Introduction

Drought is a “creeping phenomenon”, thus it is difficult to determine its onset and end (Prakash et al. 2006). It is predicted that 4 billion people—one half of the world’s population will experience severe water stress by the year 2025 (Diwan 2002). Vietnam, a tropical nation in Southeast Asia, is one of the regions most susceptible to the effects of climate change and natural disasters like droughts (Change 2013). In the period 2015–2016, Vietnam experienced the worst severe drought in the preceding 90 years, which severely harmed agricultural productivity and farmer income (Vietnam 2016). According to a recent study by Sam et al. (2019), the droughts in the Central Highlands are getting longer and more severe, and this trend is expected to continue in the near future (Sam et al. 2019). The Ba river basin, the largest river basin in the Central Highlands, faces many serious droughts, especially in the dry season, drought can take place very fiercely for a long time (Thanh et al. 2019). According to the Central Highlands Regional Hydrometeorological Center, in 2018–2019, some districts of Gia Lai Province in the Ba river basin were suffering from a severe drought. In particular, the largest hydroelectric reservoir in this basin, the Ka Nak hydroelectric reservoir, was also at its driest state ever (Tham et al. 2022).

Desertification, land degradation, and drought are all effects of human activity (Van Loon et al. 2016; Rangecroft et al. 2018). These activities complicate and raise the hazard of drought rendering its happening and severity more challenging to estimate (Diffenbaugh et al. 2015). Like other natural disasters, the effects of drought can be mitigated through early detection (Sruthi and Aslam 2015). Therefore, data collection and risk assessment in drought-prone areas are essential for the future planning of land development, drought prevention, and reducing the negative effects from it. The literature has shown that there are many methods to assess the drought hazard such as geospatial technologies (Ansari Amoli et al. 2022), remote sensing (Belal et al. 2014), GIS techniques (Zareiee et al. 2011), rainfall index (Cancelliere et al. 2007), etc. Remote sensing and GIS play a crucial role in identifying, evaluating, and managing droughts because they provide current information on both spatial and temporal scales (Abuzar et al. 2017). In addition, delineation of drought hazard zones can be implemented using different drought indices. Major drought indices use parameters like drought normalized difference vegetation index (NDVI), normalized difference water index (NDWI), normalized difference moisture index (NDMI), standard precipitation index (SPI), land surface temperature (LST) (Bhattacharya et al. 2021), rainfall index (Cancelliere et al. 2007), etc. In recent years, various studies have highlighted that multi-criteria decision analysis (MCDA) is a useful method for disaster management such as flood hazard (Dung et al. 2021a, b, 2022; Hussain et al. 2023), groundwater potential (Abrar et al. 2021; Dang and Nguyen 2021), landslide (Bahrami et al. 2021) etc. The AHP approach, one of the most popular MCDA models, has also been used to manage these events. According to Sivakumar et al. (2021), the AHP is a well-known technique for addressing challenging issues relating to drought vulnerability (Sivakumar et al. 2021). This method successfully applied in numerous research of drought risk assessment through the integration of the AHP algorithm, remote sensing (RS), and GIS techniques. Some scientists combined the AHP with GIS technology in drought risk evaluation (Palchaudhuri and Biswas 2016; Wijitkosum 2018; Wijitkosum and Sriburi 2019; Promping and Tingsanchali 2020; Foyhirun and Promping 2021; Nugrahani and Santosa 2021). Others assessed drought vulnerability by integrating AHP and GIS (Ekrami et al. 2016; Sivakumar et al. 2021; Alharbi et al. 2022), or established drought hazard map (Koem et al. 2022), analyzed spatial distribution of the drought hazard index (Olii et al. 2021) using AHP an GIS method. The results obtained from the studies show that the AHP approach has the ability to be used to conduct a spatial analysis based on a variety of parameters to determine the distribution of the region that could be affected by drought (Nugrahani and Santosa 2021) or provides a good solution for drought risk modeling (Palchaudhuri and Biswas 2014) and drought-prone area mapping (Koem et al. 2022).

