Journal of Environmental Management 225 (2018) 160–167

Contents lists available at ScienceDirect

Journal of Environmental Management

journal homepage: www.elsevier.com/locate/jenvman

Research article

Markov chains and cellular automata to predict environments subject to

T
desertification
a a a
Kelly de Oliveira Barros , Carlos Antonio Alvares Soares Ribeiro , Gustavo Eduardo Marcatti ,
Alexandre Simões Lorenzon , Nero Lemos Martins de Castro , Getulio Fonseca Domingues , ∗
a a a

a Federal University of Viçosa/UFV, Av. Peter Henry Rolfs; s/n, 36570-000, Viçosa, MG, Brazil
b Federal University of Espírito Santo/UFES, Center of Agrarian Sciences and Engineering, Alto Universitário; s/n, 29500-000, Alegre, ES, Brazil

ARTICLE INFO ABSTRACT

Keywords: The foremost objective of this study was to analyze the performance of a Markov chain/cellular automata model for predicting
Prognosis land use/land cover changes in environments predisposed to desertification. The study area is the Vieira river basin, located in
Landscape Montes Claros (MG, Brazil). Land use/land cover prognosis was performed for the year 2005 so that this result could be
Cellular automata compared with the ranked image for the same year, taken as ground truth. Kappa indices were used to evaluate the change
Markov chain
level that occurred between these two cases. Results from cellular automata were evaluated from those of the Markov chain
Semiarid
model. The latter proved to be efficient in the quantitative prediction of changes in land use/land cover. Regarding the cellular
Degradation
automata, an average per-formance was noted in the spatial distribution of classes. Specifically, with regard to desertification,
the use of the CA-Markov model was effective at estimating the total area of the most susceptible class to this process, Bare
Soil; however, it was inefficient in its spatialization. Even with the caveats related to the performance of cellular automata, the
overall prediction capacity of CA-Markov models can be considered as good.

1. Introduction
Desertification is a process of land degradation, limited to dry re-gions. Its
possible incidence is restricted to arid, semiarid, or dry sub-humid areas.
Among several factors considered as causes of desertifi-cation, climate
change and anthropic activities are highlighted (Brasil, 1998). Recognized as
one of the main causes of desertification, several human practices have
exceeded the productive land capacity, char-acterized by inappropriate soil
management, including grazing, ero-sion, salinization, continuous use of fire,
extractivism, mining, and deforestation (Adamo and Crews-Meyer, 2006;
D'Angelo et al., 2000; Huang et al., 2009; Mabbutt, 1992; McClure, 1998;
Rhodes, 1991; Santini et al., 2010).
Desertification affects over 100 countries and is present on all continents,
especially in arid and semiarid regions that occupy ap-proximately one-third
of the land surface and are subject to the oc-currence of this process. It is
worth mentioning that one-sixth of the
world's population lives in these areas (Matallo Jr, 2001). Despite these
alarming numbers, there is a surprising lack of consensus on the ap-propriate
form of evaluating desertification (Verón et al., 2006).
Owing to the intrinsically spatial nature, studies on desertification have
benefited directly from recent advancements in the areas of re-mote sensing
and geographical information systems (GIS). Using these techniques, the CA-
Markov model combines the Markov chain and cellular automata (CA),
where the probabilities of possible land use changes supplied by a Markov
chain are spatialized by CA. Several authors studied desertification employing
the CA-Markov model and considered this model as viable and efficient in
predicting trends con-cerning the behavior of this land degradation process
(Huo-ping et al., 2009; Lu et al., 2009; Muller and Middleton, 1994; Wu et
al., 2010). Recent studies highlight that the association between Markov chain
and cellular automata makes it possible to predict the probability of trends in
land use evolution, such as the phenomenon of desertification (Jiang and Lin,
2018; Joseph et al., 2018; Varghese and Singh, 2016; Xu et al.,

∗
Corresponding author. Federal University of Espírito Santo/UFES, Center of Agrarian Sciences and Engineering, Alto Universitário; s/n, 29500-000, Alegre, ES. Brazil.

E-mail addresses: kellyobarros@yahoo.com.br (K. de Oliveira Barros), cribeiro@ufv.br (C.A. Alvares Soares Ribeiro), gustavomarcatti@gmail.com (G.E. Marcatti),
alexandre.lorenzon@ufv.br (A.S. Lorenzon), nerolemos@yahoo.com.br (N.L. Martins de Castro), getulio.floresta@gmail.com (G.F. Domingues), jromario_carvalho@hotmail.com (J.
Romário de Carvalho), alexandre.santos@pq.cnpq.br (A. Rosa dos Santos).

