Satellite-based Drought Assessment: Integrating Ahp Method and Fuzzy Logic for Comprehensive Vulnerability and Risk Analysis

doi:10.21203/rs.3.rs-4873555/v1

Download PDF

Research Article

Satellite-based Drought Assessment: Integrating Ahp Method and Fuzzy Logic for Comprehensive Vulnerability and Risk Analysis

https://doi.org/10.21203/rs.3.rs-4873555/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Drought research is a timely issue, as drought is an extreme phenomenon with consequences that threaten nature, landscapes and society. Drought is typically defined as a prolonged period of abnormally low precipitation leading to water shortages in soils and water supplies. This study investigates the vulnerability and risk of the landscape to drought in the Banská Bystrica region of Slovakia, focusing on the integration of Landsat 8 satellite image analysis, fuzzy logic and Analytic Hierarchy Process (AHP) methods. The evaluation process involves the selection and processing of input factors from Landsat 8 satellite imagery that are key contributors to drought vulnerability. These methods are used to assess drought vulnerability and associated risks. The resulting drought vulnerability map was created using fuzzy logic in a GIS environment. The final drought risk map was then evaluated. The final maps were categorised into four classes, with comparisons made between drought vulnerability index (DVI) and drought risk index (DRI) at gauging stations. Our findings highlight significant differences in drought vulnerability and risk across different areas of the region. The study provides valuable insights into the comprehensive analysis of drought. Examination of the map shows that the highest levels of risk are found in both the northern and southern parts of the region. This spatial pattern highlights areas that are particularly vulnerable to drought.

drought vulnerability

drought risk

AHP method

fuzzy logic

Landsat 8

The impact of climate change on our planet is profound and multifaceted. One of its most pronounced manifestations is the increasing frequency and intensity of drought, which significantly affects aquatic and terrestrial ecosystems, as well as human communities. In this context, the scientific community studying drought aims to develop and apply methods that offer a comprehensive view of the hazard, vulnerability, and risk associated with spatiotemporal assessment and future prediction of drought.

Drought is a natural phenomenon that differs from other natural disasters due to its gradual and protracted evolution, often described as a slow process. The environmental impacts of drought are long-lasting (Cancelliere et al. 2007). Although drought has historically been associated with arid and semi-arid regions due to specific climatic conditions, it has increasingly affected temperate climates in recent decades, with its negative consequences amplifying in magnitude and intensity. The frequency and severity of droughts are expected to increase due to climate change, which is exacerbated by rising temperatures and extreme weather events (Nam et al. 2015; Fendeková et al. 2018a).

Drought is categorized into meteorological, hydrological, agronomic, and socio-economic types. Meteorological drought, resulting from a reduction in average long-term precipitation, is deemed the most critical type, with other forms gradually manifesting in its aftermath (Alamdarloo et al. 2020). Adopting a global perspective, the World Bank (2019) provides comprehensive principles and guidance for assessing drought hazard and risk. Hazard is defined as a process, phenomenon, or human activity that may cause loss of life, injury, property damage, social and economic disruption, or environmental degradation. Vulnerability is determined by physical, social, economic, and environmental factors or processes that increase the susceptibility of individuals, communities, assets, or systems to the impacts of hazards. Drought risk is defined as the potential loss of life, injury, or damage to assets that could occur to a system, society, or community in a specific period, determined probabilistically as a function between drought hazard, exposure, and vulnerability.

The Analytic Hierarchy Process (AHP) and fuzzy logic are sophisticated methods for assessing drought. AHP is particularly useful for mapping drought vulnerability due to its comprehensive inclusion of all types of drought. This method is effective, widely used, cost-efficient, and time-saving. AHP is a popular method for evaluating vulnerabilities related to natural hazards, especially droughts. It enables multi-criteria spatial decision-making through a structure-based comparison matrix. Saaty (2008) and Saini et al. (2022) have both emphasized the usefulness of AHP in this regard. This is because of its structured framework for hierarchical analysis of factors and addressing prioritization hierarchies and interrelationships among different aspects of drought. The combination of fuzzy logic and conventional methods enables the quantification of uncertainties, improving the modelling of subtle nuances present in complex environmental phenomena. This approach enhances the reliability of drought vulnerability information, making it a valuable tool in the broader context of hazard assessment and management.

Many researchers investigate drought issues using GIS. Drought assessment using AHP and fuzzy logic has been implemented in various regions, including Iran (Shiravand and Bayat 2023), India (Mahato et al. 2023; Palchaudhuri and Biswas 2016; Saha et al. 2021, 2022, 2023; Singh et al. 2019, research was conducted in China (Yuan et al. 2023), the Caribbean region - specifically Haiti (Elusma et al. 2022), Australia (Rahmati et al. 2020; Hoque et al. 2021), South Korea (Kim et al. 2019), and the Mediterranean region (Spiliotis et al. 2021). The researchers have highlighted that climate change has accelerated hydrological processes, resulting in more frequent and severe droughts with various consequences, including forest fires (Mukherjee et al. 2018). Furthermore, the combination of AHP methodology and GIS with satellite data, such as Landsat 8, has been utilised for assessing drought risk (Kumari et al. 2023).

Blauhut et al. (2022) assessed the impacts of drought events in Europe in 2018 and 2019, examining the relationship between drought management strategies and the perception of drought impact. The authors characterised the drought events objectively, without subjective evaluations. The study was based on a European survey involving 28 countries. In addition to theoretical studies that analyse historical periods and use predictive models to assess future drought development, there is a concerted effort in drought research to establish effective drought monitoring and early warning systems. Various platforms worldwide offer such information. One of the oldest is the U.S. Drought Monitor, which was developed in collaboration between the National Drought Mitigation Centre, the U.S. Department of Agriculture, and the National Oceanic and Atmospheric Administration (NOAA). Drought monitoring based on the SPI (Standardized Precipitation Index) and precipitation percentiles has also been established in south-eastern Europe through the efforts of the Drought Management Centre for South-eastern Europe.