To date, the research on drought in the Ba river basin was implemented using Landsat remote sensing data to determine how seriously climate change influences the drought in the Ba River basin (Tham et al. 2022). However, the result obtained from this study did not show the levels of drought risk spatial distribution. Besides, Thanh et al. (2019) used the Soil and Water Assessment Tool (SWAT) model to calculate the hydrological drought index and generate the drought classification map in this basin (Thanh et al. 2019). However, this model only uses hydro-meteorological data such as temperature, rainfall, and runoff without mentioning topographic factors such as slope, elevation, or information on soil, land use, etc. which significantly affect drought risks. Therefore, the primary objective of this study is to establish a drought risk map taking into account influencing factors including physical, hydrometeorology criteria and using remote sensing, GIS, and AHP algorithms.

Research areas

The Ba River basin is located on the South Central Coast, the coastal land of South Central Vietnam, with an area of about 13,900 km2, including the provinces of Gia Lai, Dak Lak, and Phu Yen. The study area lies between latitude 12° 35′ N to 14° 38′ N and longitude 108° 00′ E to 109° 55′ E longitude (Nguyen et al. 2021). The elevation of the research region varies from 0.5 to 2051 m above mean sea level. Air temperature increases gradually from north to south, from west to east, and from upstream to downstream. The annual average temperature ranges widely, from 21.5 to 23.5 °C in the upstream areas, and to 21–26 °C in the downstream areas. The relative humidity is high, 77–86.5%. The number of sunshine hours in the Ba river basin every year is about 2180–2540 h/year. Moreover, depending on the location, the annual evaporation amount is about 1000–1500 mm (Fig. 1) (Dang and Nguyen 2021). Additionally, the average annual rainfall is about 1740 mm and is very unevenly distributed in different parts of the basin. However, in some recent years, the rainfall in the basin only reached about 60% and the rainy season ended early, so the drought occurred in the rainy season last year and lasted until next year (Dung 2017). Prolonged drought also seriously affects the operation of the hydroelectric system in the Ba River basin (Thanh et al. 2019). In the study area, there are a variety of land use patterns including rocky mountain, perennial industrial tree, residential land, specially used land, grass with shrubs, natural forest, planted forest, shrub with the wooden trees, salt sand, sandy, paddy rice, annual crops, river, lake. The features of water retention and infiltration into the ground vary depending on the type of land use (Minh et al. 2022).

Fig. 1
figure 1

The Ba river basin

Full size image

Materials and methods

Materials

Drought severity depends on the impact of various factors including meteorological, hydrological, and human criteria. Although some hydrological and meteorological variables such as groundwater depth, groundwater volume, topographic wetness index, evapotranspiration, solar radiation, etc., are also crucial for predicting the likelihood of drought hazards, spatial data on these criteria are rarely available because there are so few meteorological observatories. Moreover, some of these factors can be assessed using satellite-derived observations, but challenges in obtaining cloud-free satellite information during the monsoon season limit their applicability (Pandey et al. 2012).

In the AHP processing, choosing the factors affecting drought occurrence plays an important role. According to Cabała (2010), when using the AHP approach, it is advised to utilize no more than nine criteria at most because psychologists believe that an individual can compare and reliably rank up to nine objects at once (Cabała 2010). Furthermore, the determining of affected factors relies not only on the physical geography characteristics of the study region but also on the availability of data (Dung et al. 2021a, b). The analysis of the literature on drought and the availability of data established a fundamental assumption underlying this method that there are six main criteria including rainfall, slope, drainage density, land use, temperature, and humidity contributing to drought occurrence in the Ba river basin.

In this study, a variety of data types were used for mapping drought hazards. To create the criteria layers, the data and information were gathered from a variety of sources and used GIS techniques. Landsat 8 Operational Land Imager (OLI), for the upstream and midstream area with path = 123, row = 50, the downstream area with path = 123, row = 51 was used to acquire land use data (Table 1). Besides, information on rainfall, temperature, and humidity obtained from the National Center for Hydro-Meteorological Forecasting was used to generate maps of those factors. In addition, DEM with a 30-m resolution created from the topographic map was utilized to produce slope and drainage density maps.

Table 1 Data used in the present study
Full size table

Methods

There are two main steps to solve the objective of the paper including (1) calculating the weights for the various factors using AHP; and (2) creating maps illustrating various forms of drought utilizing GIS. In the present research, the AHP was used to determine each parameter's weight based on how it influenced the drought occurrence. Figure 2 illustrates the flowchart of the research. To create a drought hazard map, the AHP method was applied to create a pairwise matrix of conditional variables, rank and weight the criteria in agreement with the importance of each factor governing the drought hazard evaluation. This approach arranged all the factors hierarchically to enable pairwise comparison. The relative relevance of each parameter was determined using a comparison scale proposed by Saaty (1987), which used integer numbers from 1 to 9, where 1 represents equal importance or the least contribution and 9 indicates extreme importance or the highest contribution (Saaty 1987; Hussain et al. 2021). These values are prepared following the recommendations of experts who have remarkable experience in drought studies and extensive knowledge of the causes of droughts as well as their impacts on the research areas.