https://doi.org/10.1016/j.jenvman.2018.07.064
Received 13 April 2018; Received in revised form 20 June 2018; Accepted 18 July 2018
0301-4797/ © 2018 Elsevier Ltd. All rights reserved.
K. de Oliveira Barros et al.
Fig. 1. Location of vieira river basin in Montes Claros (MG, Brazil).
2016, 2014).
Given the climatic delimitation of Brazil, the northeast region and the
north of Minas Gerais state are considered susceptible to desertifi-cation.
These are the semiarid and dry sub-humid areas of the country; therefore, they
are prone to the occurrence of this degradation process (Brasil, 2005, 1998).
Hence, the foremost objective of this study was to analyze the performance of
a CA-Markov model for predicting land use/ land cover changes in
environments predisposed to desertification.
2. Material and methods
2.1. Study area
The Vieira river basin is located in the municipality of Montes Claros in
the northern region of Minas Gerais state, Brazil. The study area covers
2
approximately 580 km and is located at 43° 54′ 38″ long-itude and 16° 32′
52” latitude (Fig. 1).
The climatic ranking of integrating municipalities of dry sub-humid areas
in the country includes the municipality of Montes Claros; con-sequently,
according to the desertification concept, this municipality is also considered
susceptible to the occurrence of this process (Brasil, 1998, 2010).
The adopted methodological processes are shown in the flowchart (Fig. 2)
and described below.
2.2. Study area delimitation
®
The Vieira river basin was delimited using ArcGis software, from a
dataset (1° × 1°) of the Shuttle Radar Topography Mission (SRTM) base
(NASA, 2009). To eliminate existing spurious depressions in these da-tasets,
we used the Fill command. Next, we applied the Flow direction command to
the filled out digital elevation model (DEM) to obtain the drainage direction
for each cell. Then, the Flow Accumulation command was used to determine
the flow accumulated, which corresponds to the number of cells situated
upstream of a certain cell. Finally, the Vieira river basin was delimited by
employing the Watershed command, from an interactively defined point,
corresponding to the place in which the Vieira river flows into the Verde
Grande river. All these tools are in the ArcToolbox, Spatial Analyst option, in
® 0.01–0.20 is considered bad, 0.21–0.40 is considered reasonable,
the Hydrology set of ArcGis software.
161
Journal of Environmental Management 225 (2018) 160–167
2.3. Land use and occupation
To determine the land use/land cover of the Vieira river basin, we selected
two images from the LANDSAT-5 satellite, TM sensor, and orbit/point
218/72 (NASA Landsat Program, 2012). Dated September 9, 1995,
September 22, 2000, and July 18, 2005, these images have 30 m cells.
Bands 3, 4, and 5 were used for RGB composition of images. These
bands, which originally had geographical coordinates and World Geodetic
System (WGS-84) datum, were then projected to the Universal Transverse
Mercator (UTM) coordinate system, Zone 23 S, and hor-izontal datum
SIRGAS 2000.
For co-registration of images, another image from the LANDSAT-5
satellite, orthorectified, from June 26, 1997, was also adopted for same scene.
This orthorectified image presented better quality georeferen-cing when
compared with the others and was taken as a reference to adjust the other four
images. Similar to the other images, bands 3, 4, and 5, RGB composition, and
UTM coordinates system, datum SIRGAS 2000, and Zone 23 S were also
used. The tool used for co-registration was Georeferencing in ArcGis .
®
Next, the Vieira river basin was extracted from LANDSAT compo-sitions
and used as a mask, using the Extract by Mask tool of ArcToolbox. These
new images were then submitted to a supervised classification. Using the
toolbar Image classification, training samples were collected in each
previously determined class: Urban Area, Forest, Cropland/Pasture, and Bare
Soil. Note that for Forest, we considered natural vegetation and eucalyptus
culture areas. Maximum likelihood estimation was adopted for the supervised
ranking, which performs a pixel-by-pixel ranking at a 5% probability level.
Two other maps were adopted as reference for the ranking. The first is in
the study carried out by Leite et al. (2011), where a land use map was
prepared for the same area of the Vieira river basin. The second was the
native land cover and reforestation mapping carried out in 2005 by the State
Institute of Forests (IEF, MG) in partnership with the Federal University of
Lavras (UFLA) (Scolforo and Carvalho, 2006).
The accuracy of rankings was evaluated by Kappa index, using
®
training samples, in ArcGIS . Kappa can be qualitatively evaluated ac-
cording to the ranking proposed by Landis and Koch (1977), where a
Kappa coefficient lower than 0.00 is considered awful quality,
0.41–0.60 is considered good, 0.61–0.80 is considered very good, and
K. de Oliveira Barros et al.
Fig. 2. Methodological flowchart of the steps performed in the study area.
0.81–1.00 is considered excellent.
After ranking and evaluation of the accuracy, images were used as input
data for the projection of land use/land cover and, subsequently, to associate
the behavior of each class with susceptibility to desertifi-cation. This stage
®
was performed in the Idrisi software.
2.4. Markov chain and cellular automata
To employ the Markov chain, a pair of images with properly ranked dates
is necessary. The year to which the scenery will be projected corresponds to
the period between the two images in the study added to the corresponding
year to the second image. Images of years 1995 and 2000 were selected to
apply the Markov chain. Since the year of pro-jection for this combination is
2005, it is possible to evaluate the ob-tained result by comparing it with the
ranked image of 2005, taken as ground truth.
Given that the Markov chain presents only the transition prob-abilities
among the different soil use classes, geographical reference was added to the
process using cellular automata (CA). The CA has a
162
Journal of Environmental Management 225 (2018) 160–167
contiguity filter that applies a spatial weighting factor to each class, so that the
spatial dependence between data is considered properly. This filter will
indicate higher or lower aptitude of the territory for each soil class in the
study. The filter used was the 5 × 5 pattern of the CA-Markov model.
The Kappa indices were created using ranked images of 1995 and 2000.
®
We used the CrossTab tool in ArcGIS , which supplies coefficients that
indicate the land use change level between two dates.
After obtaining the Markov chain results, the CA-Markov model was
applied. Faced with the absence of more significant information on the forces
driving the evolution of land use/land cover, an isotropic cost surface was
adopted, assigning a value of 1 to all cells of the Vieira river basin. The
subjectivity and complexity naturally associated to the se-lection process of
factors and assignment of weights to describe their relative importance were
removed. The decisive rule of land use/land cover was based only on the
transition probability produced by the Markov chain, considering uniform
properties irrespective of direction or land use.
K. de Oliveira Barros et al.
2.5. CA-Markov model validation
To validate the result obtained in the CA-Markov process, cross
®
tabulation was performed in ArcGIS software. Since this procedure indicates
