Using Remote Sensing Techniques to Assess Land-Cover Change and Degradation in the Deserts of the Southeast Iberian Peninsula

Abstract

Many drylands around the world have seen both soil and vegetation degradation around watering points. It can be seen in spaceborne imagery as radial brightness belts that fade with distance from the water areas. The study's primary goal was to characterize spatio-temporal land degradation/rehabilitation in the drylands of the southeast Iberian Peninsula. The brightness index of Tasseled Cap was discovered to be the best spectral transformation for enhancing the contrast between the bright-degraded areas near the points and the darker surrounding areas far from and in-between these areas. To comprehend the spatial structure present in spaceborne imagery of two desert sites and three key time periods, semi-variograms were created (mid-late 2000s, around 2015, and 2020). In order to assess spatio-temporal land-cover patterns, a geostatistical model (kriging) was used to smooth brightness index values extracted from 30 m spatial resolution images. To assess the direction and intensity of changes between study periods, a change detection analysis based on kriging prediction maps was performed. These findings were linked to the socioeconomic situation prior to and following the EU economic crisis. The study discovered that degradation occurred in some areas as a result of the region's agricultural activities being exploited.

Full text

Article
Using Remote Sensing Techniques to Assess Land-
Cover Change and Degradation in the Deserts of the
Southeast Iberian Peninsula
Emilio Ramírez-Juidias 1*, Antonio Madueño-Luna 2, José Miguel Madueño-Luna 3, Miguel
Calixto López-Gordillo 3 and Jorge Luis Leiva-Piedra 4
1 Instituto Universitario de Arquitectura y Ciencias de la Construcción (IUACC), Universidad de Sevilla, 2
Reina Mercedes Avenue, 41012 Seville, Spain
2 Departamento de Ingeniería Aeroespacial y Mecánica de Fluidos, Universidad de Sevilla, 41092 Seville,
Spain
3 Departamento de Ingeniería Gráfica, Universidad de Sevilla, 41092 Seville, Spain
4 Laboratorio de Investigación en Teledetección, Universidad Tecnológica del Perú, Intersección Avenida El
Progreso s/n-Vía de Evitamiento, Chiclayo 14001, Peru
* Correspondence: erjuidias@us.es
Abstract: Many drylands around the world have seen both soil and vegetation degradation around
watering points. It can be seen in spaceborne imagery as radial brightness belts that fade with
distance from the water areas. The study's primary goal was to characterize spatio-temporal land
degradation/rehabilitation in the drylands of the southeast Iberian Peninsula. The brightness index
of Tasseled Cap was discovered to be the best spectral transformation for enhancing the contrast
between the bright-degraded areas near the points and the darker surrounding areas far from and
in-between these areas. To comprehend the spatial structure present in spaceborne imagery of two
desert sites and three key time periods, semi-variograms were created (mid-late 2000s, around 2015,
and 2020). In order to assess spatio-temporal land-cover patterns, a geostatistical model (kriging)
was used to smooth brightness index values extracted from 30 m spatial resolution images. To assess
the direction and intensity of changes between study periods, a change detection analysis based on
kriging prediction maps was performed. These findings were linked to the socioeconomic situation
prior to and following the EU economic crisis. The study discovered that degradation occurred in
some areas as a result of the region's agricultural activities being exploited.
Keywords: Andalusia; remote sensing; desert of Tabernas; Sierra Alhamilla; Almería; mathematical
models
1. Introduction
Overgrazing, according to the United Nations Environment Program (UNEP), is the practice of
allowing a much greater number of animals to graze at a location than it can actually support. As a
result, in terms of plant density, plant chemical content, community structure, and soil erosion,
overgrazing by various types of livestock is perhaps the most significant anthropogenic activity that
degrades rangelands and causes desertification [1]. Overgrazing degrades approximately 75 million
ha of land globally, destroying the original biotic functions [2]. According to [3], Land (soil and
vegetation) degradation is particularly associated with areas surrounding natural or artificial water
sources, such as wells or boreholes, in arid and semi-arid environments.
During grazing, domestic animals tend to concentrate near watering points, with the
concentration gradually decreasing as distance from water increases [4-5]. Typically, grazing occurs
about 5 km of the watering point, but this can increase up to 20 km in extreme conditions [3, 6]. It is
also interesting to specify that the radial grazing pattern, known as the "piosphere", was named after
the Greek root "pios" meaning "drink" [3]. On the other hand, and according to [7], "grazing gradient"
refers to the spatial patterns in soil or vegetation resulting from grazing and can indicate land
Disclaimer/Publisher’s Note: The statements, opinions, and data contained in all publications are solely those of the individual author(s) and
contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting
from any ideas, methods, instructions, or products referred to in the content.
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 12 June 2023 doi:10.20944/preprints202306.0781.v1
© 2023 by the author(s). Distributed under a Creative Commons CC BY license.