Fig. 2
figure 2

The methodological flow chart used in this study

Full size image

In this study, the questionnaire surveys on comparison ratings on a scale of 1–9 were created and sent to 45 knowledgeable individuals who are national and international experts in the areas of surveying, soil, geomatics, meteorology, geology, water resources management, and hydrology. The pairwise comparison matrix of this study was created by the opinions of the specialists in the drought-related fields. In the paper, the eigenvector method is utilized for the calculation of the largest eigenvalue to determine the weights. To verify the accuracy of the weights of the drought criteria, consistency indicators such as consistency index, and consistency ratio were calculated based on Eqs. 1, and 2. The matrix has a good consistency when CR is less than 0.10; otherwise, the matrix must be altered (Saaty 1987).

$${\text{CR}} = \left( {\lambda_{\max } {-}n} \right)/\left( {n - 1} \right),$$
(1)
$${\text{CI}} = {\text{CR}}/{\text{RI}},$$
(2)

where λmax the eigenvalue, CI is the consistency index, CR is the consistency ratio, RI is the random inconsistency index (depending on the number of factors or the order of the matrix), n is the number of parameters.

After determining the weights, a weighted overlay analysis was performed on the layers. The drought hazard zones were detected in five classes based on the index value calculated using Eq. 3.

$${\text{DHZI }} = {\text{ Ra}}_{{\text{r}}} {\text{Ra}}_{{\text{w}}} + {\text{ Dr}}_{{\text{r}}} {\text{Dr}}_{{\text{w}}} + {\text{ Sl}}_{{\text{r}}} {\text{Sl}}_{{\text{w}}} + {\text{ So}}_{{\text{r}}} {\text{So}}_{{\text{w}}} + {\text{ Ve}}_{{\text{r}}} {\text{V}}_{{\text{w}}} + {\text{ La}}_{{\text{r}}} {\text{La}}_{{\text{w}}} ,$$
(3)

where DHZI is drought hazard zone index, Ra is rainfall, Dr is drainage density, Sl is slope, So is soil, Ve is vegetation, and La is land use. The suffixes r and w stand for the rank and weight of each layer, respectively.

Results and discussions

Determining the weights of criteria affecting drought hazard

In this paper, an expert questionnaire survey was conducted to ensure the importance and priority of criteria in drought hazard analysis. The selected experts specialized in surveying, soil, geomatics, geology, water resources management, and hydrometeorology from academics, universities, and research institutions. Factors affecting drought have different roles and importance, so it is necessary to assess quantitatively the significance of these influencing criteria. Table 2 is the summary result of the priority of parameters influencing drought in the study area. In this table, the minus sign (−) shows the less importance (less influence) of the preceding factor compared to the following one in the pair of compared criteria.

Table 2 The summary result of the priority of factors affecting drought
Full size table

Here, the pairwise comparison matrix is generated as shown by the 6 × 6 matrix in Table 3, in which the component values in each row were provided for each factor affecting drought formation based on their relative importance.

Table 3 Pairwise comparison matrix of drought occurrence in the Ba river basin
Full size table

AHP analysis determines the weights for the primary parameters and Table 4 provides the list of weights for each criterion. Rainfall has the largest weight with a value of 39.7%, followed by other factors as illustrated in Table 4. The accuracy of the obtained weights can be checked by using consistency indicators. As a result, the consistency ratio (CR) of this study is 0.07 or 7%, which is less than the permissible level of 10%. This indicates that the CR of the weighted coefficient is acceptable (Saaty 1987, 2008).

Table 4 The weight for criteria
Full size table

Generating the thematic maps

In the present study, we have used satellite image data and numerous collateral information to generate spatial maps pertaining to 6 factors, which were produced in a GIS environment. Using GIS, the thematic maps were analyzed, combined, and multiplied with, respectively, appropriate weights and ratings to create a drought hazard map for the research area. The rank and weightage of individual thematic layers are given in Table 5. According to (Dibyosaputro 1999), the hazard interval class can be determined by the following equation:

Table 5 Rank and weightage of individual thematic layers for drought hazard zonation
Full size table
$$i=\frac{h-l}{c}$$
(4)

In which i the number of distance interval class, h the number of the highest value, l the number of the lowest value, c the number of classes desired.