the coincidence between the classes of two images, it was possible to evaluate
the result obtained by the Markov chain by cal-culating the Kappa index
produced between the two dates. The image obtained in this process in the
combination of years 1995 and 2000, in other words, the projected image for
2005, and the image ranked with properly standardized numbers of each class
with a date of 2005, was used. This procedure was repeated with the projected
image of 2005 and the ranked image of 2000.
To validate the result obtained by CA, a class-by-class comparison was
performed between the ranked images of 2005 and the image ob-tained by the
CA-Markov model, verifying the hits of the CA-Markov image for each case.
It was possible to evaluate the CA performance in the spatial reference
determination for each class.
2.6. Susceptibility to desertification relation with land use/land cover
classes
A susceptibility relation was established between the defined classes and
occurrence of desertification in which the Bare Soil class was in-dicated as
that presenting highest susceptibility to this process, fol-lowed by the
Cropland/Pasture class, and then Forest and Urban Area classes.
3. Results and discussion
3.1. Image ranking
It was possible to notice in the three dates in this study that Forest is the
most representative class in the Vieira river basin, followed by
Cropland/Pasture, Urban Area, and lastly by Bare Soil (Table 1 and Fig. 3).
The Kappa index was calculated, and the obtained value cor-responds to 0.89,
0.88, and 0.91 for years 1995, 2000, and 2005. Ac-cording to Landis and
Koch (1977), such values reveal an excellent quality ranking, with great
coincidence between the ground truth and the ranked image.
3.2. Markov chain
Transition probabilities among land use classes for 2005 are in listed in
Table 2.
It was noted that all classes presented a higher probability of re-maining
unchanged. When comparing the probability of the Urban Area class altering
to other classes, it was possible to notice that there was a higher trend of this
class turning into Forest (20%). City tree planting or even urban horizontal
expansion of the municipality, with frequent land covered backyards, can also
justify the probability found for this transformation between Urban Area and
Forest being the second most representative (França and Soares, 2007). The
probability of Urban
Table 1
Area of each land use class for dates of Sept. 9, 1995; Sept. 22, 2000, and Jul.
18, 2005, in the Vieira River basin, Montes Claros (MG, Brazil).
Classes Dates
09/09/1995 09/22/2000 07/18/2005
2
Area (km )
Urban Area 58 65
Forest 280 308
Cropland/Pasture 194 155
Bare Soil 48 52
Journal of Environmental Management 225 (2018) 160–167
Area changing into Cropland/Pasture was 10%. In the ranking, it was possible
to notice a higher confusion of the rater regarding some shadow pixels or
even Bare Soil, which in the RGB composition pre-sented a gray shade,
probably with similar spectral behavior to that of the pixels indicated by the
sample as Urban Area. Thus, this value can be justified by the confusion that
existed in the ranking.
For Bare Soil, it was observed that the value indicated as probability for
this class continuing unchanged is quite close to that indicated as this class
changing into Cropland/Pasture. Therefore, it was possible to confirm the
hypothesis that part of the areas identified as Bare Soil were simply
momentarily without any agriculture and the identified Bare Soil is not
necessarily degraded soil; in other words, it could not be used for agriculture
or rangeland or even that vegetation could not re-generate. The change
probability from Bare Soil to Forest was very small when compared with the
other three possibilities, that is, only 5%. The Bare Soil and Forest classes can
be considered opposite si-tuations when associated with degradation. This
low value of Bare Soil changing into Forest indicates the precarious capacity
of recovery of the area without anthropic intervention, and consequently, a
higher prob-ability of these areas, which are, in fact, Bare Soil, remaining as
such; in other words, a higher inclination for desertification. In the case of the
transition possibility between Bare Soil and Urban Area, the value of 20%
was observed, which can be justified by the higher use of Bare Soil areas
close to the Urban Area.
Among the factors that may be acting directly in the transition from
pasture to degraded soils, we can highlight the trampling of the animals and
the lack of adequate pasture management, since in Brazil the an-imal
production system is extensive with free animals in the pasture (Vzzotto et al.,
2000). These factors contribute to changes in soil phy-sical properties that
will lead to problems of penetration of the plant root system and infiltration of
water into the soil, which can cause surface runoff and soil surface loss of
organic matter (Tollner et al., 1990; Vzzotto et al., 2000). This can lead to a
shortage of nutrients in the soil and, consequently, affect the physiology of
the plant, leading to death and causing bare soil (Tollner et al., 1990; Torres et
al., 2012).
The Forest class presented the highest probability of remaining unchanged
when compared with the other possible changes among classes. The highest
revealed tendency was for vegetation changing to Cropland/Pasture followed
by Urban Area and Bare Soil, respectively. Such a fact reveals that not only
were humans a transforming agent of the landscape but also anthropic
pressure is, in great part, responsible for the suppression of land cover, which
has intensified over the years mainly because of increased agriculture.
When comparing the chances of Urban Area and Forest classes suffering
alterations, it is considered that Urban Area presents a higher probability for
smaller alterations, given that the land cover can suffer modifications faster
due to climatic events, fires, or any human actions. Despite Urban Area
having presented a high value of remaining un-changed in every year of
analysis, it was not superior to the value presented by Forest at any point. The
presented values, in these two cases, can possibly be explained by the
characteristics of these two classes in the images not ranked yet. In the case of
Forest, it is possible to consider this class with lower degree of variability
among its pixel values. In contrast, in the case of Urban Area, the pixels
present con-siderable variation in values in its corresponding area. This pixel
value variation is associated with variability itself of what composes an urban
area: construction, asphalt, vegetation, water, and bare soil. The pos-sibility of
transition from Forest to Urban Area was only not lower than the change
probability from Forest to Bare Soil. These low values are also justified by the
low occurrence of Forest in the surroundings of Urban Area, where Bare Soil
and Cropland/Pasture classes prevail around the urban spot. Values of
possible transformations from the Cropland/Pasture class to Urban Area and
from Bare Soil to Urban Area were superior when the Forest probability was
80 changed to Urban Area. When compared with Cropland/Pasture, Bare Soil
295 had a higher prob-ability of changing into Urban Area, and such values are
146 justified by the
59
163
K. de Oliveira Barros et al. Journal of Environmental Management 225 (2018) 160–167

