Technical report No 40
CORINE land cover technical guide
– Addendum 2000
Prepared by:
M. Bossard, J. Feranec and J. Otahel
May 2000
Project manager:
Chris Steenmans
European Environment Agency
2
Cover design: Rolf Kuchling, EEA
Legal notice
The contents of this report do not necessarily reflect the official opinion of the European
Commission or other European Communities institutions. Neither the European Environment
Agency nor any person or company acting on behalf of the Agency is responsible for the use that may
be made of the information contained in this report.
Additional information on the European Union is available on the Internet.
It can be accessed through the Europa server (http://europa.eu.int)
©EEA, Copenhagen, 2000
Reproduction is authorised provided the source is acknowledged
Printed in Copenhagen
Printed on recycled and chlorine-free bleached paper
This report can be found on the internet at http://www.eea.eu.int
European Environment Agency
Kongens Nytorv 6
DK-1050 Copenhagen K
Denmark
Tel: +45 33 36 71 00
Fax: +45 33 36 71 99
E-mail: eea@eea.eu.int
Homepage: www.eea.eu.int
3
Contents
Introduction ....................................................................................................6
Part I State-of-play production methods of the CORINE
land cover database
1. Introduction..............................................................................................6
2. Hardcopy inventory..................................................................................7
3. Softcopy inventory ...................................................................................8
3.1. GIS................................................................................................................ 8
3.2. Text, images and icon information .............................................................. 8
3.3. System customisation................................................................................... 8
3.4. Required functionalities ............................................................................... 9
3.5. Co-pilot software ......................................................................................... 9
4. Conversion of (semi-) automated classifications.....................................10
4.1. Methodological aspects of a conversion ................................................... 10
4.2. Specific conversion quality issues .............................................................. 11
5. Validation methodology.........................................................................13
5.1. Sampling methodology.............................................................................. 13
5.2. Determination of the number of sampling points ..................................... 13
5.3. Random sampling technique ..................................................................... 14
5.4. Interpretation of the random sampling points .......................................... 17
5.5. The raising.................................................................................................. 18
6. References..............................................................................................21
Part II Addendum to the CORINE land cover
nomenclature illustrated guide
1. Introduction............................................................................................23
2. Methodology used for the compilation of the addendum .....................24
3. Characteristics of the CORINE land cover classes ..................................26
4 References...........................................................................................105
4
Introduction
This technical guide is an addendum to the CORINE Land Cover Technical Guide,
published by the European Commission within the frame of the CORINE Programme
(CEC, 1994).
This addendum 2000 was prepared by the European Topic Centre on Land Cover and
Phare Topic Link on Land Cover to support the land cover monitoring activities of the
European Environment Agency (EEA Multi-annual Work Programme 1998-2002, project
1.3.5 Land Cover).
The purpose of this addendum is:
•= to provide those in charge of collecting and updating CORINE Land Cover data with
additional information which was not included in the first CORINE Land Cover
Technical Guide (CEC, 1994);
•= to describe new methodological developments to produce the CORINE Land Cover
database, especially the use of soft copy tools for interpretation of satellite data on
video screen and (semi-) automated conversions of pixel-based clarification towards
CORINE Land Cover nomenclatures;
•= to enhance the description of the CORINE Land Cover nomenclature, including the
particularities relevant to the extension of the CORINE Land Cover inventory to
Central and Eastern Europe;
•= to extend the CORINE Land Cover class descriptions with a number of generalisation
rules for mapping at scale 1:100 000;
•= to improve the understanding of the thematic contents of the land cover classes
regarding landscape characteristics.
5
Part 1
State-of-play production methods of the
CORINE land cover database
6
1. Introduction
Initially the CORINE Land Cover (CLC) Programme applied a method for land cover data
collection based on a hardcopy inventory from satellite image printouts. This proved to be
the most feasible approach in the mid-eighties, the starting period of the Programme. This
method made only limited use of image processing and GIS software. The procedure has
proven its merits and is still valuable.
However, this procedure inevitably introduces errors during digitisation time, and asks for
two intermediate hardcopy products (transparencies and satellite photo) before obtaining
digital results. Moreover border matching between sheets and regional boundaries was
more critical since it depended on variables such as the quality of the transparencies and
the accuracy of the interpretation.
However, technical developments have made it possible to introduce computer
technologies throughout the process of building the inventory (softcopy). Moreover, it is
more comfortable to have data sets on screen, enabling more efficient performance rates,
and hence reducing costs. For this reason an overview is given of the different approaches
that can lead to the production of a CLC database.
In some countries experiments have been set up with statistical classification procedures
based on spectral classification of a limited number of land cover classes. The results of
these complete coverages are worth translating into the CLC nomenclature. However, it
should be emphasised that due to the physiographic nature of the CLC classes, and to a
limited extent the functional distinctions that are introduced in the nomenclature, it is
hardly imaginable to fully match the CLC nomenclature starting from an automated
classification procedure, without additional human interpretation work.
Many variants of switching between, or combining, analogue and digital procedures were
used in the different Member States, but it should be stressed that both analogue and
digital approaches, as well the conversion methods are still valid for the CLC production
process and its updates.
7
2. Hardcopy inventory
In the first CLC Technical Guide (CEC, 1994), the procedure is described to perform a
computer aided photo-interpretation. At the very beginning of the CLC project, the idea
was to use digital techniques as much as possible. However, technical constraints and costbenefit
considerations made it unrealistic at that time to propose this approach as ‘the one
and only to be followed’. As a result, the digital image processing system was initially used
to prepare the basic satellite image data in such a way that geometric, spectral and
radiometric characteristics match the required optima to produce high quality hardcopy
prints of the images in order to enable correct photo-interpretation.
Hence, the accepted common methodology for CORINE Land Cover consists of computerassisted
photo-interpretation of satellite images, with the simultaneous consultation of
ancillary data, classifying data into classes of the CORINE Land Cover nomenclature.
Interpretation of the images is based on transparencies overlaid on 1/100.000 hardcopy
prints of satellite images. Ancillary data are essential to help identify and confirm the
contents of certain land cover features, which have been detected on the images.
It is useful to remember in this process the principal characteristics of a CLC feature:
•= the feature should be considered as a homogeneous zone or should comprise a
combination of zones with a certain recognizable structure;
•= it should present a significant part of the surface in the used scale;
•= it should be easily limited from its neighbouring zones.
In some cases, the digital image processing system was used during the interpretation
process, as a more flexible visualisation tool, enabling further focus on the areas of
interest. From the digitisation onwards, one fully re-entered the digital world. In most cases
the image processing system was addressed again to evaluate geometrical precision of the
results, and to verify the thematic contents of the interpretation.
8
3. Softcopy inventory
Technical developments have made it feasible to interpret and digitise in a single passage
through the use of softcopy digitisation. Some general rules should be respected, with
regard to the different man-machine interface. General considerations can be given on
integrated GIS systems that are designed specifically for such tasks.
3.1. GIS
These integrated GIS systems are mostly provided by a vector database manager, designed
to work together with raster data. A proper design of the GIS engine for CORINE Land
Cover mapping is required to reach the high performance levels one needs in operations
typical of an integrated system as computer aided photo-interpretation, or the updating of
a vector based database on basis of imagery.
One of the most useful options one has to look for when deciding on such an integrated
system, is the capability to manage simultaneously several geographic data sets in different
windows or in the same window, if they have the same frame reference and to link to each
window a different set of colour or B&W images.
In order to manage a wide range of data sets from different origins, it is advisable to look
for support of several data types.
Also the import and export options should cover a wide range of raster file formats, such as
BSQ, BIL, BIP, TIFF, geoTIFF. The same remark is valid for the vector formats, eventually
including the data base info as well, although less evident (e.g. Arc/Info export, Arc/Info
ungenerate, DXF).
After years of commercial battles on file formats, it is worth mentioning that both for raster
data and vector data, at last a trend has started to try to identify a common exchange
formats. In case of raster data, major software developers and data providers agreed on the
geoTIFF standards, whereas for vector data, discussions are still ongoing.
3.2. Text, images and icon information
More generally spoken, many different kinds of information may be connected to a
geographic objet, consisting not only on alphanumeric attributes, logic and topologic
relations with other objects, but also image documents, icons and text.
Image documents and icons are very different from layered GIS information because they
are not necessarily geo-referenced. Image documents may be pictures or drawings showing
the features of a geographic object, and its neighbours. It is a asset when the software
provides text and icon linkages with each type of object, and interactive selection and
display.
Moreover texts related to objects should be editable either through standard word
processing applications, or through OS related text editors.
3.3. System customisation
Various software packages are available, some of them on PC-based operating systems,
some of them in typical workstation environments. It is suggested to verify carefully the
architecture of the software, and especially add-ons such as specific development kits,
and/or the modularity of the package and the ease with which one can address separate
modules, in order to enable easy development of advanced user interfaces or extra
functionality.
Customisations can occur at different levels depending on specific needs or the
programming know-how. A typical first level is the development of a customised user
9
interface. A second level is the development of simple applications in straightforward
programming environments such as, for instance, Visual Basic. A third level consists of the
development of advanced applications in advanced environments such as C or C++.
Although not always possible, it is advisable to aim at a development design, which enables
as far as possible matching with general standards such as the ANSI standard, in order to
facilitate porting to other OS environments.
3.4. Required functionalities
A whole series of functionalities are offered by the different software developers. However
one can identify a minimal set of functional requirements that should be matched by the
system:
•= Basic functions: interactive display and editing of objects geometry, attributes and
relations
•= Image functions: basic image management tools such as import/export, integrated
display of raster and vector data, image enhancement by means of stretch functions,
convolutions
•= Query functions: geographic queries, arithmetic queries, Boolean queries. Queries
may deal with geometry, attributes and logic and topologic relations.
•= database manager program: functions to customise de data base structure and to
define a geographic index to enable faster data access.
•= Import/export functions: programs to exchange data from other platforms, in other
file formats.
•= Documentation functions: to enable the inclusion of information on the structure of a
file in its header.
•= Geo-referencing functions: image to image geo-referencing based on polynomial
geometric transforms, geo-referencing based on digitised GCP’s in general.
•= Statistical functions: functions to provide statistical operations on images such as
principal components and spectral classifications.
An additional remark could be made concerning the use of softcopy for CLC inventory:
Considering that one should be able to display different data types in a single window, in
an overlay situation, it would be user-friendly if one could also display the different legends
related to the displayed data.
3.5. Co-pilot software
The Joint Research Centre at Ispra developed a specific software for softcopy production,
maintenance and updating of the CORINE Land Cover data base (see also Perdigao et al,
1997). This software offers an integrated GIS system, running under MS-windows 3.1x or
MS-windows95. Geographic data in Co-pilot are managed ‘at continuum, which means that
data acquired from different maps on different times are managed in one and only
geographic data base without breaks.
The minimum information unit on Co-pilot is the geographic object. From the geometric
structure viewpoint, a geographic may belong to three different classes: points, lines, or
surfaces. Any geographic object is characterised by a type (e.g. building, railway, forest).
Types are defined, for each geographic database, by the user in an information structure
called Catalogue. The object type specification contains information about geometric class,
alphanumeric attributes, logic and topologic relations with other objects and auxiliary
display attributes. The graphic display mode is defined in another information structure,
called Typelook. Primitive graphic elements are defined in another structure called:
GraphicSet. All main information structures of the system are stored in ASCII files and are
well documented and easy to edit.
The Co-pilot software is available from JRC for the national CLC Land Cover production
teams.
10
4. Conversion of (semi-)automated
classifications
As an introductory note, it is worth warning for the limitations of this approach,
intrinsically connected with the procedure of statistical classification processes. First of all
one should be aware about the variability of the image quality and contents, due to several
variables such as atmospheric conditions, technical image quality and relief. Secondly, the
already mentioned differences between land cover, sensu stricto, and functional intrusion
from land use items makes it difficult to set up such an automated conversion. Thirdly
there’s the essential difference between pixel-based classification and the human holistic
interpretation capacity, in which spatial organisation of pixels to a higher level spatial
pattern or objects enables grouping of pixels into heterogeneous or composite CLC
classes. Having these remarks in mind, it is clear that, given an attentive surveillance of the
conversion process, one can obtain valuable CLC results, and so this approach disserves
attention in this addendum to the technical guide.
The experiments for conversion carried out so far prove to be successful for the positional
accuracy, which is at the basis of the entitlement ‘spatial generalisation’ or ‘cartographic
generalisation’, and reasonably controlled as far as attribute accuracy is concerned.
Anyhow, one should be aware that the rules for conversion or ‘spatial generalisation’,
whether manual of automatic, depend upon complex interactions between the size and
patterns of ground features, the user objectives, the operator, the form of input
information and output scales and formats.
4.1. Methodological aspects of a conversion
‘Spatial generalisation’ basically reduces the complexity of the data structure, influencing
some of the components of data quality including location accuracy, attribute accuracy,
consistency and completeness. When generalising, map accuracy deteriorates in favour of
simplicity and legibility. Therefore the quality of the generalised map is related to the
specified requirements.
As an overview to the generalisation, one can discern three possible types of spatial
procedures:
•= procedures where no spatial contiguity is taken into account;
•= procedures where spatial contiguity is taken into account (land cover at point x also
depends on spatial associations with the surrounding area);
•= procedures where the attributes vary over time.
A few user controlled methods, developed for the UK, Swedish and Finnish land cover
situation can be summarised into the following steps, which they share in common:
•= reclassification: this is the regrouping of objects into the CORINE feature classes,
sharing equivalent attributes. Normally reclassification is done in a hierarchical way as a
relabelling from lower to upper levels in the hierrachy. This process, however, can
hardly cover all specifications of the 44 CLC class hierarchy;
•= aggregation: this is the merging of a group of individual small features that are in close
proximity, and the representation of this group as one continuous area;
•= amalgamation: this is the joining of contiguous features, either by merging the area to
the closest one, semantically, or by dividing the area between neighbouring features. In
the attribute amalgamation, the feature to be generalised is compared to the
neighbouring features, and the semantically nearest class is selected to fill the whole
feature. In the spatial amalgamation, the pixels in the feature that are to be generalised
are filled individually according to the nearest class in Euclidian distance, and thus the
features are divided between the neighbours. Combination of attribute and spatial
amalgamation is also possible: first perform the attribute amalgamation, and if no
11
semantically close neighbours exist any longer, perform the spatial amalgamation for
the rest of the features;
•= smoothing: this is the relocation or shifting of a boundary to remove the step-effect of
the original raster cells, and to capture the significant trends of boundaries;
•= simplification: this is the reduction of boundary line complexity by removing changes of
direction smaller than a certain threshold.
The quantitative correspondence between automated and manual CORINE land cover
maps represents the cumulative impacts of:
•= accuracy in the per-pixel classification
•= efficiency of automatic generalisation
•= accuracy of visual interpretation
•= efficiency of manual generalisation
It is stated from experimental studies that the major limitation of this approach is the
quality of the automatic per pixel classification.
4.2. Specific conversion quality issues
Land cover is an example of a categorical coverage, i.e.: a single variable taking a finite
number of discrete, nominal classes, whose value is known everywhere in the mapped
region. The observation medium for CLC data is a satellite image, measuring reflection
values of different land cover types within operational limits, defined for a certain
measurement system (satellite sensors), for certain production procedures (computer
aided visual photo-interpretation) and for certain areas (Europe).
