Local Sustainability Assessment of the Wonogiri Multipurpose Reservoir Catchment Area in Central Java Province, Indonesia

Abstract

The sustainability of watershed management is a crucial issue that must be addressed to guarantee the persistence of watershed services including agriculture, food production, and energy supply. This issue has also been addressed in Presidential Regulation No. 18/2020 concerning the National Medium-Term Development Plans for 2020–2024, which stipulate the restoration of priority watersheds, including the Upstream Bengawan Solo Watershed. This study seeks to address this information gap by assessing the local sustainability of the watershed from a temporal dynamics perspective by calculating the Local Sustainability Index (LSI), Local Moran Index, and spatial associations. Measuring sustainable development indices locally is essential because each location has different characteristics, and using specific indicators at the local level is rarely done. The enactment of the national law on village autonomy in Indonesia necessitates the formulation of sustainable development indicators at the village level. These indicators serve as the metrics and frameworks for local government policies and initiatives. Our results show that village sustainability in the social and economic dimensions has increased from 2007 to 2021, especially in urban activity center areas that serve social and economic facilities. This seems different in the environmental dimension, where the sustainability value decreased from 2007 to 2021. The concentration of low sustainability values on ecological conditions occurred in pocket areas. Environmental problems were indicated by land-use conversion and disaster areas.

1. Introduction

Disasters are events caused by natural and human behavior factors that can threaten and disrupt life and result in property loss, loss of life, and environmental damage. In the last ten years, there has been an increase in the number of natural disasters in Indonesia by 175.37%, dominated by floods, and Java Island is included in the category of >200 disaster events per year [1].

The Wonogiri Multipurpose Reservoir (WMR) is located in Wonogiri Regency, Central Java Province, Indonesia. It plays a crucial role in flood management while also serving multiple functions such as supporting fisheries, aiding the Java–Bali hydroelectric power plant, providing raw water for irrigation and other uses, and attracting tourists. Hydrologically, the WMR is an artificial reservoir that stores water from the longest river on the island of Java, contributing to the watershed system of the upper Bengawan Solo River. The government has prioritized the Bengawan Solo Watershed, as outlined in Presidential Regulation No. 18 of 2020, which addresses the National Medium Term Development Plans (RPJMN) for 2020–2024. This initiative aligns with global sustainability goals, focusing on ensuring clean water, proper sanitation, watershed conservation, and reducing forest degradation and deforestation.

The selection of priority watersheds is determined by the pressing management requirements of water catchment areas, which are increasingly facing conflicts between natural physical conditions and human development. When the WMR was operating in 1978, it experienced shallowing, decreasing capacity from year to year, with an average sediment inflow reaching 3.2 million m³/year [2]. This condition is increasingly worrying, especially in the rainy season when water discharge peaks. Hence, the spillway gate needs to be opened, resulting in floodings downstream, such as in Surakarta, Sukoharjo Regency, and Karanganyar Regency [3].

As an ecological unit, problems downstream are inseparable from conditions in the region of the middle and upstream of the watershed. Communities around the watershed, for social and economic activities, require space in the form of settlements, agriculture, trade, and services, and their needs are increasing along with growth and population development. The dynamics of spatial requirements lead to land conversion [4,5], from forest to non-forest, and from agriculture to non-agriculture [6] in watersheds [7,8,9]. Climate problems in the global environment also contribute to ecological degradation in water catchment areas/watersheds [10,11,12,13,14].

The problems of a watershed cannot be addressed from just one point of view. However, a comprehensive study is needed, which includes multi-dimensional, multi-time, and multi-stakeholders [15,16,17,18] in the context of riparian ecosystems [19,20] and sustainability [21]. Measuring sustainability remains a challenge [22]. Development sustainability has been discussed widely at global, national, and regional levels; however, it still needs to be expanded to the local level [23,24]. Local-level problems of villages need to be elaborated to prepare more detailed programs [25] and activities to be used as input for calculating village fund budget allocations.

This research aims to measure sustainability at the subdistrict/village levels over several periods and determine the spatial associations [26,27] to see connections among the surrounding areas. This research is a refinement of the previous research conducted by Setianingtias et al. (2019) [28], Rahma et al. (2019) [29], Fauzi & Oxtavianus (2014) [30], Rendrarpoetri (2024) [31] which considered the social, economic, and environmental aspects in measuring the Sustainable Development Index at the regional level. The indicators used in the research above are only available at the provincial or district level. Data on these indicators have yet to be available for the most minor administrative units or villages. With the same unit of analysis, Pravitasari et al. (2018) [32] measured the sustainability index [33] by considering the spatial interdependence relationships at the regional scale. Fadli (2017) [34] stated that most regions in the world have their characteristics, and that measuring sustainability at the village level cannot necessarily use the sustainability measurements at the regional level.