Fig. 3. Supervised ranking of the Vieira river basin, Montes Claros (MG, Brazil) for dates: September 9, 1995 (A); September 22, 2000 (B); and July 18, 2005 (C).

Table 2 Comparing the values produced by the Markov chain projection and by
Transition probabilities (%) among land use classes for year 2005 in Vieira Rver basin, the ranking (Table 3), a higher proximity of Cropland/Pasture and Bare Soil
Montes Claros (MG, Brazil), based in years 1995 and 2000. 2
class values was observed, with a difference of 6 km for the first class and 8
2
2000
km for the second class. In the case of the Forest class, this difference
2
between two areas was higher at 26 km . Among classes, Urban Area
2
Urban Area Forest Cropland/Pasture Bare Soil
presented an intermediary difference of 12 km . However, despite these
differences, the Kappa between these two cases was 0.45, which in the
ranking of Landis and Koch (1977) indicates good quality. This Kappa
Projection for 2005 ranking confirms the good prediction performed by the Markov chain
1995 concerning probabilities of each land use/land cover class. Owing to the slight
difference between the result presented by the Markov chain and the ranking,
it is possible to confirm that change probabilities in land use were not, in fact,
Urban Area 61 20 10 8 constant along the years. Nevertheless, even when Markov chain determined
Forest 2 83 11 1 the stationarity of probabilities and this did not occur, the small difference
Cropland/Pasture 5 30 52 11 between its prediction and the ranking provides this part of the CA-Markov
model with good performance.
Bare Soil 20 5 32 41