degradation. For this reason, most of the drylands used for pastoralism show degraded features
caused by grazing activities.
Studies conducted using ground measurements and/or analysis of data obtained from remote
sources have been conducted to observe the changes in grazing effects on rangelands. A variety of
biotic, abiotic, and environmental effects have been documented in the literature, such as those
discussed in references [8-11]. Studies have looked at how vegetation cover, species richness and
diversity, and the spread of biological soil crusts change depending on the distance from the watering
point. Further, research has been conducted on soil chemistry and physical properties, as well as the
effect of erosion and trampling. The majority of the ground-based measurements are taken from
samples along transects, and these samples are taken from plots of varying sizes, ranging from 1 to
100 m2. It has been demonstrated that methods which assess only limited areas are heavily influenced
by natural changes in vegetation composition and landscape features that cannot be confidently
linked to the effects of grazing [12]. Traditional ground-based measurements are laborious, relatively
costly, and distant sites are not available for regular repeatable sampling [7]. [4] postulated that
grazing effects cannot be monitored when the measurement is conducted in a limited spatial scope.
The challenges associated with semi-arid rangelands are exacerbated, yet remote sensing data can be
utilized to counter them.
Since the radial pattern around watering points is clearly visible from satellite imagery, recent
studies have been conducted by interpreting and modeling the remote sensing data [13, 14], which
can be processed in a semi-automated and repeatable manner over large and distant regions. For this
reason, different remote-sensing models that utilize geographic information systems techniques have
been created to calculate the spatial spread of various elements in the vicinity of watering points [5,
13-15].
Various studies have indicated that the circular pattern of grazing gradients around watering
points results in similar patterns for various biotic, abiotic, and environmental variables. Vegetation
cover, measured using indices such as SAVI (soil adjusted vegetation index) or NDVI (normalized
difference vegetation index) [5, 15-17], annuals production and grass [14, 18], organic content, soil pH
and bush encouragement, nitrate and phosphate [11], soil nutrient concentrations [19], and track
density [20], are commonly used variables that follow this pattern. The improvement or decline of
each variable does not change significantly after a few kilometers from the watering point. However,
in some cases, such as inversed, composite, and complex gradients [17, 21], different responses were
observed. The composite gradient is associated with non-native species replacing desirable ones,
while the inverse gradient is observed in areas with a dam and mostly woody vegetation. The
complex gradient is seen in areas where vegetation growth is decreased in runoff and erosion areas,
but increased in run-on and sediment deposits. According to [11], the grazing effects form a radial
pattern with a sacrifice zone up to 50 m from the borehole, a nutritious grass zone up to 800 m,
dominated by palatable and grass species, and a bush encroachment zone up to 2000 m. This distance
is considered as the farthest point where grazing has a significant impact.
Intensive grazing near water sources is a global occurrence. However, this research was carried
out in the arid regions of Southeast Iberian Peninsula, focusing on the socioeconomic changes that
took place in this region towards the end of the 20th century. In these desert areas, livestock farming
is a significant sector of the economy [22].
In this regard, the importance of goats must be considered, since although it is important for the
socioeconomic development of the study area, it is also important in the overgrazing process. This
fact provokes that the canopy cannot regenerate in time, and therefore erosive processes dominate
the system [23].
A study conducted in 2006 [24] used remote-sensing methods to examine local desertification
processes in the Southeast Iberian Peninsula. In this current study, the authors present a different
method for evaluating and mapping the impact of grazing around watering points. The main
objective of the study was to investigate changes in vegetation and soil patterns in the drylands of
the Southeast Iberian Peninsula over time, in relation to socio-economic changes. The study had four
specific goals: (1) to analyze the spatial structure of imagery from two desert sites during three
different time periods (mid-late 2000s, around 2015, and 2020) using semi-variograms; (2) to apply
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 12 June 2023 doi:10.20944/preprints202306.0781.v1

the kriging interpolation technique to smooth brightness index values extracted from satellite images
with a spatial resolution of 30-80 meters, in order to assess spatial and temporal land-cover patterns;
(3) to use kriging prediction maps for change detection analysis to determine the direction and
intensity of changes between study periods; and (4) to relate these findings to the socio-economic
situation before and after the EU economic crisis, which affected grazing intensity and therefore the
land-use and land-cover state of the study sites.
2. Study Area
The region under investigation is situated in the southeastern part of Spain, in the Almeria
province, as shown in Figure 1. It spans approximately 50 square kilometers and includes Tabernas,
the largest town in the area with a population of 4025 people. The nearest large urban center is the
city of Almeria, which has a population of 199237 and is located 37 km to the south of Tabernas.
Figure 1. Location of the study area (Desert of Tabernas “37º 00’ 00’’ N; 02º 27’ 00’’ W” and Sierra Alhamilla “36º
59’ 20’’ N; 02º 21’ 05’’ W”).
The central region of the study area, where human activities are concentrated, is the Tabernas
valley. It is surrounded by two mountain ranges running from east to west: the Sierra de los Filabres
to the north, and the Sierra Alhamilla to the south. In the western portion of the study area lies the
Desert of Tabernas, which is regarded as the sole authentic desert in Europe meeting the desert
criteria. Protection has been granted to both the Desert of Tabernas and a section of the Sierra
Alhamilla as natural reserves [25].
The prevailing land cover types found in the region include vegetation consisting of bushes and
grass, with or without trees [26]. Agriculture, primarily barley and some irrigated crops, is the
primary land use, although it faces significant constraints due to the area's climatic and topographical
conditions [27]. Nonetheless, within the past ten years, irrigated olive and almond plantations have
been implemented. Currently, mining only occurs in a handful of gypsum quarries, and abandoned
mines are prevalent in the area. While industry is not a major contributor to the local economy, the
movie and entertainment industry has attracted numerous visitors in recent years. As a matter of fact,
tourism is becoming increasingly important and could emerge as one of the primary economic
activities in the future.
In another vein, according to Andalusian Government [28], in the study area the primary use is
very scarce. This fact makes it possible for livestock activity to concentrate in areas where the canopy
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 12 June 2023 doi:10.20944/preprints202306.0781.v1

