Land Cover and Socioeconomic Analysis for Recommended Flood Risk Reduction Strategies in Java Island, Indonesia

Abstract

This study utilizes a novel approach by analyzing land use and socioeconomic factors to enhance flood risk reduction strategies on Java Island, Indonesia. Using datasets from inaRISK hazard profiles, GlobCover land cover data, and Indonesian national statistics, the research offers a methodology for mitigating flood risks in diverse geographic and socioeconomic landscapes. The study found flood exposure rates ranging from 1.1% to 63%, averaging 26.5% across 114 districts, and revealed a correlation between high flood exposure and socioeconomic indicators. Cluster analysis identified four types of regions with unique flood risk management needs. Socioeconomic analysis showed significant income and education level variations across clusters, with urban areas having a higher Gross Regional Domestic Product and better education levels than rural areas. This socioeconomic divide is crucial for understanding different regions’ capacities to respond to and recover from floods. Integrating socioeconomic factors with land use and flood exposure data allows for tailored disaster risk reduction strategies. For high-risk urban areas, structural interventions and community-focused initiatives are recommended, while rural areas benefit from sustainable land management practices. This study highlights the need for a combined approach to flood risk management and strategies, and provides a model adaptable to similar regions.

1. Introduction

The threat of climate change is intensifying with an increase in the frequency and intensity of extreme weather events becoming a stark reality [1]. Indonesia, as an archipelago, is vulnerable because of its unique environmental conditions—the confluence of topography, geology, weather patterns, and other natural forces—making it susceptible to various natural disasters, including floods, landslides, and coastal erosion [2]. Heavy rainfall is an ever-increasing threat that triggers devastating floods and landslides, posing significant threats to life, property, and infrastructure [3]. To mitigate these impacts, effective disaster risk reduction (DRR) strategies are essential for empowering communities to recover and minimize losses.

Many studies in various parts of the world have highlighted disaster mitigation by empowering communities. For instance, Cutter et al. proposed a methodology for incorporating a known measure of social vulnerability, the Social Vulnerability Index (SoVI), into USACE civil works planning for flood control [4]. Chen et al. presented a preliminary study on social vulnerability in the Yangtze River Delta region of China with the aim of replicating and testing the applicability of the place-based Social Vulnerability Index (SoVI) developed for the United States in the Chinese cultural context [5]. Similarly, Koks et al. examined how the combined assessment of hazard, exposure, and social vulnerability provides valuable information for the evaluation of flood risk management (FRM) strategies in the Netherlands [6].

In Indonesia, Sigit et al. conducted a study in Central Java using inaRISK, a national risk information platform, which showed a significant gap between the size of the flood-exposed economy and the number of poor individuals in different administrative regions [7]. Drawing on a case study focused on Indonesia, this research has demonstrated the importance of socioeconomic factors such as poverty, low education, and economic impact in flood risk assessment. This approach underscores the significance of addressing these factors, illustrating their relevance in enhancing local disaster risk reduction (DRR) capacity.

Although national government policy frameworks and stakeholder engagement have been cornerstones of past responses, shortcomings remain. Local governments often lack the necessary capacity for DRR, systematic learning is scarce, and mainstreaming DRR into the broader development agenda faces challenges in terms of commitment [8]. This highlights the need for a more comprehensive approach, drawing inspiration from studies that go beyond simply integrating social vulnerability and economic factors into flood risk assessments.

Looking at other factors, land use in Java has undergone a significant transformation due to population growth. The expanding residential areas consequently elevate flood risk [9]. Densely populated areas with more impervious surfaces increase this risk by reducing infiltration and increasing surface runoff, leading to a greater potential for material loss [10,11]. Therefore, the addition of land use assessment in terms of exposure is essential for developing more effective disaster risk management policies, especially in flood-prone areas, to enhance DRR capabilities. Even more so in rural areas, where population density is lower and assets are fewer compared to urban areas, it remains important to develop DRR strategies tailored to local characteristics to improve community welfare. Currently, Indonesia lacks a comprehensive hazard risk assessment method that spans urban to rural areas, considering both land and social characteristics. Given the complex interplay between land use/land cover, flood hazards, and the socioeconomic conditions of exposed communities, the poor people are the most vulnerable [12]. These issues are not unique to Indonesia, but are common to all developing countries with large socioeconomic disparities and inadequate resources for DRR strategy development assessments.

This study introduces a novel approach by strategically integrating land use and socioeconomic analysis into flood risk reduction strategies, providing a clearer understanding of flood risk characteristics at the administrative district level. This innovative methodology merges land cover/land use analysis with socioeconomic and demographic data, enabling policymakers and planners to develop sustainable, impactful methods to mitigate the effects of floods on communities. This approach bridges the gap from assessment to policymaking, facilitating the selection of DRR measures tailored to the specific risk profiles of different regions. This integration not only enhances the assessment process but also smoothens the transition to effective policy implementation, ultimately contributing to more resilient communities. Our study specifically addresses the following questions: How do different land use/land cover patterns influence flood risk in the administrative division scale? What are the key socioeconomic factors that exacerbate vulnerability to flooding in these regions? How can a combined analysis of land use and socioeconomic data inform more effective, region-specific DRR strategies?

In this study, we utilized Geographic Information Systems (GIS) to integrate flood vulnerability maps from inaRISK with population distribution and land cover datasets. Specifically, we focused on the relationship between land use and population distribution, attempting to classify 114 administrative districts into several clusters. Additionally, we evaluated socioeconomic vulnerability across various administrative regions and quantitatively analyzed exposure to flood risks. Based on these findings, tailored disaster risk reduction strategies were proposed for each region to inform policy development and enhance flood risk management practices on Java Island.