Two data models are involved in the conversion procedure to present land cover data:
•= the polygon model identifies areas of homogeneous value and is typified by the lines
drawn on a map, or its digitised equivalent;
•= the raster model identifies the class that is dominant in every cell of a regular array of
tessellation and is typified by a classified remotely sensed scene.
The difference between these two is significant from an operational point of view.
However, difference is insignificant under certain conditions when testing land cover
quality. The insignificance is due to the use of pixel sizes that do not significantly alter the
feature boundaries. Therefore, the pixel size should be well below the minimum mapping
unit, as it is the case in these conversion methods. Under this condition, the positional
errors near boundaries due to pixel size can be considered insignificant.
In a ‘phase-space’ model for continuous multidimensional variables which are partitioned
into domains of land cover classes, the mapped area can be assumed to be discretised as an
array of N cells, with rows and columns indexed by x and y. The variable that is mapped
over the area has n discrete classes, indexed by k. The probability that cell x, y falls into
class k, is given by p
k
xy. Associated with is cell is a vector of probabilities, which are
commonly replaced in remote sensing by ‘the most likely class’.
The sources of errors in the multispectral classification can be reduced to four main types:
•= confusion of spectral classes (overlap of spectral distributions)
•= mixed pixels (pixels containing 2 or more land cover types)
•= system errors (recording, geometry)
•= conceptual errors (coherence between land cover classes and separability in the
feature space, and internal variability within one class).
4.2.1. Statistical properties of land cover data
A phase-space model illustrates the interdependency between the taxonomic definitions
that prescribe the partitioning of the phase-space and the structure of the resulting map.
Such a model allows investigating the relationship between the accuracy of a land cover
database and the accuracy of measures derived from it. Many other statistical methods are
not suitable for studying the quality of land cover data, due to the discrete and nominal
scale of the classes, and their spatial differentiation. In fact, one should use mainly non-
12
parametric or distribution free methods, or simply some statistical descriptors, which are
not formal statistics.
4.2.2. Properties of the generalisation procedure
Actually none of the present quality measures are suitable as such to describe the quality of
the generalisation result. This means that one has to rely on tests with summary measures
or with a visual overlay or resulting features with the original one. A good starting point is
the properties of the different conversions. The first and most important property affecting
quality assessment is the change in scale of the data. The generalisation reduces complexity
of the data structure, and adds error to the database. Thus the quality is always
deteriorated in favour of simplicity and legibility. Most of the quality assessment
procedures do not take these errors into account and therefore do not understand that the
error rate in the generalised database includes both the degree of generalisation and the
real error. The bias in summary measures and unintended positional and attribute errors
produced by generalisation.
13
5. Validation methodology
Validation is a method to evaluate the CORINE Land Cover inventory with regard to
identification and delineation of the unit areas according to the CORINE Land Cover
methodology. The user of the data shall be informed on the reliability of the inventory. It
is not the aim of the validation to assess the data model and the method for data collection
of the CORINE Land Cover project.
During validation the data collection is repeated with the same data sources as in the
original land cover data collection. The data collection is repeated by means of random
sampling. Parts of the national area are interpreted again and then compared to the
original interpretation. The results of the comparison provide a confusion matrix by which
it is possible to differentiate for particular land cover items and to analyse for omission and
commission errors.
The user is therefore supplied with information on the validity of particular land cover
items. By building weighted sums of the separate results of the validation, it is possible to
produce results for the entire database or to produce results for different levels of the
nomenclature. The proposed method for random sampling will allow any kind of
stratification. The stratification can be applied for the different levels of the nomenclature
as well as for different regions of the land cover inventory.
5.1. Sampling methodology
The land cover items will be identified at points whose geographical locations are
determined by random sampling. A procedure of stratified random sampling will be
applied. The stratification can be done according to the land cover items and/or
according to geographical regions. The input parameters for the random sampling
procedure are the digital land cover map and the amount of random sampling points for
each stratum. In the first phase of the procedure, strata are selected and the amount of
points to be drawn in every selected stratum is determined. In the second phase the
sampling points are drawn, i.e. their geographical location within each land cover unit is
determined. The selection method guarantees that every point represents the same area
size in its stratum. The validation procedure yields the surface of the wrongly identified
area as a percentage of the total area surface for each stratum as well as for the total area to
be validated. By stratifying a sample, it is possible to minimise the standard error of the
total sample. Moreover the size of the sample can be determined for each stratum
separately according to the expected variation in every stratum.
In the following the different steps of the sampling procedure will be explained in detail.
The method was developed and tested during the national CORINE Land Cover inventory
in Germany.
5.2. Determination of the number of sampling points
The necessary amount of sampling points for each stratum h is determined on the basis of
a one-step random selection with the binomial formulation and will be calculated using the
following formula:
n
p p
h
h h
h
=
−( )1
2
σ
where nh is the number of sample points, ph is a previously estimated error rate and σh is the
accepted absolute standard error for the estimation of the percentage of incorrect
identified size of area for the stratum h.
14
Random sample size
per stratum
144
204
256
300
336
364
384
396 400
36
51
64
75 84 91 96 99 100
0
50
100
150
200
250
300
350
400
10% 15% 20% 25% 30% 35% 40% 45% 50%
Expected error rate
2.50%
5%
Points
Standard error
Figure 1: Number of sample points dependent on the expected error rate and
the accepted absolute standard error
The expected error rate can be estimated by interpreting selected basic working units
twice. The resulting digital maps have to be overlaid using GIS techniques. Summing up
the area sizes of the units with identical and with different key-codes, the error rate can be
calculated. An error rate of 50 % yields the highest number of sample points. The standard
error determines the precision of the result. Results with higher precision require a higher
amount of sample points. When the results of the validation are at hand, it is possible to
calculate the standard error on the basis of known error rate. When the error rate is higher
than previously estimated, the standard error will increase1
. On the contrary, the standard
error decreases when the error rate is lower than estimated.
For every stratum the error rate has to be estimated and the accepted standard error has to
be determined. Then the number of points, that have to be distributed, can be calculated.
When the size of the area of a stratum is low, the density of points can increase
dramatically. From the statistical point of view, this may be correct, but from the practical
point of view, it may be not sensible to place in a 1 km² unit a total number of 256 points.
Therefore the number of points may be limited to a certain amount per square unit. It is
suggested to introduce a limit of 2 points per km
²
. Therefore the total number of points
per stratum will be diminished for strata that are relatively small. This would result in a
higher standard error. However with the proposed method for the interpretation
2
of the
points, it would be possible to accept a very high density of random sampling points too.
5.3. Random sampling technique
5.3.1. The first step of random sampling
The selection probability is proportional to the area. The step-size is calculated by dividing
the total area of the stratum through the number of sample points per stratum. The
drawing of the land cover units and the calculation of the number of sample points per
1
Assuming normal that the previously estimated error rate is nowhere below 50 %
2
The points are not visible during interpretation.
15
chosen land cover unit is done in a cumulative way. The area sizes of the land cover units
of a stratum are subsequently cumulated. Then the land cover units are systematically
selected applying the step-size with a random starting point. At the same time the number
of sample points per land cover unit is calculated.
Ah: Total area of the stratum h.
Ahi: Area of the land cover unit i in the stratum h.
nh: Number of random sampling point in the stratum h.
A
n
h
h
: Step-size
Interval for random start in the size of
A
n
h
h
Figure 2: Scheme for cumulative selection of land cover units and calculation of
the number of random sampling points
In figure 2 the scheme for the selection of units and the determination of the number of
points is illustrated. The land cover units have to be systematically cumulated, e.g.
cumulation of the units starting in the northwestern edge of the map and ending in the
southwest. The selection starts at a randomly chosen value in the interval from 0 to the
step-size. The vertical arrows indicate the random sampling points. Every unit that has
been hit by the systematic selection receives a random sampling point. In the figure 2, the
units no. 1 and 3 have one point, the unit no. 5 has two points and the units no. 2 and 4 do
not get a point. This method ensures a random selection together with a spatially equal
distribution of the random sampling points. The units whose area-size is larger than the
step-size are always selected. The number of points is dependent on the size of the unit
therefore the selection is area-proportionate.
5.3.2. The second step of random sampling
In the second step the location of the random sampling points in the units are
determined. The points in the units are drawn in a systematic way combined local random
sampling. This method ensures that the points are not concentrated in one part of the
unit, but are optimally distributed in the unit.
There are three different cases in sampling:
•= If only one point has to be drawn from a given unit, pure random sampling does this.
Determining the extreme co-ordinates of the unit draws the point. These co-ordinates
define the smallest rectangle (boundary rectangle) which includes the unit completely.
Random sampling of a point within this rectangle is simple. A point drawn is valid if it
is located within the unit. Sampling continues until a valid point has been drawn.
•= If more than one point has to be drawn, systematic sampling is combined with local
random sampling.
A
n
h
h
A
n
h
h
A
n
h
h
Ah1 Ah2 Ah3 Ah4 Ah5 Ah6
16
•= If n points have to be drawn and 2n3, the boundary rectangle is subdivided into n
rectangles of the same size by straight lines that are parallel to its shorter sides. A point
is drawn once from each of these rectangles. If none of the drawing processes is valid,
this procedure is repeated. If at least one but not all drawing processes are valid, there
will be a remainder r>0 of less than n points. To distribute these points, r rectangles of
the n rectangles are randomly selected from which exactly one point each is drawn
once. If each of the r drawing processes is valid, all points have been determined,
otherwise the procedure is repeated with the new remainder of r' points, 0<r'r.
If five or more points have to be drawn, the boundary rectangle of a given unit is covered
with a regular grid of squares oriented to the north. A randomly selected point of
intersection of the grid determines the position of the grid. The width of the grid (d)
depends upon:
•= the number n of sampling units to be drawn from the unit, and
•= the maximum extension of the unit both from north to south and east to west, i.e.
the length of the boundary rectangle sides.
Figure 3 shows how the unit and the boundary rectangle are covered by the grid:
width of the grid (d)
unit
m = 27
boundary rectangle
Figure 3: Method for drawing the plots - stage one
For determining the width of the grid (d), n should be greater than or equal to the
maximum number of those rectangles of the grid, which overlap the unit. This would be
the lower level of d. However, the impact of the random component on the sampling
methodology grows as d rises, so that the value chosen for d should be ‘as close as possible’
to that lower level. If R and H are the sides of the boundary rectangle from east to west and
north to south (easting and northing, respectively),
s
R
d
H
d
=
é
+
é
+( )( )
∆ ∆
2 2
.
Here any x is written as [x] (thus differing slightly from the usual definition): if x is a
whole number, [x] stands for x-1, otherwise for the greatest whole number smaller than x.
As it is not possible to resolve the above formula so as to obtain d, another formula must be
drawn upon to determine the width of the grid. To make sure that ns, the obvious solution
would be obtaining d as the result of the quadratic equation
n
R
d
H
d
= + +( )( )
∆ ∆
2 2
.
17
However, in practice this will in many cases lead to a width, which is far to great, especially
for units, which are far from filling their boundary rectangle. Therefore, the above formula
is replaced by:
n
R
d
H
d
= + +( )( )
∆ ∆
1 1
(F),
which leads to the following solution:
d
R H
n
c c with c
R H
n
=
−
+ + =
+
−
∆ ∆ ∆ ∆
1 2 1
² ,
( )
,
If m is the number of rectangles overlapping the unit, the inequation nm will in most cases
be satisfied. After the m rectangles have been determined, the procedure used is similar to
that of case 2 (2n3). Equation
n a m r with r m= × + <, ,
defines a single whole number. In each of the m rectangles, one point is drawn a times.
Due to invalid drawing processes, there will be a remainder of r'r points. If r'm, the process
is repeated with r' instead of n. If r'<m, the remaining r' points are drawn exactly in the
same manner as in case 2. If n<m for a given unit, as is possible since formula (F) is used, it
is obvious that a=0, so that r'=n from the beginning, i.e. the total of points to be drawn (n)
is treated as a ‘residual’ right from the start.
After finishing the random sampling the geographical location of each point is
determined. Now the type of land cover for the points can be interpreted.
5.4. Interpretation of the random sampling points
The basic principle of the land cover interpretation is explained in the CORINE Land
Cover Technical Guide (see chapter ‘Computer aided-visual interpretation’). It is an
iterative process that leads to the final CORINE Land Cover inventory. First the units on
the first level of the nomenclature are delineated.
In a second step these large units are further subdivided. During this step the ancillary data
is evaluated deeply. The step of subdividing big land cover units is repeated until all land
cover units, that exceed the minimum mapping size are delineated and identified. After
this quasi-definitive delineation the field survey is conducted for resolving remaining
interpretation problems and a general verification of the interpretation.
All in all the delineation is done from the outside boundaries of large and complex land
cover units to the inner boundaries while subdividing further. The size of the units
decreases during the interpretation. On the level of the land cover nomenclature, the
identification is done from higher to lower levels.
A fundamental difference between the original data collection and the validation is, that
the type of land cover is determined at one point. The interpretation is done in the
opposite direction compared to the original data collection. The photo-interpreter tends
to determine the type of land cover at the sampling point without regard of the
surrounding area. It is however vital to the validation process that the identification is done
for a larger area around the sampling points, sticking to the minimum mapping size. The
ignorance of the minimum mapping size would lead to a completely different
interpretation. The application of the CORINE Land Cover methodology during the
validation process is a fundamental principle.
To support the application of the CORINE Land Cover inventory, the identification should
18
be done in an eccentric circle with 1 km radius around the random sampling point. The
interpreter should not know the location of the random sampling point. The
interpretation should be done by a person who was not directly involved in the original
data collection of the particular basic working unit but who is familiar with the CORINE
Land Cover (interpretation) methodology. The total circle should be interpreted
according to the CORINE Land Cover interpretation method.
5.4.1. Conduction of the validation
The interpreter should be equipped with the data sources that were used in the original
data collection. It is suggested to do the mapping directly on-screen. As experiences have
shown, the geometric distortions when placing a second interpretation transparency on
the photographic print of the satellite image are too big. Therefore the interpretation of
the circles should be done on the background of the digital satellite image on-screen.
The display should be fixed to a certain scale that should not exceed the value 1/40.000.
When zooming in the image the interpreter looses the overview of the image. Additionally
sticking to a scale similar to the CORINE Land Cover inventory scale ensures that the
geometrical accuracy is similar to the accuracy of the original data. It is possible to display
parts of the original land cover data in areas where no sampling points are located. This
helps the photo-interpreter getting familiar with the CORINE Land Cover interpretation
method. Due to financial restrictions it is not possible to conduct a field survey during
validation. Therefore it is vital to have a good documentation of the field survey, that was
performed in the original data collection. Moreover a good meta data documentation is
very helpful for the interpreter performing the validation.
It is very important to consider the minimum mapping size, since this cannot be
systematically verified as in the original data collection. Units that are located at the border
of the circle can be smaller than 25 ha when they can be extended outside the circle to a
size larger than 25 ha. The boundaries of the outline coverage of the satellite image are
used as reference for the delineation. Allcircles are subsequently interpreted. It is sensible
to divide the total area of the national CORINE Land Cover database into smaller units.
Therefore it is suggested to apply the principle of the basic working units to the validation
process too.
5.4.2. Verification of the interpretation results
The GIS operator has to check the validation map for topology errors. The checks are
described in the chapter on verification of the digital data in the original data collection.