INTERSUCHO, a drought monitoring system, has been operational in the Czech Republic since April 2014. It focuses on assessing soil moisture and vegetation condition. The system was established through collaboration between the Institute of Climate Change Research of the Czech Academy of Sciences, Mendel University, and Masaryk University in Brno. Weekly updates are provided, including a seven-day forecast based on GFS model data. In 2015, the Slovak Hydrometeorological Institute joined the project to extend drought monitoring to cover both the Czech Republic and Slovakia. Furthermore, the institute has developed its own drought monitoring system, which focuses on meteorological drought. This monitoring is updated weekly and evaluates SPI and SPEI indices on a daily basis, along with the CMI index on a weekly basis (Fendeková et al. 2018a).

Slovakia has experienced several significant drought periods, including in 1947, 1982–1983, 1992–1994, 2003, 2012, and 2015 (Fendeková et al. 2018b). Fendeková et al. (2017) assessed the occurrence of drought in surface and groundwater in the Kysuca River basin, taking into account the global climate situation and specific meteorological conditions in Slovakia. Slivová et al. (2016) evaluated drought from the perspective of groundwater over three hydrological years (2013–2015). Janáčová et al. (2018) investigated meteorological drought in lowland regions of Slovakia between 1981 and 2010, utilising SPI indices. Fendeková et al. (2011) assessed surface and hydrogeological drought in the Torysa River basin in Slovakia and its impact on aquatic habitat quality. Fendeková et al. (2018a) evaluated the prognosis and development in Slovakia using detailed knowledge of the country's climatic and hydrological conditions, aided by modelling tools, to create scenarios for the forecasted development of elements of the hydrological balance until 2100.

This contribution evaluates the vulnerability and risk of drought in the Banská Bystrica region of Slovakia. The AHP method and fuzzy logic are integrated within the ArcGIS environment, utilizing Landsat 8 satellite imagery. Satellite images are incorporated into the drought assessment process as factors for calculating the vulnerability of the territory to drought. These advanced analytical tools enable the quantification of drought dynamics and the assessment of drought risk for a specific area.

The Banská Bystrica Region is the largest territorial division within Slovakia, covering an area of 9,455 km² (Fig. 1). It comprises 13 districts: Banská Bystrica, Banská Štiavnica, Brezno, Detva, Krupina, Lučenec, Poltár, Revúca, Rimavská Sobota, Veľký Krtíš, Zvolen, Žarnovica, and Žiar nad Hronom. These districts contain a total of 516 municipalities, including 24 towns.

The Banská Bystrica Region displays a remarkable diversity of geomorphology, ranging from the towering mountain ranges of the Nízke Tatry Mountains in the north, to a landscape of alternating mountain massifs and valleys in the central part, and to undulating and flat terrain in the southern region. The highest point in the Banská Bystrica region is Ďumbier Mount in the Nízke Tatry, which stands at 2043 m above sea level, and the lowest point is 126 m above sea level, located in the village of Ipeľské Predmostie.

The region's climatic conditions are significantly influenced by its morphologically diverse terrain, which is marked by instability and variability. The northern part of the territory falls within a cold climatic zone, where temperature inversions occur in the valleys. Conversely, the Southern Slovak and Zvolen basins experience warmer and drier weather patterns, as reported in the Atlas of Landscape SR (2002).

The region's diverse orographic and geomorphic features are reflected in its varied geological

The hydrogeological conditions in the area are influenced by various factors, including the geological structure of the terrain, tectonic disturbances, geomorphological features, hydrology, and climate. The Banská Bystrica region contains all three hydrogeological regions found in Slovakia, categorized based on the predominant type of permeability (Atlas of Landscape SR, 2002).

Intergranular permeability is predominantly found in the northwest around the Hron River and to the south and southeast of the area.
Karst and karst-fracture permeability are predominantly found in the north and east of the area.
Fracture permeability is found in the central part, west, and southeast of the Banská Bystrica region.
The most significant hydrogeological units are composed of Mesozoic rocks and Quaternary cover formations.

3.1 Drought Hazard

The methodological approach for the estimation of the drought hazard index (DHI) is based on the analysis of the SPI calculated for the characterisation of drought in the region. The calculation of the SPI is based on a long-term series of monthly precipitation data (a recommended minimum of thirty years of data), transformed into a time series with a normal frequency distribution using a theoretical probability distribution, most commonly the gamma distribution. Average monthly precipitation was downloaded in ASCII format from the NASA Prediction of Worldwide Energy Resources (NASA POWER) online platform for the period 1981 to 2020, covering 34 gauging stations. In this study, SPIs were calculated with 6- and 12-month timescales. The drought hazard was calculated based on the probability of drought occurrence. The DHI maps can be categorised into four classes, namely low, moderate, severe and very severe.

3.2 Drought Vulnerability

The drought vulnerability assessment consisted of several steps (Fig. 2), which are described below. First, it was necessary to calculate the input parameters using satellite images, then the parameters were weighted using the AHP method, and finally the fuzzy method was used to create a drought vulnerability map.

3.2.1 Satellite Data

The Landsat mission has provided consistent, moderate-resolution Earth observation data from space for nearly 50 years. Landsat 8, launched on 11 February 2013, is the latest operational satellite in the Landsat series. Landsat 8 carries two sensors: The Operational Land Imager (OLI) and the TIR Sensor (TIRS). The TIRS sensor has two thermal bands (Band 10 and Band 11), while the OLI sensor has nine reflectance bands with a spatial resolution of 30 metres. Although the native spatial resolution of the TIR bands is 100 metres, the USGS publishes them at 30 metres through resampling (Sekertekin and Bonafoni 2020). Landsat 8 data were obtained freely from the USGS website (https://earthexplorer.usgs.gov/). Landsat 8 satellite imagery was used to calculate the vulnerability indices.