2. Materials and Methods

2.1. Study Area

This study focused on Java Island, as shown in Figure 1, where most of the national population is distributed [13]. In this study, only the main island of Java was included in the analysis; the islands surrounding Java were excluded. The exclusion of small islands from the analysis resulted in errors between the statistics presented for each administrative district and the GIS tabulations. However, these errors did not significantly affect the results. In addition, owing to the exclusion of small islands, the DRR strategy derived in this study could not be applied to them. As a result, there were 80 regencies (Kabupaten in Indonesian) and 34 cities (Kota in Indonesian) in Java, and the characteristics of each of these 114 administrative districts in the study area were analyzed.

The Republic of Indonesia consists of more than 17,000 islands distributed between Southeast Asia and Australia. Its land area is approximately 2 million square kilometers. Its main islands are Java, Bali, Sumatra, Kalimantan, Sulawesi, and Papua [14]. Like other main islands, Java Island in Indonesia is characterized by a tropical climate with high temperatures and humidity, and it receives substantial rainfall throughout the year [15]. The climate is divided into a rainy season from November to March, marked by intense rainfall, and a dry season from April to October, marked by less precipitation. Flooding risks are exacerbated by these climatic conditions combined with inadequate urban planning and insufficient drainage systems, particularly in urban areas and near rivers [16]. Additionally, rapid water flows from mountainous regions and storm surges in low-lying coastal areas further heighten flood risks, making Java highly susceptible to flooding.

2.2. Method Framework and Materials

2.2.1. Outline

The outline of this study is shown in Figure 2. This study proposes a flood disaster risk analysis method using four major types of open data, considering the flood hazards and socioeconomic characteristics of each administrative district in the target area. The four groups were hazardous area, population, land cover and land use, and socioeconomic data. Except for the socioeconomic data, the data are open data distributed as raster or vector data that can be analyzed using GIS software version 3.14.16-Pi. The population and socioeconomic data used in this study were statistical data published on the Internet by national organizations, which were recompiled for each regency and city.

Three analyses were performed, as indicated by the blue boxes in the outlined figure. Flood exposure analysis analyzed the distribution of the population living in areas at risk of potential flood inundation, where flood vulnerability was obtained from the risk mapping analysis using inaRISK. Sigit et al. [7] used the inaRISK flood hazard vulnerability map [2], which is the natural disaster risk information published by the Indonesian government, as hazard information and combined it with open data on the estimated population distribution [17] to obtain the following information on potential disasters: the authors defined the potentially exposed population (P_ep) as the value obtained by multiplying the number of residents per population distribution pixel (population per pixel, P_pp) by the risk value (0–1) presented in the inaRISK vulnerability map. In this study, we used the same definition. We treated the value of P_ep aggregated by the administrative division divided by the total population of the administrative division as the exposure rate, R_exp.

Population distribution characteristic analysis used open land cover data to analyze land cover distribution characteristics and the relationship between land use and the residential population exposed to flood hazards in the target area. We examined the classification of land cover data, which are globally maintained by the Food and Agriculture Organization (FAO) [18], focusing on the relationship between land cover and population distribution for each administrative district. In the comprehensive analysis of socioeconomic vulnerability, we examined the correlations among the indicators and discussed the relationship between socioeconomic indicators and land cover classification, as well as the distribution characteristics of the disaster-exposed population. Finally, based on these analysis results, we proposed effective DRR measures tailored to the socioeconomic characteristics of exposure to flood hazards in each administrative district.

2.2.2. Materials

Table 1. Data groups, names, and types.

Group	Name	Data Type	Target Year
Hazardous area data	inaRISK floodhazard vulnerability (layer_bahaya_banjir_bandang) 1	Raster dataset, resolution: 3 arc-second	-
Population data	Indonesia 100 m population 2	Raster dataset, resolution: 3 arc-second	2020
Land cover	Land cover of Indonesia—GlobCover (22 classes) 3	Polygon dataset, resolution: 300 m	2005
Socioeconomic data	Total population Percentage of population by age group (0–14, 15–64, 65–) Population of poor people 4	Text data Read and input the numbers in the table shown in the PDF of statistical reports published for each administrative division.	2021
	Expenditures per month in B40, M40, and T20 categories 5 Percentage of population aged 15 and over per educational rank 6
	Percentage of population aged 15 and over per educational rank in B40, M40, and T20 categories	Excel download data from website	2020
	Gross Regional Domestic Product (administrative district per person)	Excel download data from website	2020

^1. Data for some of the layers viewable in the inaRISK web GIS are available for download. ^2. 2020 estimates of the number of people per pixel (P_pp), with national totals adjusted to match the UN population division estimates. ^3. Land cover data with a Land Cover Classification System (LCCS) legend are available. This land cover product was published in 2008 as a result of an initiative launched in 2004 by the European Space Agency (ESA). GlobCover is currently the most recent (2005) and resolved (300 m) dataset for land cover globally. ^4. The “poor” are those who have an average monthly per capita expenditure below the poverty line. To measure poverty, the Indonesian Central Bureau of Statistics uses the concept of the ability to meet basic needs (the basic needs method). With this approach, poverty is seen as an economic inability to fulfill basic food and non-food needs, measured in terms of expenditure. ^5. The B40, M40, and T20 classifications represent the Bottom 40%, Middle 40%, and Top 20% of income earners, respectively. This study includes monthly expenditure values for households B40, M40, and T20. ^6. The educational backgrounds of household heads were tabulated into four ranks. 1: no education, 2: finished elementary school, 3: finished junior school, and 4: high school or higher.

lists the data used in this study. The data are open datasets maintained globally for most types of data. Information on socioeconomic status is likely to vary by country but is common, at least in the Republic of Indonesia. In some countries, more indicators of socioeconomic conditions may be available.

