Identifying Trends in Land Use/Land Cover Changes in the Context of Post-Socialist Transformation in Central Europe: A Case Study of the Greater Olomouc Region, Czech Republic
Tomáš Václavík1
Center for Applied Geographic Information Science (CAGIS), Department of Geography and Earth Sciences, University of North Carolina at Charlotte, 9201 University City Blvd, Charlotte, North Carolina 28223
John RoganGraduate School of Geography, Clark University, 950 Main Street, Worcester, Massachusetts 01610
Abstract: Many countries in Central and Eastern Europe have been undergoingmarked economic changes following the collapse of the former “Eastern Bloc” andtotalitarian regimes. In the Czech Republic, this transition has had a profound effecton land use management that subsequently results in widespread land cover changes.This study analyzes trends in land use/land cover changes (LULCC) in the context ofpolitical and economic transformation of the Czech Republic, using the greater Olo-mouc region in the period between 1991 and 2001 as a case study. Landsat ThematicMapper (TM) and Enhanced Thematic Mapper Plus (ETM+) images from 1991 and2001 were acquired and processed for maximum likelihood classification to produceland use/land cover maps for both times with overall map accuracies between 0.8and 0.84. Major land changes were identified using post-classification comparisonand trend surface analysis. Results showed significant marginalization of intensiveagricultural activities (12%), a shift in forest composition from mixed to deciduousforest (6%), and an overall increase in residential development on arable land(3.5%). Our findings are consistent with recent socioeconomic and political studiesthat describe post-socialist land change drivers in Central and Eastern Europe, suchas decreased need for intensive agriculture, shift to ecological management of for-ested areas, or increasing suburbanization.
INTRODUCTION
The Czech Republic (CR) is currently undergoing transformation from a central-ized regime of communist dictatorship (1948–1989) toward a modern democraticstate. The Olomouc region in the eastern CR thus has experienced significant changes
in the last two decades, triggered by the “Velvet Revolution” in 1989. Although thepolitical and socioeconomic transition is generally recognized as an important driverof land use change (Ptáček, 2000), few studies have assessed and quantified land use/land cover changes (LULCC) in the context of the post-socialist transformation inCentral and Eastern Europe (Bičík et al., 2001; Fanta et al., 2004; Zemek et al., 2005;Kuemmerle et al., 2006). In this study, we present an approach for identifying majorLULCC in the Olomouc region, by applying remote sensing techniques to comparetwo sets of multispectral Landsat Thematic Mapper (TM) and Enhanced ThematicMapper Plus (ETM+) data acquired 2 and 12 years after the revolution in 1989. Wepay close attention to specific trends in LULCC within the post-socialistic period:changes in agricultural areas, forested areas, and residential development. We baseour analysis on the assumptions that two years after the political transition is such ashort time that the 1991 land cover is fairly representative of the conditions in the pre-vious regime. On the other hand, 12 years after the revolution is a sufficient time forsocioeconomic changes to be reflected in the land-cover composition.
The fall of the Iron Curtain and the consequent breakdown of totalitarian regimesin the former “Eastern Bloc” have brought dramatic changes to the political and eco-nomic systems in most Central and Eastern European countries, including the CR(Bičík et al., 2001). Alterations of former socioeconomic structures, such as the resti-tution of private property previously nationalized under Communism or the privatiza-tion of agricultural co-operatives, have impacted a number of factors that ultimatelyshape the environment and have resulted in widespread modification of land manage-ment and land use decision making. Such a transition from politically to economicallydriven land use provides a unique opportunity to study the effect of broad-scale polit-ical and socioeconomic factors on LULCC (Kuemmerle et al., 2006).
The objective of this research was to analyze Landsat imagery from 1991 and2001 in order to empirically assess changes in the land use/land cover that occurredover a large area in the CR in the post-socialistic period. We identified the major landuse/land cover transitions and trends using post-classification comparison based on across-tabulation technique and trend surface analysis. We focused on the trends inLULCC generally recognized as most significant in Central and Eastern Europe:changes in agricultural areas, forest cover, and urban development. Moreover, wequantified these changes, localized their occurrence by trend surface analysis, andinterpreted them in the context of post-socialist transformation of the CR.
BACKGROUND
The general trends in LULCC have been described by various studies focusingon conditions in agriculture, forestry, and urban development. The most importantprocess in the agricultural sector has been the change in land ownership and propertystructure (Takács-György et al., 2007). Implementation of land reforms resulted inphysical fragmentation of arable land due to the splitting of large parcels, managed bystate co-operatives, into smaller privatized farmlands (Csaki, 2000; Sabates-Wheeler,2002; van Dijk, 2003). Land abandonment has also occurred extensively, as manylandowners withdraw completely from farming, and former areas of intensive agricul-ture are thus being converted to grassland and forest (Augustyn, 2004; Ioffe et al.,2004). In addition, marginalization of agricultural land leads to secondary afforestation.
56 VÁCLAVÍK AND ROGAN
A minor increase in forest extent, but especially the change in forest structure, hasbeen documented for the post-1989 era (Palang et al., 1998; Kozak, 2003; ÚHÚL,2006). In some countries (e.g. Poland and the CR), a certain amount of agriculturalland is currently being lost due to the advancing processes of suburbanization andconversion of arable land to newly developed areas (Ptáček, 1998; Jackson, 2002;Kreja, 2004).
In the CR specifically, Fanta et al. (2005) recognized three main events in the last50 years that had profound consequences for the country and its land use. The firstevent was the communist coup d’état and subsequent collectivization of land in the1950s, which encompassed the introduction of large-scale collective farming, espe-cially intense in the Olomouc region, aiming to maximize agricultural production.The second event was the abolition of the totalitarian political system in 1989, fol-lowed by restitution of private land ownership in the 1990s, re-introduction of democ-racy and a market economy, and development of market-driven forms of land use.The third event was the preparation of the CR for accession to the European Union(EU) in 2004, including complete implementation of EU environmental and agricul-tural policies. Moreover, processes of (1) partial privatization of state property, (2)increasing environmental consciousness, and (3) transformation of agricultural co-operatives had a direct impact on LULCC after 1989. These processes influenced par-ticularly the agricultural sector, forestry, and the rate and type of new development(Bičík et al., 2001).