tendency of reduction of voids existing within and around city limits, making
this area more densely occupied. This was observed visually on the classified
images when compared with the most recent classification from 2005 (Fig.
3C).
The Forest class presented a lower change probability to Bare Soil. It is
necessary to consider that in this ranking, the type of land cover was not
specified, and as observed in the land cover mapping prepared by the IEF and
UFLA, deciduous forest occurs in the Vieira river basin, in other words, forest
in which trees lose most of their leaves in periods of drought (Carvalho and
Felfili, 2011), providing less land cover to the soil during this period of
climatic severity. Since a single class was determined, irrespective of
vegetation type and its characteristics, the demonstrated probability is
considered acceptable, which points to a great deal of soil protection by the
vegetation, regardless of specificity.
The second highest probability after Forest not changing in any other
class was the transition to Cropland/Pasture. Despite values showing the
pressure that these activities apply on land cover, the probability value of the
Cropland/Pasture class changing into Forest was higher than the opposite
situation. This situation shows a higher probability that cropland and
rangeland areas develop vegetation, probably due to issues around areas being
deserted than to the reverse situation, which would correspond to pressure on
this last class.
3.3. CA and CA-Markov model validation
When comparing Fig. 3C, from 2005 taken as ground truth, with Fig. 4,
the result of the projection for the same year, it can be noticed that the rater
confusion regarding some pixels of Bare Soil or shadow, classifying them as
Urban Area, was reduced, although not removed completely. This observation
is associated with the role of the CA of the model that, due to the tendency to
expand border spots, irrespective of class, ended up reducing this rater
confusion since it was a matter of small groups or even individual pixels. This
behavior of the CA is si-milar to the natural behavior of the environment, in
which there is a higher change probability in transition areas of classes than
when compared with change in the interior of a spot of a determined class
distant from the border. Hence, in the visible comparison between the image
produced by CA (Fig. 4) and the ranking image of 2005 (Fig. 3C), Bare Soil
areas, in parts already existing in the image of 2000 (Fig. 3B), became larger
instead of being distributed in smaller areas, but more scattered in the Vieira
river basin, as noticed in the ranking image of 2005 (Fig. 3C).
When the image produced by CA (Fig. 4) was compared with the ranked
image of 2000 (Fig. 3B), it was possible to notice higher simi-larity between
these images than between CA images (Fig. 4) and the 2005 ranking (Fig.
3C). Thus, the Kappa was produced between the ranked image of 2000 and
the projection for 2005. The obtained Kappa was 0.50, which is higher than
the Kappa produced between the CA result and the 2005 ranking (0.45).
These values show that, in fact, there is a greater similarity between the CA
result and the 2000 ranking than with the 2005 ranking. This fact can be
justified because the year

Table 3 demonstrates values obtained for each class by the Markov chain
with projection for 2005 and for ranking in the same year. Fig. 4 shows
spatial distribution of land use/land cover classes of the Vieira river basin for
2005 by the projection performed by the CA-Markov model.
164
K. de Oliveira Barros et al. Journal of Environmental Management 225 (2018) 160–167

Fig. 4. Projection of land use/land cover classes for 2005 for the Vieira river basin, Montes Claros (MG, Brazil).

Table 3 3.4. Susceptibility to desertification and land use/land cover classes in the
Areas of land use/land cover classes obtained by Markov chain projected for year 2005 Vieira river basin
and for supervised image ranking of the same year.

2 The existence of a narrow relationship between desertification and soil is

Area (km )
known in which the use given to this resource, when in regions of arid,
semiarid or dry sub-humid climate, influences its original patterns; in
Markov Chain Ranking Difference (%) addition, the reduction or loss of fertility as well as salinization may occur
Classes (Souza et al., 2009). Wind or water erosion, deterioration of soil properties,
either physical, chemical or biological character, and re-moval of vegetation
for long periods are among the possible causes of desertification directly
Urban Area 68 80 15
associated with soil (Souza et al., 2009; UNCCD, 1992). Hence, when bare
Forest 321 295 8
soil occurs in dry lands, a significant potential for the development of
Cropland/Pasture 140 146 4 desertification is assumed for this class. The expansion of these areas was
Bare Soil 51 59 13 considered by Guerra et al. (2010) as an aid in the identification of the
desertification process in their study area, in Jaguaribe, Ceará, Brazil.