2. Materials and Methods

2.1. Study Area

A river basin is defined as a unit for managing water resources that encompasses one or more river basins and/or small islands with an area of 2000 km² or less (Regulation of The Minister of Public Works and Public Housing of The Republic of Indonesia 04/PRT/M/2015 regarding Criteria and Determination of River Areas). According to this definition, the Bengawan Solo River Basin comprises several watersheds and sub-watersheds, with some directing their flow north toward the Java Sea and others flowing south toward the Indian Ocean.

The study area is situated in the Central Java Province on the island of Java, Indonesia (Figure 1a). It forms part of the upstream Bengawan Solo River Basin, as illustrated in Figure 1b, where water flow predominantly heads north toward the Java Sea, while some flows south toward the Indian Ocean. This flow direction is influenced by the topography and slope gradient, moving from upstream to downstream. The boundaries of the study area align with the river basin which serves as the water catchment area for the Wonogiri Multipurpose Reservoir (WMR) (Figure 1c). Administratively, the area encompasses 193 villages across 20 districts, all located within Wonogiri Regency, Central Java Province, Indonesia (Figure 1d).

2.2. Data Collection

This research used Village Potential in Numbers (Potensi Desa) and Districts in Numbers (Kecamatan Dalam Angka) from the years 2007 to 2021 as issued by the Central Statistics Agency (BPS), as well as literature studies from various sources such as the Ministry of Environment and Forestry, the Ministry of Agrarian Affairs and Spatial Planning/National Land Agency, and the Ministry of Public Works and Public Housing and regional governments. These data encompass the social, economic, and environmental elements for 193 communities. The social dimension is constructed from 10 factors that pertain to drinking water, sanitation, social facilities, and organizations. The economic component consists of 9 factors relating to economic activities, poverty alleviation, and economic infrastructure. Meanwhile, the environmental dimension comprises variables such as disasters, pollution, and land degradation. Details of the variables and indicators for each dimension are given in

Table 1. List of variables.

Code	Themes	Variables	Indicators	References *
Social Dimensions
SOCDIM1	drinking water	number of households served by sustainable safe drinking water	percentage of household in each district served by Local Water Company (PDAM)	[35,36,37,38,39,40,41,42,43,44,45,46]
SOCDIM2	sanitation	number of villages whose residents have access to sanitation facilities	percentage of villages in each district whose households have toilets	[35,36,37,38,39,40,41,42,44,45]
SOCDIM3	facility	number of health facilities	number of health facilities	[32,35,36,39,45,46]
SOCDIM4	density	population density	population density (people/hectare)	[36,45,46]
SOCDIM5	organization	Number of social institutions/organizations	Number of social institutions/organizations	[35,36,37,38,39,41,42,44,45,46]
SOCDIM6	organization	number of sports groups	number of sports groups	[35,38,39,45,46]
SOCDIM7	facility	number of sports facilities	number of sports facilities	[32,35,36,38,39,44,45,46]
SOCDIM8	organization	the number of villages that have the habit of mutual cooperation	Number of efforts to maintain security	[38,39,41,45,46]
SOCDIM9	facility	Number of village security facilities	Number of village security facilities	[36,38,39,41,45,46]
SOCDIM10	facility	Number of worship facilities	Number of worship facilities	[35,36,37,39,40,41,42,44,45,46]
ECONOMIC DIMENSIONS
ECODIM1	facility	number of small and medium industries	number of small and medium industries	[32,35,36,37,38,39,40,41,42,43,44,45,46,47]
ECODIM2	facility	Number of economic facilities	Number of cooperatives	[32,35,38,39,44,46,47]
ECODIM3	economic performance	Number of poverty reduction programs	Number of poverty reduction programs	[32,35,36,37,38,39,42,44,46,47]
ECODIM4	economic performance	number of villages with the main commodity/sub sector of agriculture (rice)	amount of rice production	[32,35,38,39,46,47]
ECODIM5	utilization of water resources	the number of villages whose residents use rivers for irrigation of agricultural land	agricultural land area	[41,45,46]
ECODIM6	Economic activities	number of credit facilities	number of credit facilities	[39,41,45,46]
ECODIM7	Economic activities	the number of villages whose residents use the river for transportation	the number of villages whose residents use the river for transportation	[39,41,45,46]
ECODIM8	Economic activities	number of shops/grocery number of shops/grocery	number of shops/grocery number of shops/grocery	[32,35,38,39,44,46,47]
ECODIM9	Economic activities	Number of markets and minimarkets	Number of markets and minimarkets	[36,40,43]
ENVIRONMENTAL DIMENSIONS
ENVDIM1	degradation	area of converted forest	forest area	[32,36,38,39,40,41,45,46,47]
ENVDIM2	disaster	landslide disaster	The frequency of landslide disasters	[32,36,39,40,41,46]
ENVDIM3	disaster	floods	Frequency of flood disasters	[32,36,39,40,41,46]
ENVDIM4	disaster	forest and land fires	Frequency of forest and land fires	[32,36,39,40,41,46]
ENVDIM5	pollution	number of villages where land pollution occurs due to industry	Number of types of environmental pollution	[32,35,36,37,38,39,40,41,42,43,44,45,46,47]
ENVDIM6	disaster	disaster mitigation	Number of disaster mitigation efforts	[38,39,41,44,46]
ENVDIM7	degradation	land unsuitability	Extent of land cover incompatibility with protective functions	[32,36,38,39,40,41,45,46,47]
ENVDIM8	degradation	land unsuitability	Extent of land cover mismatch with buffer function	[32,36,38,39,40,41,45,46,47]
ENVDIM9	degradation	land unsuitability	Critical and very critical land area	[32,36,38,39,40,41,45,46,47]
ENVDIM10	degradation	land unsuitability	Built-up land area (settlement)	[37,39,41,42,44,47]
ENVDIM11	degradation	land unsuitability	Plantation land area	[37,39,41,42,44,47]