Political transition in the CR led to marginalization of intensive agricultural areasin a process driven by a combination of socioeconomic and environmental factors.Due to marginalization, farming ceased to be viable in many places, resulting inunprecedented abandonment of agricultural land (Fanta et al., 2004). Extensive areasof previously cultivated land in the country are currently fallow or have converted tosecondary grasslands—meadows and pastures (Bičík et al. 2001).
Forest covers approximately 33% of the total area of the CR (ÚHÚL, 2006). For-est cover increased to this extent in the 20th century since the time of its minimumextent at the end of the 18th century. Most wooded land is far from its native compo-sition. Monocultures of Norway spruce (Picea abies) were planted over large areasthat are now used predominantly for timber production. However, the boom in envi-ronmental awareness and the incorporation of sustainability concepts into Czechlegislation in the early 1990s caused a distinctive tendency toward alternativeapproaches in forest management that take into account the natural species composi-tion and potential native vegetation (Neuhäuslová, 1998; ÚHÚL 2006).
The extent of built-up areas in the CR increased considerably after 1989 (Bičík etal., 2001). As in other parts of Europe, the issue of suburbanization has been identi-fied in the CR since the 1990s (Ptáček, 1998; Jackson, 2002). However, suburbaniza-tion in the CR is represented by relatively fine-scale residential development in thevicinities of larger cities, and does not bear the typical traits and negative effects ofthe large-scale suburban sprawl of the United States or countries of Western Europe(Václavík, 2004; EEA, 2006).
Although the political and socioeconomic factors driving LULCC have been welldocumented and the general trends of environmental changes recognized (Bičík et al.,2001), explicit spatial analysis of LULCC in Central and Eastern Europe is scarce.Few studies have applied geographic information systems (GIS) and remote sensing
LAND USE/LAND COVER CHANGES 57
techniques to quantify specific land use/land cover transitions in the post-socialist era.For example, in the Eastern Carpathians, analysis of Moderate Resolution ImagingSpectroradiometer (MODIS) data showed that the forest cover in the Upper Tiszawatershed declined on average by 5% between 1992 and 2001, while there was a 10–20% increase in the eastern sub-catchments, caused by different forest managementpractices in Romania, Ukraine, and Slovakia (Dezso et al., 2005). In the Orawa regionin southern Poland, the comparison of historical maps and contemporary satelliteimages from Landsat and Advanced Spaceborne Thermal Emission and ReflectionRadiometer (ASTER) imagery revealed that the proportion of forest cover increasedfrom 25% to 40% over the last 180 years (Kozak, 2003). In the Biesczady Mountainsin Poland, significant changes in village structure and extensive farmland abandon-ment were identified by visual comparison of historical maps and Landsat imagesfrom the late 1990s (Angelstam et al., 2003). In Slovakia, agricultural intensification(7% of the total area) and processes of deforestation and urbanization were detectedfrom the assessment of the European Union Coordination of Information on the Envi-ronment (CORINE) land cover (CLC) database between the 1970s and 1990s(Feranec et al., 2003). A cross-border comparison of land cover and landscape pat-terns was conducted for the Slovak, Polish, and the Ukrainian parts of the CarpathianMountains using Landsat TM and ETM+ imagery from 2000. Marked differencesbetween countries were discovered. For example, Slovakia and Poland had 20% moreforest cover than Ukraine, Slovakia had more deciduous forests than Poland andUkraine, and Ukraine experienced larger land abandonment and agricultural fragmen-tation than Poland and Slovakia (Kuemmerle et al., 2005, 2006).
Existing spatially explicit studies that have assessed land use/land cover issues inCentral and Eastern Europe have employed either small study areas (Angelstam et al.,2003; Kozak, 2003), combined data from different sources that are difficult to com-pare (e.g., historical maps and satellite imagery; Angelstam et al., 2003), or did notasses changes over time (Kuemmerle et al., 2005, 2006). In the CR, very few pub-lished remote sensing studies have addressed only the issues in forestry and forestmanagement, particularly damage to, and health conditions of, certain forest types(Škapec et al., 1994; Stoklasa, 1995), detection of gradual changes of forests innational parks (Šíma, 1995), or the potential of remote sensing applications in vegeta-tion monitoring (Kučera, 1999). Therefore, the major goal of this study was toanalyze remotely sensed data acquired in 1991 and 2001 to assess and quantify thespatial and temporal changes in land use/land cover composition over a large area inthe CR. We identified and interpreted the locations, types, and trends of the majorland use/land cover transitions in the Olomouc region that occurred in a span of 10years in the early post-socialist period.
METHODS
Study Site
We studied the greater Olomouc region located in the northeastern CR (Fig. 1).The study area (5012 km2) covers the majority of the Olomouc county administrationunit, one of the 14 administration units in the CR. The northeastern section of thestudy area overlaps Moravskoslezský County. The central region is comprised of the
58 VÁCLAVÍK AND ROGAN
Fig.
1. S
tudy
are
a of
the
grea
ter O
lom
ouc
regi
on lo
cate
d in
the
north
east
ern
Cze
ch R
epub
lic.
LAND USE/LAND COVER CHANGES 59
wide alluvial plain of the Morava River, surrounded by undulating hills of theZábřežská and Drahánská uplands to the west and the Nízký Jeseník mountain rangeto the northeast, with altitudes ranging from 200 to 800 m above sea level. The low-lands are highly urbanized and include the major cities of Olomouc, Litovel,Prostějov, Zábřeh, Šumperk, and others. Due to favorable climate and fertile soils,lowlands historically and currently represent substantive agricultural areas in the CR.Despite its intensive development, the core of the Olomouc region is constituted bythe Litovelské Pomoraví Protected Landscape Area with natural complexes of flood-plain forests. The other major forested habitats in the Olomouc region are located inthe northeastern uplands, predominantly composed of coniferous and mixed standsand used for timber production. We decided to study the greater Olomouc regionbecause it is formed by relatively homogenous environment, yet it includes ampleproportions of all land use/land cover categories that we intended to examine. Both inthe past and at present, the greater Olomouc region has been a major agriculturalcenter but still includes substantive forest cover and residential areas of all sizes anddensities from larger cities to small villages. Although our study site experienced nosignificant population growth after the revolution in 1989, it has been strongly influ-enced by socioeconomic changes (Ptáček, 2000). These characteristics make thegreater Olomouc region an ideal study site for comparative assessment of land use/land cover changes in the context of post-socialist transformation in the CR.