Bare Soil in the Vieira river basin presented higher susceptibility to

2000 was used as one of the input years for the Markov chain, and it was last
desertification among the classes and demonstrated growth in the analyzed
chronologically, had the greatest influence on the result. De-spite the
time interval from 1995 to 2005. This was similarly reflected in the performed
proximity of these values, while evaluating the spatial re-ference marked by
projection, in which for all cases, this class maintained continuous growth.
the CA, it is possible to notice that some classes presented a low hit value
When the soil is deprived of land cover for a pro-longed time, erosive
when comparing images of this method and with those from the 2005 ranking,
processes to property alterations are subject to occur to such an extent that a
class by class.
level of degradation is achieved. The suppression of vegetation for farming
Forest had higher coincidence between the CA result and the 2005
activities, when carried out with a lack of soil conservation techniques, will
ranking (83%), followed by Urban Area (56%), and Cropland/Pasture (49%).
lead to the appearance of bare soil and may even accelerate the appearance of
The worst hit level was for Bare Soil, which was only 30% (Table 4). It can
such areas. Such activities, while affecting the fertility of this resource, may
be observed that the role of the CA regarding class spot size, in which higher
also induce an exodus of the rural population. This situation of land cover
size spot tends to expand over the others. Thus, the Forest class presented
reduction and the appearance of new bare soil areas were also noticed in this
several great spots, followed by Urban Area, which despite also presenting
study area, with scenery that shows the cause-and-consequence re-lationship
small spots has a concentrated urban spot of significant size. The spots for
of desertification (Pachêco et al., 2006; Sousa et al., 2008).
Cropland/Pasture and Bare Soil classes are similar in being well distributed
all over the basin area and also for having varied sizes. However, in most
It is assumed that susceptibility to desertification in the Vieira river basin
cases, the Cropland/Pasture class presents larger spots when compared with
has grown over time, given that the most sensitive class to the process
Bare Soil spots (Fig. 4).
presented a continuous growth trend. It must still be empha-sized that despite
the decline in agriculture and cattle raising activities, in the case of the study
Table 4 area, this increase was always associated with a reduction in vegetation area
CA performance for each land use/land cover class compared with the 2005 ranking reduction and increase in bare soil. Nevertheless, despite a reduction of the
image. Cropland/Pasture class occur-ring in the years 2000 and 2005, this decline
2 2 was not directly propor-tional to Forest growth, as the growth of Bare Soil
Class Area (km )/Ranking Area (km )/Cellular Hits (%)
class Automata class and Urban Area