* References: [32] A. E. Pravitasari et al. (2018); [35] Nijkamp (1999); [36] Department of Economic and Social Affairs. Division for Sustainable Development United Nations (2001); [37] Boggia dan Cortina (2010); [38] Tanguay et al. (2010); [39] Shen et al. (2011); [40] The Ministry of Environment and Forestry (2014); [41] Strezov et al. (2017); [42] Boggia et al. (2018); [43] The Ministry of Home Affairs (2018); [44] Nogués et al. (2019); [45] OECD (2019); [46] European Commission. Joint Research Centre. (2021); [47] Hély and Antoni (2019).

.

2.3. Analysis Methods

2.3.1. Measurement of the Local Sustainable Index (LSI)

Factor analysis (FA) selects the main performance indicators/variables that form the LSI for each dimension. Factor analysis is exploratory, without distinguishing between the independent and dependent variables. In this analysis, the requirement to become a new factor is that the eigenvalue is >1. The eigenvalue shows how much influence a variable has on the formation of the characteristics of a matrix. The next stage is determining the factor score by combining several variants in one factor (summated scale). Variables that have a factor loading value > 0.7 are characteristic indicators. Several things influence a variable’s factor loading value, namely the variable’s correlation coefficient and variations in the indicator data. High data variations have a strong potential to have high factor loadings. Indicators with low factor loadings could be more robust in explaining a latent construct. So, the characteristic indicators in this study are different for each year.

Realizing the importance of considering the economic, social, and environmental aspects as comprehensive aspects in the concept of sustainable development, the LSI was developed by considering local spatial interdependence, which can be used to measure sustainable development performance at the rural level of the WRM catchment area. The Local Sustainable Development Index was determined by calculating the factor scores and new factor eigenvalues resulting from factor analysis in the following equation [48]:

$${LSI}_{ki}=\sum _{m=1}^{n_k}{E}_{km}.S_{kmi}$$

where LSIki = LSI for k-th Dimension on i-th village (desa); k = Dimension (k = 1: economy; k = 2: social; k = 3: environment); Ekm = Eigenvalue for k-th Dimension on m-th factor; Skmi = Factor score for k-th Dimension, m-th factor on i-th village (desa); i = 1, 2, 3,…, n.

To standardize the LSI value (LSIki (std)) on a scale of 0–100, we used the following formula:

$${LSI}_{ki}\left(std\right)=\left({LSI}_{ki}-{LSI}_{ki}\left(min\right)\right).\left({\displaystyle \frac{100}{{LSI}_{ki}\left(max\right)-{LSI}_{ki}\left(min\right)}}\right)$$

The LSI measurement results for each economic, social, and environmental dimension were divided into four categories, namely not sustainable (0.00–25.00), less sustainable (25.01–50.00), reasonably sustainable (50.01–75.00), and sustainable (75.01–100.00).