has reached a higher level of growth. It is evident that overgrazing, as well as livestock trampling,
along with seasonal variation in rainfall regimes, leads to a progressive loss of existing vegetation,
and as a result, a progressive loss of soil.
In reference to the loss of sustainability of the study area, it is of great interest to highlight the
difference between the erosive and desertification processes that occur. According to Andalusian
Government [28], the concept of desertification is more functional, starting to be considered as a
disturbance that occurs in arid climates and that leads the system (human beings - natural resources)
to an irreversible loss of sustainability. On the other hand, erosion consists of the loss of soil by
uprooting, transport and subsequent accumulation, either by the action of wind or water. This
process can be understood as a form of soil resource degradation, and therefore, an effect or symptom
of desertification, but although it is not the only one. In this regard, and as is well known, although
on a world scale the main form of erosion associated with desertification is wind, in the Iberian
Peninsula water erosion predominates, constituting an endemic environmental problem in most of
Mediterranean Spain. and, in particular, of the peninsular southeast (Desert of Tabernas and Sierra
Alhamilla).
Regarding the climate of the study area, it should be noted, according to the Köppen and Geiger
classification, that it is of the BSk type (local steppe).
The main climatic feature of the study area is its Mediterranean character, with mild
temperatures and marked aridity. This is because the Betic mountains intersect the Atlantic fronts
and leave this area in a rain shadow. The average annual precipitation is 239 mm and the number of
rainy days per year ranges from 25 to 55, although only 6% of the rainy episodes exceed 20 mm. The
average annual temperature is 17.9°C, the average minimum of the coldest month is between 3ºC and
10ºC, with the maximum exceeding 40°C in summer (sometimes reaching 48).
As is known, the study area is made up of loose sediments [29], highly salinized and easily
washed away by rainwater. This fact, along with the sparse vegetation (Figure 2c) and torrential rains,
caused the soil in these areas to erode almost completely (Figure 2b and e), excavating large gullies
and ravines separated by steep slopes. On these slopes, erosion is intense and allows few plant species
to take root, which is why they are often bare and give the landscape as a whole the characteristic
desert appearance (Figure 2a, b, d and e).
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 12 June 2023 doi:10.20944/preprints202306.0781.v1

Figure 2. Desert of Tabernas in Sentinel-2 L2A RGB bands (432) combination (a and b). Sierra Alhamilla in
Sentinel-2 L2A False color bands (843) combination (d and e). Typical vegetation (c) of the study area.
3. Materials and Methods
A total of 528 Landsat pictures were obtained from the Earth Explorer platform
(https://earthexplorer.usgs.gov/) and utilized to cover the two research locations. These images were
captured by the Thematic Mapper (TM) sensor of the Landsat 5 and Landsat 8 satellites at three
different time intervals, as indicated in Table 1. The only land-use method in these areas is livestock
grazing.
Table 1. Satellite images used in this study (both in the desert of Tabernas and in the Sierra Alhamilla).
Satellite and sensor
Nº of images
Date Path/row (WRS-2)
Landsat 5 TM 188 Mid-late 2000s (January 1st,
2005 – December 31th, 2010) 199 and 200/034
Landsat 8 LC 204 Around 2015 (January 1st
, 2014
– December 31th, 2016) 199 and 200/034
Landsat 8 LC 136 2020 (January 1st, 2019 – 31th,
2020) 199 and 200/034
The process of image processing commenced by transforming the digital numbers of the image
into reflectance values [30-31]. From these values, various vegetation indices such as NDVI [32], SAVI
[33], MSAVI [34], perpendicular vegetation index or PVI [35], and greenness and brightness indices
[36] derived from tasseled cap were computed for each image subset. To determine the most
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 12 June 2023 doi:10.20944/preprints202306.0781.v1

appropriate index for distinguishing degraded and non-degraded land, spectral separability analysis
was carried out using the mean and standard deviation values of extreme classes in each scene, for
all the indices [37-39]. The results of this analysis showed that the brightness index (BI) had the
highest separability value, and hence, it was selected for further analysis (see Eq. 1).
 =  · 


(1)
According to [36-37], BI can be inferred for different sensors using spectral band numbers (Bi)
and appropriate BI coefficients (αi). One of the benefits of using the BI index is its ability to compare
different sensors that have different spectral bands. This is possible because the BI values can be
normalized to create a single layer of data, making comparisons between sensors possible. Further,
the BI index was originally designed to analyze soil properties, and there have been statistically
significant results using this method.
The following process involved averaging each window of 6 x 6 pixels in the TM (Landsat 5)
and LC (Landsat 8) images, which reduced the resolution by a factor of 6. This resulted in a new pixel
size of 171 meters, which was small enough to assess changes within the 6 km range from the
watering point where degradation is expected. The subset size was reduced from around 1.5 million
pixels to only 40000 pixels, which made the data more manageable for processing. Note that even
though the pixel size was reduced, image-to-image geometric correction was applied for better
accuracy in the change detection process.
In another vein, the geostatistics is based on the concept of a regionalized variable, which is a
variable that can be characterized based on a number of spatial measurements. The fundamental idea
behind geostatistics is that when spatial continuity is assumed, adjacent samples are expected to be
more similar to each other than samples that are further apart. This spatial dependence can be
statistically analyzed and described using parameters derived from a semi-variogram, which is a
function that relates the semi-variance to the distance and direction between two samples. The semi-
variance, defined as half the mean-squared difference between two samples that are a certain distance
apart in a given direction, is used to quantify this spatial dependence. This approach provides a way
to analyze and understand the spatial variation in a given variable (Eq. 2).
(ℎ)=1
2· (ℎ) · ( − )
()

(2)
As is well known, the Eq. (2) referred to earlier uses a vector h to define both the direction and
distance between two samples, which is commonly known as the lag.
In Eq. (2), the semi-variance at lag h is denoted by γ(h), and N(h) represents the number of pairs
of samples that are separated by a distance of h. The value of the regionalized variable at a given
location i is represented by Zi. In addition to the lag, the variogram is characterized by three other
parameters: the nugget, range, and sill. The nugget accounts for variability at zero distance and is
attributable to errors in both sampling and analysis. The range represents the distance, often denoted
as "a" beyond which spatial autocorrelation between sampling sites becomes negligible. The sill
represents the variability of samples that are spatially independent. Empirical semi-variograms can
be computed from a set of observations, and then a theoretical model can be fitted [40]. The
exponential model is a popular theoretical model and is expressed in the form shown in Eq. (3), where
C0 is the nugget, C0 + C1 is the sill, a is the range and h is the lag.
(ℎ)= +·[1− (·||
)]
(3)
To better understand the spatial structure of imagery for a given date and location, variogram
analysis was used in the studied sites. The decision to use semi-variograms was based on the
similarity in spatial structure of most of the variables, which gradually increased or decreased as the
distance from the watering point increased, until grazing effects were no longer observed. Semi-
variograms were introduced to remote sensing by [41], who found that variogram parameters could
be directly linked to image features. In this study, the presence or absence of a sill in the variogram
could indicate the presence of a radial grazing pattern around watering points, while the level of the
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 12 June 2023 doi:10.20944/preprints202306.0781.v1