3.2.2 Vulnerability indices

Normalized Difference Vegetation Index (NDVI) – it is vegetation index that gives an indication of differences in green vegetation cover and can be used to estimate the density of vegetation over an area of land (Aouragh et al. 2023). This index was calculated from a Landsat 8 satellite image by dividing the difference in near-infrared (NIR) and red (R) reflections (visible wavelength observations) and then adding these measures together using the following formula:

$$\:NDVI=\:\frac{(NIR+Red)}{(NIR-Red)}$$

The NDVI itself varies between − 1.0 and 1.0. A negative value indicates abnormality and the presence of drought (Kumari et al. 2023). Deviation of NDVI from the long-term mean can be used to understand changes in vegetation due to climate impacts (Arab et al. 2022).

Top of Atmosphere (TOA) - spectral radiance (TOA) was calculated by multiplying the multiplicative rescaling factor (0.000342) of the thermal infrared (TIR) bands by the corresponding TIR band and adding the additive rescaling factor (0.1) (Mohan et al. 2021):

$$\:TOA=MX\:·{Q}_{cal}+AX$$

Where: MX means multiplicative rescaling factor, which is equal to the value 0.0003342,

Q_cal is Band 10,

AC is additive rescaling factor, which is equal to the value 0.1,

X means band number (Mohan et al. 2021).

Brightness Temperature (Bt) – corresponds to the temperature of a blackbody emitting an equivalent amount of radiation at a given wavelength. The calculation involves the inverse solution of the Planck function (Sekertekin and Bonafoni, 2020). It is determined by applying the thermal conversion constants found in the Landsat 8 image metadata along with the radiance (Mohan et al. 2021):

$\:Bt=(K2/(ln\:\left(K1/TOA\right)$ + 1)) – 273.15 (3)

Where: K1 (= 774.8853) and K2 (= 1321.0789) are thermal conversion constant,

TOA is Top of Atmosphere (Radiance).

Land Surface Temperature (LST) - is a crucial parameter for understanding the condition of the land surface. It is often used to calculate water and energy budgets at the surface-atmosphere interface and can also be used as a proxy for quantifying evapotranspiration and vegetation water stress in various contexts (Alharbi et al. 2022). LST provides insights into global climate change, hydrological processes, agricultural activities, and urban land use/land cover patterns. The computation of LST involves the use of the infrared spectral band of Landsat 8 and utilizes TIR band data. LST is derived from Brightness Temperature (BT) measurements obtained through the infrared spectral channels of geopositioned satellites. LST is an approximation that depends on factors such as albedo, vegetative cover, and soil moisture. It plays a crucial role in both surface energy and water balance considerations (Kumari et al. 2023). The computation of LST is described as follows (Mohan et al. 2021):

$$\:LST=Bt\:/(1+(\lambda\:·Bt)/(c2\:·\text{ln}\varepsilon\:\:)$$

The Land Surface Temperature (LST) for Landsat 8 is calculated using the measured radiance from Band 10 and its corresponding wavelength value (λ). The Land Surface Temperature (LST) for Landsat 8 is calculated using the measured radiance from Band 10 and its corresponding wavelength value (λ). The Land Surface Temperature (LST) for Landsat 8 is calculated using the measured radiance from Band 10 and its corresponding wavelength value (λ). The brightness temperature and emissivity are also utilized in the calculation process, as described by Mohan et al. (2021):

$$\:LST=(Bt/1+(0.00115·Bt/1.4388)·ln{E}_{c}$$

Normalized Difference Water Index (NDWI) - is the most appropriate index for mapping water bodies. This is because water bodies have strong absorbance and low radiation across visible to infrared wavelengths. NDWI uses the green and NIR bands in remote sensing images based on this characteristic. It effectively enhances water information in most scenarios but may overestimate water bodies due to its sensitivity to built-up land. The calculation of NDWI, according to Eq. (6), is as follows (Taloor et al. 2021):

$$\:NDWI=\:\frac{(NIR-Green)}{(NIR+Green)}$$

Normalized Difference Drought Index (NDDI) - is a drought-sensitive index used to monitor drought parameters across various global regions (Dobri et al. 2021). It is calculated using the equation below, which involves additional indices, namely NDVI and NDWI, both of which are responsive to the presence of vegetation and water in a given area (Artihanaur et al. 2022):

$$\:NDDI=\:\frac{(NDVI-NDWI)}{(NDVI+NDWI)}$$

Where NDVI means Normalized Difference Vegetation Index,

NDWI is Normalized Difference Water Index.

Proportion Vegetation Index (PVI) – is a method that is similar to a difference vegetation index, but it is sensitive to atmospheric variations. It can be calculated using Eq. 8 (Anindita et al. 2023):

$$\:PVI=\:\frac{(NDVI-\:{NDVI}_{min})}{({NDVI}_{max}-{NDVI}_{min})}$$

Soil Moisture Index (SMI) – is a measure of soil moisture that has a significant impact on the water cycle, both directly and indirectly. The SMI is calculated using an empirical parameterization that establishes a relationship between land surface temperature (LST) and normalized difference vegetation index (NDVI), as demonstrated in Saha et al. (2019):

$$\:SMI=\:\frac{({LST}_{max}-LST)}{({LST}_{max}-{LST}_{min})}$$

Where LST_max and LST_min represent the maximum and minimum surface temperature for a given NDVI, and LST is the land surface temperature.