Rainfall

Rainfall or precipitation is the most important climate factor (Wijitkosum 2018). Many scientists believe that rainfall is the most remarkable factor that affects drought phenomenon (Koem et al. 2022). In the present study, the results in Table 5 also show that rainfall accounts for the highest weight among the six factors affecting drought hazard. Areas with low rainfall are more likely to experience a severe drought hazard compared with regions receiving a higher amount of precipitation (Pandey et al. 2012). Regions that are prone to drought are characterized by low rainfall and humidity (Esfahanian et al. 2017). In this study, the inverse distance weighted (IDW) method was applied to interpolate the point-based datasets for generating a spatial layer of annual rainfall in GIS (Ullah and Zhang 2020). The rainfall map for the study basin was classified into five categories having minimum and maximum precipitation of 1200 and 2200 mm, respectively. Based on the smallest, highest rainfall, and the number of classes desired, applying formula (4), the number of distance interval class can be determined as 200 mm. Figure 3a shows the hierarchical map of the impact levels of rainfall on drought risk (five hazard levels), in which the highest drought hazard areas are assigned to the parts of the basin receiving the least amount of rainfall, for example in some parts of Ayun Pa, Phu Thien, Iapa, Ea H’leo, Krong Pa districts. Whereas, the rainfall magnitude was found creasing toward the Southern and the rating thus assigned lowest.

Fig. 3
figure 3

Hierarchical map of the influence levels on drought hazard of factors: a rainfall; b drainage density

Full size image

Drainage density

Drainage density is calculated by dividing the length of the streamline by the basin’s areas (Koem et al. 2022). Drainage density was selected as a relevant factor for mapping drought hazard. According to Pandey et al. (2012), compared to a region with little drainage, an area with high drainage density has more water contact locations (Pandey et al. 2012). Therefore, it can be remarked that the regions with a high density of drainage systems are less likely to experience droughts and vice versa (Jose et al. 2015). The digital elevation model with 30 × 30 m resolution generated from the 10-m interval contour of the topographical map was used to produce the drainage density map. This obtained map was categorized into five classes by the line density analysis tool of ArcGIS software as Fig. 3b. The range of drainage density of the study area is from 0.095 to 2.79 m/m2 and it was classified into five classes for the study area. Based on the smallest and highest values of drainage density as well as the number of classes, the number of distance interval class can be calculated by Eq. (4) as 0.539.

Slope

According to Hoque et al. (2020), vegetation, soil properties, and evapotranspiration rate are influenced by the exposure of the slope to sunshine and wind (Hoque et al. 2020). Moreover, there is an inverse relationship between the slope of the land surface and infiltration, where a decrease in slope would increase in infiltration and vice versa (Ekrami et al. 2016; Al-Masnay et al. 2022). Therefore, the slope was also utilized in drought hazard evaluation. Like the drainage density map, the slope distribution map was also derived by DEM which was generated from a topographical map using the Arc GIS 10.2 spatial analysis tool. To zone the drought risk in the basin, it is necessary to classify the slope according to the drought-affecting level. According to (Cu 2003), when researching scientific arguments for finding solutions in avoiding and limiting the consequences of floods in the Ba River basin, terrain slopes are divided into areas with slope levels as follows:

  • Very gently inclined (slope < 3°).

  • Gently inclined (slope from 3° to 8°).

  • Moderately inclined (slope from 8° to 15°).

  • Steep (slope from 15° to 25°).

  • Very steep (slope > 25°).

The slope of research areas ranges from 0° to 43.6°. For the effect of assigning ranks, the slope of the study areas is categorized into five classes. The lowest rating was assigned to the flat terrain with a slope value of 0–3 degrees based on the classification of Cu (2003). This rating gradually increased as the slope value rose (Table 5). The hierarchical map of the impact levels of slope on drought risk with five classifications is shown in Fig. 4a.

