739.186.791.7100.035.0
Omission error31.360.913.38.30.065.0
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Fjellstad, W.; Hościło, A.; Krøgli, S.O.; Rizzi, J.; Chmielewska, M. The Potential of the Copernicus Product “Imperviousness Classified Change” to Assess Soil Sealing in Agricultural Areas in Poland and Norway. Land 2025, 14, 794. https://doi.org/10.3390/land14040794

AMA Style

Fjellstad W, Hościło A, Krøgli SO, Rizzi J, Chmielewska M. The Potential of the Copernicus Product “Imperviousness Classified Change” to Assess Soil Sealing in Agricultural Areas in Poland and Norway. Land. 2025; 14(4):794. https://doi.org/10.3390/land14040794

Chicago/Turabian Style

Fjellstad, Wendy, Agata Hościło, Svein Olav Krøgli, Jonathan Rizzi, and Milena Chmielewska. 2025. "The Potential of the Copernicus Product “Imperviousness Classified Change” to Assess Soil Sealing in Agricultural Areas in Poland and Norway" Land 14, no. 4: 794. https://doi.org/10.3390/land14040794

APA Style

Fjellstad, W., Hościło, A., Krøgli, S. O., Rizzi, J., & Chmielewska, M. (2025). The Potential of the Copernicus Product “Imperviousness Classified Change” to Assess Soil Sealing in Agricultural Areas in Poland and Norway. Land, 14(4), 794. https://doi.org/10.3390/land14040794

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop