A GIS-based approach to determine the priority area for rainwater harvest in Kupang

Kupang, a pivotal economic and administrative hub in Nusa Tenggara Timur, Indonesia, faces recurring droughts resulting in crop failures, food insecurity, and health problems. Addressing this water crisis entails rainwater harvesting as a potential solution. This study establishes a comprehensive framework for prioritizing rainwater harvesting areas in Kupang. Using the weighted overlay method and open-source GIS software, our approach involves three stages: identifying drought and flood priority zones, defining suitability indicators, and quantifying the potential of rainwater harvesting systems. By leveraging climatological, geological, and hydrological information, we determine priority areas for drought and flood management. Our analysis of rainwater harvesting suitability integrates climatological information, land use and land cover mapping, runoff potential assessment, and water use index. Quantification of rainwater harvesting system requirements hinges on rooftop area and population data. The entire wards in Kupang are classified as high priorities for drought mitigation. Therefore, rainwater harvesting is a viable strategy for all wards. The most prioritized and in-demand wards, Oesapa, Sikumana, Lasiana, and Liliba, have the potential to install a minimum of 12,000 rainwater harvesting systems, each with a 1500-liter volume barrel.


Introduction
Kupang is emerging as a hub for economic, educational, and socio-cultural activities in East Nusa Tenggara Province (after this, referred to as NTT).However, this growth faces a significant challenge: a rising demand for clean water.A previous study conducted in Kupang found that existing water sources struggle to meet long-term needs [1].Approximately 17-40% of the local population spends their income acquiring clean water [2].The semi-arid classification and limited water resources exacerbate this issue.A semi-arid region is characterized by a prolonged dry season and a short rainy season.Areas under this climatic zone typically experience favorable temperatures for crop cultivation but unreliable soil moisture [3].
Given these considerations, optimizing the utilization of limited rainfall within reservoirs and rainwater harvesting systems (RWHS) becomes imperative.Rainwater harvesting serves as a compensatory measure for water scarcity [4].This type of RWHS is characterized by using the structure of the roofs or patios to collect water in tanks.The World Health Organization (WHO) highlights its viability as an alternative water source, provided designs incorporate precipitation, catchment area, roof material, and losses [5].Numerous studies have highlighted the advantages of RWHS implementation.Harvested rainwater can fulfill crop water demands in areas adjacent to greenhouses in Turkey [6], and it can curtail runoff by reducing the volume of water entering storm drainage systems, thus mitigating flood risks.Opting for RWHS in rural or urban communities where access to quality water is limited by scarcity and contamination brings benefits such as economic savings since rainwater is a free resource [7].
The selection of RWHS location relies on six crucial factors, i.e., hydrology, climate, soil, agronomy, topography, and socio-economic conditions [8].Considering all these aspects can escalate the complexity of site selection, particularly in cases involving vast watersheds.The Geographical Information System (GIS) and remote sensing technology have simplified the site selection procedure for rainwater harvesting structures by dropping the number of proposed sites and selecting only optimal locations.Some studies have documented the selection of RWHSS sites based on GIS [9], [10].However, those studies usually involve a complex hydrogeological approach and tend to neglect the socio-economic condition [8].
This study aims to develop a framework for selecting the RWHS priority area using multiple variables involving flood vulnerability and drought potency, land use, runoff, climatological information, disaster data, population, and building information.We applied our approach to the Kupang city area, which lies in the Benain-Noelmina Watershed in Timor Island of NTT.This approach can emphasize the role of rainwater harvesting as one of the solutions for drought in semi-arid regions.

Site Location
The study area is in Kupang (Figure 1), comprising six districts and 51 wards (or villages).Alak holds the largest expanse, while Kota Lama is the smallest.The southern part of Kupang exhibits elevated terrain, ranging from 100 to 350 m above mean sea level (ASL).This region lies in the Benain-Noelmina watershed, with several prominent rivers coursing through Kupang, including Liliba, Dendeng, and Merdeka.Notably, most tributaries experience intermittent flow, often drying up during the dry season.

Methodology
Table 1 presents the data used, including thematic map, topographic, demographic, climatological, disaster, and building information, and Figure 2 illustrates the framework.We analyzed RWHS potential in subsections A-C within an R-Studio framework, which was influenced by a Python-based study for spatial data preprocessing [11].The initial data preprocessing steps included reprojecting input files to WGS84 UTM zone 51S, cropping them to the Kupang study area, and subsequently interpolating and resampling them to a specified resolution.The Digital Elevation Model (DEM) and building information were obtained from the raster and OSM library, respectively [12], [13].The preprocessing tasks, such as reprojection, resampling, and cropping, were executed using libraries like sf, raster, RGDAL, RGEOS, and dplyr [13]- [17].R-Studio has a history of being utilized for spatial analysis in various geophysics studies, offering cost-effectiveness and efficiency in data management and visualization [18]- [20].