Fig. 4
figure 4

Hierarchical map of the influence levels on drought hazard of factors: a slope; b land use

Full size image

Landuse

Land use is defined by the way humans dominate the land to meet their needs. Changes in land coverage such as deforestation may cause drought. Land use is also one of the parameters that contributes to drought occurrence (Palchaudhuri and Biswas 2016). Each land use type has various impact degrees on the drought hazard (Koem et al. 2022). In this paper, land use map data was derived from the Landsat 8 satellite image of 2019 with 30-m spatial resolution. The supervised image classification algorithm was performed to categorize and determine the type of land use. The research region consists of fifteen types of land use namely: rocky mountain, perennial industrial tree, residential land, specially used land, grass with shrubs, natural forest, planted forest, shrub with wooden trees, salt sand, sandy, paddy rice, annual crops, river, and lake. Classifying the degree of impact on drought risk depends on the characteristics of each type of land use and the water-holding capacity of each kind. The land use hierarchical map was separated into five classifications as Fig. 4b based on the influence of each land use type on the occurrence of drought (Table 5).

Temperature

Temperature is also one of the most important factors in determining drought hazard zones. According to Hoque et al. (2020), places with higher temperatures are more susceptible to drought than regions characterized by lower temperatures (Hoque et al. 2020). In other words, temperature is directly proportional to the drought hazard levels, increasing the temperature will raise the sensitivity to drought severity, so the greater weight value was allocated to the regions with the higher temperature and vice versa (Heydari Alamdarloo et al. 2020). The average temperature of the study areas varied from 21.7 to 26.5 °C was classified into five classes affecting drought occurrence (Table 5). Based on the lowest and highest temperature as well as the number of classes, the number of distance interval class can be determined by Eq. (4) as 0.9 °C. The map of the annually average temperature was produced using the IDW interpolation algorithm. The temperature hierarchical map with five hazard categories is shown in Fig. 5a.

Fig. 5
figure 5

Hierarchical map of the influence levels on drought hazard of factors: a temperature; b humidity

Full size image

Humidity

One of the crucial effective parameters in drought occurrence is humidity (Koem et al. 2022). It cannot be denied that drought hazard is typically more prevalent in low-humidity regions. That means the higher the humidity, the lower the drought hazard. Therefore, humidity has an inverse relationship with drought occurrence. In the study area, the annual average relative humidity varies between 77 and 86.5%. Weights for each humidity class were assigned depending on the level of drought. The map of humidity has been generated by using the IDW interpolation algorithm. Through the number of classes, the maximum and minimum humidity, and the number of distance interval class can be calculated as 1.9% using Eq. (4). The map of humidity is divided into five categories based on how they affect the drought formation as shown in Fig. 5b.

Drought hazard map

The weights of the six factors were utilized as input data for the GIS model. The relative importance of different thematic layers and their corresponding classes were used for creating the drought hazard map. By calculating the drought hazard zone index (DHZI) according to the formula (3), we obtained index values and classified them into five classes corresponding to five hazard levels. Overlaying the boundary map with the final result map, the drought hazard range will be determined by areas. The drought hazard zonation map generated by this method is shown in Fig. 6.

Fig. 6
figure 6

The drought hazard zone map of the Ba river basin

Full size image

The area of the hazard levels is shown in Table 6. The final output of the model for the drought occurrence indicated that approximately 5478.735 km2 (39.2%) of the total regions was prone to a moderate drought hazard. Notably, nearly half of the basin area is at high and very high drought hazard (accounting for 46.08%), while only 1.36% and 13.36% are considered to have a very low and low hazard, respectively. The severity of drought in the southwest part of the study area is very low and low on account of higher rainfall, thicker drainage density, greater humidity, and lower temperature. Besides, extreme drought is detected to be mainly focused in the central part of the basin, especially in Ayunpa, Krong Pa, Phu Thien, and Iapa districts and some scattered portions of the Kong Chro, Chu Se, Dak Po districts where the obtained rainfall is the lowest in the basin. Although, these areas have good drainage density and some types of land use had a moderate and high water holding capacity such as salt sand, sandy, paddy rice, annual crops, grass with shrubs, and natural forest, these places had high temperatures and especially the annual average precipitation was the least in the whole basin. This explains why drought in these regions was considered to be relatively high and very high.