sill could be used to assess the homogeneity (i.e., lower variance) of an area. The lag distance in the
variogram could also be related to the walking distance of livestock, with observations becoming
increasingly independent beyond a certain distance (the range).
In each subset of the data, the empirical semi-variogram was calculated using 40000 pixels, and
a theoretical model that best fit the semi-variogram was estimated. The model parameters were then
determined by minimizing the squared differences between the empirical semi-variogram values and
those of the theoretical model. Once the best-fitting model was chosen, several criteria were applied
to assess its accuracy and to fine-tune its parameters: 1) it was carried out a cross-validation scatter
plot; 2) it was inferred the mean estimation error using the Eq. (4); and 3) it was obtained the mean
standardized squared estimation error through Eq. (5).
1
· ( −
∗
)

 =1
· 

 ≈ 0
(4)
1
· [ −
∗
∗]


 =1
· (
)


 ≈ 1
(5)
In order to minimize the influence of local effects and obtain a more even representation of
surface brightness values across the study sites, researchers employed a linear geostatistical method
called ordinary kriging interpolation. This method, according to [42], is useful in revealing spatial
phenomena. The approach estimates the mean of the values within a searching neighborhood as a
constant.
The study employed the BI differencing method as a post-processing change detection
technique, which is a modified version of the vegetation index differencing method [43]. Its purpose
was to evaluate the primary patterns of degradation (in magenta color) or rehabilitation (in green
color) in the study areas. To achieve this, two sets of kriging maps (i.e., 2016-2005 and 2020-2016)
were subtracted for each site (desert of Tabernas and Sierra Alhamilla).
4. Results and Discussion
Figure 3 displays the relevant BI products, which were computed using Eq. (1) and the
corresponding coefficients. The bright areas that are scattered across the mid to late 2000s images,
indicating watering areas, are much less visible in the 2015 and 2020 images. In the more recent 2020
image of the desert of Tabernas and Sierra Alhamilla, many watering areas are absent, but there is a
large bright area present. The brightness index (BI) values, which indicate bare, degraded soil, were
calculated using the same method for all sensors, and the brightness levels were uniformly stretched.
The desert of Tabernas has BI values ranging from 0.66 to 0.95, while the Sierra Alhamilla sites have
BI values ranging from 0.20 to 0.60.
The geostatistical analysis utilized all of the BI images. Initially, empirical semi-variograms were
created for each of the two sites for the three time periods. Field observations did not reveal any
anisotropic patterns that could potentially govern the direction of grazing, such as linear dunes or
other barriers. Therefore, an isotropic distribution was assumed in all cases. The results of the cross-
validation analyses are outlined in Table 2. The slope coefficient and the intercept coefficient of the
exponential model chosen were very close to unity and zero, respectively, indicating that the model
was able to accurately replicate the observed values.
Following the examination of several theoretical models [40], the exponential model (Eq. 3) was
chosen as it yielded the best results during cross-validation. The fit of the model was assessed using
the least-squares measure, and the variogram parameters can be found in Table 2. Each variogram
was processed with 17 lags, each spanning a distance of 800 m. The lag values were determined
through a trial and error process to optimize the aforementioned criteria. When observed visually,
the variograms from the mid-late 2000s and around 2015 appear quite similar, exhibiting a typical
variogram shape. Two notable features can be observed. Firstly, the sill of the variogram around 2015
is lower compared to the mid-late 2000s, suggesting a decrease in variance in the latter year. Further,
the range in the mid-late 2000s is 4800 m, while in around 2015 it is 6400 m, which corresponds to the
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 12 June 2023 doi:10.20944/preprints202306.0781.v1

reported walking distance of livestock from their drinking source [44]. In contrast, the variogram for
2020 only reaches a similar sill level after a distance of 12 km. This range does not appear to be
indicative of the grazing pattern in the region.
Figure 3. Brightness index (BI) in the study area at different date. All images of each study area, corresponding
to a determined period of time, are the images resulting from the treatment carried out with all the images of
each period.
Table 2. Variogram parameters in the study area.
Study area Date Nugget
Sill Range
(m) Spatial variance
Tabernas and
Alhamilla
Mid-late 2000s
0 0.00088
4800 0.00088
Around 2015 0.0001 0.00064
6400 0.00054
2020 0.0001 0.0008
12000
0.0007
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 12 June 2023 doi:10.20944/preprints202306.0781.v1

Later, kriging interpolation techniques were utilized, employing exponential models (see Figure
4) with parameters from Table 2. The final outcomes illustrating the distribution of BI values for the
three periods and both sites are displayed in Figure 5. The initial maps, specifically during the mid-
late 2000s, reveal the presence of bands encircling the watering points, signifying the progressive
degradation of land extending from the wells. These bands denote the grazing gradient or the impact
of grazing. Regions depicted in dark red in the images indicate areas where grazing impact
predominates, exerting a strong influence on spatial variation. Such regions are referred to as the
"sacrifice zone" by [11]. The adjacent belts in light red, orange and yellow represent a mixed zone
where the effects of grazing and natural variability coincide or establish a stable balance. This zone
can be likened to the outskirts of the grazing impact and highlights the primary migration routes of
livestock. The zone exhibiting green shades signifies an area where natural variability surpasses the
impact of grazing and is termed the "grazing reserve" by [11].
Figure 4. Model variograms obtained from Brightness Index (BI) in the Desert of Tabernas and Sierra Alhamilla.
The BI differencing method, explained in the concluding paragraph of Section 3, was utilized to
analyze changes in grazing patterns over two specific time periods. Figure 6 displays the maps
illustrating these changes. In the change detection map for the period 2016-2005, a large portion of
the area appears grey, indicating that there were no noticeable alterations in grazing patterns. The
remaining areas are represented by magenta, indicating degradation, and green tones, representing
rehabilitation. Consequently, between the mid-late 2000s and around 2015, the overall trend was a
positive one, with land-cover conditions improving or undergoing rehabilitation. It is worth noting
that around 2015, the area became more uniform, as indicated by the reduced variation (lower sill) in
Figure 4 and the narrower range of colors in Figure 6a and b. Conversely, during the second period
(2020-2016), a significant portion of the area experienced degradation processes (magenta colors),
while the area undergoing rehabilitation (green tones) diminished considerably (Figure 6a and b).
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 12 June 2023 doi:10.20944/preprints202306.0781.v1