Data Collection and Preparation
Landsat-5 TM and Landsat-7 ETM+ images were acquired for this study, as theyprovide an appropriate and cost-efficient source of information for a wide range ofapplications, including land change mapping (Rogan and Chen, 2004). The TM datacomprised one scene 018-385 (path 190, row 25) from 10 September 1991; the ETM+data included two scenes 036-343 and 036-344 (path 190, row 25 and path 190, row26) from 24 May 2001. Two ETM+ images from 2001 were needed because the studyarea was located in the overlap of two Landsat-7 ETM+ swaths. Both data sets wereobtained from the Global Land Cover Change Facility [http://glcf.umiacs.umd.edu/data/], having a ground resolution element of 28.5 × 28.5 m. We did not employ thethermal bands in our analysis due to their coarser spatial resolution and weak signal tonoise ratio (Jensen, 2004). The TM image was previously georegistered to the ETM+images using 12 ground control points, first-order linear transformation, and nearestneighbor interpolation. Based on the given set of control points, the total root meansquare error (RMSE) was computed to be lower than the threshold of 0.3 pixels thatwe defined as acceptable. A set of scanned and georeferenced black-and-white aerialphotographs from 1991 and a set of true color orthophotographs from 2002, both withthe spatial resolution of 1 m, were obtained from the Litovelské Pomoraví ProtectedLandscape Area Administration to serve as reference data for the classification andmap accuracy assessment processes. Vector data of the CR boundary and the pro-tected area were acquired from the Czech Environmental Information Agency(CENIA) ArcIMS server [http://geoportal.cenia.cz].
Seven land cover categories were defined in the Olomouc region: water, decidu-ous forest, coniferous forest, mixed forest, developed, agricultural areas, and mead-ows. Mixed forest was defined as not having a dominating share of either coniferous
60 VÁCLAVÍK AND ROGAN
or deciduous tree species, i.e., more than 75% (ÚHÚL, 2006). Built-up landscapes,including all urban, industrial, and residential areas, were categorized as developed.The agriculture class included areas of intensive farming. The meadows category rep-resented all grassland ecosystems: pastures, periodically mowed meadows, or second-ary grasslands of abandoned farmland. Secondary succession vegetation, shrub cover,and early stage stands were visually assessed and labeled as either a certain type offorest or a meadow, in cases where the wood cover was sparse (threshold of approxi-mately 20% of cover). To facilitate land use/land cover classification, polygonal areasof interest (AOI) were selected as training sites to represent characteristic land use/land cover types in the study area. The AOIs consisted of approximately 100 pixelsfor each category. In order to develop representative spectral signatures for each class,training sites were digitized based on the combination of ground visits of the studyarea and digital reference data (georeferenced aerial photographs from 1991 andorthophotographs from 2002).
The Land Change Modeler module in IDRISI software was utilized for landchange detection and trend surface analysis. The Visual Basic for Applications (VBA)macro created by Pontius and Silva (Silva, 2006) was used for the stratified samplingdesign and for the computation of error matrices in order to assess the classificationaccuracy.
Image Processing
Figure 2 presents the steps of image processing, classification, and comparativeassessment that were needed to achieve the defined study objectives. First, weassessed the satellite data for their image quality. While both ETM+ images did notexhibit any significant radiometric noise in the entire scene, the TM image containeda small amount of haziness in the northeastern portion and a subtle striping through-out the entire area. As there were no data available on the sensor spectral profile orthe atmospheric properties for the time of TM image acquisition, absolute atmo-spheric correction was not possible. Instead, principal components analysis (PCA)was applied to reduce the haze and striping dimensions of the data and improve thesignal to noise ratio (Zhao and Maclean, 2000). PCA transforms the original data intoa set of uncorrelated variables, where the first components represent the majority ofvariance from the original data sets and the subsequent orthogonal componentsaccount for less variance and a higher proportion of noise (Eastman and Fulk, 1993).We ran PCA using a standardized variance/covariance matrix and all six optical TMbands as inputs. PCA created six principal component images, where the first fourexplained over 98% of the total variance and the remaining two contained most of thenoise. The original noise-free bands were then restored through the inverse PCA tech-nique, retaining only the first four components with meaningful information. Second,the study area was located on two ETM+ images that were overlapping by approxi-mately 20% of their area. A mosaic of the images was created by spatially orientingthem and balancing their numeric characteristics. Edge feathering with the histogram-matching algorithm was used to adjust the brightness values and blend the seams inthe overlapping images.
LAND USE/LAND COVER CHANGES 61
Fig.
2. F
low
dia
gram
of i
mag
e pr
oces
sing
, cla
ssifi
catio
n, a
nd c
ompa
rativ
e as
sess
men
t.
62 VÁCLAVÍK AND ROGAN
Image Classification
Rogan and Chen (2004) suggested that supervised classification methods may bemore intuitive for land cover change detection if the required prior information aboutthe landscape is gained through personal knowledge of the study area, in addition to acombination of ground visits and aerial photography interpretation. We used the max-imum likelihood classification to derive seven land use/land cover categories in theOlomouc region. The maximum likelihood classification is based on the probabilitydensity function that is associated with a particular training site signature. All pixelsare assigned the label of the most likely category based on an evaluation of the subse-quent probability that the pixel belongs to the class with the highest probability ofmembership (Atkinson and Lewis, 2000; Jensen, 2004). Although the maximum like-lihood method assumes a normal distribution of the data, it is still considered as oneof the most useful classifiers, as it does not always require large training data sets andits performance is comparable to other algorithms if the training sites are of goodquality or limited size (Wu and Shao, 2002; Franklin et al., 2003).