Population data were derived from the 2020 estimates of the number of people per pixel (P_pp), with the national totals adjusted to match the UN population division estimates. The estimated number of persons per grid square is shown for each mesh with a horizontal resolution of approximately 100 m. Because the population distribution data were estimated for 2020, the following socioeconomic data were also collected for 2020: Figure 3a.

This study used the flood vulnerability assessment map of inaRISK to identify flood-prone areas. The flood hazard vulnerability map is shown in Figure 3b. InaRISK is an ArcGIS-based portal version 10.8 for disaster risk assessment launched in 2016 with support from the United Nations Development Program (UNDP). It provides data on disaster-prone areas, affected populations, and potential financial and environmental losses and is integrated into disaster management plans for risk reduction. It uses a background layer from the 2012 General Guidelines for Disaster Risk Assessment [19], ensuring transparency in the risk evaluation process. InaRISK calculates vulnerability ratings by analyzing four factors: the exposed population, physical and economic losses, environmental damage, and vulnerability class. Despite procedural complexities, these guidelines are the official sources of disaster risk assessment methods.

Land cover data with Land Cover Classification System (LCCS) legends are available [17]. The land cover map is shown in Figure 3c. This land cover product was published in 2008 as a result of an initiative launched in 2004 by the European Space Agency (ESA).

Socioeconomic data for each administrative district were extracted from the annual statistical reports published by the Indonesian government’s statistical agency [20]. Sigit et al. [7], in their analysis of Central Java, Indonesia, found a gap of up to 3.7 times in the poor population ratio per administrative district and a strong negative correlation between poverty and education level. They also found that the average income per capita differs by a factor of up to 5.8. This study also focused on the proportion of the poor and their incomes and introduced information on their expenditures. The data on expenditures and the education level of the household head were based on the B40, M40, and T20 Income Classifications recently adopted by the Indonesian government. These classifications represent the Bottom 40%, Middle 40%, and Top 20% of income earners. This study included the monthly expenditure values for B40, M40, and T20 households.

GIS version 3.28.13-Firenze, an open-source desktop geographic information, was used for the GIS-based analysis. Statistical analysis was performed using the general-purpose statistical computing software R x64 4.3.2 and Microsoft Excel version 2302 with Power Pivot.

GlobCover is currently the most recent (2005) and resolved (300 m) dataset for land cover globally. In this study, we focused on this dataset as open data on land cover that are maintained globally by a highly reliable organization. However, because the year analyzed for the satellite data was 2005, the data may not reflect land modification in the nearly 20 years that have passed since then.

2.3. Analysis

2.3.1. Flood Exposure Analysis

Sigit et al. [7] proposed a simple method for assessing the population’s potential for exposure to flood hazards by multiplying the inaRISK flood hazard vulnerability value by the resident population. The same method was used in this study. Maps of the flood disaster vulnerability assessment results from the inaRISK geospatial database were used as information on potential flood-hazardous areas. The results of the vulnerability assessment for flooding were expressed as a 0–1 value for each grid in the raster data, with higher numbers indicating greater vulnerability. We defined the potentially exposed population (P_ep) by multiplying the vulnerability assessment values (ranging from 0 to 1) for each pixel by the estimated P_pp and aggregating them by administrative divisions. Figure 3a illustrates the distribution of P_pp in the study area, and Figure 3b displays the flood vulnerability assessment results generated by inaRISK.

P_ep can be expressed by the following equation (Equation (1)), assuming that the estimated population per pixel is P_pp and the vulnerability assessment values for floods are V_f., where the subscript i denotes each pixel, and A denotes the classification of the aggregation area. In this study, P_ep was aggregated using polygons that indicated the boundaries of the regency and city.

$$P_{ep A }=\sum _{i \in A}{P}_{pp i}·V_{f i}$$

(1)

Furthermore, the exposure rate R_exp is defined as shown in Equation (2). When the exposure rate for each administrative division was defined as R_exp, R_exp was defined as P_ep divided by the total population, denoted as P_all. R_exp for each administrative division calculated in this study was used for comprehensive analysis.

$$R_{exp A }=P_{ep A }/ P_{all A }$$

(2)

2.3.2. Comprehensive Analysis of Exposed Socioeconomic Vulnerability

Information on the population, poor population, representative economic indicators, education level, etc., from each administrative division published by public agencies was collected and organized into a socioeconomic database. Sigit et al. [7] proposed two indicators that focused on the characteristics of communities exposed to flooding, particularly the presence of poor people and the size of the economy, which are considered proportional to assets. The number of poor people exposed to disaster P_epp was calculated by multiplying the poor population by the exposure rate, R_exp. The economy that may be exposed to a disaster, E_exp, was calculated by multiplying the Gross Regional Domestic Product (GRDP), which was obtained by multiplying the total population by the average income using the exposure rate, R_exp. The socioeconomic database also combined the results of the exposure rates R_exp obtained from the flood exposure analysis, the number of poor populations exposed to disaster P_epp, and the economy that may be exposed to disaster E_exp. A list of the 51 variables in the socioeconomic database is presented in

Table 2. Data groups, names, and types from the socioeconomic database.