3.2.3 The analytic hierarchy process (AHP method)

The Analytic Hierarchy Process (AHP) is used to analyse complex problems, drawing on principles from both mathematics and psychology. The AHP method was developed by Professor Thomas L. Saaty in 1970 and has since found extensive applications, including in the assessment of drought vulnerability (Alharbi et al. 2022). The methodology consists of three fundamental steps: constructing a pairwise matrix of conditional variables, assigning weights to individual input parameters, and evaluating the consistency of the resultant matrix. Input parameters are assessed using Saaty's scale, which expresses the degree of preference between two factors (Table 1). The input factors are compared pairwise in an 8 x 8 matrix. Each row (Ai) represents a factor that is compared to a factor in the corresponding column (Bi). Significance is determined using Saaty's scale, which transforms qualitative assessments into numerical quantities. The higher the value, the more significant the element in row Ai compared to the element in column Bi. If factor Bi is considered more important than Ai, it is necessary to assign the reciprocal value (Zarei et al. 2021). The weight of input factors is calculated from values in the pairwise matrix.

Table 1

Scales for pairwise comparisons (Saaty 2000)
Intensity of importance	Definition
1	equal importance
3	moderate importance of one over another
5	essential importance
7	demostrated importance
9	absolute importance
2, 4, 6, 8	intermediate values between the two adjacent judgments

The precision of the matrix can be assessed through the consistency ratio (CR), expressed as (Saaty 1980, 2000):

$$\:CR=\:\frac{CI}{RI}$$

where CI denotes the consistency index and is expressed as:

$$\:CI=\:\frac{\lambda\:-n}{n-1}$$

where $\:\lambda\:$ is highest values of matrix,

N means number in elements in matrix,

RI express random index.

In general, an acceptable consistency ratio is less than 0.1. If the value is higher, it means the result is not consistent and it is necessary to recalculate the matrix.

3.2.4 Fuzzy method

Fuzzy logic is an approach that calculates the 'degree of truth' rather than adhering to the binary 'true or false' structure of Boolean logic. The fuzzy theory incorporates the concept of membership functions to operate within a numeric range spanning from 0 to 1, indicating the level of certainty regarding membership. This framework acknowledges not only the extremes of truth (0 and 1) but also various intermediary states. In other words, fuzzy logic enables partial membership, as expressed mathematically by (Dalay et al. 2018):

µ_A (x) → [0, 1] (12)

In Eq. 12 µ_A (x) represents the membership grade of element x in a fuzzy set A, and the X denotes the universal set defined within the specific problem context.

When constructing a fuzzy logic-based model, it is crucial to carefully select the appropriate membership function. In this study, fuzzy membership functions are used to convert the input raster to a 0–1 scale, using a designated fuzzification algorithm. Fuzzification, as defined by Jain (2014), is the process of transforming an input into a fuzzy value. When there is uncertainty, ambiguity, or vagueness that causes indistinctness, a variable can potentially be fuzzy and represented by a membership function (Saha et al. 2021). The degree of membership is determined through the fuzzification process. The MSSmall algorithm was used for the fuzzification process, which calculates membership by considering the mean and standard deviation of the input data. In this approach, smaller values exhibit higher membership, and it serves as the default membership type. The resulting output may resemble the Small function, depending on the definitions of the mean and standard deviation multipliers. The default values for these multipliers are 1 for the mean multiplier and 2 for the standard deviation multiplier (Dalay et al. 2018).

After the input variables have been fuzzified, the next step is to execute the fuzzy overlay operation. This involves combining fuzzy membership raster data using a selected overlay type. For this study, we used the GAMMA overlay method, which multiplies the increasing fuzzy SUM and decreasing fuzzy PRODUCT effects, each raised to the power of gamma. The mathematical expression for fuzzy GAMMA is as follows (Alamdarloo et al. 2020):

µ_A (x) = (Fuzzy algebraic sum)^γ · (Fuzzy algebraic production)^{1− γ} (13)

where µ_A (x) is the fuzzy operation gamma equation and γ represents a number between 0 and 1. The optimal selection of the gamma value relies on the user to ensure a balanced compromise between the 'decreasing' and 'increasing' tendencies of fuzzy PRODUCT and fuzzy SUM, respectively.

3.3 Drought risk

The level of drought risk in a specific area is determined by both the hazard and vulnerability. The DRI is usually calculated by multiplying the DVI and the DHI. This index provides a comprehensive representation of the physical characteristics of drought, primarily driven by meteorological variables, along with socioeconomic factors influenced by human activities. The DRI serves as a quantitative measure of the consequences resulting from drought hazard. It provides insight into the extent of risk in a specific area during a drought period. The calculation of DRI follows the formula presented in Eq. 14 (Singh et al. 2019):

DRI = DHI ·DVI (14)

where DRI is drought risk index,

DHI represents drought hazard index,

DVI means drought vulnerability index.

The map depicting the risk of drought is created through map algebra in the ArcMap environment and is divided into five categories.

4.1 Vulnerability

The input factor maps were obtained from LANDSAT 8 and calculated following the methodology described in Chap. 3.2.1. These maps formed the basis for understanding various environmental parameters crucial for assessing drought vulnerability. In total, eight input maps were prepared, each representing distinct facets of the landscape's characteristics, ranging from vegetation health to surface temperature. The maps were created in the ArcGIS environment.

The input parameters underwent an evaluation process using the AHP method. This method determined the relative importance of each input factor in the context of drought vulnerability assessment by assigning weights to these factors. The significance of these factors in influencing the vulnerability of an area to drought events was quantified.