Spectral signatures of individual land use/land cover classes were developedbased on selected training sites and assessed for their separability using feature spaceimages and scatterplots. Unambiguous classes were retained; spectrally similarclasses representing the same land use/land cover categories (e.g., agricultural fieldswith crops and agricultural fields with bare soil) were merged. To distinguish betweenspectrally similar classes representing different land use/land cover categories (e.g.,developed areas and agricultural areas with bare soil), various combinations of con-textual measures were tested on all bands. A texture image created 10 with the domi-nance index and kernel window of 5 × 5 pixels applied to the near-infrared (NIR)band produced the most satisfactory result, after it was used as an additional inputwith original bands in the classification process. In addition, some areas were maskedand classified manually because the 2001 ETM+ image was acquired in the springseason, when certain types of crops (e.g., cereals) are in a phenological stage thatexhibit similar spectral response as secondary grassland.
We applied a stratified random sampling strategy to select land use/land coversample sites for assessment of the classification accuracy (Jensen, 2004; McCloy,2006). Both maps for 1991 and 2001 produced by the maximum likelihood classifica-tion were subdivided into individual land use/land cover strata. Sample locationswere then randomly distributed throughout the study area for each land use/land covermap using a random number generator in GIS. Single TM and ETM+ pixels were theunits of assessment and the same proportion of sample locations (n = 20) in each stra-tum was the sample size (total n = 140). The x,y coordinates of the sample pixels werethen identified in the digital reference data. Two sets of georeferenced aerial photo-graphs from 1991 and orthophotographs from 2002 were used as the main sources ofreference information. We determined the land use/land cover classes throughdetailed visual examination of the photographs and visited those sites that were notdistinguishable on the 2002 images.
The error matrices were constructed for both classified maps to provide the basisfor characterizing errors by cross-tabulating the classified land cover categories insample locations against those observed in reference data (Smits et al., 1999; Foody,2002). We computed overall classification accuracy for both maps, as well as
LAND USE/LAND COVER CHANGES 63
producer’s and user’s accuracies to measure omission and commission errors for indi-vidual land use/land cover categories. Because the traditional error matrix presentsinformation on sampled locations only, we adjusted the overall accuracy by takinginto account the proportion of each stratum (land use/land cover category) in the clas-sified maps (Silva, 2006). In this way, we estimated the total proportions of pixels thatwere classified correctly and incorrectly in both maps.
Land Use/Land Cover Map Comparison
We applied the post-classification comparison technique to characterize landchanges between 1991 and 2001. The IDRISI Andes software provides an efficienttool for rapid assessment of LULCC and their implications based on cross-tabulationprinciples. The Land Change Modeler (LCM) for Ecological Sustainability allows auser to evaluate gains and losses in land cover classes, land cover persistence, andspecific transitions between selected categories. Using the classified land-use/land-cover maps from 1991 and 2001 as input parameters, this tool was applied to identifythe locations and magnitude of the major land use/land cover changes and persis-tence. Additionally, we estimated the spatial trends of major transitions between landuse/land cover categories of special interest in the study area, using trend surfaceanalysis (TSA). TSA is an interpolation procedure that disaggregates the broadregional patterns from the non-systematic, fine-scale variation in the data (Chorleyand Haggett, 1965; Eastman, 2006). It is designed to extract the regional componentfrom a map, such as general location of a specific land change trend, from the residualcomponent (Gittins, 1968). This empirical, least-square technique assumes the gen-eral spatial trend, inherent to the data, can be reasonably represented by a polynomialsurface of closest fit to the observations, minimizing the difference between the inter-polated value at a data location and its original value (Gustafson, 1998). It may bedefined mathematically as:
where Z is the areally distributed variable, in this case, the transition between twoselected land use/land cover categories, αs are the polynomial coefficients, and U andV are the locational coordinates. The TSA surfaces are calculated by coding the pixelsof a specific transition as 1 and pixels of no change as 0, and treating them as if theywere continuous values (Chorley and Haggett, 1965; Eastman, 2006). To help visual-ize the general locations of land use/land cover transitions in the period between 1991and 2001, we employed the sixth-order polynomial TSA for three specific categoricaltransitions: from agriculture to meadows, from agriculture to developed, and frommixed forest to deciduous forest.
RESULTS
Figure 3 presents the results of maximum likelihood classification for 1991 and2001. The error matrices for both classified maps were constructed to assess the clas-sification accuracy (see Table 1). In the 1991 map, the proportion of agreementbetween land use/land cover categories in the classified map and the reference data
64 VÁCLAVÍK AND ROGAN
was 79%. When the proportion of agreement was estimated for the entire landscape,the overall accuracy was 80%. The producer’s accuracy was lowest for the agricul-tural class (65%), as some of the agricultural areas were misclassified as meadows ordeveloped. Producer’s accuracy for mixed forest was 71%, as some pixels were mis-classified as pure coniferous or deciduous forest. The user’s accuracy was lowest formixed forest (60%), as some sites were identified as coniferous or deciduous in thereference data, and the category of developed (70%), as some sites were identified asagricultural.
Fig. 3. Land use/land cover maps derived from 1991 (A) and 2001 (B) Landsat images.
LAND USE/LAND COVER CHANGES 65
Ta
ble
1. E
rror
Mat
rices
for C
lass
ified
Lan
d U
se/L
and
Cov
er M
aps f
rom
199
1 an
d 20
01
Cat
egor
yW
ater
Dec
iduo
usC
onife
rous
Mix
edfo
rest
Dev
elop
edA
gric
ultu
ral
Mea
dow
sTo
tal
Use
r’s
accu
racy
Land
cov
er m
ap 1
991
Gro
und
refe
renc
e da
ta—
aeria
l pho
togr
aphs
199
1W
ater
190
00
10
020
0.95
Dec
iduo
us0
170
10
02
200.
85C
onife
rous
00
164
00
020
0.80
Mix
ed fo
rest
02
612
00
020
0.60
Dev
elop
ed1
00
014
50
200.
70A
gric
ultu
ral
00
00
117
220
0.85
Mea
dow
s0
00
00
416
200.
80To
tal
2019
2217
1626
2014
0Pr
oduc
er’s
acc
urac
ya0.
950.
890.
730.
710.
880.
650.