Figure 5. Kriging maps inferred by interpolation in the Desert of Tabernas and Sierra Alhamilla.
As a result of the 2007 economic crisis, there was a progressive decrease in the livestock census
in the study area due to increased costs, which is why many farmers had to abandon economic
activity due to lack of profitability. This fact had a positive influence on the environment, since it led
to the recovery of those areas where livestock pressure was highest. This is one of the reasons why,
from the mid-late 2000s to around 2015, a recovery of the study area can be observed (see Figure 6a).
Despite the above, and although in 2016 the Andalusian Government started the Recovery Plan for
the Desert of Tabernas and Sierra Alhamilla [45], the increase in degradation (Figure 6b) is mainly
due to climatic causes.
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 12 June 2023 doi:10.20944/preprints202306.0781.v1

Figure 6. Result of the BI differencing method in the Desert of Tabernas (left side) and Sierra Alhamilla (right
side) for both 2016-2005 (a) and 2020-2016 (b) periods.
To explain the climatic effect in the study area, Table 3 shows the data corresponding to the sum
of both average monthly temperature (in Celsius) and monthly precipitation (in mm), in each of the
years subjected to study. All climatic data were downloaded from Version 4 of the CRUTS monthly
high-resolution gridded multivariate climate dataset [46].
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 12 June 2023 doi:10.20944/preprints202306.0781.v1

Table 3. Climatic data in the study area from 2005 to 2020.
Year
Tª (ºC)
P (mm)
Year
Tª (ºC)
P (mm)
2005
179.6 226.6 2013
185.1 385.6
2006
180.7 392.3 2014
186.1 293.5
2007
182.2 410.8 2015
187.1 338.4
2008
183.7 407.3 2016
189.9 320.5
2009
185.0 428.2 2017
190.6 293.3
2010
185.3 646.6 2018
191.6 490.7
2011
184.2 333.0 2019
194.6 340.8
2012
184.8 386.1 2020
193.7 359.5
From Table 3, a new aridity index was obtained to help explain the results shown in Figure 6.
Said index, which we have called the RAMALL (RAmírez, MAdueño, López and Leiva) Aridity
Index. Eq. (6) specifies how the RAMALL index is calculated.
   = ∑


∑

 +∑


(6)
where:
i = months of each year.
j = year in which you want to calculate RAMALL aridity index.
MPj = monthly precipitation (mm) in the year “j”.
AMTj = average monthly temperature (ºC) in the year “j”.
AMTj+1 = average monthly temperature (ºC) in the year “j+1”.
Values greater than 1 indicate low aridity index, while values lower than 1 indicate low aridity
index. Figure 7 shows the evolution of this new RAMALL aridity index in the study area between
2005 and 2020.
Figure 7. RAMALL aridity index during the study period in the Desert of Tabernas and Sierra Alhamilla
(∑AMT2021 = 186.51 ºC).
After analyzing the data shown in Table 3, as well as Figure 7, the RAMALL aridity index
presents mean values of 1.05 (low aridity index) between 2005 and 2015, and 0.94 (high aridity index)
between 2016 and 2020.
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 12 June 2023 doi:10.20944/preprints202306.0781.v1

As can be seen, the 2007 economic crisis had a positive effect on the average aridity during the
2016-2005 period (RAMALL aridity index value greater than 1). During this period, the water erosion
phenomena [28] were minimized not only by the decrease in livestock pressure, but also as a
consequence of the increase in plant variety that, being present in the soil seed bank, only colonized
the area study as a result of said decreased grazing.
However, in the second period under study (2020-2016 period), whose RAMALL aridity index
values are lower than 1, the progressive increase in the average annual temperature (Figure 8), along
with the almost disappearance of grazing that occurred at the end of the first period (2016-2005),
favored the plant variety loss in the study area, and therefore the shrub stratum was dominant. This
fact has, as a serious consequence for the environment of the study area, an increase in erosive
processes, as well as both the soil and habitat loss over time if the global warming trend continues
[28].
Figure 8. Average annual temperature evolution in the study area between 2005 and 2020.
Finally, it is necessary to emphasize that, both in the desert of Tabernas and Sierra Alhamilla,
the influence of the arid climate has given rise, over time, to a landscape of mainly erosive
morphology, whose main signs are the slopes (of soft materials) in the form of gullies. With respect
to the drainage network, it is necessary to indicate the action of the existing channels on the
smoothing of the relief, resulting in a predominance of erosive action at the headwaters, and
therefore, the boxing in of the channels, as well as a predominance of sedimentation in its lower part.
The impact caused by raindrops on bare soil gives rise to a wide variety of microforms in the
landscape, a key characteristic of soil loss through surface runoff.
5. Conclusions
The aim of the present study was to analyze the changes in vegetation patterns over time in the
desert of Tabernas and Sierra Alhamilla (Almería, Spain), focusing on land degradation and
rehabilitation. Specifically, the research aimed to understand how these changes were influenced by
socio-economic factors following the economic crisis in 2007. To achieve this, the tasseled cap-derived
brightness index (BI) was employed as a tool to depict the spatial distribution of land surface
characteristics. The BI was chosen for its superior ability to differentiate among various spectral
indices, its original purpose of examining soil properties, and its capability to compare different
sensors with distinct spectral bands by normalizing them into a single BI layer.
The use of geostatistical analysis, specifically semi-variance analysis, was determined to be a
suitable approach for understanding the spatial structure within imagery at a specific date and
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 12 June 2023 doi:10.20944/preprints202306.0781.v1