Number_Variable Name	Mean	S.D.	Description (Unit, Definition)	Official Data	Analysis Result
X01_Pop_fix	1,294,377	855,313	Total population by administrative district based on statistical data (people)	Yes
X02_Age0–14	22.1	2.4	Percentage of population under 14 years old (%)	Yes
X03_Age15–64	69.3	2.2	Percentage of population from 15 to 64 years old (%)	Yes
X04_Age65–	8.6	2.8	Percentage of population over 65 years old (%)	Yes
X05_Pop_15-	998,405	635,750	Population over 15 years old (people)	Yes
X06_Poor_Pop	116,628	74,635	Population of poor people (people)	Yes
X07_Poor_ratio	9.3	3.4	Poor population ratio (%, X07 = X06/X01)
X08_Pop_sum	1,332,225	1,016,048	Total population from flood exposure analysis		Yes
X09_Pep_sum	359,683	386,969	Potentially exposed population (Pep) from flood exposure analysis		Yes
X10_expose_ratio	26.5	16.0	Exposure ratio for flood hazard (%, X10 = X09/X08)		Yes
X11_Pop_Ed_NoSch	134,010	101,533	Population aged 15 and over without an elementary school graduation certificate (people)
X12_Pop_Ed_El_Sch	271,562	202,275	Population aged 15 and over with elementary school graduation (people)
X13_Pop_Ed_JH_Sch	226,010	149,897	Population aged 15 and over with junior high school graduation (people)
X14_Pop_Ed_SH_Sch	369,061	309,812	Population aged 15 and over with high school graduation and above (people)
X15_Pop_B40_NoSch	68,417	49,745	Population aged 15 and over without an elementary school graduation certificate (people), B40 class
X16_Pop_B40_El_Sch	125,381	94,096	Population aged 15 and over with elementary school graduation (people), B40 class
X17_Pop_B40_JH_Sch	97,602	69,240	Population aged 15 and over with junior high school graduation (people), B40 class
X18_Pop_B40_SH_Sch	108,857	92,817	Population aged and 15 over with high school graduation and above (people), B40 class
X19_Pop_M40_NoSch	50,416	39,843	Population aged 15 and over without an elementary school graduation certificate(people), M40 class
X20_Pop_M40_El_Sch	110,221	83,378	Population aged 15 and over with elementary school graduation (people), M40 class
X21_Pop_M40_JH_Sch	91,572	59,572	Population aged 15 and over with junior high school graduation (people), M40 class
X22_Pop_M40_SH_Sch	148,048	129,989	Population aged and 15 over with high school graduation and above (people), M40 class
X23_Pop_T20_NoSch	16,337	14,242	Population aged and 15 over without an elementary school graduation certificate (people), T20 class
X24_Pop_T20_El_Sch	37,471	29,468	Population aged 15 and over with elementary school graduation (people), T20 class
X25_Pop_T20_JH_Sch	37,665	25,074	Population aged 15 and over with junior high school graduation (people), T20 class
X26_Pop_T20_SH_Sch	108,655	86,766	Population aged and 15 over with high school graduation and above (people), T20 class
X30_B40_Cost_Rp	589,141	169,366	Expenditures per household member in class B40 per month (Rp)	Yes
X31_M40_Cost_Rp	1,162,257	402,502	Expenditures per household member in class M40 per month (Rp)	Yes
X32_T20_Cost_Rp	2,810,395	1,192,257	Expenditures per household member in class T20 per month (Rp)	Yes
X33_Total.Annual.Expenditure_BRp	19,647	16,957	Total annual expenditure of the total population (billion Rp)
X34_GRDP_BRp	81,311	126,834	Gross Regional Domestic Product (BRp)	Yes
X35_Income_per_person_TRp	60,893	84,374	GRDP divided by population (TRp)	Yes
X36_Income_Expenditure_Ratio	3.9	5.5	Ratio of total income to total expenses (X36 = X34/X33)
X40_Ed_NoSch	13.3	7.1	Percentage of the population aged 15 and over without an elementary school graduation certificate (%)	Yes
X41_Ed_El_Sch	26.1	9.7	Percentage of the population aged 15 and over with elementary school graduation (%)	Yes
X42_Ed_JH_Sch	22.2	2.6	Percentage of the population aged and 15 over with junior high school graduation (%)	Yes
X43_Ed_SH_Sch	38.3	15.4	Percentage of the population aged 15 and over with high school graduation and above (%)	Yes
X44_B40_NoSch	17.3	8.4	Percentage of the population aged 15 and over without an elementary school graduation certificate (%), B40 class	Yes
X45_B40_El_Sch	30.1	9.3	Percentage of the population aged 15 and over with elementary school graduation (%), B40 class	Yes
X46_B40_JH_Sch	24.0	3.1	Percentage of the population aged 15 and over with junior high school graduation (%), B40 class	Yes
X47_B40_SH_Sch	28.6	13.3	Percentage of the population aged 15 and over with high school graduation and above (%), B40 class	Yes
X48_M40_NoSch	12.4	7.0	Percentage of the population aged 15 and over without an elementary school graduation certificate (%), M40 class	Yes
X49_M40_El_Sch	26.4	10.7	Percentage of the population aged 15 and over with elementary school graduation (%), M40 class	Yes
X50_M40_JH_Sch	22.6	3.2	Percentage of the population aged 15 and over with junior high school graduation (%), M40 class	Yes
X51_M40_SH_Sch	38.6	17.0	Percentage of the population aged 15 and over with high school graduation and above (%), M40 class	Yes
X52_T20_NoSch	8.0	5.5	Percentage of the population aged 15 and over without an elementary school graduation certificate (%), T20 class	Yes
X53_T20_El_Sch	18.2	9.5	Percentage of the population aged 15 and over with elementary school graduation (%), T20 class	Yes
X54_T20_JH_Sch	18.4	4.7	Percentage of the population aged and 15 over with junior high school graduation (%), T20 class	Yes
X55_T20_SH_Sch	55.4	17.3	Percentage of the population aged 15 and over with high school graduation and above (%), T20 class	Yes
X60_Pepp	30,114	25,998	Potentially exposed poor population (people, X60 = X06 × X10)		Yes
X61_Eexp_BRp	27,374	58,742	Exposed economy (BRp, X61 = X34 × X10)		Yes