80
Land
cov
er m
ap 2
001
Gro
und
refe
renc
e da
ta—
aeria
l pho
togr
aphs
200
2W
ater
200
00
00
020
1.00
Dec
iduo
us0
160
30
01
200.
80C
onife
rous
00
164
00
020
0.80
Mix
ed fo
rest
00
416
00
020
0.80
Dev
elop
ed0
02
012
60
200.
60A
gric
ultu
ral
00
00
019
120
0.95
Mea
dow
s0
10
00
415
200.
75To
tal
2017
2223
1230
1614
0Pr
oduc
er’s
acc
urac
yb1.
000.
940.
730.
701.
000.
660.
88
a Sam
ples
acc
urac
y 0.
79; o
vera
ll ac
cura
cy 0
.80.
b Sam
ples
acc
urac
y 0.
81; o
vera
ll ac
cura
cy 0
.84.
66 VÁCLAVÍK AND ROGAN
In the 2001 map, the proportion of sample-based agreement was 81%. The over-all accuracy for the entire landscape was 84%. The producer’s accuracy was lowestfor agriculture (66%), as some farmland was misclassified as meadows or developed,and for the category of mixed forest (70%), as some pixels were misclassified as pureconiferous or deciduous forest. The user’s accuracy was lowest for the developedclass (60%), where some sites were identified as agriculture or coniferous forest inthe reference data, and meadows (75%), where some sites were identified as agricul-ture or deciduous forest.
The result of the cross-tabulation comparison of both land use/land cover maps(Fig. 4) demonstrates that there have been marked changes in all land use/land covercategories between 1991 and 2001, with the exception of the water category. Theslight increase in the category of water can be explained by the construction of theSlezská Harta dam and reservoir in the northeastern part of the study area. Severalland cover categories experienced major overall changes. The total area of meadows(grassland) increased by 428 km2 (representing an increase of 8.5% of the total studyarea), while the area of intensive agriculture decreased by 339 km2 (7% of the totalstudy area), as did the area of coniferous forests, which decreased by 158 km2 (3% ofthe total study area). The proportion of areas covered with mixed forests decreased by53 km2 (1% of the total study area), while the proportion of deciduous forestsincreased by 85 km2 (1.7% of the total study area). Additionally, the category ofdeveloped was also affected by distinct change, with a net gain of 83 km2 (1.7% of thetotal study area).
To identify the transitions between specific land-use/land-cover categories, wefocused on those that experienced an increase in their total area during the studyperiod: meadows (grassland), developed, and deciduous forest. We calculated thecontributions of other categories to their net change (Fig. 5), i.e., which classes in the1991 map were identified as meadows, developed, or deciduous in the 2001 map andwhat proportion of the total change for a class and proportion of the study area theyexplain. Agricultural areas explain the majority of the total increase in meadows(65%), which represents about 5% of the entire study area. New developmentoccurred predominantly on former agricultural areas (85%). This transition occurred
Fig. 4. Net changes in land use/land cover categories between 1991 and 2001 in percent of totalstudy area.
LAND USE/LAND COVER CHANGES 67
on approximately 1.5% of the study area. The majority of the increase in deciduousforests is explained by the transition from the mixed forest category (35%) togetherwith the category of meadows (27%), which represents over 1% of the total studyarea.
A simplified cross-classification map (Fig. 6) shows persistence in land use/landcover categories, i.e., areas where no change occurred. Areas with some transitionsbetween land use/land cover classes are depicted as change. The land change and
Fig. 5. Contributions to net changes in categories of (A) meadows (grassland), (B) deciduousforest, and (C) developed in percent of total study area.
68 VÁCLAVÍK AND ROGAN
persistence map can be difficult to interpret visually if the locations of specific landuse/land cover changes are not clustered (Rogan and Chen, 2004). The landscape ofthe greater Olomouc region is highly influenced by human activities; therefore, thepattern of land use/land cover change is complex and the broader trends cannot beeasily discerned.
We used trend surface analysis to facilitate interpretation of the complex landchange patterns by providing a means of generalization about transition trendsbetween selected categories. Three TSA maps (Fig. 7) were created showing the tran-sitions from 1991 to 2001 between categories of interest: agriculture to meadows,agriculture to developed, and mixed forest to deciduous forest. The resulting mapsdepict a simulated surface that denotes the generalized locations of transition betweenselected categories, from areas with no change to areas with marked change. The gen-eral trend of transition from intensive agriculture to meadows was located in thenortheastern part of the study area, in the highlands of Nízký Jeseník. The generallocation of new development occurring on agricultural areas was identified in the cen-tral and southern lowlands of the Morava River alluvial plain. The major transition inforest structure from mixed forest to deciduous forest was detected in the northeasternportions of the study area.
DISCUSSION AND CONCLUSIONS
This study applied remote sensing techniques to classify satellite imagery of thegreater Olomouc region of the Czech Republic from 1991 and 2001. Our objectivewas to identify the locations, types, and trends of the major LULCC in the 10 years
Fig. 6. Simplified cross-classification map representing persistence (color) and change (black)inland use/land cover categories between 1991 and 2001.
LAND USE/LAND COVER CHANGES 69
that followed the change in political system in the CR. We assumed the land coverwould reflect broad-scale socioeconomic changes that might have affected landscapeand natural resources, such as decreased need for intensive agriculture, shift to envi-
Fig. 7. Trend surface analysis of selected transitions between land use/land cover categories ofinterest. A. Agriculture to meadows. B. Agriculture to developed. C (facing page). Mixedforest to deciduous forest.
70 VÁCLAVÍK AND ROGAN
ronmentally friendly management of forested areas, or increasing development andsuburbanization.