location. This method was chosen due to the similarity between the variogram's shape and the
directional changes observed in various biotic, abiotic, and environmental variables along the grazing
gradient emanating from watering points in arid and semi-arid regions. By utilizing variogram
models from the mid-late 2000s and around 2015, it becomes possible to quantitatively generalize the
observed phenomena across the region of interest. The average range of 4800-6400 meters aligns with
the reported walking distance of livestock from their water source. Comparing variograms from
different years can provide insights into the temporal dynamics of the region, as evidenced by the
failure to reach a sill in the variogram when grazing ceased at the end of 2015.
The kriging interpolation method was employed as a smoothing filter, whereby each pixel in the
image was replaced with the solution derived from the variogram equation (specifically, an
exponential model in this case) calculated using all other pixels. This approach aimed to reduce
spatial errors and fine-scale variability, facilitating a more accurate delineation of degradation
boundaries surrounding the watering points. The resulting maps of the study area revealed a radial
pattern, indicating progressive land degradation emanating from the wells, commonly referred to as
the grazing gradient. However, this pattern appeared blurred or absent in the 2020 maps. In
comparison to other interpolation techniques, ordinary kriging is regarded as the superior linear
unbiased estimator.
The index differencing technique was successfully utilized to analyze temporal changes. This
study showcases the capability of satellite image analysis to track land-use and land-cover
modifications resulting from both the 2007 economic crisis and the significant decrease in grazing
pressure.
Author Contributions: Conceptualization, E.R.J., A.M.-L, J.M.M.L, M.C.L.G and J.L.L.P.; methodology, E.R.J.,
A.M.-L, J.M.M.L, M.C.L.G and J.L.L.P.; formal analysis, E.R.J., A.M.-L, J.M.M.L, M.C.L.G and J.L.L.P.;
investigation, E.R.J., A.M.-L, J.M.M.L, M.C.L.G and J.L.L.P.; resources, E.R.J., A.M.-L, J.M.M.L, M.C.L.G and
J.L.L.P.; writing—original draft preparation, E.R.J., A.M.-L, J.M.M.L, M.C.L.G and J.L.L.P.; writing—review and
editing, E.R.J., A.M.-L, J.M.M.L, M.C.L.G and J.L.L.P.; supervision, E.R.J., A.M.-L, J.M.M.L, M.C.L.G and J.L.L.P.;
project administration, E.R.J. All authors have read and agreed to the published version of the manuscript.
Funding This research received no external funding.
Data Availability Statement: The data used to support the findings of this study can be made available by the
corresponding author upon request.
Acknowledgments: This study has been carried out thanks to the research project with Reference "OG-017/07".
Authors are grateful for the disinterested collaboration of the Knowledge-Based Company RSV3 Remote Sensing
S.L.
Conflicts of Interest: The authors declare no conflict of interests.
References
1. Lin, X.; Zhao, H.; Zhang, S.; Li, X.; Gao, W.; Ren, Z.; Luo, M. Effects of animal grazing on vegetation biomass
and soil moisture on a typical steppe in Inner Mongolia, China. Ecohydrology. 2022, 15(1), e2350. DOI:
10.1002/eco.2350
2. Chen, T.; Tang, G.; Yuan, Y.; Guo, H.; Xu, Z.; Jiang, G.; Chen, X. Unraveling the relative impacts of climate
change and human activities on grassland productivity in Central Asia over last three decades. Science of
the Total Environment. 2020, 743, 140649. DOI: 10.1016/j.scitotenv.2020.140649
3. Hamidov, A.; Helming, K.; Balla, D. Impact of agricultural land use in Central Asia: a review. Agronomy for
sustainable development. 2016, 36, 6–23. DOI: 10.1007/s13593-015-0337-7
4. Augusteyn, J.; Rich, M.; Mitchell, C.; Mulder, E.; Nolan, B.; Lim, L.; Melzer, R. Does reducing grazing
pressure or predation conserve kowaris? A case study at Diamantina National Park. Australian Journal of
Zoology. 2022, 70(2), 56-73. DOI: 10.1071/ZO22027
5. Yang, X.; Zhang, K.; Jia, B.; Ci, L. Desertification assessment in China: An overview. Journal of Arid
Environments. 2005, 63(2), 517-531. DOI: 10.1016/j.jaridenv.2005.03.032
6. Malan, J.A.C. Offstream watering points for cattle as a method for riparian restoration, Ph. D. Dissertation.
Central Queensland University, Queensland, Australia. 2020. DOI: 10.25946/14274524.v1
7. Fokeng, R.M.; Fogwe, Z.N. Landsat NDVI-based vegetation degradation dynamics and its response to
rainfall variability and anthropogenic stressors in Southern Bui Plateau, Cameroon. Geosystems and
Geoenvironment. 2022, 1(3), 100075. DOI: 10.1016/j.geogeo.2022.100075
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 12 June 2023 doi:10.20944/preprints202306.0781.v1