The computation of weights for input factors was a crucial step in the assessment process. The relevance of each factor was analysed using the AHP method, taking into account its potential impact on drought vulnerability. The accuracy and reliability of the weight computation were verified by evaluating the pair-wise comparison matrix and employing the consistency ratio as a measure of consistency.

Upon evaluation, the consistency ratio calculated for the pair-wise matrix was calculated to be 0.03, indicating a high level of consistency in the assigned weights.

The AHP analysis produced a matrix that outlines the input parameters and their corresponding weights. This matrix, shown in Table 2, illustrates the importance assigned to each input factor in the context of drought vulnerability assessment.

Additionally, the AHP method was used to determine the hierarchical importance of input factors. The results showed that NDVI and LST were assigned the highest weight, while TOA and BT were given the lowest weight. This suggests that NDVI and LST have a greater influence on determining drought vulnerability compared to TOA and BT.

Table 2

Pair-wise comparison matrix of the input parameters using AHP method
	NDVI	LST	NDWI	SMI	NDDI	PVI	TOA	BT	Weight
NDVI	1	2	4	3	5	5	9	9	0.32
LST	0.5	1	3	2	4	4	8	8	0.22
NDWI	0.25	0.33	1	0.5	2	2	7	7	0.11
SMI	0.33	0.5	2	1	3	3	7	7	0.15
NDDI	0.2	0.25	0.5	0.33	1	2	5	5	0.08
PVI	0.2	0.25	0.5	0.33	0.5	1	4	4	0.06
TOA	0.11	0.125	0.14	0.14	0.2	0.25	1	1	0.02
BT	0.11	0.125	0.14	0.14	0.2	0.25	1	1	0.02

Following the application of the AHP method, a fuzzy method was implemented as a subsequent step. During this process, input parameters derived from the AHP process underwent fuzzification. Each parameter was assigned a value between 0 and 1, indicating the degree of susceptibility to drought vulnerability. A value of 0 represented low susceptibility, while a value of 1 denoted high susceptibility.

Figure 3 shows the fuzzified maps resulting from this process, which visually represent the varying degrees of vulnerability across the studied area.

The drought vulnerability map (shown in Fig. 4) was created using the fuzzy overlap tool, which combined all input parameters to provide a thorough evaluation of vulnerability. The map was then divided into five distinct categories to aid interpretation and analysis.

Table 3 presents the percentage representation of each vulnerability class, providing insights into the distribution of vulnerability levels across the landscape. It is noteworthy that the high vulnerability class is the predominant category, comprising 27.44% of the study area, while the very low vulnerability class exhibits the lowest representation at 11.16%.

4.2 Drought risk

The map depicting drought risk was produced through map algebra, which involved multiplying the drought hazard and drought vulnerability maps. The cumulative values of all final maps were then classified using the 'natural breaks' method, which is implemented within the ArcGIS environment.

The resulting drought risk maps were classified into five distinct classes, representing varying levels of risk: very low, low, moderate, high, and very high. The highest proportion of areas (Table 3) falls within the first degree of risk, encompassing 32.53% of the total area, while the smallest representation is attributed to the last class (very high degree), accounting for 6.07% of the total area.

Figure 5 illustrates the resultant drought risk map. Analysis of the map shows that the areas with the highest degree of risk are concentrated in the northern and southern regions of the area. This spatial distribution highlights areas of heightened risk to drought events. Table 4 shows the DRI and DVI values for the evaluated gauging stations.

Table 3

Percentage representation of the area of the evaluated territory in individual categories for the drought vulnerability and drought risk map
class	vulnerability area [%]	risk area [%]
very low	11.16	32.53
low	18.15	25.44
moderate	21.42	17.84
high	27.44	18.3
very high	21.83	6.7

Table 4

Gauging stations and their altitudes, DVI and DRI
number	gauging station	altitude (m a.s.l.)	DVI	DRI
1	Banská Bystrica	358.299	low	very low
2	Banská Štiavnica	600.876	low	moderate
3	Brezno	495.489	very low	very low
4	Čierny Balog	533.159	moderate	moderate
5	Dekýš	466.979	very high	very low
6	Dolné Plachtince	190.083	low	high
7	Hnúšťa	365.676	moderate	low
8	Horný Tisovník	480.650	high	high
9	Hriňová	515.187	high	high
10	Chopok	2023.971	low	low
11	Jasenie	531.981	moderate	low
12	Jelšava	339.392	moderate	moderate
13	Kokava nad Rimavicou	482.543	moderate	low
14	Kremnica	741.216	high	very low
15	Krupina	359.406	high	very low
16	Lipovany	214.794	low	moderate
17	Ľubietová	565.061	very high	high
18	Lúčenec	211.052	low	moderate
19	Málinec	416.289	very low	high
20	Medovárce	187.021	moderate	very low
21	Muráň	458.106	high	low
22	Nenince	170.046	low	high
23	Nová Baňa	364.253	high	very low
24	Polomka	635.133	high	very high
25	Poltár	291.080	moderate	low
26	Rátková	308.100	very high	moderate
27	Rimavská Sobota	209.369	moderate	low
28	Senohrad	580.925	very high	moderate
29	Sklené Teplice	357.281	very high	very low
30	Sliač	293.374	moderate	moderate
31	Staré Hory	540.398	high	low
32	Tisovec	836.527	high	low
33	Vigláš	402.019	high	moderate
34	Žarnovica	313.666	high	very low

The primary aim of our contribution was to implement remote sensing of land, geographic information systems, and drought risk assessment. To calculate drought risk, we needed to evaluate both drought hazard and drought vulnerability. Hodasová et al. 2024 evaluated drought hazard for the study area. The DHI was estimated from the probability of drought occurrence, which was calculated using the SPI. It can be concluded that the probability of drought occurrence has the greatest impact on DHI.