A. Analysis of flood and drought priority area
We used a scoring methodology that considered flood vulnerability, drought severity, and historical flood and drought events to determine the priority zones for flood and drought.A widely adapted weighted overlay methodology [21] was used to map flood vulnerability due to its simplicity and popularity within the GIS framework.This method incorporates multiple datasets: five years of annual rainfall data, a watershed delineation map, a soil map, a land use and land cover map, a river stream map, and digital elevation model (DEM) data.Furthermore, we determined drought severity parameters and historical disaster events (floods and droughts) occurrences through thematic maps and the data provided in Table 1, respectively.The specific scoring approach used to identify priority zones for these two types of disasters is outlined in Table 2. (5) Water Use Index (WUI), is calculated from the discharge ratio to population.This index provides insight into how water discharge relates to water demand.(6) Population Density (P) is defined as the number of individuals residing within a square kilometer of land area.
There are two categories of indicators: Poor and Good (Table 3).A sub-district (or village) attains a priority status for RWHS installation when the scores for these indicators fall within the ranges defined for the Poor category.While recognizing the significance of economic variables, precisely the number of households below the poverty line, our study makes the simplifying assumption that this factor is uniform.This assumption is based on the 2021 report, which states that 9.17% of Kupang households meet the impoverished status criteria [23].According to P.61/Menhut-II/2014, this value nearly aligns with the Poor indicator range.

C. Analysis of RWHS potential.
We utilized the OSM library to download the distribution of buildings.Only buildings with an area exceeding 25 m 2 (Table 4) were included in the calculation.Additionally, we assumed that all identified buildings were suitable for RWHS installation, including having appropriate roof design and materials.For each building, the rooftop area was assumed to be equal to the building area, and the count of buildings was aggregated for each sub-district.The determination of potential RWHS units followed the criteria in Table 4.For every additional (25-50) m 2 of building coverage area, an additional 1 unit or 1.5 m 3 volume is required.

Flood and Drought priority area
Figure 3 displays the map depicting the potential for drought and flood occurrences, along with the identified priority areas for addressing both disasters.All sub-districts or villages in Kupang were classified as top-priority zones for drought mitigation.Additionally, 17 sub-districts have been designated as high-priority zones for flood mitigation, including Alak, Batu Plat, Mantasi, Namosain, Penkase Oeleta, Lasiana, Oesapa, Oesapa Barat, Oesapa Selatan, Airmata, Bonipoi, Fontein, Kolhua, Maulafa, Naikolan, Sikumana, and Liliba.These areas are more susceptible to floods due to their proximity to river streams.Our findings are consistent with a prior study that also designates Kupang as a susceptible area to drought [24], [25], where a Standardized Precipitation Index (SPI) below -2 was observed for this region.The SPI, a widely accepted meteorological drought index, serves to quantify the divergence of precipitation in a specific location and period from its historical average.Positive SPI values suggest wetter conditions than usual, while negative values indicate drier conditions.SPI value below -2 indicates severe drought conditions.

Indicators for RWHS Suitability
The outcomes of the RWHS indicator analysis are presented in Figure 4.Among the indicators, only rainfall intensity was classified under the good category, while the remaining five indicators were predominantly classified as Poor.Wards in the southern region fall within the good category due to their land cover, being the sole wards comprised of permanent forest.Notably, Naioni and Oesapa Selatan exhibit good indicators for WUI, with Naioni also possessing a favorable indicator for population density.
Rainfall intensity was determined using the mean daily rainfall during the rainy season (December-January-February) in 2021.This data was also employed in the calculation of discharge for the WUI.The selection of the mean daily rainfall during the rainy season is justified by Kupang's classification as a semi-arid climatic region, with a pronounced dry season spanning from eight to nine months and a relatively short rainy season of three to four months.The peak of rainfall typically occurs during monsoonal activities from December to February [26].
While rainfall fitted the good indicator class, surface runoff was classified as poor for all subdistricts, indicating a substantial loss of rainwater and a high potential for flash floods.This observation is supported by a report from BPBD of Kupang, which highlights the susceptibility of the majority of Kupang's areas to flash floods [24].3, presents the potential quantity for RWHS unit installation in each district.The total rooftop area ranges from 19,000 to 810,000 m 2 .Liliba has the broadest available rooftop area, followed by Oesapa, Kelapa Lima, Oebufu, Lasiana, and Sikumana.Consequently, Liliba has the highest potential for RWHS units, exceeding 15,000.Following closely are Oesapa, Lasiana, Sikumana, and Kelapa Lima, each with an average of 13,000 units.Despite Liliba's advantageous rooftop area, the priority for RWHS installation is given to Oesapa due to its elevated flood occurrence (as indicated in Figure 2).This prioritization aims to mitigate the impact of disasters.In contrast, Mantasi offers the smallest rooftop area due to its compact size.Hence, this sub-district presents lower potential for RWHS installations.Additionally, smaller numbers of RWHS units are anticipated for Fatukoa and Naioni subdistricts owing to their lower population figures.A prior study on RWHS planning revealed that the freshwater demand for a household in the Bonipoi ward ranged from 7.02 to 11.24 m 3 per month [27].By considering a runoff coefficient of 0.85 [27] and a daily mean rainfall of 16.24 mm (Figure 4), we can deduce that, for a 30 m 2 rooftop area, the monthly water storage achievable from the designated RWHS (as outlined in Table 4) within Bonipoi amounts to 12.42 m 3 during the rainy season or 0.41 m 3 per day.
It is essential to note that our analysis does not account for harvested water during the dry season, nor does it address the suitability of water quality for safe consumption when stored to meet dry season demands.In the future, more comprehensive investigations that involve on-site assessments of building conditions, water quality analysis, and the determination of effective storage capacity will be necessary to corroborate our current findings.Nevertheless, our study successfully demonstrates the utility of a GIS approach in pinpointing priority areas for RWHS installation.