.

Correlation analysis was performed for each variable in the database to determine whether there was a statistically significant correlation between the variables. Adjusted p-values (Holm’s method) were used to test for statistical superiority, with p < 0.05 as the criterion for confirming superiority. Representative variables were analyzed for each cluster resulting from the hierarchical cluster analysis of land cover and were used to propose disaster risk reduction strategies.

3. Results

3.1. Flood Exposure Analysis Result

The exposure rates for each administrative district, as assessed by the estimated population distribution and flood hazard vulnerability using inaRISK, are shown in Figure 4. The map shows the percentage of residents potentially exposed to flood hazards in each administrative district. The minimum and maximum values are 1.1%, 63%, and 26.5%, respectively.

3.2. Population Distribution Characteristics Analysis Result

3.2.1. The Population Composition by Land Cover Type

The percentage of the population distributed by land cover type was tabulated, and the means and standard deviations for the 114 administrative divisions are shown in Figure 5a. The results showed the average percentage of land cover in which the population is distributed on Java Island. The most populated land cover type was “Mosaic cropland (50–70%)/vegetation (grassland/shrubland/forest) (20–50%)”, followed by “Rainfed croplands,” “Artificial surfaces and associated areas,” and “Closed to open (>15%) broadleaved evergreen and/or semi-deciduous forest (>5 m).” The percentages were higher in the order. These four land cover types comprised the majority of the land cover, and the majority of Java residents lived within these landscapes. Using the exposure rates obtained from the flood exposure analysis, the total population exposure rates for each land cover type were calculated for the four major land cover types (Figure 5b). Among the four major land cover types, the order of exposure was rain-fed cropland, urban areas, mosaic cropland, and shrubland. These results suggest that the characteristics of each administrative district can be determined by focusing on the type of land cover in which residents reside.

3.2.2. Cluster Analysis for the Population Composition by Land Cover Type

For each administrative district, a hierarchical cluster analysis was conducted using population composition by land cover type as a variable. As a result, the 114 administrative divisions were divided into four clusters. As shown in Figure 6, the characteristics of each cluster were explained by the typical land cover described earlier. Cluster 1: the population was mainly distributed in forests and shrubland; Cluster 2: the population was mainly distributed in mosaic croplands and vegetation; Cluster 3: the population was mainly distributed in rainfed croplands; and Cluster 4: the population was mainly distributed in urban areas.

Figure 7 shows the results of organizing the land cover composition of the 114 administrative districts into four clusters and then organizing the land cover composition where residents reside. Figure 8 shows the results of the cluster classification for each administrative district. While it is not surprising that urban areas are dominated in Cluster 4, where most residents live in urban areas, Cluster 1, where most residents live in forests and shrublands, and Cluster 3, where most residents live in rainfed croplands, also include some urban areas. If land cover data were updated to the latest status, recent urbanization would be reflected in the data, and these administrative areas might be classified into Cluster 4.

3.3. Comprehensive Analysis Result

3.3.1. Statistically Significant Correlations

The correlation analysis identified a large set of variables with statistically significant correlations. However, most sets of variables were proportional to the population size and could not be considered independent variables.

For the poverty-related variables X06_Poor_Pop and X07_Poor_ratio, it was confirmed that the population composition and education level were related, and a negative correlation was found with X03_Age15–64, which corresponds to the percentage of the working population. It showed a positive correlation with X40_Ed_NoSch and X41_Ed_El_Sch, indicating the proportion of household heads with low levels of education below the elementary school graduate level, and a negative correlation with X43_Ed_SH_Sch, indicating the proportion of household heads with a high school education or higher. These results show that poverty is related to education levels and population composition, especially the proportion of the employed population.

For income and expenditure-related indicators, both X34_GRDP_BRp, indicating the GRDP in the administrative division, and X33_Total.Annual.Expenditure_BRp, indicating the total expenditure from households in the administrative districts, were positively correlated with population-related indicators, independent of education level. For X36_Income_Expenditure_Ratio, a multiple of the GRDP to total expenditure from households, the only correlation was for X03_Age15–64. This suggests that the proportion of workers with higher education levels is directly related to the economy size and income/expenditure balance in each administrative division.

In terms of the relationship with flood hazard risk, no statistically significant correlation was found between X10_expose_ratio (R_exp) and the independent variables not involved in calculating the value.

3.3.2. Analysis Focused on Cluster Classification with the Land Cover Component

Focusing on the four cluster categories shown in Figure 6, Figure 7 and Figure 8, we tabulated the main variables in the socioeconomic database and examined their cluster characteristics.