The minimum mapping size can be checked when the land cover unit is not located at the
border of the circle.
The results of the interpretation are documented in a plot. The plot is also for
documenting the errors found during the verification. The plot is superimposed on the
photographic print of the satellite image. The leader of the validation team verifies the
land cover units on the plot. The layout of the plot is described in the annex 7.5. Finally a
summary table of all land cover units in the basic working unit is created. The table is for
documentation and verification purposes. The total size of the inner circle area is checked
with the value before starting interpretation. The key-codes are checked for plausibility.
5.5. The raising
After completing the verification of the validated data the coverages are overlaid with the
random sampling points to determine the validated land cover key-code of the points. A
confusion matrix is set up showing the omission and commission errors. The estimation of
the percentage of correctly classified area per stratum can be done by dividing the number
of correctly classified points by the total number of points in a stratum.
19
(1) o ~
~
P
X
nhh
hh
h
=
with:
o ~Phh
: Estimation of points, that belong to stratum h and were assigned to the same item
of the nomenclature in the main data collection and the validation,
h: Stratum,
Xhh
~ : Number of correctly classified points in stratum h,
nh: Number of total points in the stratum h.
Since the drawing of the points was strictly area-proportionate the estimation of the
correctly interpreted area for each stratum can be calculated by multiplying the amount of
correctly classified points with the step-size. The stratification can be done by dividing the
total area into smaller parts, e.g. dividing the total area into geographical regions. The
stratification can also be done according to the items of the CORINE Land Cover
nomenclature. In this case the estimation of accuracy can be done for each item on the
nomenclature as well as for the total area. The estimation of the overall accuracy of several
strata has to be proportional to the share of the stratum on the total area of the considered
strata.
(2) o o ~
P
A
A
P
h
hh
h
= å
with:
o ~Phh
: The estimation of the percentage of correctly classified area in stratum h,
Ah
: Total area of stratum h of the CORINE Land Cover inventory,
A: Total area of the CORINE Land Cover inventory.
When the stratification is done according to the classes on the nomenclature, it can be
interesting to aggregate the results of the validation on different levels of the
nomenclature. The estimation of the correctly classified area on the higher level is
calculated in two steps. In the first step the estimations of the correctly assigned points
according to the higher level of the nomenclature in a stratum have to be summed up.
(3) o ~
~
~ ~
P
X
nhH
hh
hh H
=
∈
å
with:
o ~PhH
: The estimation of the correctly classified area of stratum h on the aggregated level
~
H ,
Xhh
~ : Number of correctly classified points in stratum h on the aggregated level
~
H ,
nh: The total number of points of stratum h.
Now the estimation has to be aggregated for the different strata in the sample that belong
to the aggregated level H. This has to be done proportional to the size of area of each
stratum.
(4) o o~ ~P
A
A
PHH
h
H
hH
h H
=
∈
å
with:
oPHH : The estimation for the correctly classified area on the aggregated level H,
Ah: The total area in stratum h,
AH: The total area of the strata that are elements of the aggregated level H,
oPhH : The estimation of the correctly classified area of stratum h on the aggregated level
H.
20
For each estimator of correctly classified area, it is necessary to calculate its standard error.
In each stratum the standard error can be estimated according to the binomial
formulation. In the proposed formulas for the estimation of the standard errors the
specific technique for selecting the area units and the technique for locating multiple
points in the area units is not considered. Since these techniques usually increase the
precision of the estimation, it can be assumed that the standard error tends to be lower
than calculated.
The error-variances of the formula (1), (2) and (4) can be estimated applying the
following formula(s):
(1') ( ) ( )Var P
p p
nhh
hh hh
h
o o ~
~ ~
=
−
−
1
1
This is the variance for the estimator of the percentage of correctly classified points in one
stratum. For the calculation of the variance of the total accuracy the area size of each
stratum has to be considered.
(2') ( ) ( )Var P
A
A
p p
n
h hh hh
hh
o o
~ ~
=
æ −
−
å
2
1
1
with:
Ah
: Total area of stratum h,
A: Total area of all strata.
The variance for the estimator on an aggregated level is calculated in the same way.
(4') ( ) ( )Var P
A
A
p p
nHH
h
H
hH hH
hh H
o o ~
~ ~
=
æ −
−∈
å
2
1
1
The standard errors for the above mentioned estimators are the root of the variances.
21
6. References
CEC, 1994: CORINE Land Cover Technical guide, Luxembourg.
DIN, Deutsches Institut für Normung e.V. (Hrsg.), 1994: Qualitätsmanagement und Statistik:
Normen, Verfahren 3: Qualitätsmanagementsysteme, 2. Aufl., Beuth Verlag, Berlin.
Douglas, H.D., Peucker T.K., 1973: Algorithms for the reduction of the number of points required to
represent a digitised line or its caricature, in: Canadian Cartographer, 10, No.2, Dec. 1973.
Heidrich-Riske, Holger,Dr., 1995: Ziehung einer Punktstichprobe mit ARC/INFO am Beispiel der
Qualitätskontrolle im Rhamen der Realisierung des EU-Projektes CORINE Land Cover in der
Bundesrepublik Deutschland, Kranzberg.
Statistisches Bundesamt, 1994: CORINE Land Cover Datenerhebungsanleitung, Wiesbaden.
Jaakolla O., 1997: Quality and automatic generalisation of land cover data, Helsinki
Perdigao V., Annoni A., 1997: Technical and methodological guide for updating CORINE Land
Cover database, Luxembourg.
22
Part II
Addendum to the CORINE land cover
nomenclature illustrated guide
23
1. Introduction
During more than 10 years of inventory of the CORINE Land Cover (CLC) classes an
enormous experience has been built up with regard to the visual interpretation of the
Landsat and SPOT satellite images. It has been found out that for most classes the initial
definition could be refined, in order to exclude potential confusions, however without
changing the core content of the class definition.
In the CLC Technical Guide (CEC, 1994) CLC classes are defined in such a way as to
summarise land cover of Europe at scale 1:100 000 with all its particularities. Perfect
knowledge of the CLC database requires inclusion of all particularities - features of land
cover of individual countries generalised in the uniform European CLC nomenclature.
Within the first CLC Technical Guide, published in 1994, each class of the nomenclature
was explained with a short definition. This addendum provides an enhancement to this
definition by adding a number of clarifications:
•= refinement or extension of the definitions to be more specific on class description;
•= characteristics of the class contents, by providing a list of cases which should be
included or excluded from the land cover class;
•= generalised patterns, helping interpretation of images;
•= representative or typical of class photographs of classes;
•= characteristics of class particularities (Phare countries).
24
2. Methodology used for the compilation
of the addendum
Visual interpretation is a method of identifying and assessing for objects recorded in aerial
or satellite images. This method was applied creating the CLC database. It is based on
analysis of interpretation elements, of the recorded landscape objects.
Further itemisation or up dating of CLC database will mean emphasis on analysis of the
relation ‘landscape object - its manifestation in satellite image’. Analysis of this relation
assumes a perfect knowledge of landscape objects, namely their characteristics, which
decisively influence reflection and emission capability of electromagnetic radiation.
However, emphasis must be laid also on analysis of interpretation elements, by means of
which the objects are represented in images. As far as the descriptions of the land cover
and its particularities are concerned, the following part of the addendum contains graphic
and text presentation of a sum of objects and the characteristics which are part of the CLC
classes at scale 1:100 000 from the viewpoint of land cover. This part of the addendum was
compiled based on image documentation and comments prepared by experts dealing with
the issue of land cover from the European Topic Centre on Land Cover (ETC/LC) and
Phare Topic Link on Land Cover (PTL/LC). Every land cover class is characterised (CEC,
1994) and its refinement or extension if necessary, a list of dominating objects comprised
in a particular class, typical arrangement of the objects by means of a pattern (Fig. 2 and 3)
the dominating textures (Fig. 1), a representative photograph, list of particularities and
photographs. The set of characteristics of CLC classes will help the interpreters in image
analysis, making use of interpretation elements like texture and pattern. It will also
contribute to further understanding of the landscape types of Europe, especially its spatial
features.
Fig. 1 Texture – is the area variability in the tone arrangement in the
satellite images, representing objects or groups of Earth's surface
objects that are too small to be discerned as individual features; the
same objects are represented in satellite images of the same type
and scale by the same pattern of tones, e.g. very small stripes, circles,
points, etc. (Feranec et al. 1995).
25
Fig. 2 Pattern – is the spatial arrangement of objects represented in the
satellite images by different textures (Feranec et al. 1995).
Fig. 3 Pattern types. Land cover cases are determined by significant
appearance characteristics of landscape objects, manifesting
themselves in the satellite images by means of typical patterns which
are formed at least by two textures (Feranec et al. 1995).
26
3. Characteristics of the CORINE land
cover classes
Class 1: Artificial areas
Class 1.1 Urban fabric
Areas mainly occupied by dwellings and buildings used by administrative/public utilities or
collectivities, including their connected areas (associated lands, approach road network,
parking-lots).
Class 1.2 Industrial, commercial and transport units
Areas mainly occupied by industrial activities of transformation and manufacturing, trade,
financial activities and services, transport infrastructures for road traffic and rail networks,
airport installations, river and sea port installations, including their associated lands and
access infrastructures. Includes industrial livestock rearing facilities.
Class 1.3 Mine, dump and construction sites
Artificial areas mainly occupied by extractive activities, construction sites, man-made waste
dump sites and their associated lands.
Class 1.4 Artificial non-agricultural vegetated areas
Areas voluntarily created for recreational use. Includes green or recreational and leisure
urban parks, sport and leisure facilities.
Specifications
In case of cultivated areas inter-mixed with built-up areas within a patchwork system, the
minimum threshold to be considered to classify in discontinuous urban fabric is 30 % (at
least 30 % of the small parcels are urban fabric). Otherwise the area should be classified as
complex cultivation patterns.
The figures below illustrate different urban fabric density, which can be used as a template
for distinguishing:
•= agricultural areas with scattered houses and discontinuous urban fabric,
•= discontinuous and continuous urban fabric.
27
1 % 2 % 5 %
10 % 15 % 20 %
25 %
35 %
40 %
50 % 70 %
Fig. 4 Illustration of different urban fabric density
28
111 Continuous urban fabric
Most of the land is covered by structures and the transport network.
Building, roads and artificially surfaced areas cover more than 80 % of the
total surface. Non-linear areas of vegetation and bare soil are exceptional.
Extension:
80 % of the total surface at least should be impermeable.
The continuous urban fabric class is assigned when the urban structures and transport
network (i.e. impermeable surfaces) occupies more than 80 % of the surface area. This
coverage percentage pertains to real ground surface. Therefore, localisation of this cut-offpoint
requires particular attention to avoid confusion with the apparent vegetation (i.e.
visible tree crowns) and permeable surfaces under trees. For example, in the streets
bordered with trees, the real ground surface under the trees is mostly covered with asphalt
or concrete. So, the vegetation percentage has to be estimated taking into account the
shape structure and context visible on the satellite image.
This heading includes:
•= urban centre types and dense ancient suburbs where buildings form a continuous and
homogeneous fabric;
•= public services or local governments and commercial/industrial activities with their
connected areas inside continuous urban fabric when their surface is less than 25 ha;
•= interstices of mineral areas;
•= parking lots, concrete or asphalt surfaces;
•= transport network;
•= small squares, pedestrian zones, yards;
•= urban greenery (parks and grass areas) amounting to 20 % of the polygon area;
•= unvegetated and vegetated cemeteries less than 25 ha located inside continuous urban
fabric.
= Texture of buildings
Texture of parking lots
Texture of urban greens
Texture of streets and roads
Fig. 5 A generalised pattern of the class 111
Fig. 6 Representative demonstration of the quoted class on examples of
Bratislava (Slovakia) and Athens (Greece)
29
Generalisation:
•= If two adjacent polygons of discontinuous and continuous urban fabric, each of them
less than 25 ha, but in total > 25 ha, are connected to each other, they should be
mapped as one polygon, and discontinuous urban fabric is privileged.
111
112
< 25 ha
< 25 ha
112
> 25 ha
No particularity was identified in this class.
112 Discontinuous urban fabric
Most of the land is covered by structures. Building, roads and artificially
surfaced areas associated with vegetated areas and bare soil, which occupy
discontinuous but significant surfaces.
Extension:
Between 30 to 80 % of the total surface should be impermeable.
The discrimination between continuous and discontinuous urban fabric is set from the
presence of vegetation visible in the satellite image illustrating either single houses with
gardens or scattered apartment blocks with green areas between them.
The density of houses is the main criteria to attribute a land cover class to the built-up
areas or to the agricultural area (242). In case of patchwork of small agricultural parcels
and scattered houses, the cut-off-point to be applied for discontinuous urban fabric is 30 %
at least of urban fabric within the patchwork area.
This heading includes:
•= private housing estates, residential suburbs made of individual houses with private
gardens and/or small squares;
•= scattered blocks of residential flats, hamlets, small villages where numerous
unmineralised interstitial spaces (gardens, lawns) can be distinguished;
•= large blocks of flats where green spaces, parking areas and adventure playgrounds
cover significant surface area;
•= transport network;
•= sport area smaller than 25 ha included within discontinuous urban fabric;
•= buildings with educational, health care and production functions and market places
smaller than 25 ha included within this class;
•= unvegetated and vegetated cemeteries smaller than 25 ha included within
discontinuous urban fabric;
•= public utilities/communities surfaced areas less than 25 ha;
•= holiday cottage houses are included in 112 if infrastructures like houses, road network
are visible in the satellite image; they must also be connected to built-up areas;
30
•= troglodyte villages along streets and subterranean housings visible from the satellite
image.
This heading excludes:
•= holiday cottage areas, which are only used for recreational purposes and defined as a
specific unit in the satellite image should be classified as 142;
•= installations structured with a view to summer settlement with bungalows and a specific
organisation (road, facilities) which must be classified as 142;
•= scattered main and secondary residences implanted in natural or agricultural areas
when their coverage is less than 30 % of the total surface; they are classified as 242 or
243;
•= greenhouses; they are classified as 211.
Texture of buildings
Texture of gardens
Texture of urban greens
Texture of streets and roads
Fig. 7 A generalised pattern of the class 112
Fig. 8 Representative demonstration of the quoted class on example of
village Liskova (Slovakia).
Generalisation:
•= Generalisation of discontinuous built-up areas located along roads: until 300 m to
maintain land cover feature of street villages.
•= For housing with large gardens, an arbitrary 100 m – buffer is added around the
houses and infrastructure to delimit the artificialised surfaces from the surrounding
area (often agricultural area).
d > 300 m
31
•= Small discontinuous urban fabric areas less than 25 ha are grouped together if the
distance between each of them is less than 300 m in order to reach 25 ha. The exterior
contour line leans on road network.
> 25 ha
112
112
112
112
< 25 ha
road
road
•= In case of free space surrounded by discontinuous urban fabric, the free space zone is
assigned 242 if its surface is more than 25 ha. On the contrary, it should be included in
discontinuous urban fabric.
d = +100 m
32
242
112
112
242
112
< 25 ha
> 25 ha
•= Extension of villages (112) with mixed fabric of settlement/industrial and commercial
activities along the roads need to be added in 112.
road
industrial unit
< 25 ha
242
112
112
242
•= Troglodyte villages along roads, with length over 2.500 m, are mapped as linear urban
construction by artificially delimiting a 100 m – linear buffer zone including the road
from the house frontage.
road
100m wid112
d<100m
No particularity was identified in this class.