The post-classification comparison of remotely sensed data is as accurate as theclassification results used in the analysis. The need for accurate land use/land covermaps is thus evident, as every error in classification propagates in the analysis itself(Jensen, 2004). We estimated our overall accuracy of both land use/land cover mapsclose to a set standard of 80–85% (Rogan et al., 2003), and identified the producer’sand user’s accuracies for individual categories. The overall accuracy for the 2001 mapwas 4% higher than for the 1991 map. Categories of water and deciduous forest hadproducer’s and user’s accuracies for both maps over 85%, thanks to the highseparability of their spectral signatures (unimodal distribution of training data). Thecategory of agriculture was the most problematic because it represented a mixture ofvarious crops in different phenological stages as well as bare soil (plowed fields). Thespectral signature of certain crops (e.g., cereals) were mixing with the signature ofgrassland, resulting in lower producer’s accuracies (<70%) for both classified maps.Agricultural land with bare soil also exhibited a certain degree of spectral collinearitywith developed areas. Although the contextual analysis (texture images) helped sig-nificantly with distinguishing between these two classes, the user’s accuracies fordeveloped land were less than 75%. Generally, the mixed categories are more chal-lenging to classify because the class borders are drawn artificially and a conflictbetween desired thematic classes and their separability may exist (Foody, 2002).Mixed forest proved particularly difficult to classify due to its spectral heterogeneityand overlap with coniferous and deciduous forest.
Fig. 7. continued
LAND USE/LAND COVER CHANGES 71
Accuracy assessment is considered most reliable if an independent set of groundreference points is collected. However, obtaining such data was not feasible in ourcase because our analysis concerned land use/land cover situations from the past. Weused high-quality aerial photographs and orthophotographs for reference data andused stratified random sampling to generate a set of control points, independent fromthe training sites. Due to potential inaccuracies in georegistration of the referenceimages and the challenges associated with visual examination of the photographs, wecannot completely rule out a bias. Despite these challenges, the performed classifica-tion assessment should provide information about the level of uncertainty associatedwith the map comparison and thus an essential context for the interpretation ofreported results.
The results of this research support initial assumptions based on general knowl-edge of some land use drivers in the post-socialist period analyzed in recent socioeco-nomic studies (Ptáček 1998; Bičík et al., 2001; Fanta et al., 2004; Zemek et al., 2005).There have been significant losses in categories of intensive agricultural areas andconiferous forest, and gains in grassland, developed areas, and deciduous forest. Fromthe former agricultural areas (in 1991), almost 12% became meadows, pastures, orsecondary grassland in 2001, especially in the northeastern hilly part of the study site.Approximately 3.5% of arable land was developed in the central and southern low-lands of the study area. More than 6% of the previous mixed forest in the northeasternhilly part of the region was identified as deciduous forest in 2001.
This study provides no empirical evidence of direct causality between mappedland use/land cover changes and political and socioeconomic transformation of theCR. However, the spatial trends of observed land change suggest a distinct correla-tion. Concerning the transition from the category of intensive agriculture to the cate-gory of meadows, the major trend was observed in the northeastern uplands of thestudy site and in the Nízký Jeseník mountain range. This observation is consistentwith the suggestions of Zemek et al. (2005) that the marginalization of agriculturalareas occurs first at locations with unfavorable natural conditions, especially inuplands where agricultural production was previously forced by an extensive use offertilizers and pesticides. Most of these sites are not suitable as arable lands and wereconverted to farmlands during the socialist period of agricultural industrialization.After the restitution process and changes in agricultural subsidies, the majority of newlandowners ceded their shares to successor organizations of former state co-opera-tives or keep the land but have not continued the previous agricultural activities(Bičík et al., 2001).
Concerning newly built-up areas, the major conversion to the class of developedwas identified from the class of agriculture. Trend surface analysis revealed that sucha transition occurred prominently in the central and southern lowland regions of thestudy site. This observation is consistent with a general suburbanization process inCentral Europe, where the area of low-density residential development is rapidlyexpanding, although at a lower rate than in Western Europe (EEA, 2006). Progressivedevelopment of open land is associated with the booming economy of the country andwith the restitution and privatization process as well (Bičík et al., 2001). New residen-tial areas tend to be built in the form of “satellite” towns and villages in the vicinity oflarger cities, existing infrastructure, and recreational areas (Ptáček, 1998; Jackson,2002). The lowland regions in the alluvial plain of the Morava River can offer a dense
72 VÁCLAVÍK AND ROGAN
transportation network and a hinterland of several existing cities, including historicalOlomouc and Litovel. Also the Litovelské Pomoraví Protected Landscape Areaserves as an important resource for recreation activities.
Regarding the change in forest structure, the transition from mixed forest todeciduous tree cover was observed predominantly in the northeastern hilly and moun-tainous part of the study site. In these areas, the intermediate elevation and associatedenvironmental conditions potentially favor deciduous forest stands (Neuhäuslová,1998). This finding correlates with the general diversion in forest management in thelast 15 years from clear-cut practices and spruce and pine plantations to the alterna-tive use of native broadleaved species of trees in the lower and intermediate eleva-tions of the country (ÚHÚL, 2006). The causes of such a diversion stem from anincrease in environmental awareness in early 1990s, and has influenced the manage-ment of environmental resources in the country (Bičík et al., 2001). Although forestmanagement is still focused on timber production, other functions of forest ecosys-tems are being taken into account and non-native tree species are being systematicallyremoved from the stands where economic conditions permit (ÚHÚL, 2006). In addi-tion, some increase in the deciduous forest class was explained by the conversionfrom the category of meadows. We suggest that this likely indicates a process of landabandonment in mountainous areas where the secondary succession of shrubs andearly stages of broadleaved vegetation were classified as deciduous forest.
The presented findings of specific types and trends of LULCC in the greaterOlomouc region are consistent with socioeconomic processes that were describedabove as general drivers of land transformation in Central and Eastern Europe. How-ever, our results suggest that the scale and intensity of land changes do not entirelyfollow the patterns of land transformation identified in other Eastern European coun-tries. First, privatization of state property and transformation of agricultural collec-tives in the CR resulted in marginalization of farmland but at a smaller scale than inPoland or Ukraine (Sabates-Wheeler, 2002; Angelstam et al., 2003). Similarly as inSlovakia (Csaki, 2000), decollectivization of socialist farming patterns was slowedbecause a large number of landowners left their land within the successor organiza-tions of former collectives. In contrast to other eastern countries in Europe, the transi-tion of agricultural areas to grassland can be in the CR explained not only by theabandonment of farmland but also by the increasing focus of currently active farmerson biomass production and maintenance of grassland communities in the countrysideas one of the possible solutions to former agricultural overproduction (Šarapatka,2006). Second, significant secondary afforestation (up to 20% of forest cover) wasidentified in Poland, Ukraine, and Romania due to the abandonment of agriculturalland (Kozak, 2003; Dezso et al., 2005); however, our data show the opposite trend.Results of our analysis suggest slight reforestation of former grassland, but the totalnet area of forest vegetation decreased by 2% between 1991 and 2001. Third, the pro-cess of urban sprawl is associated primarily with the countries of Western rather thanEastern Europe. Due to the different socioeconomic trends and historical conditions,marked suburbanization trends were described only in Poland, Slovakia, and the CR.We identified an increase of 2% in urban development during our study period, whichis significantly lower than is the case for the Western European countries, such asGermany or France (5–15%) (EEA, 2006).