Urban Area 80 45 56
Forest 295 246 83
Cropland/Pasture 146 71 49
Bare Soil 59 18 30
165
K. de Oliveira Barros et al.
classes was involved in this process. Although for the Vieira river basin
urbanization presented only a small change in vegetation areas, this value
must not be neglected.
3.5. Evaluation of analytical tools
The CA-Markov prediction result should be analyzed separately. The
Markov chain performance in this model with regard to susceptibility to
desertification, in other words focusing on Bare Soil, was quite close in area
2 mental hazards in Jáchal, Argentina. Appl. Geogr. 26, 61–85. https://doi.org/10.
to the 2005 ranking, 51 and 59 km . Even with a reduction of this class, the
Markov chain managed to satisfactorily predict areas that have higher
susceptibility to desertification. However, when the CA performance was
evaluated, mainly associated with Bare Soil, poor performance was noted
compared with the 2005 ranking: the CA result indicated only 30% of
coincidence between areas that are in fact Bare Soil in 2005. From the
desertification point of view, it is possible to confirm that the situation pointed
by CA is more serious than that re-vealed by the ranking. It was noticed that
in the prediction (Fig. 4), Bare Soil classes, despite having a slight
distribution reduction in the study area, increased in size. Conversely, in the
case of the ranked image (Fig. 3C), Bare Soil had a scattered distribution in
the Vieira river basin and in several smaller areas. In an attempt to promote
the re-covery of a desertified area, it is assumed that combating a larger area,
in other words, with a more intense degradation nucleus, is more dif-ficult,
from the physical, biological, and economic points of view.
The climatic ranking for Montes Claros by Brasil (2004) is dry sub-humid
climate, which, in the susceptibility level to desertification, from a strictly
climatic point of view, is a level of lower susceptibility. However, as it can be
noticed, human activities in the Vieira river basin can be considered as largely
responsible for the possible occurrence of desertification, far exceeding the
capacity of the climate alone to cause desertification of the study area.
With the obtained results, we hope to contribute to the advance-ment of
investigations on desertification. The use of geoprocessing tools must be
emphasized as a great ally in the generation of databases for subsequent
decision-making. The information produced in this study for the Vieira river
basin may help in the elaboration of land use planning projects, either aiming
at preservation or conservation of susceptible areas to desertification or also
recovery or restoration of already desertifield areas.
In addition, the CA-Markov model is an interesting tool that can be used
for desertification studies and can be used in different regions of the globe. In
addition, it can serve as an indicator for decision-making to take mitigating
measures for the advancement of desertification.
4. Conclusion
The present study made it possible to evaluate the efficiency of the use of
the Markov chain and CA in the prediction of zones of suscept-ibility to
desertification. In this context, we verified that the CA model should be
evaluated separately from the Markov chain results. The latter proved to be
efficient in the quantitative prediction of changes in land use. In relation to
CA, an average performance was observed in the spatial distribution of the
classes involved.
Regarding desertification, specifically, the use of the CA-Markov model
was efficient in estimating the total area of the class with greater
susceptibility to this process, Bare Soil; however, it was inefficient in its
spatialization. Even with CA performance limitations, CA-Markov's overall
predictive capacity can be considered as a good starting point for studies
aiming to monitor areas susceptible to the desertification process and the
advancement of desert areas, regardless of the global region.
Acknowledgments
The authors are grateful to the research group Geotechnology
166
Journal of Environmental Management 225 (2018) 160–167
Applied to Global Environment (GAGEN), to the National Council for
Scientific and Technological Development (CNPq) and the Postgraduate
Program in Forest Sciences of the Federal University of Viçosa for providing
the necessary resources for implementation and execution of this work.
References
Adamo, S.B., Crews-Meyer, K.A., 2006. Aridity and desertification: exploring environ-
1016/j.apgeog.2005.09.001.
Brasil, 2010. Ministério Do Meio Ambiente - MMA, Secretaria Dos Recursos Hidricos -
SRH, Programa De Ação Estadual De Combate À Desertificação E Mitigação Dos
Efeitos Da Seca – Pae-ce, Governo Do Estado De Ceará, Srh. Ministério do Meio
Ambiente/Secretaria de Recursos Hídricos, Brasília.
Brasil, 2005. Ministério da Integração Nacional - Nova delimitação do semiárido brasi-leiro.
Brasília MIN/Secretaria Políticas Desenvolv. Reg). .
Brasil, 1998. Ministério do Meio Ambiente dos Recursos Hídricos e da Amazônia Legal -
Convenção das Nações Unidas de combate a desertificação nos países afetados por seca
grave e/ou desertificação, particularmente na África, second ed. Plano Nacional de
Combate a Desertificação, Brasília.
Carvalho, F.A., Felfili, J.M., 2011. Variações temporais na comunidade arbórea de uma floresta
decidual sobre afloramentos calcários no Brasil Central: composição, estru-tura e
diversidade florística. Acta Bot. Bras. 25, 203–214. https://doi.org/10.1590/ S0102-
33062011000100024.
D'Angelo, M., Enne, G., Madrau, S., Percich, L., Previtali, F., Pulina, G., Zucca, C., 2000.
Mitigating land degradation in Mediterranean agro-silvo-pastoral systems: a GIS-based
approach. Catena 40, 37–49. https://doi.org/10.1016/S0341-8162(99)
00063-6.
França, I.S., de Soares, B.R., 2007. EXPANSÃO URBANA EM CIDADES MÉDIAS: uma
reflexão a partir do núcleo e da área central de Montes Claros no Norte de URBAN