Particularity of class 112: Housing estates
Areas of multiflat or multistoreyed houses forming built-up areas particularly
at the outskirts of urban settlements typical for their physiognomic
uniformity.
This heading includes:
•= multiflat-multistoreyed houses
•= squares, streets
•= transport network
•= parking lots
•= vegetation within housing estates (lawns, flower beds, shrub and tree formations)
33
Texture of buildings
Texture of urban greens
Texture of streets and roads
Fig. 9 A generalised pattern of the particularity of class 112
Fig. 10 Representative of the quoted class particularity on example of
housing estates of Petrzalka - Bratislava (Slovakia)
121 Industrial or commercial units
Artificially surfaced areas (with concrete, asphalt, tarmacadam, or stabilised,
e.g. beaten earth) without vegetation occupy most of the area, which also
contains buildings and/or vegetation.
This heading includes:
•= research and development establishments;
•= security and law and order services (fire stations, penal establishments);
•= company benefit schemes (old people's home, convalescent homes, orphanages, etc.);
•= stud farms, agricultural facilities (co-operatives, state farm centres, livestock farms,
living and exploitation buildings);
•= exposition sites, fair sites;
•= nuclear power plants, military barracks, testing pistes, test fields, biological waste water
treatment plants, water houses, transformers;
•= large shopping and exposition centres;
•= hospitals, spas,;
•= universities, schools;
•= parking lots;
•= abandoned industrial sites and by-products of industrial activities where buildings are
still present;
•= water retention dam and hydroelectric dam in total >25 ha;
•= telecommunication networks (relay stations for T.V., telescopes, radar stations).
This heading excludes:
•= extractive industry (class 131);
•= oil terminals inside port activities (class 123);
•= dumps, decanting basin structures (class 132);
•= dockyards (class 123);
•= merchant departments belonging to private or public services (class 11x);
•= places of worship: convents, monasteries, etc. (class 142).
34
Textures of industrial and commercial buildings
Texture of storing area and parking lots
Texture of roads
Texture of urban greens
Fig. 11 A generalised pattern of the class 121
Fig. 12 Representative demonstration of the quoted class on examples of
aluminium factory in Ziar nad Hronom (Slovakia) and electricity power
plant in France
•= Commercial/industrial units less than 25 ha which are connected to urban fabric
larger than 25 ha should be assigned as urban fabric.
•= Urban fabric units equal or larger than 25 ha inside commercial/industrial units
should be mapped.
•= Park areas and factory approaches are mapped as industrial units (class 121) even if a
road network crosses them.
road 122>100m
park area
factory 121
35
Particularity of class 121: Agricultural farms
Areas of other than housing buildings, in-door spaces, stables, garages,
workshops, lay-by and storing areas, often also bad land with ruderal
vegetation, part of farms. The farms are often located in outskirts or close to
rural settlements with agricultural function. Concentration of agricultural
buildings in areas of various sizes was associated with collectivisation of
agriculture. The quoted areas smaller than 25 ha are included in class 112.
This heading includes:
•= buildings, in-door spaces and yards for keeping farm animals;
•= garages, workshops, production buildings, lay-by areas for agricultural machinery;
•= paved and unpaved storing areas and warehouses;
•= bad land with ruderal vegetation.
Texture of industrial and commercial buildings
Texture of storing area
Texture of lay-by area
Texture of bad land with ruderal vegetation
Texture of scattered greenery
Fig. 13 A generalised pattern of the particularity of class 121
Fig. 14 Representative of the quoted class particularity on example of
agricultural farm in Liskova (Slovakia)
122 Road and rail networks and associated land
Motorways and railways, including associated installations (stations,
platforms, embankments). Minimum width for inclusion: 100 m.
This heading includes:
•= transport networks – roads, railways, funiculars, minimum width 100 m;
•= motorway rest areas, service stations, parking areas haulage depots connected on
motorway networks, services and maintenance activities for roads, tollbooths;
•= marshalling yards, perimeter of stations, services and maintenance activities for trains;
•= compounds of large crossroads with minimum area 25 ha;
•= tramways networks;
•= cableway networks.
36
This heading excludes:
•= motorways and high-speed train under construction (133);
•= closed-down transport network (classified under the real appropriate land cover class).
Texture of railway
Texture of stores and service buildings
Texture of station buildings
Texture of road
Texture of parking lot
Texture of linear greenery
Fig. 15 A generalised pattern of the class 122
Fig. 16 Representative demonstration of the quoted class on examples of
road and rail network in the Hron valley (Slovakia) and the crossroad
in Vilnius (Lithuania)
Generalisation:
•= The 100 m minimum width refers to the real linear coverage of the network including
vegetated trenches and/or embankment slopes. The minimum real length should be
2.500 m.
embankment
122
road
embankment or trenches
motorway
•= Disused transport infrastructures (ancient stations, disused railway lines and associated
land) are assigned according to their present land cover, which predominates in this
case against land use.
•= Railways have a higher priority than urban fabric does and roads have a lower priority
than urban fabric does.
37
112
112
112
112
railway road
factory
platform
park
area
122121122
•= In case of complex transport networks and associated lands, associated land should be
restricted to those areas, which are clearly isolated by roads or railways.
122
311<25ha
122
122
311
122
122
311
<25ha
>25ha
•=
122motorway
functioning basin haulage depots
No particularity was identified in this class.
38
123 Port areas
Infrastructure of port areas, including quays, dockyards and marinas.
This heading includes:
•= commercial and military ports;
•= shipyards;
•= fishing ports;
•= yachts ports, sport and recreation ports;
•= shipping and infrastructure port facilities;
•= sea, river and lake ports;
•= harbour stations, dock houses;
•= oil terminals;
•= roads, railways and parking lots within the port area;
•= adjacent water areas shirted by quays if the area of infrastructure of the port (firm land
part) is smaller than 25 ha
This heading excludes:
•= industrial and commercial units larger than 25 ha associated with port activities (class
121).
Texture of water surface
Texture of quays
Texture of port area infrastructure (buildings)
Texture of storing and loading areas
Fig. 17 A generalised pattern of the class 123
Fig. 18 Representative demonstration of the quoted class on examples of
Bratislava Port (Slovakia) and Klajpeda Port (Lithuania)
Generalisation:
•= Port areas for which the artificial surface is covering more than 25 ha do not include
docks or water surfaces larger than 25 ha.
•= Port areas for which the artificial surface is covering less than 25 ha and characterised
by the construction of two protection arms surrounding a water area resulting in a
total area larger than 25 ha, should be generalised as port areas.
39
•= Narrow strips of industrial units below the area threshold contiguous to the harbour
are included in the harbour area.
112
121 < 25 ha
123 123 > 25 ha
112
523
No particularity was identified in this class.
124 Airports
Airports installations: runways, buildings and associated land.
Extension:
This class includes associated lands (mainly grassland).
This heading includes:
•= take-off and landing runways (concrete, grass surfaced) of civil, military and sport
airports with non concreted or asphalted runways and with installations;
•= terminals, hangars, service and storing buildings and in-door spaces;
•= flying schools used for pilot’s training programme of civil aviation;
•= parking lots and lay-by areas;
•= adjacent grass areas, or dispersed trees and shrubs within the buffer zone of airport;
•= small sport airports with non-concreted or asphalted runways used for agriculture and
forestry (e.g. spreading of fertilizers and chemical materials).
This heading excludes:
•= small sport airports with non-concreted or asphalted runways;
•= disused airport or airfield should be classified as 321.
Texture of runways
Texture of service buildings
Texture of hangars and storing buildings
Texture of parking lot
Texture of associated lawns and scattered greenery
Fig. 19 A generalised pattern of the class 124
40
Fig. 20 Representative demonstration of the quoted class on example of
airport in Vilnius (Lithuania)
Generalisation:
•= All land cover features such as terminal buildings, runways, etc larger than 25 ha,
located inside the airport territory, should be generalised as airport area 124.
grassland
•= At least a buffer zone of 100 m should be delineated around the runways of an airport.
No particularity was identified in this class.
131 Mineral extraction sites
Areas with open-pit extraction of construction material (sandpits, quarries)
or other minerals (open-cast mines). Includes flooded gravel pits, except for
river-bed extraction.
Extension:
This class includes flooded gravel pits surface of which is less than 25 ha and temporary
mining pools.
This heading includes:
•= open-pit extraction often associated with heaps of extracted building material (gravel,
sand, stone or clays) or ore and non-ore mineral material (iron, manganese ores,
magnesite, lignite, brown coal, kaolin, etc.);
•= infrastructure of buildings and installations providing for extraction, or primary
processing of the quoted material and minerals;
•= transport networks associated with areas of open-pit extraction;
•= lay-by areas;
41
•= water bodies (smaller than 25 ha), often associated with open pit extraction of gravel,
sand, etc.;
•= rock salt pits;
•= sand extraction site inside coastal dune areas;
•= inland Salinas;
•= oil fields with wells;
•= petroleum, gas and liquid petroleum gas, shale oil extraction site.
This heading excludes:
•= exploited peat bogs (class 412);
•= associated land of mines where barren materials are dumped (coal tips, slag dumps)
(class 132);
•= coastal saline (class 422);
•= scree covered areas (class 332);
•= extraction sites abandoned and reconverted to leisure areas (class 142).
Textures of open-pit extraction mineral and building materials
Texture of water surface
Texture of lay-by area
Texture of service buildings
Fig. 21 A generalised pattern of the class 131
Fig. 22 Representative demonstration of class 131 on example of brown coal
extraction in Most (Czech Republic) and a quarry in Lithuania
Generalisation:
•= Flooded mineral extraction surfaces are included if < 25 ha. The flooded areas should
be isolated and assigned as water bodies (class 512) if > 25 ha. The water surfaces and
their visible extraction surrounds are connected together to create a single polygon
131 > 25 ha.
42
512 < 25
512 > 25
131 < 25
131 131
512
How to map mineral extraction sites associated to industrial areas if both polygons are less
than 25 ha?
•= In this case, the size of each polygon should be considered.
•= In general, priority will be given to 121 industrial area if both polygons cover
approximately the same area. Otherwise, the dominating area will be represented.
•= Disused gravel pits filled with water should be classified according to their current
actual land cover.
No particularity was identified in this class.
512 324
512 142
131
121
quarry < 25 ha
stone processes
factory < 25 ha
43
132 Dump sites
Public, industrial or mine dump sites.
Extension:
This class includes dump sites of raw materials or liquid wastes.
This heading includes:
•= dump sites of public, communal waste (landfills);
•= dump sites of industrial waste - waste rock after processing of various raw materials;
•= dump sites of waste material from stations cleaning the communal waste water;
•= pools of waste water/liquid waste, products of various chemical processes;
•= protecting dikes;
•= line vegetation belts, part of buffering/protective zones around the dump sites;
•= buildings, transport networks with parking lot associated with dump site;
•= slag heaps which are unvegetated.
This heading excludes:
•= decanting basins of biological water treatment plants by means of lagoonage
processing (class 121);
•= dump sites abandoned and reconverted to leisure areas (class 142);
•= vegetated slag heaps (class 3xx).
Texture of solid waste dump site
Texture of liquid waste dump
Texture of linear greenery
Fig. 23 A generalised pattern of the class 132
Fig. 24 Representative demonstration of the quoted class on examples of
dump sites of public and industrial waste in Slovakia
No particularity was identified in this class.
44
133 Construction sites
Spaces under construction development, soil or bedrock excavations,
earthworks.
This heading includes:
•= public and industrial fabric structures, road and rail networks, water dams/reservoirs,
etc. under construction.
This heading excludes:
•= completed parts of transport networks under construction when they are larger than
25 ha.
Texture of buildings under construction
Fig. 25 A generalised pattern of the class 133
Fig. 26 Representative demonstration of the quoted class on examples of
houses and technical objects under construction in Slovakia and
Lithuania
Generalisation:
If structures for crossing and/or junctions are already present and visible on the satellite
image along traces of transport networks under construction, then the following scheme
should be applied according to the size of structure elements.
junctions > 25 ha
under construction
under construction
junctions < 25 ha
122 122133
133
No particularity was identified in this class.
45
141 Green urban areas
Areas with vegetation within urban fabric, includes parks and cemeteries
with vegetation, and mansions and their grounds.
Extension:
This class includes cemeteries with important vegetation coverage.
Green urban areas concern all vegetated areas greater than 25 ha which are either situated
within or in contact with urban fabrics. Greenery with strips of lanes and paths may be
found within these areas created for recreational use.
This heading includes:
•= parks, park basins, lawns, flower beds in settlements;
•= ornamental gardens;
•= botanical and zoological gardens situated in settlement (urban fabric - 112) or in
contact-peripheral zone of settlement;
•= city squares;
•= inner spaces of city blocks;
•= cemeteries with vegetation in settlements;
•= vegetated areas which can be used for recreational purpose even if it is not their main
utilisation such as woods in urban fabric.
This heading excludes:
•= city gardens (class 242);
•= vegetated cemeteries outside urban fabric (class 142);
•= unvegetated cemeteries inside urban fabric (class 11x).
= Textures of urban greenery
=
= Texture of pavements
= Texture of recreation and service facility
Fig. 27 A generalised pattern of the class 141
Fig. 28 Representative demonstration of the quoted class on example of the
Freedom Square in Bratislava (Slovakia)
No particularity was identified in this class.
46
142 Sport and leisure facilities
Camping grounds, sports grounds, leisure parks, golf courses, racecourses,
etc. Includes formal parks not surrounded by urban areas.
This heading includes:
•= areas of sport compounds (football stadiums with the corresponding infrastructure,
hockey halls, swimming pools and tennis courts, cycling stadiums, athletic halls and
stadiums, etc.) within settlements and out of them;
•= sport shooting-ranges;
•= cemeteries with vegetation situated out of the settlements;
•= camping sites;
•= cottage (tourist) communities used for recreation and leisure activities outside the
settlements only for temporary residence;
•= zoological and botanical gardens out of settlements;
•= compounds of disclosed/open archaeological sites;
•= golf courses;
•= racecourses;
•= ski resorts (except ski pistes);
•= motor – racing circuit;
•= forest-parks in the periphery of settlements;
•= small sport airports with non concreted or asphalted runways.
This heading excludes:
•= motor-racing circuits inside industrial unit areas used for test purposes (class 121);
•= caravaning parking used for commercial activities (class 121);
•= beaches (class 331);
•= camping areas within forests that are not specially prepared for the purpose (class
31x);
•= stud farms (class 121).
Textures of sport and service facilities
Texture of water surface
Texture of racing areas
Texture of parking lot
Fig. 29 A generalised pattern of the class 142
Fig. 30 Representative demonstration of the quoted class on example of
sport area in Bratislava (Slovakia)
47
Generalisation:
•= In case of recreational/leisure strip areas surrounding lakes are less than 25 ha, the
water surface and leisure facility areas should be grouped together to reach a 142
polygon > 25 ha.
512 512
< 25 ha
142 < 25 ha
142 > 25 ha
No particularity was identified in this class.
Class 2: Agricultural areas
Class 2.1 Arable land
Lands under a rotation system used for annually harvested plants and fallow lands, which
are permanently or not irrigated. Includes flooded crops such as rice fields and other
inundated croplands.