2.3.2. Determining Spatial Association of LSI

The interdependence of local locations characterized the measurement of the LSI; in this case, the researchers used the Geoda 1.20 software to calculate the Moran’s I index and use ArcGIS 10.8 software to visualize it into a map. To measure the autocorrelation locally, the Local Indicator of Spatial Autocorrelation (LISA) method was used, as performed by Jiao & Liu (2012) [49] and Marinda et al. (2020) [50]. The higher Local Moran Index value showed that neighboring areas had increasingly similar observed values. The Moran Index measures whether x and y are correlated with each other in one variable, for example, ki and kj, so the equation for the Moran Index is as follows [51]:

$$\begin{gathered} I={\displaystyle \frac{\sum _{i=1}^n\sum _{j=1}^n{W}_{ij}\left({LSI}_{ki}-{LSI}_k\right)\left({LSI}_{kj}-{LSI}_k\right)}{\sum _{{LSI}_k}^2\sum _{i=1}^n\sum _{j=1}^n{W}_{ij}}} \\ {I}_i={\displaystyle \frac{\sum _{j=1}^n{W}_{ij}\left({LSI}_{ki}-{LSI}_k\right)\left({LSI}_{kj}-{LSI}_k\right)}{\sum _{{LSI}_k}^2\sum _{i=1}^n\sum _{j=1}^n{W}_{ij}}} \end{gathered}$$

where I = Global Moran’s Index; I_i = Local Moran’s I or LISA statistics; LSIki = the value of LSI for k-th Dimension on the i-th village (desa); LSIk = the average value of LSIk; W_ij = contiguity matrix; representing the proximity of i-th village (desa) i’s and village (desa) j’s locations; n = the total number of villages (desa); Z2LSIk = the variance of the LSIk.

We can address spatial heterogeneity and spatial dependence, two distinct spatial effects, with the help of the locational problem. The first law of geography and spatial dependence are directly related. “Everything is related to everything else, but near things are more related than distant things”, according to the first law of geography [52]. It is feasible to confirm that the observations will be geographically agglomerated, or that there will be clusters, based on that law. Stated differently, there will be no independent location of the geographic data. Secondly, the concept of spatial heterogeneity is associated with the notion of economic behavior being unstable throughout space.

The weighting matrix (contiguity matrix) is a neighbor relationship expressed as weighting matrix W. The element of this matrix is Wij, which shows the size of the relationship between the i-th and j-th locations. According to Kosfeld & Dreger [53] and Fischer et al. [54], the neighborhood relationship rook contiguity is a matrix weighting based on the sides that touch each other, and the corners are not considered. The Moran Index value ranges from −1 to 1; if the I value < 0, it means a negative spatial autocorrelation, while if the I value > 1, it means a positive spatial autocorrelation.

3. Results and Discussions

3.1. Local Sustainability Index (LSI)

The results of the new factor analysis for social indicators in 2007 and 2021 are presented in

Table 2. Factor loading of social dimension in 2007 and 2021.

Variable		2007			2021
	Factor 1	Factor 2	Factor 3	Factor 4	Factor 1	Factor 2	Factor 3	Factor 4
SOCDIM_1	−0.021	0.517	−0.176	−0.391	0.141	0.214	0.228	0.756
SOCDIM_2	0.151	0.830	0.102	−0.069	0.366	0.842	0.063	0.134
SOCDIM_3	0.121	0.033	0.634	0.102	0.111	0.167	0.675	−0.017
SOCDIM_4	0.089	0.573	0.177	0.161	0.521	0.202	0.170	−0.061
SOCDIM_5	−0.021	0.072	0.046	0.867	0.121	0.129	0.219	−0.684
SOCDIM_6	0.919	0.032	0.096	−0.078	0.787	0.057	−0.140	0.223
SOCDIM_7	0.922	0.088	0.051	−0.034	0.803	0.137	−0.144	0.057
SOCDIM_8	−0.100	0.017	0.803	−0.056	0.136	0.204	−0.735	0.006
SOCDIM_9	0.496	0.154	−0.150	0.338	0.703	0.128	0.127	−0.211
SOCDIM_10	0.018	0.651	−0.086	0.090	0.091	0.915	−0.058	−0.050
Eigenvalue	1.997	1.748	1.163	1.079	2.238	1.756	1.189	1.161
% Total Variance	19.968	17.480	11.626	10.787	22.380	17.557	11.892	11.611
Cumulative Eigenvalue	19.968	37.448	49.074	59.861	22.380	39.937	51.829	63.440

The bold numbers are the characteristic variables, namely those with values >0.7.