Conclusions
Using GIS, it is possible to identify the priority sub-district and assess the collective potential quantity of RWHS installations at the sub-district level.The framework utilized an open-source GIS program that is cost-effective and amenable to editing by other users.Our findings indicated the viability of installing over 12,000 units in several priority sub-districts in Kupang, including Liliba, Oesapa, Kelapa Lima, Oebufu, Lasiana, and Sikumana.The storage potential of the water system is contingent upon the rooftop area, with a 30 m 2 rooftop capable of storing up to 12.42 m 3 of water per month.This framework is efficient, straightforward, and adaptable.It holds potential for application in other areas, solidifying its value as a tool to effectively mitigate flood and drought risks.

Figure 1 .
Figure 1.Administrative map of Kupang and the topography profile

Figure 2 . 1 )
Figure 2. Research frameworkB.Analysis RWHS indicatorThe analysis of RWHS indicators entails assessing suitability based on various physical, social, and economic parameters relevant to RWHS implementation.Within this context, we evaluated six key parameters: dry spell frequency (DS), rainfall intensity (I), percentage of vegetation cover (PVC), surface runoff (R), water use index (WUI), and population density (P).These indicators are included in the factors affecting rainwater harvest[22].Simplicity and ease of application guided the selection of these indicators, as they were readily accessible for data acquisition within the developed framework.Each indicator is described as follows.(1)Number of Dry Spells (DS), representing the consecutive count of dry days (days without rainfall) that occur without interruption from rainy days.(2) Rainfall Intensity (I), is defined as the total depth of rain (measured in millimeters) that falls within a specific unit of time (usually an hour).In our study, we utilized the daily average rainfall intensity for the rainy season from GPM-IMERG data.(3) The percentage of Vegetation Cover (PVC) is calculated from the ratio of the vegetated land area to the total sub-district area.Only the land cover class representing perennial plants is included in the calculation.(4) Surface Runoff (R) refers to the flow of water that occurs over the ground surface due to the saturation of the soil's infiltration capacity.

Figure 3 .
Figure 3. Disaster vulnerability and disaster priority zones in Kupang: (a) Drought severity and factual drought map, (b) Drought priority map, (c) Flood vulnerability and factual flood events map, (d) Flood priority map

Figure 4 .
Figure 4. Matrices describing the RWHS indicators.The definitions and criteria for each indicator are provided in subsection 3.2.23.3.Potential units of RWHS Figure 5 illustrates the distribution of buildings in Kupang, and Figure 6, which employs the criteria from Table3, presents the potential quantity for RWHS unit installation in each district.The total rooftop area ranges from 19,000 to 810,000 m 2 .Liliba has the broadest available rooftop area, followed by Oesapa, Kelapa Lima, Oebufu, Lasiana, and Sikumana.Consequently, Liliba has the highest potential for RWHS units, exceeding 15,000.Following closely are Oesapa, Lasiana, Sikumana, and Kelapa Lima, each with an average of 13,000 units.Despite Liliba's advantageous rooftop area, the priority for RWHS installation is given to Oesapa due to its elevated flood occurrence (as indicated in Figure2).This prioritization aims to mitigate the impact of disasters.In contrast, Mantasi offers the smallest rooftop area due to its compact size.Hence, this sub-district presents lower potential for RWHS installations.Additionally, smaller numbers of RWHS units are anticipated for Fatukoa and Naioni subdistricts owing to their lower population figures.

9 Figure 5 .Figure 6 .
Figure 5. Building coverage map used for RWHS quantity potential in each sub-district

Table 1 .
The data used in the analysis of potential RWHS quantity

Table 2 .
Scoring for the scale of priority a Score = Values of vulnerability or severity + Values of Factual events

Table 4 .
Criteria for RWHS installation