Class 2.2 Permanent crops
All surfaces occupied by permanent crops, not under a rotation system. Includes ligneous
crops of standards cultures for fruit production such as extensive fruit orchards, olive
groves, chestnut groves, walnut groves shrub orchards such as vineyards and some specific
low-system orchard plantation, espaliers and climbers.
Class 2.3 Pastures
Lands, which are permanently used (at least 5 years) for fodder production. Includes
natural or sown herbaceous species, unimproved or lightly improved meadows and grazed
or mechanically harvested meadows.
Class 2.4 Heterogeneous agricultural areas
Areas of annual crops associated with permanent crops on the same parcel, annual crops
cultivated under forest trees, areas of annual crops, meadows and/or permanent crops
which are juxtaposed, landscapes in which crops and pastures are intimately mixed with
natural vegetation or natural areas.
48
Fig. 31: Decision scheme for agricultural areas in relation to semi-natural areas
211 Non-irrigated arable land
Cereals, legumes, fodder crops, root crops and fallow land. Includes flowers
and fruit trees (nurseries cultivation) and vegetables, whether open field,
under plastic or glass (includes market gardening). Includes aromatic,
medicinal and culinary plants. Does not include permanent pastures.
Extension:
This class includes flower, fruit trees (nurseries) and vegetable cultivation. Includes other
annually harvested plants with more than 75 % of the area under a rotation system. Part of
this class are the plots of arable land with area of several hectares reaching tens (hundreds)
of ha.
This heading includes:
•= multi–year plants as asparagus and chicory;
•= flooded crops as water cross beds;
•= semi-permanent crops as strawberries;
•= temporary fallow lands (lands under three yearly rotation system);
•= drained arable land should be mapped as 211 instead of 212;
•= fragmented agricultural land use resulting in juxtaposition of different annual crops;
•= weeded crops;
49
•= non-permanent industrial crops as textile plants, oleaginous plants;
•= tobacco;
•= condiment plants;
•= sugar cane;
•= flowers under a rotation system;
•= industrial flower crops as lavender species;
•= nurseries-garden (seedlings of fruit trees and shrubs);
•= dispersed, mostly line vegetation;
•= abandoned irrigated arable land even the irrigation channel network is still visible in
the satellite image.
This heading excludes:
•= city gardens (class 242);
•= lands which lie fallow for at least three years (class 231 or 32x);
•= hop plantations (class 222);
•= rice field (class 213);
•= forest tree nurseries with non-commercial purposes located in forest areas (31);
•= fruit trees and berry plantation under glass greenhouses (class 222);
•= osier trees for wicker production (class 222);
•= permanent plantations of roses (class 222);
•= wine-growing nurseries (class 221).
Textures of agricultural annual crops
Texture of glass and plastic greenhouses
Texture of scattered greenery
Fig. 32 A generalised pattern of the class 211
Fig. 33 Representative demonstration of the quoted class on example of
arable land in Western part of Slovakia
50
Particularity of class 211: Abandoned (fallow) land
Areas of arable land not used for 1-3 years. Identification of these areas
requires data obtained by field research as well as topical statistic data on
area of arable land in particular part of the country.
This heading includes:
•= abandoned agricultural land;
•= arable land, where the above quoted agricultural crops are cultivated;
•= scattered (especially line) greenery.
Textures of agricultural crops
Texture of abandoned agricultural land
Fig. 34 A generalised pattern of the particularity of class 211
Fig. 35 Representative of the quoted class particularity on example of
abandoned land in Latvia
212 Permanently irrigated land
Crops irrigated permanently or periodically, using a permanent
infrastructure (irrigation channels, drainage network). Most of these crops
cannot be cultivated without an artificial water supply. Does not include
sporadically irrigated land.
Extension:
This class excludes drainage network areas, which are assigned to 211, 231 or 242, applied
for pumping infrastructure and irrigation systems from superficial water supplies.
This heading includes:
•= recently abandoned irrigation systems; decision must be taken based on the satellite
image spectral reflectance showing if soils are still wet and the infrastructure;
•= sown grassland (as part of crop rotation) if the irrigation infrastructure is permanently
present.
51
This heading excludes:
•= drainage network intended to clean up wet soils (classes 211, 231 or 242);
•= crops under greenhouses (classes 211 or 222);
•= underground irrigation pipes and above ground pipes and furrows (other
cultivation classes);
•= spray sprinkler line (other cultivation classes);
•= rotary sprinkler (other cultivation classes);
•= rice fields (class 213).
Texture of irrigation channels
Textures of agricultural crops
Fig. 36 A generalised pattern of the class 212
Fig. 37 Representative demonstration of the permanently irrigated land on
example from Greece
No particularity was identified in this class.
213 Rice fields
Land prepared for rice cultivation. Flat surfaces with irrigation channels.
Surfaces periodically flooded.
Extension:
Abandoned rice fields are not included. One or two yearly rotation is applied for rice
fields, therefore the land cover is mapped according to the presence at the time of satellite
data acquisition.
This heading includes:
•= rice fields;
•= irrigation channels.
This heading excludes:
•= ancient rice fields with irrigation channels should be mapped according to their actual
land cover (mainly classes 211 or 231);
52
•= abandoned rice fields (class 2xx).
Texture of irrigation channels
Texture of rice fields
Fig. 38 A generalised pattern of the class 213
Fig. 39 Representative demonstration of the rice field on example from
Portugal
No particularity was identified in this class.
221 Vineyards
Areas planted with vines.
Extension:
Vineyard areas are classified as 221 if the vineyard parcels exceed 50 % of the area and/or
they determine the land use of the area.
This heading includes:
•= vine-growing nurseries inside vineyard areas;
•= vineyards for wine production;
•= vineyards for consumer grapes and raisins;
•= complex cultivation pattern areas where vineyards parcels cover at least 50 % of the
area.
This heading excludes:
•= vines mixed with arable land and/or meadows within a single parcel (class 241);
•= vines (single parcels (25 ha)) mixed with arable land and/or meadows interspersed
with significant natural vineyard parcels covering less than 40 % of the area (class 243).
53
Texture of vineyards
Texture of orchard
Texture of agricultural crop
Texture of scattered greenery
Fig. 40 A generalised pattern of the class 221
Fig. 41 Representative demonstration of the quoted class on example of
vineyards in surroundings of Modra (Slovakia)
Generalisation:
How to map vines associated with fruit trees within a single parcel?
•= In this case, the dominance of each permanent crop should be considered. In general,
priority will be given to vineyard if dominances are approximately the same. Otherwise
the dominating permanent crop will be represented.
fruit trees vines
vine cover = 50 % 221
vine cover > 50 %
vine cover < 50 %
222
221
•= In case of vines associated to olive trees within a single parcel, priority will be given to
class 221.
No particularity was identified in this class.
54
222 Fruit trees and berry plantations
Parcels planted with fruit trees or shrubs: single or mixed fruit species, fruit
trees associated with permanently grassed surfaces. Includes chestnut and
walnut groves.
Extension:
This class includes ligneous crops and chestnut and walnut tree orchards intended for fruit
production.
This heading includes:
•= hop plantations;
•= plantations of berry shrubs, black and/or red currants, raspberries, gooseberries,
blackberry crops;
•= willow plantation for wicker production;
•= fruit trees under greenhouses;
•= abandoned orchards which still preserve characteristic alignments;
•= fruit, orchards of apples, pears, plums, apricots, peaches, cherries, figs, quinces and
other rosaceae;
•= ligneous crops : chestnut, walnut, almond, hazel, pistachio groves;
•= permanent florist plantations of roses;
•= plantation of vines associated to fruit trees within the same parcel where vines cover at
least 40 % of the surface;
•= tropical fruit trees : avocados, bananas, guavas, mango, kiwis, passion fruits, papayas,
pineapples, pomegranates, brazil nuts, cashew nuts, coconuts, nutmegs;
•= citrus fruit trees : oranges, lemons, mandarins, tangerines, grape fruits, pomelos;
•= fruit tree greeneries inside fruit tree plantations;
•= permanent industrial plants : coffee, cacao, mulberry, tea;
•= recently abandoned orchards where characteristic plantation structures (espaliers and
climbers) are still visible;
•= scattered greenery.
This heading excludes:
•= strawberries (class 211);
•= olive groves (class 223);
•= vineyard (class 221);
•= fruit trees nurseries (class 211);
•= carob trees (class 311);
•= chestnut and walnut grove forests intended for wood production (class 311);
•= abandoned orchards where plantation structures have disappeared (class 324);
•= orchards located in permanently irrigated lands (class 212);
•= multi–year plants as asparagus (class 211).
Texture of fruit trees
Texture of berry plantations
Texture of scattered greenery
Fig. 42 A generalised pattern of the class 222
55
Fig. 43 Representative demonstration of the quoted class on example of
apple orchard in surroundings of Cachtice (Slovakia)
Generalisation:
When fruit trees are associated to vines on the same parcel, the dominance of each
permanent crop should be considered.
•= In case of equal density, priority will be given to vineyard (class 221). The
generalisation scheme adopted for class 221 should be also applied for class 222.
•= In case of fruit trees associated to olive trees on the same parcel, the dominance of
each permanent crop should be evaluated and priority will be given to class 223 when
no dominance is visible.
fruit trees olive trees
fruit trees cover
= 50 %
223
fruit trees cover
< 50 %
fruit trees cover
>50 %
222
223
Particularity of class 222: Hop plantations
Areas of arable land where hop is cultivated requiring especially tall
supporting construction.
This heading includes:
•= hop-garden plots.
Texture of hops
56
Fig. 44 A generalised pattern of the particularity of class 222.
Fig. 45 Representative of the quoted class particularity on example of hop
plantations in the Western part of Slovakia
223 Olive groves
Areas planted with olive trees, including mixed occurrence of olive trees and
vines on the same parcel.
Extension:
This class includes Mediterranean plantations of Olea europaea ssp. europaea.
This heading includes:
•= olive groves shading herbaceous layer.
This heading excludes:
•= olive trees (Olea europaea ssp. sylvestris) as part of evergreen forest areas (class 311);
•= wild olive trees (Oleaster spp.) as part of sclerophyllous vegetation areas (class 323);
•= abandoned olive groves (class 323).
Texture of olive trees
Texture of vines
Texture of scattered greenery
Fig. 46 A generalised pattern of the class 223
57
Fig. 47 Representative demonstration of the olive groves on example from
Greece
No particularity was identified in this class.
231 Pastures
Dense grass cover, of floral composition, dominated by graminacea,
not under a rotation system. Mainly for grazing, but the fodder may
be harvested mechanically. Includes areas with hedges (bocage).
Extension:
Grazing used by cattle.
Pastures can be described as extensively used grasslands with presence of farm structure
such as: fences, shelters, enclosures, watering places, drinking trough, or regular
agricultural works: mowing, drainage, hay making, agricultural practices, manuring.
This heading includes:
•= temporary and artificial pastures not under a rotation system which become
permanent grasslands five years after ploughing. Significant number of natural
vegetation species are present (as Taraxacum officinale, Ranunculus spp.,
Chrisanthemum leucantemum, Knautia arvensis, Achillea millefolium, Salvia spp.,
etc.);
•= abandoned arable land not under a rotation system used as pastures (after 3 years);
•= pastures may include patches of arable land which do not cover 25 % of the total
surface;
•= humid meadows with dominating grass cover. Sedges, rushes, thistles, nettles, cover
less than 25 % of the parcel surface;
•= scattered trees and shrubs (10–20% of surface).
This heading excludes:
•= military exercising grass fields (without grazing) (class 321);
•= salt meadow located in intertidal flat areas (class 423);
•= lawns inside sport and leisure facility areas (class 142);
•= high-productive natural alpine meadows far from houses and/or crops (class 321);
•= fodder crops (class 211);
58
•= derelict grassland where semi-ligneous/ligneous vegetation cover at least 25 % of the
parcel (class 322/324);
•= strong humid meadows where hygrophyle plant species cover at least 25 % of the
parcel (class 411);
•= herbaceous grass cover composed of non-palatable and undesirable species for cattle
as Molinia spp. and Brachypodium spp. (class 321).
Texture of grassland
Texture of scattered trees and shrubs (10-20% of surface)
Fig. 48 A generalised pattern of the class 231
Fig. 49 Representative demonstration of the quoted class on example of
grassland in Central part of Slovakia
Particularity of class 231: Pastures on abandoned land
Grassland developed by not using arable land for more than three years.
Identification of the quoted grassland requires application of data obtained
by field checking.
This heading includes:
•= areas of grassland representing succession of natural overgrowing of arable land by
prevailingly herbaceous vegetation;
•= areas of sporadically occurring shrubs.
Textures of abandoned arable land (more than 3 years)
Texture of scattered shrubs
Fig. 50 A generalised pattern of the particularity of class 231
59
Fig. 51 Representative of the quoted class particularity on example of
pasture on abandoned land in Latvia
Particularity of class 231: Wooded meadows
Meadows where dispersed trees and shrubs occupy up to 50% of surface of
the area. These meadows are characterised by rich floristic composition.
This heading includes:
•= areas of grassland, partially covered by tree crowns;
•= areas of scattered trees and shrubs.
Texture of grassland
Texture of scattered trees and shrubs (40-50% of surface)
Fig. 52 A generalised pattern of the particularity of class 231
Fig. 53 Representative of the quoted class particularity on example of
wooded meadow in Virtsu-Laelatu (Estonia)
60
241 Annual crops associated with permanent crops
Non-permanent crops (arable land or pasture) associated with permanent
crops on the same parcel.
Extension:
Permanent crops are either in juxtaposition with arable land/pastures or located along the
border of the parcels. The occupation rate of non-permanent crops is more than 50 %.
This heading includes:
•= non-permanent crop areas in which they are shaded by a fairly closed canopy of fruit
trees or olive trees or vines;
•= non-permanent crop areas which are bordered by a reticulated structure of fruit tree
lines, vine lines;
•= some parcels of permanent crops more or less irregular with annual crops/pastures
less than 25 ha and inserted into a dominating non-permanent crop whole where none
of these crops represents more than 75 %.
This heading excludes:
•= permanent crops associated with fruit trees (classes 22x);
•= non-permanent crops associated with forest trees (class 244);
•= natural grasslands shaded by permanent crops (class 324);
•= pastures planted with trees (class 231).
Texture of permanent agricultural crops
Textures of annual agricultural crops
Texture of scattered greenery
Fig. 54 A generalised pattern of the class 241
Fig. 55 Representative demonstration of the olive trees associated with
arable lands on example from Portugal
61
Generalisation:
According to the bio-climatic zone, the heading could be described under two illustrations:
•= reticulated landscapes.
211
211
211
211231
231
fruit trees vines
vines
non permanent crop
fruit trees
241
•= small plot and orchard patterns.
211
211
211
211
211231
231 231211 211
grassland
fruit trees
f it t
241
No particularity was identified in this class.
242 Complex cultivation patterns
Juxtaposition of small parcels of diverse annual crops, pasture and/or
permanent crops.
Extension:
This class includes juxtaposition of small parcels of annual crops, city garden pastures,
fallow land and/or permanent crops eventually with scattered houses or gardens.