There was no significant difference in the populations for each cluster (Figure 9a). However, the poverty rate shown in Figure 9b was significantly lower for Cluster 4. The education level of the heads of households was also higher in Cluster 4 (Figure 9c). Cluster 2 had the lowest average educational level. The low level of education in Clusters 1, 2, and 3, especially in Cluster 2, may reflect the poor accessibility of educational institutions and higher education in rural areas.

Figure 9d, which shows total household expenditures, and Figure 9e, which shows the GRDP, were similarly significantly higher in Cluster 4. Still, Figure 9e, which shows the ratio of the GRDP to total household expenditures, was even more noteworthy. In Clusters 1, 2, and 3, which belong to rural and forest areas, the GRDP was 3 to 4 times the total household expenditure, while in Cluster 4, which is urban, the GRDP was on average eight times, indicating that large economic activities are conducted by firms and other entities. The highest cluster was the Special Province of Jakarta, the capital of the country. Urban centers like Jakarta are hubs of multinational corporations, advanced technology industries, and financial services, offering relatively high-income levels and diverse job opportunities. These cities also boast higher educational levels, with numerous higher education institutions.

The results of the exposure ratios R_exp, potentially exposed poor population P_epp, and exposed economy E_exp evaluated based on the results of the flood hazard exposure analysis were tabulated by cluster classification and are shown in Figure 10a, Figure 10b, and Figure 10c, respectively. In Figure 9, which is compiled for socioeconomic indicators, only Cluster 4 showed different characteristics, whereas in Figure 10, which shows the exposure characteristics to flood hazards, Clusters 1–3 also showed different trends. The flood hazard R_exp was higher in Cluster 3, which was dominated by rainfed croplands, and Cluster 4, which was dominated by urban areas. Cluster 3 had the highest P_epp value, indicating that the poorest population is potentially exposed to floods.

4. Discussion

Java Island, Indonesia’s political and economic center, is home to approximately half of the country’s population. Socioeconomic diversity on the island is pronounced, with significant disparities between urban and rural areas. Urban centers such as Jakarta, Surabaya, and Bandung are hubs of corporations, advanced industries, and financial services, offering relatively high-income levels and diverse job opportunities. These cities also have institutions with higher educational levels. Contrarily, the rural regions of Java are predominantly agricultural, and residents primarily engage in traditional farming. These areas are characterized by markedly lower incomes than urban areas, with limited educational opportunities and inadequate basic infrastructure. This discrepancy contributes to disparities in health and quality of life. These socioeconomic disparities are particularly salient when flood risks are considered. Areas with low income or insufficient infrastructure face increased vulnerability during disasters, complicating both immediate response and long-term recovery. Thus, the socioeconomic diversity of Java significantly affects its resilience to natural disasters, influencing both risk exposure and recovery capabilities.

Based on the analytical approach proposed in this study, the above general argument was embodied in terms of the relationship between flood hazards and the socioeconomic characteristics of each administrative district. An analysis using cluster classification focusing on the relationship between land cover and population distribution effectively depicted the socioeconomic characteristics of exposure to flood disasters.

FAO land cover data are comprehensive and global but have drawbacks, such as a coarse resolution that misses fine-scale changes and being outdated, as it is from 2008. These data may not reflect current land conditions, especially given urban expansion and deforestation over the past decade. Additionally, their global scope might overlook important regional land cover dynamics for local flood risk assessment. These limitations can affect study results, including flood risk accuracy, the misrepresentation of vulnerable populations, and policy relevance. Despite these issues, integrating flood risk, socioeconomic, and land cover analysis is crucial for comprehensive flood risk assessment, targeted disaster risk reduction (DRR) interventions, and efficient resource allocation by policymakers. This method remains feasible and relevant, particularly for poor and developing countries. It is practical with similar data constraints, scalable to larger areas, adaptable to other data sources, and helps build local DRR capacity. Thus, despite its limitations, this integrated approach is valuable for understanding and reducing flood vulnerability in developing countries.

To develop effective DRR strategies that consider regional characteristics, decision-makers involved in disaster prevention and mitigation should have a common view of the characteristics of the region and features of the various strategies. The results of this study provide useful insight for this purpose. As a basis for developing a strategy for each administrative division, attention should be paid to the exposure rate (Figure 4), the poor population exposed to flood hazards, and the size of the economy (Figure 11a,b). This information provides a bird’s-eye view of the distribution of flood risk over a wide area in social and economic terms. The four cluster classifications shown in Figure 6, Figure 7 and Figure 8, which focus on the relationship between population distribution and land cover, have a clear correspondence not only with the main land cover where the population is distributed or landscape characteristics but also with socioeconomic conditions, such as poverty, education level, and income (Figure 9 and Figure 10).

Sigit et al. [7] advocated multifaceted flood risk management strategies involving various stakeholders, emphasizing the integration of development organizations and disaster management agencies. They recommended enhancing early warning systems by aligning strategies with local strengths and vulnerabilities, promoting investment in infrastructure, and fostering community collaboration. Moreover, they stressed the importance of incorporating indigenous local knowledge (ILK) into early warning systems, which would bolster community resilience through sustainable practices and ecosystem-based approaches. Effective communication strategies that use both traditional and modern channels are essential to ensure that these improvements reach and effectively engage all affected communities. The results of this study can substantiate this argument.

Sigit et al. [7] provided suggestions for a comparative analysis of the advantages and weaknesses of flood risk management strategies (

Table 3. Comparative analysis of flood risk management strategies.