. In 2007 and 2021, four new factors were formed. Eigenvalues were calculated for each resulting factor to indicate the variations in the data. These latent factors can explain the variables they represent for 59.86% of the dataset variance in 2007 and 63.44% in 2021 (Table 2).

The variables of socdim 2, socdim 6, socdim 7, and socdim 8 are the same characteristic variables in 2007 and 2021 because these variables have a factor loading value of >0.7, namely the number of households served by sanitation (having toilets) (socdim 2), the number of types of sports group activities (socdim 6), the number of sports facilities (socdim 7), and the number of efforts to maintain security (socdim 8). Meanwhile, the new characteristic variables in 2021 are the number of households served by sustainable drinking water (Socdim 1), safety facilities (Socdim 9), and the number of worship facilities (Socdim 10).

The results of the new factor analysis for economic indicators in 2007 and 2021 are presented in

Table 3. Factor loading of economic dimension in 2007 and 2021.

Variable	2007			2021
	Factor 1	Factor 2	Factor 3	Factor 1	Factor 2	Factor 3	Factor 4
ECODIM_1	0.413	0.346	0.409	0.111	−0.032	−0.057	0.905
ECODIM_2	0.151	0.827	−0.154	0.941	0.192	−0.022	0.057
ECODIM_3	−0.013	0.526	−0.009	−0.008	−0.139	0.601	−0.236
ECODIM_4	0.982	0.084	0.042	0.163	0.880	−0.192	−0.064
ECODIM_5	0.982	0.087	0.046	0.065	0.920	0.133	0.051
ECODIM_6	0.170	0.182	0.655	0.941	0.192	−0.022	0.057
ECODIM_7	−0.035	0.058	−0.585	0.201	−0.073	−0.649	−0.386
ECODIM_8	0.142	0.796	0.152	0.364	0.084	0.603	−0.013
ECODIM_9	0.214	0.338	−0.520	0.501	−0.245	0.351	0.008
Eigenvalue	2.217	1.879	1.259	2.238	1.788	1.328	1.037
% Total Variance	24.637	20.879	13.987	24.871	19.866	14.758	11.518
Cumulative Eigenvalue	24.637	45.516	59.503	24.871	44.737	59.495	71.013

The bold numbers are the characteristic variables, namely those with values >0.7.

. In 2007, three new factors were formed; in 2021, four new ones were formed. Eigenvalues were calculated for each resulting factor to indicate the variations in the data. The cumulative eigenvalue that forms the new economic dimension can explain the variables they represent for 59.50% of the data set variance in 2007 and 71.01% in 2021 (Table 3).

The variables of ecodim 2, ecodim 4, and ecodim 5 are the same characteristic variables in 2007 and 2021 because these variables have a factor loading value of >0.7, namely the variables for the number of cooperatives (ecodim 2), the amount of rice production (ecodim 4), and the area of agricultural land (ecodim 5). Meanwhile, the new characteristic variable in 2021 is the number of credit facilities (ecodim 6).

The results of the new factor analysis for environmental indicators in 2007 and 2021 are presented in

Table 4. Factor loading of environmental dimension in 2007 and 2021.

Variabel	2007						2021
	Factor 1	Factor 2	Factor 3	Factor 4	Factor 5	Factor 6	Factor 1	Factor 2	Factor 3	Factor 4	Factor 5	Factor 6
ENVDIM_1	−0.744	0.011	−0.039	0.028	0.044	−0.125	0.153	−0.688	0.160	0.043	−0.005	0.042
ENVDIM_2	−0.213	0.708	−0.150	0.100	0.042	0.115	0.100	0.703	0.221	−0.025	−0.067	0.065
ENVDIM_3	0.460	0.491	−0.226	−0.279	0.039	−0.183	0.311	0.105	0.613	−0.128	−0.123	0.017
ENVDIM_4	0.485	−0.118	0.031	0.106	0.030	−0.033	0.634	0.371	−0.141	0.003	−0.018	0.066
ENVDIM_5	0.194	0.041	0.820	−0.056	−0.041	−0.066	−0.661	0.236	−0.092	−0.039	−0.061	−0.025
ENVDIM_6	−0.417	−0.002	0.561	0.086	0.185	0.103	−0.476	0.131	−0.162	0.399	0.369	0.089
ENVDIM_7	0.214	0.174	−0.145	0.059	−0.521	0.011	0.009	0.100	0.037	0.179	−0.816	−0.086
ENVDIM_8	0.177	0.131	−0.089	0.037	0.844	0.002	0.014	0.122	0.043	−0.795	0.213	−0.098
ENVDIM_9	0.050	0.033	−0.013	−0.022	−0.007	0.974	0.117	0.079	0.082	0.182	0.178	0.819
ENVDIM_10	0.073	0.045	−0.019	0.960	−0.017	−0.026	0.221	0.159	0.142	0.432	0.391	−0.549
ENVDIM_11	0.013	−0.661	−0.257	0.002	0.080	0.028	0.165	0.063	−0.811	−0.047	−0.050	−0.019
Eigenvalue	1.343	1.246	1.159	1.038	1.032	1.028	1.286	1.250	1.194	1.066	1.058	1.009
% Total Variance	12.205	11.326	10.539	9.437	9.384	9.345	11.687	11.364	10.850	9.693	9.616	9.177
Cumulative Eigenvalue	12.205	23.530	34.069	43.507	52.891	62.236	11.687	23.051	33.901	43.594	53.210	62.387