This heading includes:
•= mixed parcels of permanent crops (fruit trees, berry plantations, vineyards and olive
groves);
•= interstices of non-mineralised free spaces in discontinuous urban fabric < 25 ha;
•= complex cultivation pattern areas with scattered houses inserted within a patchwork
structure when built-up parcels cover less than 30 % of the patchwork area;
•= summer settlement areas if infrastructure/road network is not visible;
•= hobby/city gardens;
•= parcels of grassland.
This heading excludes:
•= market gardening (class 211);
•= nurseries cultivation (class 211);
•= in spite of strong fragmentation, the areas with more than 75 % of area under rotation
system (class 211);
•= complex cultivation patterns areas with scattered houses when they occupy more than
30 % of the patchwork area (class 112).
Textures of annual crops
62
Textures of permanent crops
Texture of grassland
Fig. 56 A generalised pattern of the class 242
Fig. 57 Representative demonstration of the quoted class on example of
complex cultivation patterns in Central part of Slovakia
Particularity of class 242: Complex cultivation patterns with scattered houses
Alternation of small plots (smaller than 25 ha) of arable land with annual or
permanent crops with scattered garden huts or scattered houses. They are
usually situated in proximity of rural or urban settlements and are used for
growing agricultural crops, fruit, and vegetable for the particular
households.
This heading includes:
•= parcels of arable land smaller than 25 ha with one-year agricultural crops and various
kinds of vegetables;
•= parcels of permanent crops (smaller than 25 ha), fruit trees, vineyards, berry
plantations;
•= parcels of grassland;
•= garden huts and scattered, sporadically occurring houses.
Textures of agricultural annual crops and vegetables
Texture of permanent crops
Texture of grassland
Texture of garden huts or scattered houses
Fig. 58 A generalised pattern of the particularity of class 242
63
Fig. 59 Representative of the quoted class particularity on example of
complex cultivation patterns with scattered garden huts in the
Western part of Slovakia
243 Land principally occupied by agriculture, with significant areas of natural
vegetation
Areas principally occupied by agriculture, interspersed with significant
natural areas.
Extension:
This class includes land occupied by agriculture with areas of natural or semi-natural origin
(including wetlands and water bodies, out crops).
This heading includes:
•= parcels of arable land (smaller than 25 ha);
•= parcels of orchards, vineyards and berry plantations (smaller than 25 ha);
•= parcels of the rests of natural forests, groups of trees and shrubs (smaller than 25 ha);
•= small areas of water bodies;
•= sporadically occurring houses of rural settlement, or farm buildings
•= linear structures of trees organised for truffle production;
•= hortillonage (vegetable crops and canals);
•= agriculture and scattered heaps of stones.
This heading excludes:
•= agricultural land associated with small plots of fruit trees/olive groves without natural
vegetation (class 242);
•= small islands of 243 made by mapping the forest units < 25 ha with a buffer of
agricultural land to reach units > 25 ha;
•= areas in which the share of agricultural areas is above 75 % (classes 21x, 22x or 23x);
•= areas in which semi-natural areas predominate (more than 75 %) (classes 3xx).
Textures of annual agricultural crops
Texture of orchards
Texture of grassland
Texture of scattered greenery
Texture of water surfaces
Fig. 60 A generalised pattern of the class 243
64
Fig. 61 Representative demonstration of the quoted class on example of
arable land interspersed with natural vegetation in the central part of
Slovakia
Generalisation:
•= Class 2.4.3 should be used for example in case of small agricultural units located within
forest areas.
311 311
243
class 2xx (each < 25 ha)
> 25 ha
No particularity was identified in this class.
25 % < 311 < 75 %
25 % < 211 < 75 %
d > 300 m
d < 100 m
< 25 ha
65
244 Agro-forestry areas
Annual crops or grazing land under the wooded cover of forestry species.
Extension:
This class includes annual crops or grazing land and fallow land covering less than 50 % of
the surface.
This heading includes:
•= areas of forest trees imbricated with fruit trees/olive trees while neither of the two
kinds of trees dominates;
•= carob trees shading agricultural lands;
•= agricultural land shaded by palm trees in Mediterranean context.
Texture of forestry species
Textures of annual or permanent agricultural crops
Fig. 62 A generalised pattern of the class 244
Fig. 63 Representative demonstration of the agro–forestry area on example
from Portugal
No particularity was identified in this class.
66
Class 3: Forest and semi-natural areas
Class 3.1 Forests
Areas occupied by forests and woodlands with a vegetation pattern composed of native or
exotic coniferous and/or deciduous trees and which can be used for the production of
timber or other forest products. The forest trees are under normal climatic conditions
higher than 5 m with a canopy closure of 30 % at least. In case of young plantation, the
minimum cut-off-point is 500 subjects by ha.
The 30 % minimum threshold to be considered can be illustrated by the three following
figures.
Figs. 65 A and C correspond to random distribution of the coverage and Fig. B illustrates a
regular distribution.
30 %
A B C
Fig. 64 Random distribution of the forest trees coverage (A, C) and regular
distribution (B).
Class 3.2 Shrubs and/or herbaceous vegetation associations
•= Temperate shrubby areas with Atlantic and alpine heaths, sub Alpine bush and tall
herb communities, deciduous forest re-colonisation, hedgerows, dwarf conifers.
•= Mediterranean and sub-Mediterranean evergreen sclerophyllous bush and scrub
(maquis, garrigue, mattoral, phrygana sensu lato), re-colonisation and degradation
stages of broad-leaved evergreen forests.
•= Dry thermophilous grasslands of the lowlands, hills and mountain zone. Poor Atlantic
a sub-Atlantic mat-grasslands of acid soils; grasslands of decalcified sands; Alpine and
sub Alpine grasslands. Humid grasslands and tall herb communities; lowland and
mountain mesophile pastures and hay meadows.
Class 3.3 Open spaces with little or no vegetation
Natural areas covered with little or no vegetation, including open thermophile formations
of sandy or rocky grounds distributed on calcareous or siliceous soils frequently disturbed
by erosion, steppic grasslands, perennial steppe-like grasslands, meso- and thermoMediterranean
xerophile, mostly open, short-grass perennial grasslands, alpha steppes,
vegetated or sparsely vegetated areas of stones on steep slopes, screes, cliffs, rock fares,
limestone pavements with plant communities colonising their tracks, perpetual snow and
ice, in land sand-dune, coastal sand-dunes and burnt areas.
67
311 Broad-leaved forest
Vegetation formation composed principally of trees, including shrub and
bush understoreys, where broad-leaved species predominate.
Extension:
This class includes areas with a crown cover of more than 30 % or a 500 subjects/ha
density for plantation structure, broad-leaved trees represent more than 75 % of the
planting pattern. In case of young plants or seedlings the proportion of broad-leaved
plants to be considered is at least 75 % of the total amount of plants.
This heading includes:
•= plantations of eucalyptus;
•= young plantations of deciduous trees;
•= walnut trees and chestnut trees used for wood production included into forest area
context;
•= sparse broad-leaved forests with a 30 - 60 % bracket of crown cover;
•= evergreen broad-leaved woodlands composed of sclerophyllous trees (mainly Quercus
Ilex, Quercus Suber, Quercus Rotondifolia);
•= arborescent mattoral with sclerophyllous species;
•= olive-carob forests dominated by Olea europaea sp. sylvestris, Ceratonia siliqua;
•= palm groves woodlands (one single case found in Greece);
•= holly woods dominated by Ilex aquifolium;
•= tamarix woodlands;
•= broad-leaved wooded dunes;
•= transitional woodland areas when the canopy closure of the trees cover more than 50
% of the area and if their average breast diameter is at least 10 cm;
•= denuded spots and grassland;
•= clear–cuts (applied for European Union countries)**.
This heading excludes:
•= burnt areas inside forest areas (classes 32x or 334);
•= non-evergreen coniferous trees dominated by larix species (class 312);
•= woodland areas composed of broad-leaved trees smaller than 5 m high (class 322);
•= vegetated areas where the crown cover of the broad-leaved trees is less than 25 % (class
324);
•= forest nurseries specialised in reproduction situated inside broad-leaved wooded areas
(class 324);**
•= clear-cuts (class 324, applied for Phare countries)**;
•= forest nurseries outside forests for commercial purpose (class 211);
•= wooded parks (class 141).
** See the general remarks concerning heading 31x.
Texture of broad-leaved forest
Texture of shrubs
Texture of grassland and area without trees
Fig. 65 A generalised pattern of the class 311
68
Fig. 66 Representative demonstration of the quoted class on example of
broad-leaved forest in Rouge (Estonia)
No particularity was identified in this class.
312 Coniferous forest
Vegetation formation composed principaly of trees, including shrub and
bush understoreys, where coniferous species predominate.
Extension:
Coniferous trees represent more than 75 % of the formation. In case of young plants or
seedlings, the proportion of coniferous plants to be considered is at least 75 % of the total
amount of plants and their texture is very similar to a surrounding coniferous forest
texture.
This heading includes:
•= non-evergreen coniferous trees woodland composed of larch trees (Larix spp.);
•= young plantations of coniferous trees;
•= coniferous wooded dunes;
•= arborescent mattoral with dominating Juniperus oxycedrus/phoenica;
•= coniferous wooded land;
•= Christmas tree plantations;
•= denuded spots and grassland;
•= clear-cuts (applied for European Union countries)**.
This heading excludes:
•= dwarf coniferous trees as Pinus mugo (class 322);
•= sclerophyllous trees (class 311);
•= vegetated areas where the crown cover of coniferous trees is less than 30 % (class 324,
231, 321);
•= forest nurseries specialised in reproduction situated inside coniferous wooded areas
(class 324);**
•= clear-cuts (class 324, applied for Phare countries).**
** See the general remarks concerning heading 31x.
Texture of coniferous forest
Texture of grassland and area without trees
Texture of shrubs
Fig. 67 A generalised pattern of the class 312
69
Fig. 68 Representative demonstration of the quoted class on examples of
coniferous forest in Taheva (Estonia) and the High Tatras (Slovakia)
No particularity was identified in this class.
313 Mixed forest
Vegetation formation composed principally of trees, including shrub and
bush understoreys, where neither broad-leaved nor coniferous species
predominate.
Extension:
Mixed forests with a crown cover of more than 30 % or a 500 subjects/ha density for
plantation structure. The share of coniferous or broad-leaved species does not exceed 25
% in the canopy closure.
This heading includes:
•= mixed-forest wooded dunes;
•= denuded spots and grassland;
•= sporadically occurying shrub formations;
•= clear–cuts (applied for European Union countries)**.
This heading excludes:
•= young plantations (class 324)**;
•= forest nurseries specialised in reproduction situated inside mixed-forest areas (class
324)**;
•= clear-cuts (class 324, applied for Phare countries)**;
•= burnt areas inside mixed-forest areas (class 3.3.4);
•= woodlands with mixed species trees smaller than 5 m high (class 3.2.2);
•= vegetated areas where the crown cover of mixed species trees is less than 30 %
(class 3.2.4, 231, 321).
** See the general remarks concerning heading 31x.
Texture of broad-leaved trees
Texture of coniferous trees
Texture of grassland and area without trees
70
Fig. 69 A generalised pattern of the class 313
Fig. 70 Representative demonstration of the quoted class on examples of
mixed forest (created by alternation of single trees and stands of
trees) in Stiavnicke vrchy (Slovakia)
Generalisation:
•= Mixed-forest area is formed by alternation of plots or single trees of broad-leaved and
coniferous trees.
311
312
< ha25 311
313
312
311
<25 ha 312
313
311
313
312
single trees
311
71
312
313
311
312
single trees
No particularity was identified in this class.
321 Natural grassland
Low productivity grassland. Often situated in areas of rough, uneven
ground. Frequently includes rocky areas, briars and heathland.
Extension:
Natural grasslands are areas with herbaceous vegetation (maximum height is 150 cm and
gramineous species are prevailing) which cover at least 75 % of the surface covered by
vegetation which developed under a minimum human interference (not mowed, fertilized
or stimulated by chemicals which might influence production of biomass); here belong for
instance grass formations of protected areas, karstic areas, military training fields, etc.
(even though the human interference cannot be altogether discarded in quoted areas, it
does not suppress the natural development or species composition of the meadows), areas
of shrub formations of scattered trees.
This heading includes:
•= saline grasslands grown on temporary wet areas of saline soils;
•= humid meadows where sedges, rushes, thistles, nettles cover more than 25 % of the
parcel;
•= natural grasslands with trees and shrubs if they do not cover more than 25 % of the
surface to be considered;
•= high-productive Alpine grasslands far from houses, crops and farming activities;
•= herbaceous military training areas;
•= grasslands which can be grazed, never sown and not otherwise managed by way of
application of fertilizers, pesticides, drainage or reseeding except by burning;
•= grasslands with a yearly productivity less than 1.500 units of fodder/ha;
•= herbaceous grass covered composed of non-palatable gramineous species such as
Molinia spp.and Brachypodium spp.;
•= derelict natural grassland where ligneous vegetation cover less than 75 % of the area;
•= grasslands found on calcareous soils with a high proportion of calcicole species of
limestone, chalk Machair or Karst;
•= grasslands dotted with bare rock areas which represent less than 25 % of the surface.
This heading excludes:
•= grey dunes (class 331);
•= swampy grassland (class 411);
•= fallow land (class 211).
72
Texture of natural grassland
Texture of scattered trees and shrubs
Fig. 71 A generalised pattern of the class 321
Fig. 72 Representative demonstration of the quoted class on example of
natural grassland in the Morava floodplain (Slovakia)
Generalisation:
•= At high altitude, class 321 might be present as altitude formation between heathlands
(322) or class 31x and sparsely vegetated areas (333).
333
322
321
312
312
322
321
333
altitude
73
Particularity of class 321: Alpine meadows
Grass formations which occur in high mountains above the upper timberline.
The most extensive areas of this particularity as far as the Phare countries
are concerned, are in the mountains of the Alps, the Carpathian Arch, etc.
This heading includes:
•= natural grassland;
•= rocky formations;
•= dwarf pines.
Texture of natural grassland
Texture of rocky formations
Texture of dwarf pines
Fig. 73 A generalised pattern of the particularity of class 321
Fig. 74 Representative of the quoted class particularity on example of alpine
meadow in the High Tatras (Slovakia)
Particularity of class 321: Grass formations of innundated alluvial plains and
coastal plains, lowlands
Human influence is very low with regard to natural conditions – of grass
formations - extreme soil humidity and seasonal inundated.
This heading includes:
•= natural grassland;
•= water bodies;
•= shrub formations and scattered trees.
74
Texture of natural grassland
Texture of water surfaces
Texture of scattered trees and shrubs
Fig. 75 A generalised pattern of the particularity of class 321
Fig. 76 Representative of the quoted class particularity on examples of lower
coastal grassland and upper coastal grassland (Rumpo and Vaemla,
Estonia)
322 Moors and heathland
Vegetation with low and closed cover, dominated by bushes, shrubs and
herbaceous plants (heather, briars, broom, gorse, laburnum, etc.).
Extension:
This class includes temperate shrubby area vegetation (climax stage of development):
includes dwarf forest trees with a 3 m maximum height in climax stage.