LAND USE/LAND COVER CHANGES 73
Few studies to date have attempted to examine and quantify the major land use/land cover changes in the CR in the context of post-socialist transformation. Classifi-cation of multispectral satellite data and comparison of produced land use/land covermaps represent an effective alternative to the use of conventional data. Traditionaldata sources—e.g., historical maps, cadastre maps, agricultural censuses, or demo-graphic statistics—differ in scale and accuracy, making their use difficult (Kuem-merle et al., 2006). We have demonstrated that the application of remotely sensed dataand techniques of geographical analysis (e.g., trend surface calculation), in combina-tion with knowledge provided by relevant socioeconomic studies, can be a valid com-ponent in a complex understanding of the consequences of broad-scale political andeconomic changes for the environment.
REFERENCES
Angelstam, P., Boresjo-Bronge, L., Mikusinski, G., Sporrong, U., and A. Wastfelt, A.,2003, “Assessing Village Authenticity with Satellite Images: A Method to Iden-tify Intact Cultural Landscapes in Europe,” Ambio, 32:594–604.
Atkinson, P. M. and P. Lewis, 2000, “Geostatistical Classification for Remote Sens-ing: An Introduction," Computers & Geosciences, 26:361–371.
Augustyn, M., 2004, “Anthropogenic Pressure in the Environmental Parameters ofthe Bieszczady Mountains,” Biosphere Conservation, 6:43–53.
Bičík, I., Jeleček, L., and V. Štěpánek, V., 2001, “Land-Use Changes and Their SocialDriving Forces in Czechia in the 19th and 20th centuries,” Land Use Policy,18:65–73.
Chorley, R. J. and P. Haggett, 1965, “Trend-Surface Mapping in GeographicalResearch,” Transactions of the Institute of British Geographers, 37:47–67.
Csaki, C., 2000, “Agricultural Reforms in Central and Eastern Europe and the formerSoviet Union—Status and Perspectives,” Agricultural Economics, 22:37–54.
Dezso, Z., Bartholy, J., Pongracz, R., and Z. Barcza, 2005, “Analysis of Land-Use/Land-Cover Change in the Carpathian Region Based on Remote Sensing Tech-niques,” Physics and Chemistry of the Earth, 30:109–115.
Eastman, J. R., 2006, IDRISI Andes Guide to GIS and Image Processing, Worcester,MA: Clark University, Clark Labs, IDRISI Productions.
Eastman, J. R. and M. Fulk, 1993, “Long Sequence Time Series Evaluation usingStandardized Principal Components,” Photogrammetric Engineering andRemote Sensing, 59(8):1307–1312.
EEA (European Environment Agency), 2006, Urban Sprawl in Europe: The IgnoredChallenge, Copenhagen, Denmark: European Environment Agency, 58 p.
Fanta, J., Prach, K., and F. Zemek, F., 2004, Status of Marginalisation in CzechRepublic: Agriculture and Land Use. EUROLAN, National Report—CzechRepublic, České Budějovice, Czech Republic, Faculty of Biological Sciences,University of South Bohemia, and Institute of Landscape Ecology Academy ofSciences of the Czech Republic, 29 p.
Fanta, J., Zemek, F., Prach, K., Heřman, M. and Boucníková, E., 2005, Strengtheningthe Multifunctional Use of European Land: Coping with Marginalization.EUROLAN, National Report—Czech Republic, České Budějovice, CzechRepublic, Faculty of Biological Sciences, University of South Bohemia, and
74 VÁCLAVÍK AND ROGAN
Institute of System Biology and Ecology, Academy of Sciences of the CzechRepublic, 11 p.
Feranec, J., Cebecauer, T., Otahel’, J., and M. Suri, 2003, “Assessment of the SelectedLandscape Change Types of Slovakia in the 1970s and 1990s,” Ekologia (Bratis-lava), 22:161–167.
Foody, G. M., 2002, “Status of Land Cover Classification Accuracy Assessment,"Remote Sensing of Environment, 80:185–201.
Franklin, J., Phinn, S. R., Woodcock, C. E., and H. Rogan, 2003, “Rationale and Con-ceptual Framework for Classification Approaches to Assess Forest Resourcesand Properties,” in Methods and Applications for Remote Sensing of Forests.Concepts and Case Studies, Wulder, M. and S. E. Franklin (Eds.), Dordrecht, TheNetherlands: Kluwer Academic, 279–300.
Gittins, R., 1968, “Trend-Surface Analysis of Ecological Data,” The Journal of Ecol-ogy, 56(3):845–869.
Gustafson, E. J., 1998, “Quantifying Landscape Spatial Pattern: What is the State ofthe Art?,” Ecosystems, 1:143–156.
Ioffe, G., Nefedova, T., and I. Zaslavsky, 2004, “From Spatial Continuity to Fragmen-tation: the Case of Russian Farming,” Annals of the Association of AmericanGeographers, 94:913–943.
Jackson, J., 2002, “Urban Sprawl,” Urbanismus a územní rozvoj (Urbanism andUrban Development), 5(6):21–28.
Jensen, J. R., 2004, Introductory Digital Image Processing, 3rd ed., Upper SaddleRiver, NJ: Pearson Prentice Hall, 544 p.
Kozak, J., 2003, “Forest Cover Change in the Western Carpathians in the Past 180Years—a Case Study in the Orawa Region in Poland,” Mountain Research andDevelopment, 23:369–375.