Table 6 Drought hazard zone
Full size table

The findings of this study are relatively consistent with the results of Thanh et al. (2019) which are related to mapping drought zone in the Ba river basin in the context of climate change (Thanh et al. 2019). This study showed that the most severe drought areas were concentrated in the central parts of the basin, while (Thanh et al. 2019) also assessed this area at a very high and high hazard of drought. There are some places where the results of the two methods differ slightly in terms of hazard levels due to different usage data. Because the study of Thanh et al. (2019) used data on average temperature, precipitation, and flow, this study, in addition to using temperature and rainfall data, also utilized humidity information and topographic data such as slope, land use, and drainage density. In addition, the obtained findings can be verified based on the actual drought situation in the study area. First of all, the April 2010 drought took place on a large scale in the districts of Kong Chro, Krong Pa, Kbang, and Dak Po, causing the crop area to be affected by drought up to 7000 ha (Tuyet 2010). These are also areas shown on the map with a high and very high hazard of drought. In March 2016, some provinces of Central Highlands, Vietnam, including Gialai (mainly in the districts of Chu Se, An Khe, Ayun Pa, Dak Po, and Krong Pa) experienced the most severe drought in history. This “thirst” had caused nearly 50,000 ha of crops to be influenced and lost, and damaged more than 150 billion VND (Express 2020). Also in this area, in March 2020, the drought lasted for a long time, the small lakes were almost dry or at dead water levels. Many irrigation works have dried up, the reservoir capacity is only 10–30% compared to the original one. At that time, some hydropower had to operate in moderation or even stop working (Express 2020). Notably, in September 2021, some districts of Gia Lai province such as Kbang, Ia Pa, Kong Chro, Dak Po, and Krong Pa were affected significantly by a severe drought. The reason is that the weather has been dry and hot for a long time and there has been no rain since May (the first month of the rainy season), leading to a drought situation in the middle of the rainy season (Hien 2021). The areas that were severely influenced and significantly damaged during the droughts mentioned above are at in high and very high-hazard regions on the drought hazard zoning map. This proves that the obtained results by the used method are accurate and reliable.

The obtained results indicated that AHP can be used as an efficient approach to evaluate and generate the drought hazard zoning map in the GIS environment. The drought occurrence is strongly determined by important factors like rainfall, slope, land use, temperature, humidity, and drainage density. Furthermore, this research demonstrated that, by using this type and quantity of data, it is possible to delineate the drought hazard zones for the areas with similar physical conditions and regions that lack data on factors affecting drought hazards. This study suffered from the absence of data about hydrological data (groundwater volume, topographic wetness index) as well as meteorological data (solar radiation, evapotranspiration), therefore we also could not determine evapotranspiration which affects significantly on drought formation. Moreover, socio-economic factors (population, cultivators) were not also considered in depth, which is another possibility to take into account during conducting such research.

Nevertheless, the six parameters that were used for the present study are appropriate for delineating the drought hazard zones in the Ba river basin. To improve the accuracy of the identification of areas prone to drought, it is necessary to choose factors that are suitable to the physical conditions of the study area. However, the number of factors could be considered a limitation (less than 9 criteria), because the more criteria there are, the more challenging it is to solve a pairwise matrix. In addition, finding experts in relevant fields for pairwise comparison should also be emphasized as their judgments mainly determine the quality of the final product. However, in studies with spare data, and limited information, the AHP may be the best approach to solve various drought management problems.

Conclusions

In the present study, the combination of the AHP method in a GIS environment has been successfully applied to assess the spatial extent and characteristics of drought in the Ba river basin. Weights for each criterion were determined using a pairwise comparison method in AHP. The parameters were weighted and ranked based on their importance and the opinions of relevant experts. The obtained results indicated that the rainfall factor has the highest weight with a value of 0.39, followed by drainage density, temperature, humidity, land use, and slope. According to the final map, the research area could be categorized into five distinct drought hazard zones including very low, low, moderate, high, and very high. The findings revealed that high and very high drought hazard regions were predominantly located in the central part of the basin with 46% of the study area. Moreover, 14.72% of the total areas were identified in low and very low drought hazard categories. Moderate hazard regions spread over the catchment area and covered 39.2% of the study area. Because actual data on drought events during the drought hazard map creation period was not available, it was not utilized to compare with the drought hazard map. However, the findings can be verified by comparing them with existing and proven published research. In addition, the levels of drought hazard in the resulting map can be checked by comparing it with typical droughts that occurred in the past. The obtained result and validation confirmed that the drought hazard map by AHP and GIS is dependable and reliable. This research could assist decision-makers in developing mitigation strategies toward lessening the influence of drought on production and life.