The vulnerability of a country to drought is a crucial factor that affects the ecological and economic stability of regions. In this study, we assessed the vulnerability of a specific area using data from the LANDSAT 8 satellite and analytical methods such as AHP and fuzzy logic method. Drought, caused by meteorological drought, is increasingly affecting surface and groundwater, even in temperate climate zones. The assessment of drought in the three driest years of the 21st century in the Kysuce watershed in Slovakia revealed that the spread of drought in the hydrological system depends not only on precipitation-temperature ratios but also on the physical-geographical conditions of the area and the degree of water saturation (Fendeková et al. 2017). We evaluated eight factors in the weighting process using the AHP method. The analysis revealed that NDVI and LST are the most significant factors affecting vulnerability, which is consistent with previous research. Kumari et al. (2023) examined the relationship between different indices for analyzing factors from satellite imagery in drought assessment. The computed factors were visualized and combined in the resulting vulnerability map, which was subsequently categorized into four classes. In our specific context, we identified the NDWI as a potentially important factor, as it considers the presence of water in the soil. These findings emphasize the complex nature of the impact of environmental factors on dry conditions and highlight the need to integrate multiple factors when assessing vulnerability.

The AHP was chosen as the expert-based approach for weighting due to its reliance on subjective judgments. The process of determining weights involves constructing matrices to compare the relative importance of factors. The accuracy of these matrices is assessed by the consistency ratio. AHP can assign weights to both categories and parameters collectively. However, the AHP is considered one of the most intricate methods due to the several steps involved in weight determination. Firstly, partial matrices are compiled to compare categories of factors, followed by the main factor matrix where parameters are assessed holistically. Additionally, consistency ratios must be calculated before finalizing the weights. It is important to note that AHP is inherently subjective, which can significantly influence the calculations. Although some studies have reported high success rates for AHP (Shiravand and Bayat, 2023; Saha et al. 2021), its overall effectiveness is still uncertain. Therefore, the usefulness of AHP depends on the expert's knowledge of the causal factors, with efforts made to ensure objectivity in the results.

Comparison with existing literature demonstrates that our method of combining AHP and fuzzy logic provides a robust and reliable approach to assessing a country's vulnerability to drought. This combination of methods enables the quantification of the impact of individual factors and provides a visual representation of vulnerability in the form of map layers, which is crucial for effective management and planning of adaptation to dry conditions. Mahato et al. (2023) applied the AHP and fuzzy method to assess vulnerability to drought, which yielded good results. The authors used fuzzification of weighted input parameters to transform values into fuzzy values. Subsequently, the input data were processed through overlay fuzzy method to create a drought vulnerability map. The fuzzy logic method offers a versatile tool for vulnerability assessment that takes into account uncertainty, integrates different types of data, and when combined with the AHP method, also permits a certain degree of subjective assessment.

The vulnerability and risk of drought vary significantly within the studied area (Fig. 4 and Fig. 5). Identifying priority areas for adaptive measures and interventions can be crucial, and understanding specific vulnerability and risk patterns can aid in this process. However, high drought risk was present in regions where the DHI and DVI were high, confirming the findings of Nasrollahi et al. 2018. However, only 21% of the total evaluated area is considered to have a very high level of vulnerability, while a mere 6.07% is classified as having a very high level of risk (refer to Table 3). This suggests that risk is not directly proportional to vulnerability, but rather the product of hazard and vulnerability. Table 4 presents the variation between the DVI and the DRI across the different rainfall stations that were evaluated. Shiravand and Bayat (2023) found no direct relationship between precipitation and drought exposure. They also demonstrated that areas with high precipitation can still be susceptible to drought. In addition, their research revealed that areas with higher urbanization ratios and lower precipitation levels exceed the level of adaptation capacity to drought. In this case, it is evident that the correlation differences between individual rainfall stations are due to variations in elevation (refer to Table 4) and, therefore, dissimilar precipitation distribution.

In general, our study makes a significant contribution to comprehending the country's susceptibility to drought in the specified region. Our discoveries can be used as a foundation for future research, which can concentrate on more precise modelling of vulnerability and the creation of measures to alleviate the adverse effects of drought.

Conflict of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper

Funding

This research was funded by the Slovak Research and Development Agency under contract No. APVV–14–0174, as well as the Ministry of Education, Science, Research and Sport of the Slovak Republic under contract No. VEGA 1/0302/21.