This heading includes:
•= wet heath distributed on humid or semi-peaty soils (peat depth < 30 cm) with Erica
tetralix/ciliaris, Sphagnum spp. and Molinia spp.;
•= Pinus mugo coverage above the upper tree limit in the Alpine zone or in the bottom of
large depressions with temperature inversion;
•= maritime, prostrate, wind-swept and cushiony heaths with maritime ecotypes;
•= heath and scrub formation in Atlantic, sub-Atlantic and sub-continental areas with
gorse (Ulex spp.), vaccinium heaths (Calluna vulgaris, Vaccinium spp.), heather (Erica
spp.), bracken or gorse (Genista spp.), bilberry heaths (Vaccinium myrtillus), briar patch
(Rubus spp.);
•= moors in supra-Mediterranean area with box trees and gorse, hedgehog-heaths (Buxus
spp., Astragalus spp., Bupleurum spp., etc.);
•= sub Alpine tall herbs with dominating bushy facies (Calluna spp., Vaccinium spp., Rubus
spp., Juniperus nana, etc.);
•= arctic moors areas with moss, lichen, gramineous coverage and small dwarf or
prostrate shrub formations (Betula nana, Salix lapponum, Salix glauca, Juniperus alpina,
Dryas spp.);
•= thickets and brush woods in temperate climate areas (box, bramble thickets, broom
fields, gorse thickets, braken fields, common juniper-scrubs);
•= brush woods and bush-like forest in Alpine area with dwarf mountain pine scrub or
green alder scrub (Pinus mugo ssp. mughus and Alnus spp.) Alpine willow brush, etc.,
accompanied by Rhododendron spp.;
75
•= thickets and bush-like forest in arctic area with Betula nana and Salix lapponum/glauca
spp.;
•= abandoned crops where ligneous/semi-ligneous species cover more of 25 % of the
surface;
•= coastal dunes (so-called brown dunes) covered and fixed with shrubs (Hippophae spp.,
Empetrum spp., Salix spp.);
•= herbaceous coverage formations mainly composed of non-palatable gramineous
species such as Molinia spp., Brachypodium spp., etc.
This heading excludes:
•= low maquis/mattoral vegetation (class 323);
•= heathland under recolonizing process where tree-like species cover more than 30 % of
the surface (class 324).
Texture of bushes, shrubs and herbaceous plants
Texture of grassland and area without shrubs
Fig. 77 A generalised pattern of the class 322
Fig. 78 Representative demonstration of the quoted class on example of
heather moor (Määvli, Estonia) and alvar with juniper (Rumpo,
Estonia)
Generalisation:
•= At high altitude class 322 should be used as altitude range formation according to the
vegetation gradient between classes 321 and 324/31x.
321
324
322
312
312
324
322
321
altitude
76
Particularity of class class 322: Dwarf mountain pine scrub (climax stage of development)
These shrub formations are 2-2.5 m high and form a compact canopy. Dwarf
mountain pine scrub is found on large areas in the mountains ranges of the
Alps, the Carpathian Arch, etc. It is also often artificially planted, for instance
in the coastal dunes as protection against deflation (for instance, in
Lithuania). In case of shrub vegetation areas composed of sclerophyllous
species such as Juniperus oxycedrus and heathland species such as Buxus
spp. or Ostrya carpinifolia with no visible dominance (each species occupy
about 50% of the area), priority will be given to sclerophyllous vegetation
and the whole area will be assigned class 323.
This heading includes:
•= dwarf mountain pine scrub (Pinus mugo spp. mughus);
•= rocky formations;
•= sporadic areas of grassland;
•= sporadic tree enclaves.
Texture of dwarf pines
Texture of rock formations
Texture of natural grassland
Texture of scattered trees
Fig. 79 A generalised pattern of the particularity of class 322
Fig. 80 Representative of the quoted class particularity on example of dwarf
mountain pine shrubs (High Tatras, Slovakia)
77
323 Sclerophylous vegetation
Bushy sclerophyllous vegetation, includes maquis and garrigue. In case of
shrub vegetation areas composed of sclerophyllous species such as
Juniperus oxycedrus and heathland species such as Buxus spp. or Ostrya
carpinifolia with no visible dominance (each species occupy about 50% of
the area), priority will be given to sclerophyllous vegetation and the whole
area will be assigned class 323.
Extension:
This class includes evergreen sclerophyllous bushes and scrubs which compose maquis,
garrigue, mattoral and phrygana.
This heading includes:
•= mattoral of arid zone with pre-desert brushes and tall Ziziphus lotus;
•= laurrel mattoral with Laurus nobilis;
•= cypress mattoral with native or planted cypressus;
•= tree-spurge formation with dense stands of Euphorbia dendroides in thermoMediterranean
area;
•= palmetto brush formations with dominating Chamaerops humilis;
•= pre-desert scrub with halo-nitrophyllous scrubs and gypsum scrubs: jujube brush
(Ziziphus lotus), shrubes of African affinities (spiny brush formation of accacia);
•= abandoned olive groves.
This heading excludes:
•= arborescent mattorals which are a pre- or post-broad-leaved evergreen forest formation
with more or less dense arborescent cover with a usually thick high evergreen shrub
stratum organised around evergreen oaks (Quercus suber/ilex/rotundifolia) olive trees or
pines the crown cover of which is more than 30 % (class 3.1.1). If the crown cover is
less than 30 %, it is assigned 3.2.4.
Texture of bushy sclerophyllous vegetation
Texture of a few isolated trees
Texture of sand rocks and soil without vegetation
Fig. 81 A generalised pattern of the class 232
78
Fig. 82 Representative demonstration of the sclerophyllous vegetation
(maquis composed of Quercus coccifera) on examples from FYROM.
The 3
rd
photos illustrates a vegetation association of Juniperus
oxycedrus mixed with Ostrya carpinifolia or Buxus spp. with no
dominance to be assigned class 323
No particularity was identified in this class.
324 Transitional woodland/shrub
Bushy or herbaceous vegetation with scattered trees. Can represent either
woodland degradation or forest regeneration/recolonisation.
Extension:
Areas of natural developmental forest formations (young broad–leaved and coniferous
wood species with herbaceous vegetation and dispersed solitary trees) for instance; in
abandoned meadows and pastures or after calamities of various origin, part of this class
may be also various degenerative stages of forest caused by industrial pollution, etc.
This heading includes:
•= arborescent mattorals which are pre- or post-formation of broad-leaved evergreen
forest with a usually thick evergreen shrub stratum composed of evergreen oaks
(Quercus suber/ilex/ rotundifolia), olive trees, carob trees or pines the crown cover
density of which is less than 30 % of the surface;
•= agricultural lands (classes 2xx) under recolonizing process with occurrence of forest
trees which cover more than 30 % of the surface (scattered trees or small plots of
forests);
•= abandoned fruit tree plantations and orchards;
•= clear cuts in forest areas;
•= young plantations;
•= forest nurseries inside forests areas;
•= natural grassland areas with small forests < 25 ha and/or with trees intermixed which
cover more than 30 % of the surface;
•= open clear-felled or regeneration areas with regrowing during transition stage which
last for maximum 5-8 years;
•= forest burning areas which do not show black tones any more in the satellite image but
are still visible;
•= heavily damaged forests by wind, snow-brake or acid rains and other pollution with
more than 50 % dead trees;
•= marginal zones of bogs with a vegetation composed of shrubs and pine bogs which
79
cover more than 50 % of the surface;
•= bare rocks with scattered trees that cover more than 10 % of the surface.
This heading excludes:
•= transitional woodland areas when the area has been overgrown with forest vegetation.
The canopy closure which is at least 50 % and if the average breast diameter of trees is
at least 10 cm (class 311);
•= abandoned olive groves (class 323);
•= agricultural lands (classes 2xx) with occurrence of forest vegetation with an
overgrowing rate less than 50 % (class 243);
•= stable/climax tree-like forest formations with a tree height less than 4 m and Pinus
mugo forests (class 322);
•= arborescent mattoral with trees of which the crown cover is more than 30 % (class
311).
Texture of forest regeneration or degradation
Texture of herbaceous plants
Fig. 83 A generalised pattern of the class 324
Fig. 84 Representative demonstration of the quoted class on examples of
forest degradation stage (damaged forest) in the Jizerske Mts. (Czech
Republic) and natural developmental forest formations (Kysucka
vrchovina, Slovakia) and woodland degradation in Slovenia
Generalisation:
•=Class 324 within an evolution stage of natural grasslands.
80
stage1 stage 2
stage 3
trees
shrubs
gramineousnatural grassland: class321 class324
shrubs >25%
trees
<30%
trees
>30%
class312
Particularity of class 324: Forest clear–cuts
These areas contain e.g. regular belts which originated by alternation of
natural or planted young stand with rests of forest. Young stands were
distinguished by their physiognomic appearance (height of the trees and the
canopy of crowns) in time of taking the image.
This heading includes:
•= planted or natural young stands after logging.
Texture of young stands natural or planted after cutting
Fig. 85 A generalised pattern of the particularity of class 324
81
Fig. 86 Representative of the quoted class particularity on example of forest
clear–cut (Kysucka vrchovina, Slovakia)
Particularity of class 324: Wooded fen, bog and transitional bog
These shrubby-herbaceous formations with scattered tree forms marginal
zones of peat bogs.
This heading includes:
•= shrubs and herbaceous vegetation with scattered trees (Betula pubescens, Alnus glutinosa,
Picea abies, Pinus silvestris, Salix spp.).
Texture of wooded fen, bog or wooded transitional bog
Texture of herbaceous plants
Fig. 87 A generalised pattern of the particularity of class 324
Fig. 88 Representative of the quoted class particularity on examples of
wooded fen, wooded transitional bog and wooded bog (Palase,
Avaste and Alam-pedja NR, Estonia)
82
331 Beaches, dunes, and sand plains
Beaches, dunes and expanses of sand or pebbles in coastal or continental
locations, including beds of stream channels with torrential regime.
Extension:
This class includes supra-littoral beaches and dunes developed at the back of the beach
from high water mark towards land.
This heading includes:
•= river dune formation in the immediate vicinity of great rivers;
•= inland and lacustrine dunes;
•= shifting dunes with mobile, unvegetated or open grasslands (white dune);
•= grey dunes fixed, stabilised or colonised by more or less closed perennial grasslands;
•= machair formations (nature coastal sand-plain with more or less surface and grassland
vegetation);
•= ergs (continental dune field located in desert);
•= accumulation of gravels along lower section of Alpine rivers.
This heading excludes:
•= inland dune heaths (crowberry and heather brown dunes) (class 322);
•= inland dunes thickets occupied by dense formations of shrubs including seabuckthorn,
privet, elder, willow, gorse or broom often festooned with creepers (class
322);
•= dune juniper thickets and woods (class 32x);
•= dune sclerophyllous scrubs (class 323);
•= wooded dune (class 31x);
•= humid dune-slacks (class 411);
•= unvegetated gravels on steep Alpine mountain side (class 332);
•= vegetated islands inside stream beds (class 3xx).
Texture of beach and river banks
Texture of sparse grass
Fig. 89 A generalised pattern of the class 331
Fig. 90 Representative demonstration of the quoted class on examples of
sandy bed of Morava river (Slovakia) and coastal dunes in Lithuania.
83
Generalisation:
•= Zoning of dune vegetation from the sea toward supra-littoral landscapes.
523 331 331 331521 411 32X 31X
sea beach white dune lagoon grey dune inland marsh
brackish
water
fresh
water
black dune
No particularity was identified in this class.
332 Bare rock
Scree, cliffs, rock outcrops, incuding active erosion, rocks and reef flats
situated above the high-water mark.
This heading includes:
•= unvegetated abandoned extraction sites;
•= sparsely vegetated areas where 75 % of the land surface is covered by rocks;
•= stable rocks with limestone pavements, block litter and mountain-top-debris;
•= unvegetated lapiaz;
•= sites and products of recent volcanic activities, volcanic ash and lapilli fields, barren
lava fields;
•= unvegetated supra-littoral rocky zones.
This heading excludes:
•= white dunes (class 331);
•= mediolittoral rocky sea beds (class 423);
•= bare rocks with scattered trees that cover more than 10 % of the surface (class 324).
Texture of cliff relief forms and talus cones
Texture of grassland
Fig. 91 A generalised pattern of the class 332
84
Fig. 92 Representative demonstration of the quoted class on example of the
cliff relief forms in the High Tatras (Slovakia)
No particularity was identified in this class.
333 Sparsely vegetated areas
Includes steppes, tundra and badlands. Scattered high-altitude vegetation.
Extension:
Scattered vegetation is composed of gramineous and/or ligneous and semi-ligneous
species for determining the ground cover percentage, excluding cryptograms.
This heading includes:
•= sparsely vegetated and unstable areas of stones, boulders, or rubble on steep slopes
where the vegetation layer covers between 15 % and 50 % of the surface;
•= sub-desertic steppes with gramineous species (Artemisia spp.) mixed with alfa (Stipa
spp.) when they cover between 15 % and 50 % of the surface;
•= vegetation of ‘lapie’ areas or limestone paving;
•= bare soils inside military training areas;
•= karstic areas of gramineous, ligneous and semi-ligneous vegetation.
This heading excludes:
•= windblown part of dune areas (class 331);
•= areas where ground cover more than 85 % of the surface (class 332);
•= areas where the vegetation layer covers more than 50 % of the surface (class 321);
•= dense alfa (Stipa ssp.) coverage (class 321).
Texture of karstic relief
Texture of rocky slopes of plains and cliffs
Texture of talus cones
Texture of shrubs
Texture of herbaceous plants
Fig. 93 A generalised pattern of the class 333
85
Fig. 94 Representative demonstration of the quoted class on example of
talus cone in the High Tatras (Slovakia)
Generalisation:
•= For class 333, a quantitative scheme should be applied:
Class 321 Class 333 Class 332
vegetation > 50 %
and
ground cover < 50 %
10 % < vegetation < 50 %
and
50 % < ground cover < 90 %
ground cover > 90 %
and
vegetation < 10 %
•= ‘Lapie’ areas:
Sparsely vegetated areas should also be applied for reticulated landscapes as ‘lapie’ or
limestone paving where vegetation is characterised by line/reticular distribution on
ground cracking substratum.
333
bare rocks
reticulated
vegetation
No particularity was identified in this class.
86
334 Burnt areas
Areas affected by recent fires, still mainly black.
Extension:
This class includes burnt forest areas, moors and heathlands, transitory forest-shrub
formations, areas with sparse vegetation.
This heading includes:
•= burns which are younger than three years and still visible in the satellite images;
•= all natural and semi-natural vegetated areas.
This heading excludes:
•= human farming management by burning arable lands (class 211).
Texture of trees
Texture of burnt trees
Fig. 95 A generalised pattern of the class 334
Fig. 96 Representative demonstration of the quoted class on examples of
burnt forest areas in Slovakia, Lithuania and Portugal
No particularity was identified in this class.
87
335 Glaciers and perpetual snow
Land covered by glaciers or permanent snowfields.
This heading includes:
•= glaciers and perpetual snow;
•= bare rocks.
Texture of glaciers and perpetual snow
Texture of bare rocks
Fig. 97 A generalised pattern of the class 335
Fig. 98 Representative demonstration of the glaciers on example from Arolla
(Switzerland)
Generalisation:
•= glaciers and perpetual snow =50 %
•= bare rocks =50 %
•= glaciers and perpetual snow >50 % 335
•= glaciers and perpetual snow <50 % 332
No particularity was identified in this class.