8. Bakhshi, J.; Javadi, S.A.; Tavili, A.; Arzani, H. Study on the effects of different levels of grazing and
exclosure on vegetation and soil properties in semi-arid rangelands of Iran. Acta Ecologica Sinica. 2020, 40(6),
425-431. DOI: 10.1016/j.chnaes.2019.07.003
9. Ahlborn, J.; von Wehrden, H.; Lang, B.; Römermann, C.; Oyunbileg, M.; Oyuntsetseg, B.; Wesche, K.
Climate–grazing interactions in Mongolian rangelands: Effects of grazing change along a large-scale
environmental gradient. Journal of Arid Environments. 2020, 173, 104043. DOI: 10.1016/j.jaridenv.2019.104043
10. Ma, L.; Derner, J.D.; Harmel, R.D.; Tatarko, J.; Moore, A.D.; Rotz, C.A; ...; Wilmer, H. Application of grazing
land models in ecosystem management: Current status and next frontiers. Advances in Agronomy. 2019, 158,
173-215. DOI: 10.1016/bs.agron.2019.07.003
11. Van der Weyde, L.K.; Theisinger, O.; Mbisana, C.; Gielen, M.C.; Klein, R. The value of pastoral ranches for
wildlife conservation in the Kalahari. Wildlife Research. 2021, 49(3), 215-226. DOI: 10.1071/WR21048
12. Henriques, M.; McVicar, T.R.; Holland, K.L.; Daly, E. Riparian vegetation and geomorphological
interactions in anabranching rivers: A global review. Ecohydrology. 2022, 15(2), e2370. DOI: 10.1002/eco.2370
13. Donohue, R.J.; Mokany, K.; McVicar, T.R.; O'Grady, A.P. Identifying management-driven dynamics in
vegetation cover: Applying the Compere framework to Cooper Creek, Australia. Ecosphere. 2022, 13(3),
e4006. DOI: 10.1002/ecs2.4006
14. Rivera-Marin, D.; Dash, J.; Ogutu, B. The use of remote sensing for desertification studies: A review. Journal
of Arid Environments. 2022, 206, 104829. DOI: 10.1016/j.jaridenv.2022.104829
15. Wang, G.; Li, J.; Ravi, S. A combined grazing and fire management may reverse woody shrub encroachment
in desert grasslands. Landscape Ecology. 2019, 34, 2017-2031. DOI: 10.1007/s10980-019-00873-0
16. El Aich, A.; Pueyo, Y.; Barrantes Díaz, O.; Alados Lopez, C. Management and restoration of pastures and
rangelands in the Mediterranean basin from the Northern and the Southern perspective. Options
Méditerranéennes. 2021; 125, 257-270.
17. Dara, A.; Baumann, M.; Freitag, M.; Hölzel, N.; Hostert, P.; Kamp, J.; ...; Kuemmerle, T. Annual Landsat
time series reveal post-Soviet changes in grazing pressure. Remote Sensing of Environment. 2020, 239, 111667.
DOI: 10.1016/j.rse.2020.111667
18. Ganem, K.A.; Xue, Y.; Rodrigues, A.D.A.; Franca-Rocha, W.; Oliveira, M.T.D.; Carvalho, N.S.D.; ...;
Shimabukuro, Y.E. Mapping South America’s Drylands through Remote Sensing—A Review of the
Methodological Trends and Current Challenges. Remote Sensing. 2022, 14(3), 736. DOI: 10.3390/rs14030736
19. Li, C.; de Jong, R.; Schmid, B.; Wulf, H.; Schaepman, M.E.. Spatial variation of human influences on
grassland biomass on the Qinghai-Tibetan plateau. Science of the Total Environment. 2019, 665, 678-689. DOI:
10.1016/j.scitotenv.2019.01.321
20. Nyamuryekung'e, S.; Cibils, A.F.; Estell, R.E.; VanLeeuwen, D.; Spiegal, S.; Steele, C.; ...; Cao, H. Movement,
activity, and landscape use patterns of heritage and commercial beef cows grazing Chihuahuan Desert
rangeland. Journal of Arid Environments. 2022, 199, 104704. DOI: 10.1016/j.jaridenv.2021.104704
21. Asmare, B. A review of sensor technologies applicable for domestic livestock production and health
management. Advances in Agriculture. 2022, 1-6. DOI: 10.1155/2022/1599190
22. Giménez-Anaya, A.; Bueno, C.G.; Fernández-Llario, P.; Fonseca, C.; García-González, R.; Herrero, J.; ...;
Rosell, C. What Do We Know About Wild Boar in Iberia?. In Problematic Wildlife II, 1st ed.; Angelici, F.,
Rossi, L., Eds.; Springer: Cham, Switzerland, 2020; Volume 2, pp. 251-271. DOI: 10.1007/978-3-030-42335-
3_9
23. Lozano Cantero, F.J. Informe sobre el desierto de Tabernas. Diputación Provincial de Almería. 2006, 50 pp.
Available online:
http://www.almeria.es/Servicios/Informacion/Informacion.nsf/aff1dd050488813ac1256ae10033dfd3/5d0fdf
b5dc132372c1257384004d0958/$FILE/Estudio.pdf (Accessed on 15 March 2023).
24. Alemayehu, T.; Recatalá, L.; Fabbri, A. G.; Sánchez, J. Land use change detection as a basis for analysing
desertification processes: A case study in Tabernas (Almeria, Spain). In Desertification in the Mediterranean
Region. A Security Issue, 1st ed.; Kepner, W.G., Rubio, J.L., Mouat, D.A., Pedrazzini, F., Eds.; Springer:
Netherlands, 2006; Volume 1, pp. 341-352. DOI: 10.1007/1-4020-3760-0_15
25. Andalusian Government. Decreto 172/2016, de 8 de noviembre, por el que se declaran las Zonas Especiales
de Conservación de la Red Ecológica Europea Natura 2000. 2016. Available online:
https://www.juntadeandalucia.es/medioambiente/portal/landing-page-normativa/-
/asset_publisher/5WqqWSjN2vgc/content/decreto-172-2016-de-8-de-noviembre-por-el-que-se-declaran-
las-zonas-especiales-de-conservaci-c3-b3n-de-la-red-ecol-c3-b3gica-europea-natura-2000-karst/20151.
(Accessed on 15 March 2023).
26. Verhoeven, V.B.; Dedoussi, I.C. Annual satellite-based NDVI-derived land cover of Europe for 2001–2019.
Journal of Environmental management. 2022, 302, 113917. DOI: 10.1016/j.jenvman.2021.113917
27. Lázaro, R.; Calvo-Cases, A.; Rodriguez-Caballero, E.; Arnau-Rosalén, E.; Alexander, R.; Rubio, C.; ...;
Puigdefábregas, J. Biocrusts and catchment asymmetry in Tabernas Desert (Almeria, Spain). Geoderma.
2022, 406, 115526. DOI: 10.1016/j.geoderma.2021.115526
28. Andalusian Government. Valores ambientales de la zona de especial conservación Desierto de Tabernas
(ES0000047). 2016. Available online:
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 12 June 2023 doi:10.20944/preprints202306.0781.v1