The bold numbers are the characteristic variables, namely those with values >0.7.

. In 2007 and 2021, six new factors were formed. Eigenvalues were calculated for each resulting factor to indicate the variations in the data. The cumulative eigenvalue that forms a new environmental dimension factor can explain the variables they represent for 62.23% of the data set variance in 2007 and 62.38% in 2021 (Table 4).

The variables of envdim 2 and envdim 9 are the same characteristic variables in 2007 and 2021 because these variables have a factor loading value of >0.7, namely the frequency of landslides (envdim 2) and the area of critical and very critical land (envdim 9). Meanwhile, the new characteristic variables in 2021 are envdim 7 and envdim 11, namely the area of land cover mismatch with the protection function and the location of plantation land [55],while the missing variable in 2021 is the area of land cover mismatch with the buffer function (envdim 8).

The results of LSI analysis on the social dimension show that, in 2007, the average LSI was at 24.09 (not sustainable) and, in 2021, this increased to 42.21 (less sustainable). There was a decrease in the number of villages in the non-sustainable category of LSI by 43% from 2007 to 2021. It can be explained by the characteristic variables of sanitation, sports activities, facilities [56,57,58], and efforts to maintain security that there was an increase in the index but it remained at a low level. Figure 2 explains the distribution of spatial values for each LSI social dimension in 2007 and 2021. The increase in LSI social dimension values is concentrated in the southeast, southwest, and northeast regions of WMR. This condition means that these villages have better social performance than others (see Figure 2a). This condition shows the increasing development of social activities in the area, especially in village-scale activity centers such as Pracimantoro Village and Baturetno. Based on data from the Central Statistics Agency from 2007 to 2021, there was an increase in the number of hospitals, sanitation facilities, and toilets such that the entire population had access to them (100%) (

Table 5. Average LSI in 2007 and 2021.

Dimension	Category of LSI	2007			2021
		Average of LSI	Sum of Villages	%	Average of LSI	Sum of Villages	%
Social Dimension	Non Sustainable	14.67	108	55.96%	14.92	24	12.44%
	Less Sustainable	32.24	75	38.86%	39.76	128	66.32%
	Reasonably Sustainable	58.99	8	4.15%	60.41	32	16.58%
	Sustainable	87.78	2	1.04%	85.16	9	4.66%
	Average/Total of Sosical Dimension	24.09	193	100.00%	42.21	193	100.00%
Economic Dimension	Non Sustainable	15.96	117	60.62%	20.32	26	13.47%
	Less Sustainable	32.45	69	35.75%	37.34	126	65.28%
	Reasonably Sustainable	59.50	6	3.11%	59.08	34	17.62%
	Sustainable	100.00	1	0.52%	86.68	7	3.63%
	Average/Total of Economic Dimension	23.65	193	100.00%	40.67	193	100.00%
Environmental Dimension	Non Sustainable	10.95	4	2.07%	17.62	37	19.17%
	Less Sustainable	42.03	175	90.67%	31.95	154	79.79%
	Reasonably Sustainable	57.92	11	5.70%	73.33	1	0.52%
	Sustainable	94.47	3	1.55%	100.00	1	0.52%
	Average/Total of Environmental Dimension	43.11	193	100.00%	29.77	193	100.00%

).