Strategy	Advantage	Weakness
Early Warning Systems	Timely alerts, Potential for reducing casualties.	Dependence on infrastructure, False alarms.
Infrastructure Development	Enhanced resilience, Protection of assets.	High cost, Environmental impact.
Community Engagement and Preparedness	Empowering communities, Local knowledge.	Variable community response, Resource constraints.
Land Use Planning and Zoning	Mitigation of exposure, Sustainable development.	Implementation challenges, Resistance.
Insurance and Risk Transfer Mechanisms	Financial protection, Incentive for risk reduction.	Limited coverage, Affordability.
Ecosystem-Based Approaches	Natural flood defenses, Ecological benefits.	Time-intensive, Potential conflicts.

). Thus, each strategy provides a distinct understanding of its implementation. Considering the results of a comprehensive analysis of flood risk on Java and the clustering of administrative districts based on socioeconomic characteristics and land cover, procedures for developing effective DRR strategies that consider the socioeconomic characteristics of exposure to flooding are important. Considering the aforementioned advantages and weaknesses of the general DRR strategies, we can formulate a targeted approach to mitigate flood hazards while addressing socioeconomic vulnerabilities. The proposed procedure is illustrated in Figure 12. The contents are as follows:

• Step 1: Identify the poor population and economic size exposed to flood hazards in each administrative district. Utilize socioeconomic data to identify vulnerable communities and economic assets at risk of flood damage.
• Step 2-1: Examine the general direction of DRR measures with reference to land cover types. Tailor DRR strategies for predominant land cover types within each cluster by considering each strategy’s specific advantages and weaknesses.
• Step 2-2: Obtain detailed information on each district or city from the socioeconomic database. Analyze socioeconomic indicators, such as poverty rates, education levels, and economic activities, to identify additional vulnerabilities and prioritize DRR interventions.
• Step 3: Assignment of specific measures at the land use planning level based on the relationship between population distribution and hazard risk. Develop targeted land use planning and zoning regulations that mitigate exposure to flood hazards while promoting sustainable development and community resilience.

By following Step 2, integrating a combination of flood disaster risk management strategies tailored to the socioeconomic characteristics and type of land cover in each cluster shown in

Table 4. Recommended flood risk reduction strategies for Java and their relationship to land cover clusters where the strategies are applicable.

Strategy	Consideration	Cluster
Early Warning Systems	Early warning systems offer timely alerts and the potential to reduce casualties, making them valuable tools for disaster preparedness. However, their effectiveness is contingent upon robust infrastructure and the risk of false alarms. In urban areas, where populations are densely concentrated and infrastructure is relatively well developed, early warning systems can play a pivotal role in minimizing flood impacts. In rural areas, it is expected that the response will take the form of indigenous local knowledge (ILK) rather than modern early warning systems.	Cluster 4
Infrastructure Development	Infrastructure development, such as flood barriers and drainage systems, can enhance resilience and protect assets against flood hazards. In urban areas, there are many assets exposed to flooding, so it is worthwhile to invest in preventing this. Nevertheless, the high cost and environmental impact associated with large-scale infrastructure projects pose significant challenges. In agricultural areas, where populations are distributed across mosaic croplands and rainfed croplands, targeted infrastructure investments can bolster flood resilience while supporting livelihoods.	Clusters 1, 2, 3, and 4
Community Engagement and Preparedness	Engaging communities in disaster preparedness initiatives empowers residents and leverages their invaluable knowledge of the terrain. However, community responses may vary, and resource constraints can impede effective implementation. In rural areas, where populations rely heavily on agriculture and have limited access to formal resources, community-based approaches can foster resilience and facilitate rapid response during flood events.	Clusters 1, 2, and 3
Land Use Planning and Zoning	The rainfed croplands and urban areas tend to have higher exposure rates to flooding. Land use planning and zoning measures, including floodplain regulations and green infrastructure implementation, can mitigate exposure to flood hazards while promoting sustainable development in floodplains and urban areas. Nevertheless, implementation challenges and resistance from stakeholders may hinder effective land use planning efforts. In forested areas and urban fringes, where ecosystems play a crucial role in flood regulation, integrated land use planning strategies can balance conservation with development needs.	Clusters 3 and 4
Insurance and Risk Transfer Mechanisms	Insurance and risk transfer mechanisms provide financial protection against flood-related losses and incentivize risk reduction practices. However, limited coverage and affordability issues may constrain their effectiveness, particularly in rural and low-income areas. In economically vibrant urban centers, where financial resources are relatively abundant, insurance schemes can serve as an additional layer of protection for businesses and homeowners.	Cluster 4
Ecosystem-Based Approaches	Ecosystem-based approaches harness the natural flood mitigation capacities of ecosystems, such as wetlands and mangroves, to reduce flood risk and enhance ecological resilience. Nevertheless, these approaches require time-intensive planning and may encounter conflicts with existing land uses. In rural and forested areas, where ecosystems provide critical ecosystem services, ecosystem-based approaches can complement traditional DRR measures and safeguard biodiversity.	Clusters 1, 2, and 3

, policymakers and decision-makers can choose an effective DRR strategy for their region. Although more specific measures are considered in Step 3, selecting an effective basic strategy package as a result of Step 2 will allow for a more efficient consideration of specific measures.