Class 4: Wetlands
Class 4.1 Inland wetlands
Areas flooded or liable to flooding during the great part of the year by fresh, brackish or
standing water with specific vegetation coverage made of low shrub, semi-ligneous or
herbaceous species. Includes water-fringe vegetation of lakes, rivers, and brooks and of fens and
eutrophic marshes, vegetation of transition mires and quaking bogs and springs, highly
oligotrophic and strongly acidic communities composed mainly of sphagnum growing on peat
and deriving moistures of raised bogs and blanket bogs.
Class 4.2 Coastal wetland
Areas which are submerged by high tides at some stage of the annual tidal cycle. Includes
salt meadows, facies of saltmarsh grass meadows, transitional or not to other communities,
vegetation occupying zones of varying salinity and humidity, sands and muds submerged
for part of every tide devoid of vascular plants, active or recently abandoned salt-extraction
evaporation basins.
332
332
332
332
335
88
411 Inland marshes
Low-lying land usually flooded in winter, and more or less saturated by
water all year round.
Extension:
This class includes non-forested areas of low-lying land flooded or liable to flooding by
fresh, stagnant or circulating water. Covered by specific low ligneous, semi-ligneous or
herbaceous vegetation.
This heading includes:
•= fens and transitional bogs without peat deposition or on peaty ground (peat layer is
less than 30 cm thick) with specific vegetation composed of reeds, bulrushes, rushes,
willows, sedges and tall herbs, sphagnum hummocks, often with alder or willows and
other water plants;
•= marsh vegetation located in margin zones of raised bogs;
•= water-fringe vegetation of reed beds, sedge communities, fen-sedge beds, tall rush
swamps, riparian cane formations;
•= high floating vegetation;
•= inland saline (alkali) marshes (prevailing arheic).
This heading excludes:
•= humid meadows (water logging of between 10 and 30 cm depth) (class 231);
•= rice fields (class 213);
•= free water space in wetlands (class 512);
•= salt marshes (class 421);
•= salt meadows in intertidal zone (class 421);
•= polders with reticulated channels bordered by hydrophilic vegetation (class 2xx);
•= humid forests with a crown cover more than 30 % (class 31x);
•= low floating aquatic vegetation (class 512).
Texture of water surface
Texture of hydrophilous herb vegetation
Texture of scattered trees and shrubs
Fig. 99 A generalised pattern of the class 411
Fig. 100Representative demonstration of the quoted class on examples of
inland marsh in Morava floodplain (Slovakia) and high reed vegetation
of Danube delta (Romania)
89
Generalisation:
•= When a marsh surrounds a small lake < 25 ha, the marsh area should be connected to
water bodies to make up to 25 ha polygon of 512.
411 < 25 ha
512
< 25 ha 512 > 25 ha
Particularity of class 411: Treeless fens and transitional bogs sometimes with
more than a 30 cm–thick peat layer
These are located in inland throughflow basins, in river flood valleys, areas
of springs, and margin zones of raised bogs. Surface of peatlands is plain or
concave with small microforms - hummocks and tussocks.
This heading includes:
•= areas of hydrophilous herb vegetation (Cares spp., Comarum palustres, Menyanthes
trifoliata, Phragmites australis, Trychophorum alpinum, Oxycoccus spp.).
Texture of hydrophilous herb vegetation
Fig. 101A generalised pattern of the particularity of class 411
Fig. 102Representative of the quoted class particularity on examples of
treeless fen and treeless transitional bog (Lahemaa NP and Tuhu NR,
Estonia)
90
412 Peatbogs
Peatland consisting mainly of decomposed moss and vegetable matter. May
or may not be exploited.
This heading includes:
•= minerotrophic peat bogs fed by ground water or streams with mosses (Drepanocladus
spp.) and Carex spp. or schoenus in alcaline bogs with occurrence of Calix spp., Betula
spp. and Alnus spp.;
•= ombrotrophic peat bogs fed only by direct precipitation with sphagnum species which
are abundant and dominant with other acidophilous plants such as Eriophorum
vaginatum, Scirpus spp., Carex spp., Vaccinium oxicoccos, Andromeda spp., Drosera spp. and
lichens;
•= blanket bogs with sphagnum species and Narthecium spp., Molinia spp., Scirpus spp.,
Shoenus spp., Erophiorum spp.;
•= boreal peat bogs with reticulated structure (aapa) with Sphagnum spp., Empetrum spp.,
Vaccinium spp., Betula nana, Salix nana, Carex spp., Eriophorium spp., Utriculara spp.,
Drosera spp.;
•= peat extracting areas;
•= fossil arctic peat bogs (palsa) with Vaccinium spp., Betula nana, Salix lapponum and Salix
glauca, lichens and Carex spp.
This heading excludes:
•= bog eye > 25 ha : large pool or lake occurring near the centre of raised bogs (class
512);
•= transitional bogs on peaty soils (< 30 cm thick peat) (class 324);
•= wooded peat bogs (class 31x);
•= drained peat bogs (class 411);
•= abandoned peat milling areas (class 32x);
•= upland areas of blanket peat bogs where peat does not accumulate dominated by
nardus or other deciduous grasses (class 321).
Texture of herbaceous and semi-ligneous vegetation
Texture of water surfaces
Texture of scattered trees and shrubs
Fig. 103A generalised pattern of the class 412
Fig. 104Representative demonstration of the quoted class on examples of
natural open bog, natural open bog with pools and natural dwarf
shrub bog (Marimetsa, Nätsi-Võlla and Laukasoo, Estonia)
312
91
Generalisation:
•= The figures below illustrate the two possible schemes to be applied for mapping peat
bogs according to their geographic context.
324
312
412
311
313
324
313
412
311
312
312
open areas
= eyes of peat bogs
transitional
woodland belt
forest areas
Particularity of class 412: Explored peat bogs
Areas with extraction of peat.
This heading includes:
•= uncovered peat bog;
•= exploited peat bog.
Texture of uncovered peat-bog
Texture of exploited peat-bog
Fig. 105A generalised pattern of the particularity of class 412
Fig. 106Representative of the quoted class particularity on example of the
explored peat bog in Lithuania
92
421 Salt marshes
Vegetated low-lying areas, above the high-tide line, susceptible to flooding
by seawater. Often in the process of filling in, gradually being colonized by
halophilic plants.
This heading includes:
•= intertidal sand, silt or mud-based habitats colonized by halophytic grasses such as:
Puccinelia spp., Spartina spp., rushes such as Juncus spp. and Blismus rufus and herbs such
as Limonium spp., Aster tripolium, Slicornia spp. Includes all flowering plant communities
which are submerged by high tides at some stage of the annual cycle;
•= salt meadow shep areas.
This heading excludes:
•= inland salt marshes with halophile and gypsophile communities (classes 333 or 411);
•= humid meadows of low vegetation dominated by Juncus gerardis, Carex divisa, Hordeum
marinum or Trifolium spp. and Lotus spp. of the edge of brackish lagoons (class 411).
Texture of halophilic vegetation
Texture of brackish or saline water
Texture of salty soil without vegetation
Fig. 107A generalised pattern of the class 421
Fig. 108Representative demonstration of the quoted class on example of salt
marshes Rumpo (Estonia) and in France
No particularity was identified in this class.
93
422 Salines
Salt-pans, active or in process of abandonment. Sections of salt marsh
exploited for the production of salt by evaporation. They are clearly
distinguishable from the rest of the marsh by their parcellation and
embankment systems.
This heading includes:
•= salinas organised for breeding shellfish, fishes;
•= salt pans;
•= sea water.
This heading excludes:
•= inland salinas (class 131).
Texture of sea water
Texture of salt-pans
Texture of industrial buildings
Texture of grassland
Fig. 109A generalised pattern of the class 422
Fig. 110Representative demonstration of the saline on example from France
No particularity was identified in this class.
94
423 Intertidal flats
Generally unvegetated expanses of mud, sand or rock lying between high
and low water marks. 0 m contour on maps.
Extension:
Warning: 0 m marine contour on maps is not always based on the same reference system
and might differ up to 2 m between European countries.
This heading includes:
•= intertidal seaweed-covered boulders, unvegetated shores, covered by shattered rocks or
boulders, cliffs and outcropping base-rocks.
This heading excludes:
•= salt marshes (class 421);
•= broadening of rivers entering the sea (class 522);
•= part of lagoon area directly connected to the sea which is artificially separated (class
521).
Texture of mud and sand
Texture of sea water
Fig. 111A generalised pattern of the class 423
Fig. 112Representative demonstration of the intertidal flat on example from
Aveiro Belgium
Generalisation:
•= morphological contexts of main erosion/deposit material processing where high water
mark from topographic maps might be modified.
421
95
dune
embayment
reentrant trap
331
423
current
331
dune
423
331
331
331
423
96
331
331
331
423
current
331423 423
523
123
sand
deposit seaport
current
321
island
331
423
land
"tail of comet"
current
321
island
331
423
land
"tombolo"
current
No particularity was identified in this class.
97
Class 5: Water bodies
Class 5.1 Inland waters
Lakes, ponds and pools of natural origin containing fresh (i.e non-saline) water and
running waters made of all rivers and streams. Man-made fresh water bodies including
reservoirs and canals.
Class 5.2 Marine waters
Oceanic and continental shelf waters, bays and narrow channels including sea lochs or
loughs, fiords or fjords, rya straits and estuaries. Saline or brackish coastal waters often
formed from sea inlets by sitting and cut-off from the sea by sand or mud banks.
511 Water courses
Natural or artificial water-courses serving as water drainage channels.
Includes canals. Minimum width for inclusion: 100 m.
This heading includes:
•= sand or gravel accumulations along streams < 25 ha,
•= rivers which have been canalised.
This heading excludes:
•= water bodies areas connected to watercourses (class 512),
•= hydroelectric plant located on watercourses > 25 ha (class 121).
Texture of water courses and channels
Fig. 113A generalised pattern of the class 511
Fig. 114Representative demonstration of the quoted class on example of
Danube river (Slovakia)
Generalisation:
•= Water body areas along streams are artificially separated from class 511 and assigned
class 512 even if they are connected to them.
98
water
512
511
•= Sand and gravel accumulations inside a streambed area are connected to the stream.
331
331
331
331 < 25 ha
511
511
331
331
331
511
511
331
331 > 75 %
331
How to map watercourses with significant presence of Arundo donax or Phragmites ssp. (class
411) located within the watercourse bed?
•= In this case, it is more important to preserve the continuity of watercourses without
interruption of 411 areas.
511
511
511
411
411 > 25 ha
•= Hydroelectric plant located on water-course is an exception of the continuity rule
normally applied for class 511.
99
plant
511
511 511
511
121
No particularity was identified in this class.
512 Water bodies
Natural or artificial stretches of water.
This heading includes:
•= low floating aquatic vegetation with species such as Nuphar spp., Nymphaea spp.,
Potamageton spp. and Lemna spp.;
•= archipelago of lakes inside land areas;
•= water surfaces used for fresh-water fish-breeding activities.
This heading excludes:
•= surface plant species characteristic for standing water (e.g. Typha latifolia, Carex riparia,
Glyceria maxima, Sparganium erectum and Phragmites communis (class 411);
•= liquid waste (class 132).
Texture of water surface
Fig. 115A generalised pattern of the class 512
Fig. 116Representative demonstration of the dam infrastructure on example
from Auvergne (France)
Generalisation:
Dam infrastructure should be isolated and assigned class 121 if its surface area is more than
25 ha.
100
dam
121
water
512
512
> 25 ha
< 25 ha
•= In case of a group of small lakes (each water surface < 25 ha), linking principle may be
applied if :
1) the resulting water polygon is > 25 ha,
2) the new zone created is composed of 75 % of free water.
•=
d > 300 m
512
311
each lake < 25 ha
sum of lakes > 75 %
No particularity was identified in this class.
101
521 Coastal lagoons
Stretches of salt or brackish water in coastal areas which are separated from
the sea by a tongue of land or other similar topography. These water bodies
can be connected to the sea at limited points, either permanently or for
parts of the year only.
Extension:
This class includes
This heading includes:
•= only water surface, vegetation fringe should be separated;
•= estuarine lagoon;
•= salt or brackish water surface remaining at low tide;
•= lagoons organised for breeding shellfish.
This heading excludes:
•= salt marshes (class 421);
•= water courses (class 511);
•= beaches (class 331);
•= fresh water bodies along shoreline (class 512).
Texture of salt or brackish water
Fig. 117A generalised pattern of the class 521
Fig. 118Representative demonstration of the quoted class on example of
coastal lagoon (Baltic Sea, Lithuania)
Generalisation:
•= In case of estuarine lagoon shape as described below, priority should be given to
lagoon class.
102
estuary
dune
dune
521
331
331
511
•= In case of water bodies along the shoreline, its geomorphologic position regarding to
the water table must be taken into account. Normally:
523
521
521
411
512
331
medium level of sea
sea
dune brackish
water
fresh
water
fresh watersalt water
No particularity was identified in this class.
103
522 Estuaries
The mouth of a river within which the tide ebbs and flows.
This heading includes:
•= the water and the channel bed with the fringing vegetation zone < 25 ha.
This heading excludes:
•= bays and narrow channel (class 523);
•= fjords or fiards, ryas and straits (class 523);
•= fringing vegetation along the estuary channel bed > 25 ha (class 421).
Texture of estuary water surface
Fig. 119A generalised pattern of the class 522
Fig. 120Representative demonstration of the quoted class on example of
estuary (in Aber Wrach)
Generalisation:
•= In practice, upstream maritime influence is stopped by the first floodgate; downstream,
the estuary limit is arbitrary.
No particularity was identified in this class.
104
523 Sea and ocean
Zone seaward of the lowest tide limit.
This heading includes:
•= sea water.
This heading excludes:
•= archipelago of lands located inside sea/ocean areas;
•= sea water areas as part of port areas which include sea water to reach a zone > 25 ha.
Texture of sea water
Fig. 121A generalised pattern of the class 523
Fig. 122Representative demonstration of the quoted class on example of
Baltic Sea (Lithuania)
Generalisation:
•= The same generalisation rule as for archipelago of lakes (class 512) should be applied
on two conditions:
1) resulting island polygon > 25 ha,
2) the new zone created is composed of 75% of land.
d > 300 m
512
island > 25 haisland
islands < 25 ha
3XX > 25 ha
523
sum of islands > 75 %
No particularity was identified in this class.
105
4 References
Ferancec J., Otahel’ J., Pravda J. (1995). Proposal for a methodology and nomenclature scale
1:50 000. Final report. Contract No 94-0893. Bratislava (Institute of Geography, SAS).
Ferancec J., Otahel’ J., Pravda J. (1996). Land cover of Slovakia identified by the CORINE land
cover method. Geographia Slovaca, 11. Bratislava (Institute of Geography, SAS).
CEC 1994). CORINE land cover. Technical guide. Luxembourg (Office for Official
Publications of European Communities).
The authors wish to express their thanks to: H. Dufourmont (Belgium) and C. Seabra
(Portugal) for their contributions to Part 1 of the addendum, K. Aavikso, A. Mainer, V.
Sagris, L. Jeeser, J. Paal, V. Andermann, V. Eltermann, E. Puurmann (Estonia), D.
Zdenkova (Czech Republic), H. Baranovs (Latvia), R. Kaulakys (Lithuania), Chris
Steenmans (EEA) for providing input and photos of national land cover particularities, to
H. Contrerasova for a help with translation of the text into the English.