The results of LSI analysis on the economic dimension show that, in 2007, the average LSI was at 23.65 (less sustainable) and, in 2021, this increased to 40.67 (less sustainable). There was a decrease in the number of villages in the non-sustainable category of LSI by 47% from 2007 to 2021. The characteristic variables of the number of cooperatives, the amount of rice production, and the area of agricultural land indicate that they have not been able to meet the region’s needs. Figure 2 explains the distribution of spatial values for each LSI economic dimension in 2007 and 2021. The LSI economic dimension values were concentrated in the southeast, southwest, west, northwest, and northeast of the WMR (see Figure 2b). This condition means that these villages have better economic performance than other villages. There is an increasing development of economic activity in the region, especially in activity centers that are aligned with district and provincial roads [59], namely villages in Pracimantoro, Eromoko, Baturetno, Wuryantoro, Nguntoronadi, Ngadirojo, Jatisrono, Slogohimo Districts. Furthermore, there was a 53% increase in the number of economic facilities, such as markets, minimarkets, and restaurants. The agriculture, forestry, and fisheries sectors were the most significant contributors to gross regional domestic product from 2010 to 2021, at 28% to 36% [60] (Table 5).

The results of LSI analysis on the environmental dimension show that, in 2007, the average LSI was at 43.11 (less sustainable) and, in 2021, this decreased to 29.77 (less sustainable). There was an increase in the number of villages in the non-sustainable category of LSI by 17% from 2007 to 2021. The characteristic variables such as the frequency of landslides and the area of critical and very critical land indicate the need for serious attention through disaster mitigation and environmental recovery programs. Figure 2 explains the spatial distribution of each LSI environmental dimension in 2007 and 2021. The decrease in LSI environmental dimension values were concentrated in the southeast, southwest, northwest, and northeast of the WMR (see Figure 2c). This condition means that these villages have worse environmental performance than other villages. This condition indicates environmental damage in the area, such as critical land and a high incidence of landslides. Most of the areas in Wonogiri Regency have a high landslide disaster risk. This is because 49% of the soil is lithosol, which is very sensitive to erosion, and the risk is increased by the steep slopes in each village [61]. Drought-prone disaster risk in the Wonogiri Regency is in the high category [62], so water source availability and river normalization programs are important for residents (Table 5).

In 2021, the total level of sustainability in the study area was in the low category, namely environmental in 154 villages (79.79%), social in 128 villages (66.32%), and economic in 126 villages (66.32%). The number of villages in the sustainable category for social, economic, and environmental, respectively, were nine villages (4.66%), seven villages (3.63%), and one village (0.52%), with an increase of only around 3% from 2007. No villages had a high sustainability status regarding social, economic, and environmental dimensions. Only two villages had high social and economic dimensions, namely Baturetno Village and Pracimantoro Village (Table 5).

3.2. Spatial Association of LSI

To measure the spatial information, neighborhood relations are needed, namely assessing the relative location of one location with other places. Based on these results, in 2021, the Moran social, economic, and local environmental index values were 0.09, 0.35, and 0.42, respectively, which indicated the existence of positive spatial autocorrelation; thus, it can be interpreted that in these areas, there is the same level of spatial association. Sustainable development between the neighboring villages and spatial patterns tends to be clustered.

In 2007, the spatial distribution of high–high (HH) category villages for the social dimension was concentrated in the east, southeast, and southwest of the Wonogiri Multipurpose Reservoir (WMR). By 2021, these HH category villages were found in the southwest, east, southeast, and northwest regions of the WMR, indicating that areas with high sustainability values are clustered together. Conversely, in 2007, the low–low (LL) category villages were situated in the east, southeast, southwest, and northeast of the WMR. In 2021, the LL category was present in the east, northeast, and southwest, demonstrating that the villages with low sustainability values surrounded one another (Figure 3a).

For the economic dimension, in 2007, the HH category villages were located in the east, southeast, and southwest of the WMR. By 2021, their distribution expanded to include the southeast, south, southwest, west, northwest, and east. Conversely, in 2007, the LL category was found in the east, southeast, and northeast of the WMR, remaining consistent in 2021 across the east, southeast, and northeast areas (Figure 3b).

Regarding the environmental dimension, in 2007, the HH category villages were similarly located in the east, southeast, and southwest of the WMR. By 2021, their presence shifted to the east, southwest, and northwest. The LL category in 2007 was distributed across the east, southeast, northeast, and southwest, and this distribution remained unchanged in 2021 (Figure 3c).

Based on the spatial association results in Figure 3, the next step is to display a composite that combines the HH and LL categories for LSI [63] in 2007 and 2021 for the social, economic, and environmental dimensions (Figure 4). The first category (1st) shows the best local sustainability performance condition, namely with the HH index value that persisted from 2007 to 2021. The second category (2nd) is the local HH sustainability condition, namely only at one point in the year of either 2007 or 2021. The third category (3rd) is the LL sustainability condition, namely only at one point in the year of either 2007 or 2021. The fourth category (4th) is the worst local sustainability condition, namely there is no change in sustainability which is still low/LL from 2007 to 2021.