By focusing on the characteristics of the land cover component, it has become possible to distinguish the socioeconomic characteristics exposed to flood hazards in each administrative division. However, the method proposed by this study, which primarily assesses the potential magnitude of exposure by looking at response populations, does not fully account for the actual flood damage, which can vary widely and is influenced by numerous factors [21,22]. To develop effective DRR strategies, it is crucial to consider the actual damage caused by floods. We need to evaluate various aspects, such as the extent of flood damages, the resilience to these damages, the likelihood of their occurrence, and the sensitivity of affected areas. Here are detailed practical steps, examples, and case studies to implement the recommended flood risk reduction strategies (

Table 5. Practical steps to implement DRR strategies against flooding.

Strategy	Practical Steps	Examples and Case Studies
Early Warning Systems	Assessment and Planning: Conduct a comprehensive risk assessment to identify flood-prone areas and establish a baseline of current early warning capabilities.	Indonesia Tsunami Early Warning System (InaTEWS) [23] Bangladesh’s Cyclone Preparedness Programme [24]
	Technology Deployment: Install advanced weather monitoring and forecasting systems, such as radar and satellite-based technologies.
	Communication Networks: Develop robust communication networks to disseminate alerts quickly to the affected population.
	Community Training: Train local communities and authorities on how to respond to warnings and conduct regular drills.
Infrastructure Development	Site Selection: Identify high-risk flood zones using topographical and hydrological data.	Giant Seawall Jakarta [25] Netherlands’ Delta Works [26]
	Design and Construction: Design flood barriers, levees, and drainage systems tailored to local conditions. Ensure designs are resilient to extreme weather events.
	Environmental Impact Assessment: Conduct thorough environmental assessments to minimize negative impacts on ecosystems.
	Maintenance Plans: Establish long-term maintenance plans to ensure the infrastructure remains effective.
Community Engagement and Preparedness	Community Mapping: Engage communities in mapping flood-prone areas and identifying vulnerabilities.	Thailand’s Community-Based Disaster Risk Management (CBDRM) [27] Philippines’ Barangay DRR Committees [28]
	Educational Programs: Develop and implement educational programs on flood risks and preparedness measures.
	Local DRR Committees: Establish local disaster risk reduction (DRR) committees to coordinate community efforts.
	Resource Mobilization: Facilitate access to resources and support for community-driven projects.
Land Use Planning and Zoning	Floodplain Mapping: Develop detailed floodplain maps to guide land use planning.	New York City’s Zoning for Coastal Flood Resiliency [29] Germany’s Integrated River Basin Management [30]
	Zoning Regulations: Implement zoning regulations that restrict development in high-risk areas and promote green infrastructure.
	Public Participation: Engage stakeholders in planning processes to ensure compliance and address concerns.
	Incentive Programs: Create incentives for developers to incorporate flood-resistant designs and practices.
Insurance and Risk Transfer Mechanisms	Risk Assessment: Conduct detailed assessments to determine flood risk levels and appropriate insurance premiums.	Mexico’s FONDEN [31] Turkish Catastrophe Insurance Pool (TCIP) [32]
	Product Development: Develop affordable insurance products tailored to different socioeconomic groups.
	Awareness Campaigns: Raise awareness about the benefits of insurance and how to access it.
	Public-Private Partnerships: Encourage collaboration between governments, insurance companies, and communities to expand coverage.
Ecosystem-Based Approaches	Ecosystem Restoration: Restore and protect natural floodplains, wetlands, and mangroves to enhance their flood mitigation capacities.	Vietnam’s Mangrove Restoration [33] Costa Rica’s Payment for Ecosystem Services (PES) [34]
	Sustainable Land Management: Promote agricultural and forestry practices that support ecosystem health and resilience.
	Integrated Planning: Incorporate ecosystem-based approaches into broader land use and development plans.
	Community Involvement: Engage local communities in ecosystem management and conservation efforts.

).

For the planning of more concrete measures, we must move to Step 3. This step involves developing specific strategies at the level of land use planning, which take into account the distribution of the population and the associated hazard risks. Step 3 also includes the creation of land use planning and zoning regulations aimed at reducing vulnerability to flood hazards. Additionally, these regulations should support sustainable development and enhance community resilience. Ultimately, the findings of this study contribute significantly to providing the appropriate direction for Step 3, ensuring that the strategies developed are well informed and effectively targeted at reducing disaster risks.

5. Conclusions

This study explored integrated flood risk reduction strategies for Java Island, Indonesia, by combining land cover and socioeconomic analyses. The increasing severity and frequency of floods due to climate change pose significant risks to densely populated islands, necessitating a robust approach to DRR. The methodological framework of this research includes a detailed analysis of flood exposure using inaRISK hazard data combined with land cover data from the GlobCover database and socioeconomic data from Indonesian national statistics. This integrated analysis provides a nuanced understanding of flood risk across different regions of Java, considering both physical vulnerabilities and socioeconomic resilience. The study findings emphasize the stark disparities in flood risk and socioeconomic conditions across Java. This indicates that urban areas, while economically more robust, are highly vulnerable to flooding owing to intense land use and poor water management. Rural areas, although less economically developed, face significant challenges owing to their reliance on agriculture and lack of robust infrastructure. Overall, this study underscores the need for a holistic approach to flood risk management in Java. By integrating land cover and socioeconomic data, policymakers can better understand the multifaceted nature of flood risks, develop effective strategies to mitigate them, and enhance the resilience of vulnerable communities. This study contributes to the broader discourse on integrating socioeconomic factors into flood risk management by offering insights applicable to other regions with similar geographic and socioeconomic characteristics. The approach and findings can serve as a model for further research in other parts of Indonesia and similar environments around the world, highlighting the importance of tailored, data-driven strategies to address the challenges posed by climate change and urbanization.