Author contributions

All authors contributed to the study conception and design. Material preparation, data collection and analysis was performed by Kamila Hodasová. The first draft of the manuscript was written by Kamila Hodasová and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Aouragh MH, Ijlil S, Essahlaoui N, Essahlaoui A, Hmaidi AE, Ouali AE, Mridekh A (2023) Remote sensing and GIS-based machine learning models for spatial gully erosion prediction: A case study of Rdat watershed in Sebou basin, Marocco. Remote Sens Appl Soc Environ 30:1–18. https://doi.org/10.1016/j.rsase.2023.100939
Alamdarloo EH, Khosravi H, Nasabpour S, Gholami A (2020) Assessment of drought hazard, vulnerability and risk in Iran using GIS techniques. J Arid Land 12:984–1000. https://doi.org/10.1007/s40333-020-0096-4
Alharbi RS, Nath S, Faizan OM, Hasan MSU, Alam S, Khan MA, Bakshi S, Sahana M, Saif M (2022) Assessment of drought vulnerability through an integrated approach using AHP and geoinformatics in the Kangsabati river basin. J King Saud Univ SCI 34:1–10. https://doi.org/10.1016/j.jksus.2022.102332
Anindita R, Martono E, Nurjani E (2023) Urban cooling island vs urban heat island, who is the winner? study of the green spaces effect in coolingdown of urban heat in Kapanewon Depok, Sleman, Yogyakarta, Indonesia. E3S Web of Conferences. https://doi.org/10.1051/e3sconf/202346810012 ICST UGM 2023
Arab ST, Noguchi R, Ahamed T (2022) Yield loss assessment of grapes using composite drought index derived from landsat OLI and TIRS datasets. Remote Sens Appl Soc Environ 26:1–20. https://doi.org/10.1016/j.rsase.2022.100727
Artihanur SD, Widiatmaka, Setiawan Y, Marimin (2022) Normalized Difference Drought Index (NDDI) computation for mappping drought severity in Bojonegoro Regency, East Java, Indonesia. IOP Conf. Series: Earth Environ. Sci 1109. 10.1088/1755-1315/1109/1/012027
Atlas of Landscape of the Slovak Republic (2002) 1st edition. Ministry of Environment SR, Bratislava, Slovak Environmental Agency, Banská Bystrica
Blauhut V, Stoelzle M, Ahopelto L, Brunner MI, Teutschbein C, Wendt DE, Akstinas V, Bakke SJ, Barker LJ, Bartošová L, Briede A, Cammalleri C, Kalin KC, Stefano LD, Fendeková M, Finger DC, Huysmans M, Ivanov M, Jaagus J, Jakubínský J, Krakovska S, Laaha G, Lakatos M, Manevski K, Andersen MN, Nikolova N, Osuch M, Oel P, Radeva K, ROmanowicz RJ, Toth E, Trnka M, Urošev M, Reguera JU, Sauquet E, Stevkov A, Tallaksen LM, Trofima I, Loon AF, van Vliet MTH, Vidal J, Wanders N, Werner M, Willems P, Živkovic N (2022) Lessons from the 2018–2019 European droughts: a collective need for unifying drought risk management. Nat Hazards Earth Syst Sci 22(6):2201–2217. https://doi.org/10.5194/nhess-22-2201-2022
Cancelliere A, Di Mauro G, Bonaccorso B, Rossi G (2007) Drought forecasting using the Standardized Precipitation Index. Water Resour Manage 21:801–819. 10.1007/s11269-006-9062-y
Dayal KS, Deo RC, Apan AA (2018) Spatio-temporal drought risk mapping approach and its application in the drought-prone region of south-east Queensland, Australia. Nat Hazards 93:823–847. https://doi.org/10.1007/s11069-018-3326-8
Dobri RV, Sfica L, Amihaesei VA, Apostol L, Timpu S (2021) Drought Extent and Severity on Arable Lands in Romania Derived from Normalized Difference Drought Index (2001–2020). Remote Sens 13(1478):1–22. https://doi.org/10.3390/rs13081478
Elusma M, Tung C, Lee C (2022) Agricultural drought risk assessment in the Caribbean region: The case of Haiti. Int J Disaster Risk Reduct 83:1–20. https://doi.org/10.1016/j.ijdrr.2022.103414
Fendekova M, Demeterova B, Slivova V, Macura V, Fendek M, Machlica A, Gregor M, Jalcovikova M (2011) Surface and groundwater drought evaluation with respect to aquatic habitat quality applied in Torysa river catchment. Slovakia Ecohydrol Hydrobiol 11:49–61. https://doi.org/10.2478/v10104-011-0036-1
Fendeková M, Gauster T, Slivová V, Vrablíková D, Fendek M, Labudová L, Danáčová Z (2017) Surface and groundwater drought in the Kysuca river basin in three dryiest years of the 21st century. Podzemná voda 23(1):25–37
Fendeková M, Blaškovičová L, Bochníček O, Damborská I, Fendek M, Horvát O, Pekárová P, Slivová V, Vrábliková D (2018a) Prognosis of hydrological drought development in Slovakia. Comenius University in Bratislava, Faculty of Natural Science
Fendeková M, Gauster T, Labudová L, Vrábliková D, Danáčová Z, Fendek M, Pekárová P (2018b) Analysing 21st century meteorological and hydrological drought events in Slovakia. J Hydrol Hydromech 66(4):393–403. 10.2478/johh-2018-0026
Hoque MAA, Pradhan B, Ahmed N, Sohel SI (2021) Agricultural drought risk assessment of Northern New South Wales Australia using geospatial techniques. Sci Total Environ 756:1–15. https://doi.org/10.1016/j.scitotenv.2020.143600
Jain S (2014) Design and simulation of fuzzy membership functions for the fuzzification module of fuzzy system using operational amplifier. Int J Syst Control Commun 6:69–83. https://doi.org/10.1504/IJSCC.2014.062805
Janáčova T, Labudová L, Labuda M (2018) Meteorological drought in the parts of Slovakia with lowland features in 1981–2010. Geographia Cassoviensis 12(1):53–64
Kim JS, Jain S, Lee JH, Chen H, Park SY (2019) Quantitative vulnerability assessment of water quality to extreme drought in a changing climate. Ecol Ind 103:688–697. https://doi.org/10.1016/j.ecolind.2019.04.052
Kumari M, Kumar D, Vaishnavi (2023) Dynamic drought risk assessment and analysis with multi-source drought indices and analytical hierarchy process. Int J Environ Sci Technol 20:2839–2856. https://doi.org/10.1007/s13762-022-04041-x