https://www.juntadeandalucia.es/medioambiente/portal_web/web/temas_ambientales/espacios_protegid
os/01_renpa/canales_figuras_proteccion/Red_Natura/2016_06_valores_ambientales_resumenes/nuevos/0
000047_desierto_tabernas.pdf (Accessed on 15 March 2023).
29. Veneranda, M.; Lopez-Reyes, G.; Manrique-Martinez, J.A.; Sanz-Arranz, A.; Medina, J.; Pérez, C., ...; Rull,
F. Raman spectroscopy and planetary exploration: Testing the ExoMars/RLS system at the Tabernas Desert
(Spain). Microchemical Journal. 2021, 165, 106149. DOI: 10.1016/j.microc.2021.106149
30. Bannari, A.; Al-Ali, Z.M. Assessing climate change impact on soil salinity dynamics between 1987–2017 in
arid landscape using Landsat TM, ETM+ and OLI data. Remote Sensing. 2020, 12(17), 2794. DOI:
10.3390/rs12172794
31. ED Chaves, M.; CA Picoli, M.; D. Sanches, I. Recent applications of Landsat 8/OLI and Sentinel-2/MSI for
land use and land cover mapping: A systematic review. Remote Sensing. 2020, 12(18), 3062. DOI:
10.3390/rs12183062
32. Tucker, C.J. Red and photographic infrared linear combinations for monitoring vegetation. Remote Sensing
of Environment. 1979, 8(2), 127-150. DOI: 10.1016/0034-4257(79)90013-0
33. Huete, A.R. A soil-adjusted vegetation index (SAVI). Remote Sensing of Environment. 1988, 25(3), 295-309.
DOI: 10.1016/0034-4257(88)90106-X
34. Qi, J.; Chehbouni, A.; Huete, A.R.; Kerr, Y.H.; Sorooshian, S. A modified soil adjusted vegetation index.
Remote sensing of environment. 1994, 48(2), 119-126. DOI: 10.1016/0034-4257(94)90134-1
35. Richardson, A.J.; Wiegand, C.L. Distinguishing vegetation from soil background information.
Photogrammetric engineering and remote sensing. 1977, 43(12), 1541-1552.
36. Kauth, J.; Thomas, G.S. The tasseled cap—a graphic description of the spectral-temporal development of
agricultural crops as seen by LANDSAT. In: Proceedings of the Symposium on Machine Processing of
Remotely Sensed Data. Purdue University of West Lafayette, Indiana, EEUU, 1976 (01 01 1976).
37. Remer, L.A.; Levy, R.C.; Mattoo, S.; Tanré, D.; Gupta, P.; Shi, Y.; ...; Holben, B.N. The dark target algorithm
for observing the global aerosol system: Past, present, and future. Remote sensing. 2020, 12(18), 2900. DOI:
10.3390/rs12182900
38. Llorens, R.; Sobrino, J.A.; Fernández, C.; Fernández-Alonso, J.M.; Vega, J.A. A methodology to estimate
forest fires burned areas and burn severity degrees using Sentinel-2 data. Application to the October 2017
fires in the Iberian Peninsula. International Journal of Applied Earth Observation and Geoinformation. 2021, 95,
102243. DOI: 10.1016/j.jag.2020.102243
39. Ramírez-Juidías, E.; Cabello-Franco, E.J. New Remote Sensing Technologies Applied to the Prediction of
Spontaneous Forest Fires. In Natural Hazards-New Insights, 1st ed.; Mokhtari, M., Ed.; Publisher:
IntechOpen, 2023. DOI: 10.5772/intechopen.110501
40. Lu, C.; Song, Z.; Wang, W.; Zhang, Y.; Si, H.; Liu, B.; Shu, L. Spatiotemporal variation and long-range
correlation of groundwater depth in the Northeast China Plain and North China Plain from 2000∼2019.
Journal of Hydrology: Regional Studies. 2021, 37, 100888. DOI: 10.1016/j.ejrh.2021.100888
41. Qiu, S.; Zhu, Z.; Olofsson, P.; Woodcock, C.E.; Jin, S. Evaluation of Landsat image compositing algorithms.
Remote Sensing of Environment. 2023, 285, 113375. DOI: 10.1016/j.rse.2022.113375
42. Ramírez-Juidías, E.; Amaro-Mellado, J.L.; Leiva-Piedra, J.L. Influence of the Urban Green Spaces of Seville
(Spain) on Housing Prices through the Hedonic Assessment Methodology and Geospatial Analysis.
Sustainability. 2022, 14(24), 16613. DOI: 10.3390/su142416613
43. Goswami, A.; Sharma, D.; Mathuku, H.; Gangadharan, S.M.P.; Yadav, C.S.; Sahu, S.K.; ...; Imran, H. Change
detection in remote sensing image data comparing algebraic and machine learning methods. Electronics.
2022, 11(3), 431. DOI: 10.3390/electronics11030431
44. Cowley, R.; Walsh, D. Modelling pasture growth and utilization in a large multi-watered paddock, 1st ed.;
Northern Territory Government: Australia, 2023; pp. 1-52.
45. Boletín Oficial de la Junta de Andalucía (BOJA). BOJA 245, Friday, December 23, 2016. Available online:
https://www.juntadeandalucia.es/medioambiente/portal/documents/20151/808488/4_anexo9_porn_alham
illa_tabernas_boja.pdf/654776f2-3228-9901-f582-7485caf4a92c?t=1511693637000 (accessed on 10 May 2023).
46. Harris, I.; Osborn, T.J.; Jones, P.; Lister, D. Version 4 of the CRU TS monthly high-resolution gridded
multivariate climate dataset. Scientific data. 2020, 7(1), 109. DOI: 10.1038/s41597-020-0453-3
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those
of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s)
disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or
products referred to in the content.
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 12 June 2023 doi:10.20944/preprints202306.0781.v1