EXPANSION IN MEDIUM CITIES: a reflection from the nucleos and central area of
Montes Claros in the North of Minas Gerais. Geo UERJ 2, 47–63.
Guerra, M.D.F., Souza, M.J.N., Lustosa, J.P.G., 2010. Desertificação em áreas semiáridas do
Nordeste brasileiro: o caso do município de Jaguaribe, Ceará. In: Rev. Geogr. VIII Simpó,
pp. 67–80.
Huang, Y., Wang, N., ang, He, T., Chen, H., Zhao, L., 2009. Historical desertification of the mu
Us desert, northern China: a multidisciplinary study. Geomorphology 110, 108–117.
https://doi.org/10.1016/j.geomorph.2009.03.020.
Huo-ping, D., Jian-ping, C., Gong-wen, W., 2009. 2009. A model for desertification evolution
employing GIS with cellular automata. In: Proceedings - 2009 International Conference on
Computer Modeling and Simulation, ICCMS. IEEE, pp. 324–328.
https://doi.org/10.1109/ICCMS.2009.39.
Jiang, M., Lin, Y., 2018. Desertification in the south Junggar Basin, 2000–2009: Part I.
Spatial analysis and indicator retrieval. Adv. Space Res. 62, 1–15. https://doi.org/10.
1016/J.ASR.2017.11.038.
Joseph, O., Gbenga, A.E., Langyit, D.G., 2018. Desertification risk analysis and assessment in
Northern Nigeria. Remote Sens. Appl. Soc. Environ. 11, 70–82. https://doi.org/10.
1016/J.RSASE.2018.04.012.
Landis, J.R., Koch, G.G., 1977. The measurement of observer agreement for categorical data.
Biometrics 33 (159). https://doi.org/10.2307/2529310.
Leite, M.E., Santos, I., de S. Almeida, J.W.L., 2011. Mudança de Uso do Solo na Bacia do Rio
Vieira. In: Montes Claros/MG. Rev. Bras. Geogr. Física, vol. 4. pp. 779–792.
Lu, L., Li, X., Veroustraete, F., Kang, E., Wang, J., 2009. Analyzing the forcing mechan-isms
for net primary productivity changes in the Heihe River Basin, north-west China. Int. J.
Rem. Sens. 30, 793–816. https://doi.org/ https://doi.org/10.1080/ 01431160802438530.
Mabbutt, J.A., 1992. Desertification Natural Background and Human Mismanagement, second
ed. Land Use Policy Springer-Verlag, Berlin/Heidelberg, Germany, pp. 90048–90052.
https://doi.org/10.1016/0264-vol. 8377(92.
Matallo Jr., H., 2001. INDICADORES DE DESERTIFICAÇÃO : histórico e perspectivas,
Unesco. UNESCO, Brasília.
McClure, B.C., 1998. Policies related to combating desertification in the United States of
America. Land Degrad. Dev. 9, 383–392 https://doi.org/10.1002/(SICI)1099-145X
(199809/10)9:5 < 383::AID-LDR303 > 3.0.CO;2-A.
Muller, M.R., Middleton, J., 1994. A Markov model of land-use change dynamics in the
Niagara Region. Landsc. Ecol. 9, 151–157.
NASA (National Aeronautics and Space Administration), 2009. Shuttle Radar Topography
Mission. U.S. Releases Enhanc. Shuttle L. Elev. Data. http://www2.jpl.nasa.gov/ srtm/.
NASA Landsat Program, 2012. LANDSAT-5, TM, Orthorectified. (accessed 6.17.12).
http://glcf.umd.edu/data/landsat/.
Pachêco, A.P., Freire, N.C.F., Borges, U.N., 2006. A transdisciplinaridade da deserti-
ficação. Geografia 15, 5–34.
Rhodes, S.L., 1991. Rethinking desertification: what do we know and what have we
learned? World Dev. 19, 1137–1143. https://doi.org/10.1016/0305-750X(91) 90062-
M.
Santini, M., Caccamo, G., Laurenti, A., Noce, S., Valentini, R., 2010. A multi-component GIS
framework for desertification risk assessment by an integrated index. Appl. Geogr. 30,
394–415. https://doi.org/10.1016/j.apgeog.2009.11.003.
Scolforo, J.R., Carvalho, L.M.T. de, 2006. Mapeamento e inventário da flora nativa e dos
reflorestamentos de Minas Gerais, second ed. Ufla. UFLA, Lavras (MG).
Sousa, R.F., de, Barbosa, M.P., Júnior, S.P. de S., Nery, A.R., Lima, A.N. de, 2008. Estudo
K. de Oliveira Barros et al.
da evolução espaço-temporal da cobertura vegetal do município de Boa Vista-Pb. In:
Utilizando Geoprocessamento. Rev. Caatinga 21. pp. 22–30.
Souza, B.I., Suertegaray, D.M.A., Lima, E.R.V. de, 2009. Desertificação e seus efeitos na
vegetação e solos do cariri paraibano (Desertification and its effects over the vege-tation
and soils of the cariri region of Paraíba – Brazil). Mercat 8 No 16 mai./ago.
https://doi.org/10.4215/RM2009.0816.0017.
Tollner, E.W., Calvert, G.V., Langdale, G., 1990. Animal trampling effects on soil physical
properties of two Southeastern U.S. ultisols. Agric. Ecosyst. Environ. 33, 75–87. Torres, J.L.R.,
Rodrigues Junior, D.J., Sene, G.A., Jaime, D.G., Vieira, D.M. da S., 2012. Resistência à
penetração em área de pastagem de capim tifton, influenciada pelo
pisoteio e irrigação. Biosci. J. 28, 232–239.
UNCCD (United Nations Convention to Combat Desertification), 1992. United Nations
Convention to Combat Desertification. (accessed 1.20.12). http://www2.icnf.pt/
portal/pn/biodiversidade/ei/unccd-PT/a-convencao/unccd.
Varghese, N., Singh, N.P., 2016. Linkages between land use changes, desertification and human
development in the Thar Desert Region of India. Land Use Pol. 51, 18–25.
167
Journal of Environmental Management 225 (2018) 160–167
https://doi.org/10.1016/J.LANDUSEPOL.2015.11.001.
Verón, S.R., Paruelo, J.M., Oesterheld, M., 2006. Assessing desertification. J. Arid
Environ. 66, 751–763. https://doi.org/10.1016/j.jaridenv.2006.01.021.
Vzzotto, V.R., Marchezan, E., Segabinazzi, T., 2000. Effect of cattle trampling on lowland soil
physical properties. Ciência Rural. 30, 965–969.
Wu, J., Liu, Y., Wang, J., He, T., 2010. Application of Hyperion data to land degradation
mapping in the Hengshan region of China. Int. J. Rem. Sens. 31, 5145 –5161. https://
doi.org/10.1080/01431160903302940.
Xu, D., Li, C., Song, X., Ren, H., 2014. The dynamics of desertification in the farming-pastoral
region of North China over the past 10 years and their relationship to climate change and
human activity. Catena 123, 11–22. https://doi.org/10.1016/J.CATENA. 2014.07.004.
Xu, D., Song, A., Tong, H., Ren, H., Hu, Y., Shao, Q., 2016. A spatial system dynamic model
for regional desertification simulation – a case study of Ordos, China. Environ. Model.
Software 83, 179–192. https://doi.org/10.1016/J.ENVSOFT.2016.05.017.