Mahato S, Mandal G, Kundu B, Kundu S, Joshi PK, Kumar P (2023) Comprehensive drought vulnerability assessment in north-western Odisha: a fuzzy logic and analytical hierarchy process integration approach. Water 15:3210. https://doi.org/10.3390/w15183210
Mohan K, Ramasamy A, Varghese J (2021) Drought severity assessment using automated land surface temperature retrieval technique. Arab J Geosci 14:2358. https://doi.org/10.1007/s12517-021-08672-1
Mukherjee S, Mishra A, Trenberth KE (2018) Climate change and drought: a perspective on drought indices. Curr Clim Change Rep 4:145–163. https://doi.org/10.1007/s40641-018-0098-x
Nam WH, Hayes MJ, Svoboda MD, Tadesse T (2015) Drought hazard assessment in the context of climate change for South Korea. Agric Water Manag 160:106–117. https://doi.org/10.1016/j.agwat.2015.06.029
Nasrollahi M, Khosravi H, Moghaddamnia A, Malekian A, Shahid S (2018) Assessment of drought risk index using drought hazard and vulnerability indices. Arab J Geosci 11:606. https://doi.org/10.1007/s12517-018-3971-y
Palchaudhuri M, Biswas S (2016) Application of AHP with GIS in drought risk assessment for Puruliya district, India. Nat Hazards 84:1905–1920. 10.1007/s11069-016-2526-3
Rahmati O, Falah F, Dayal KS, Deo RC, Mohammadi F, Biggs T, Moghaddam DD, Naghibi SA, Bui DT (2020) Machine learning approaches for spatial modelling of agricultural droughts in the south-east region of Queensland Australia. Sci Total Environ 669:134230. https://doi.org/10.1016/j.scitotenv.2019.134230
Saaty TL (1980) The analytical hierarchy process. McGraw–Hill, New York
Saaty TL (2000) Decision making for leaders: the analytical hierarchy process for decisions in a complex world. RWS, Pittsburgh
Saaty TL (2008) Decision making with the analytic hierarchy process. Int J Serv Sci 1:83–98
Saha A, Patil M, Goyal VC, Rathore DS (2019) Assessment and impact of soil moisture index in agricultural drought estimation using remote sensing and GIS techniques. Proceeding 7(1):2, https://doi.org/10.3390/ECWS-3-05802
Saha S, Kundu B, Paul GC, Mukherjee K, Pradhan B, Dikshit A, Maulud KNA, Alamri AM (2021) Spatial assessment of drought vulnerability using fuzzy-analytical hierarchical process: a case study at the Indian state of Odisha. Geomat Nat Hazards Risk 12(1):123–153. https://doi.org/10.1080/19475705.2020.1861114
Saha A, Pal SC, Chowdhuri I, Roy P, Chakrabortty R, Shit M (2023) Vulnerability assessment of drought in India: Insights from meteorological, hydrological, agricultural and socio-economic perspectives. Gondwana Res 123:68–88. https://doi.org/10.1016/j.gr.2022.11.006
Saini D, Singh O, Sharma T, Bhardwaj P (2022) Geoinformatics and analytic hierarchy process based drought vulnerability assessment over a dryland ecosystem of north-western India. Nat Hazards 114:1427–1454. https://doi.org/10.1007/s11069-022-05431-7
Sekertekin A, Bonafoni S (2020) Sensitivity analysis and validation of daytime and nighttime land surface temperature retrievals from Landsat 8 using different algorithms and emissivity models. Remote Sens 12:2776. https://doi.org/10.3390/rs12172776
Shiravand H, Bayat A (2023) Vulnerability and drought risk assessment in Iran based on fuzzy logic and hierarchal analysis. Theor Appl Climatol 151:1323–1335. https://doi.org/10.1007/s00704-022-04323-x
Singh GR, Jain MK, Gupta V (2019) Spatiotemporal assessment of drought hazard, vulnerability and risk in the Krishna River basin, India. Nat Hazards 99:611–635. https://doi.org/10.1007/s11069-019-03762-6
Slivová V, Kullman E, Gavurník J (2016) Analysis and occurrence of groundwater drought in hydrological years 2013, 2014 and 2015. Podzemná voda 22(1):40–51
Spiliotis M, Iglesias A, Garrote L (2021) A multicriteria fuzzy pattern recognition approach for assessing the vulnerability to drought: Mediterranean region. Evol Syst 12:109–122. https://doi.org/10.1007/s12530-020-09332-7
Taloor AK, Manhas DS, Kothyari GC (2021) Retrieval of land surface temperature, normalized difference moisture index, normalized difference water index of the Ravi basin using Landsat data. Appl Comput Geosci 9:100051. https://doi.org/10.1016/j.acags.2020.100051
Yuan B, Wang S, Guo L (2023) Drought vulnerability assessment of winter wheat using an improved entropy-comprehensive fuzzy evaluation method: a case study of Henan province in China. Atmosphere 14(5):779. https://doi.org/10.3390/atmos14050779
Zarei AR, Moghimi MM, Koohi E (2021) Sensitivity Assessment to the Occurrence of Different Types of Drought Using GIS and AHP Techniques. Water Resour Manag 35:3593–3615. https://doi.org/10.1007/s11269-021-02906-3

Download PDF

Reviewers agreed at journal
15 Aug, 2024
Reviewers invited by journal
07 Aug, 2024
Editor assigned by journal
07 Aug, 2024
First submitted to journal
07 Aug, 2024

You are reading this latest preprint version

Satellite-based Drought Assessment: Integrating Ahp Method and Fuzzy Logic for Comprehensive Vulnerability and Risk Analysis

Status:

Version 1

Abstract

Figures

1. INTRODUCTION

2. STUDY AREA

3. MATERIAL AND METHODS

3.1 Drought Hazard

3.2 Drought Vulnerability

3.2.1 Satellite Data

3.2.2 Vulnerability indices

3.2.3 The analytic hierarchy process (AHP method)

3.2.4 Fuzzy method

3.3 Drought risk

4. RESULTS

4.1 Vulnerability

4.2 Drought risk

5. DISCUSSION AND CONCLUSION

Declarations

Funding

Author contributions

References

Status:

Version 1