Kreja, K., 2004, “Changes in Spatial Patterns of Urban Consumption in Post-SocialistCities: New Large-Scale Retail Development in Warsaw,” paper presented at theConference: Winds of Societal Change: Remaking Post-Communist Cities,University of Illinois, Urbana–Champaign, June 18–19, 2004.
Kučera, T., 1999, “Monitorování změn vegetace s využitím družicových snímků(Monitoring of Vegetation/Land Cover Change with Remote Sensing Data),”Zprávy České Bot. Společ, Praha (Reports of the Czech Botanical Association,Prague), 34(17):141–151.
Kuemmerle, T., Hostert, P., Damm, A. and V. Radeloff, 2005, “Cross-border Monitor-ing of Landscape Dynamics in Eastern Europe,” Göttinger GeographischeAbhandlungen, 113:24–30.
Kuemmerle, T., Radeloff, V. C., Perzanowski, K., and P. Hostert, P., 2006, “Cross-border Comparison of Land Cover and Landscape Pattern in Eastern EuropeUsing a Hybrid Classification Technique,” Remote Sensing of Environment,103:449–464.
McCloy, K. R., 2006, Resource Management Information Systems: Remote Sensing,GIS and Modeling, second ed., Boca Raton, FL: CRC Press, 575 p.
Neuhäuslová, Z., et al., 1998, Map of Potential Natural Vegetation of the CzechRepublic, Prague, Czech Republic: Academia, 341 p.
Palang, H., Mander, U., and A. Luud, 1998, “Landscape Diversity Changes in Esto-nia,” Landscape and Urban Planning, 41:163–169.
LAND USE/LAND COVER CHANGES 75
Ptáček, P., 1998, “Suburbanizace—měnící se tvář zázemí velkoměst (Suburbaniza-tion—the Changing Face of the Metropolitan Hinterland),” Geografické rozhledy(Geographical Views), 7(5):134–137.
Ptáček, P., 2000, “Networking and Local Culture: Local Community Transformationafter 1989 on the Example of Olomouc, Czech Republic,” Acta UniversitatisPalackianae–Geographica, 36:59–64.
Rogan, J. and D. Chen, 2004, “Remote Sensing Technology for Mapping and Moni-toring Land Cover and Land-use Change,” Progress in Planning, 61:301–325.
Rogan, J., Miller, J., Stow, D., Franklin, J., and L. C. Levien, 2003, “Land-CoverChange Monitoring with Classification Trees Using Landsat TM and AncillaryData,” Photogrammetric Engineering & Remote Sensing, 69(7):793–804.
Sabates-Wheeler, R., 2002, “Consolidation Initiatives after Land Reform: Responsesto Multiple Dimensions of Land Fragmentation in Eastern European Agricul-ture,” Journal of International Development, 14:1005–1018.
Šarapatka, B., 2006, “Permanent Grassland in Organic Farming and its Productiveand Non-productive Functions,” in Proceedings of Bioacademy 2006 Confer-ence, 6th European Summer Academy on Organic Farming.
Silva, L. L., 2006, A Stratified Random Sampling Strategy to Measure the Land CoverMap Accuracy in Minas Gerais, Brazil, Master’s thesis, International Develop-ment, Community and Environment, Clark University, Worcester, MA, 22 pp.
Šíma, M., 1995, “Studium postupných změn lesního krytu Krkonoš pomocí dat Land-sat MSS a TM 1979 (Study of Gradual Changes in Forest Cover in the GiantMountains Using Landsat 6 MSS and TM 1979),” in Sborn. ref. IV. konfer. odálk. průzk. (Proceedings of the IVth conference on remote sensing), Brno, CzechRepublic, 26–32.
Škapec, L., Pauknerová, E., and P. van der Horst, P., “Geografický informační systéma management území národních parků (Geographic Information Systems andNational Parks Management),” Ochrana přírody (Nature Conservation),(49):14–19.
Smits, P. C., Dellepiane, S. G., and R. A. Schowengerdt, 1999, “Quality Assessmentof Image Classification Algorithms for Land-Cover Mapping: A Review andProposal for a Cost-Based Approach," International Journal of Remote Sensing,20(8):1461–1486.
Stoklasa, M., 1995, “Informační systém zdravotního stavu lesů z kosmických snímkůLANDSAT-TM (Information System on the Health of Forests from LANDSAT-TM Images),” in Sborn. konfer. GIS ve státní správě (Conference Proceedings—GIS in State Administration), Chrudim, Czech Republic.
Takács-György, K., Bandlerova, A. and Sadowski, A., 2007, “Land Use and LandReform in Former Central and East European Countries,’ in Proceedings, 104th(joint) EAAE-IAAE Seminar on Agricultural Economics and Transition: “WhatWas Expected, What We Observed, the Lessons Learned, September 6–8, 2007,Budapest, Hungary.
ÚHÚL (Ústav pro hospodářskou úpravu lesa [Forest Management Institute]), 2006,Report on the State of Forests and Forestry in the Czech Republic 2006, Brandýsnad Labem, Czech Republic: ÚHÚL, 128 p.
76 VÁCLAVÍK AND ROGAN
Václavík, T., 2004, The Use of GIS in Ecological Planning (A Case Study of MountDesert Island), Master’s thesis, Department of Ecology and EnvironmentalSciences, Palacky University, Olomouc, Czech Republic, 82 p.
van Dijk, T., 2003, “Scenarios of Central European Land Fragmentation,” Land UsePolicy, 20:49–158.
Wu, W. and G. Shao, 2002, “Optimal Combinations of Data, Classifiers, andSampling Methods for Accurate Characterization of Deforestation,” CanadianJournal of Remote Sensing, 28(4):601–609.
Zemek, F., Heřman, M., Mašková, Z., and J. Květ, 2005, “Multifunctional LandUse—a Chance of Resettling Abandoned Landscapes? (A case study of the Zhůříterritory, the Czech Republic),” Ecology, 24(1): 96–108.
Zhao, G. and A. Maclean, 2000, “A Comparison of Canonical Discriminant Analysisand Principal Component Analysis for Spectral Transformation,” Photogrammet-ric Engineering & Remote Sensing, 66(7):841–847.