The combination of spatial associations in the social dimension shows that the first category is located to the east of the WMR. For the second category, it is located in the southeast, southwest, northwest, and northeast of the WMR. This indicates that there is a high social sustainability association around the urban areas that are served by district and provincial roads. It looks different from categories 3 and 4 which are in villages that are relatively far from highways such as Guno Village which have a constant low sustainability association value from 2007 to 2021 (Figure 4a and Figure 5a).

The combination of spatial associations in the economic dimension shows that the first category is located in the southeast, south, southwest of the WMR. For the second category, it is located in the southeast, southwest, northwest, and northeast. This indicates that there is a high economic sustainability association around the urban areas that are passed by district and provincial roads. It looks different from the sustainability of categories 3 and 4, which are in the villages that are relatively far from road and catchment area of the WMR, and which have a constant low sustainability association value from 2007 to 2021 (Figure 4b and Figure 5b).

The combination of spatial associations in the environmental dimension shows that no villages fall into the first category. For the second category, it is located in the east, southeast, south, southwest, and northeast. This indicates that there is a high environmental sustainability association in several villages in the catchment area. However, it looks different from the sustainability association category 3, which is in a village that is relatively far from the highway and is a catchment area of the WMR. Category 4, which has a low sustainability association from 2007 to 2021, is in a village that is a catchment area of the WMR (Figure 4c and Figure 5c).

According to the combined category map in Figure 4, Figure 5 illustrates the categories in a quadrant format. Category I villages, deemed the best, maintained a high–high (HH) sustainability association from 2007 to 2021 and are located in Quadrant I (top right). In contrast, Category IV villages, considered the worst, exhibited a low–low (LL) sustainability association during the same period, placing them in Quadrant III (bottom left). Categories II and III consist of villages that demonstrated sustainability association conditions in just one year of either HH in 2007 or 2021 (Category II) or LL in 2007 or 2021 (Category IV).

4. Conclusions

Village sustainability in the social and economic dimensions increased from 2007 to 2021, especially in the urban activity center areas that serve social and economic facilities. Meanwhile, a low sustainability score indicated pockets of social and economic problems, namely non-urban areas relatively far from the national, provincial, and district road networks. This seems different in the environmental dimension, where the sustainability value decreased from 2007 to 2021. High sustainability values were concentrated in suitable environmental conditions (disasters rarely occurred with land cover that was still maintained), while poor environmental conditions occurred in pocket areas. Environmental problems were indicated by land conversion and disaster areas.

In order to maintain the sustainability of the area, it is necessary to improve the management program. For the social dimension, program improvements should be prioritized in villages that have static sustainability, namely in category 4 (worst), especially improvements in sanitation (ownership of toilets), increasing sports group activities, adding sports facilities, and security guard programs. The next priority should be applied to categories 3, 2, and 1.

For the economic dimension, program improvements should be prioritized in villages that have static sustainability, namely in category 4 (worst), by developing the number of cooperatives, improving rice productivity, and improving/increasing the area of agricultural land. The next priority should be applied to categories 3, 2, and 1.

For the environmental dimension, program improvements should be prioritized in villages with static sustainability, namely in category 3, in the form of a program to improve landslide areas, critical and very critical land, and the monitoring and regulation of locations of non-conformity of function in protected areas and plantations. The next priority is then applied to categories 2 and 1.

This study is a refinement of the previous studies on measuring watershed sustainability, as conducted by Setianingtias et al. (2019) [28], Rahma et al. (2019) [29], Fauzi & Oxtavianus (2014) [30], which considered the social, economic, and environmental aspects of measuring the Sustainable Development Index at the regional level. The indicators used in the research above are only available at the provincial or district level. Data on these indicators have yet to be available for the most minor administrative units or villages. With the same unit of analysis, Pravitasari et al. (2018) [32] measured the sustainability index by considering the spatial interdependence relationships at the regional scale. Fadli (2017) [34] stated that most of the regions in the world have their characteristics and that measuring sustainability at the village level cannot necessarily use sustainability measurements at the regional level.

The findings of this study emphasize numerous critical insights that reflect the complexity of the catchment challenges, which require coordinated planning and management by cross-agency groups and stakeholders. This study has limitations, thus more research is needed for identifying and mapping the influence of smaller regions at the village level, researching suitability through spatial planning, and establishing policies and programs for spatial usage in the watershed. Furthermore, policies and actions are required to preserve the watershed, such as erosion and sedimentation control, sustainable forest and land management, water resource management, and social institutional management.