Determinants of rainwater harvesting practices in rural communities of Limpopo Province, South Africa

ABSTRACT Most rural parts of South Africa’s Limpopo Province are faced with limited water access, provision, and inadequate infrastructure for water reticulation. With over 80% of the provincial population living in rural areas, water availability for domestic and agricultural use has become a source of concern. Rainwater harvesting (RWH) has emerged as an important source of alternative water supply for underserved communities during the summer rainfall season. RWH refers to the collection, storage and conservation of surface runoff for agricultural production and domestic uses. This study assessed the socio-economic characteristics and determinants of RWH practices in rural communities of Limpopo Province. Descriptive statistics were used to describe households’ socio-economic characteristics and logistic regression was used to examine the determinants of RWH practices in rural communities. The analysis was based on a survey of 478 households selected using convenience and purposive sampling techniques. The key findings indicate that a significant number of households (63.8%) practised RWH to augment available water for domestic use, albeit few households used the harvested water for portable use. Further analysis showed that a majority of those who adopted RWH practices were female (64.24%). The results of the empirical models indicate that there is a statistically significant positive relationship between perception, age and level of education in determining the adoption of RWH practices. The findings underline the importance of understanding the adoption of RWH in rural communities and highlight the vital role of education, knowledge, and income in rural communities with limited or no access to a water supply. The study highlights the need for policy to promote the benefits of harvesting rainwater for domestic use.


Introduction
Water security already affects every continent, and this often results in a scarcity of resources for rural communities (United Nations World Water Development Report (UN-Water, 2018).Water is a finite resource, and its availability is central to human life, sustainable development and the proper functioning of ecosystems and biodiversity.It is estimated that water use has been growing globally at more than twice the rate of population increases in the last century, and an increasing number of arid regions are struggling to sustainably deliver water services (Kummu et al., 2016).Globally, water security for livelihood, development and agricultural use has received eminence as an important area of research (Amos, Rahman, & Gathenya, 2016).As a developing region, southern African water-related challenges are rather diverse in their nature, ranging from access, financing and infrastructure, and quality issues as evidenced by the cholera outbreak over the last decade.The Africa Centers for Disease Control reported an unprecedented cholera outbreak of varying magnitude in the Democratic Republic of Congo, Mozambique, and Angola in the year 2018 while in Zambia and Zimbabwe reports were in 2017 and 2009, respectively.According to UN-Water (2011), deteriorating water quality has become a global issue of concern as the human population grows, industrial and agricultural activities expand, and climate change threatens to cause major alterations to the hydrological cycle.
Rural drinking water supplies across much of Sub-Saharan Africa are facing increasing challenges as water infrastructure, operation, maintenance, and financial sustainability are inadequate (Faye, 2021;Heidari, Arabi, Warziniack, & Sharvelle, 2021).Many rural communities are dependent on rainwater for domestic and agricultural activities; however, climate change and irregular rainfall patterns are likely to increase water demand while shrinking water supplies (Lebek & Krueger, 2023).To compound this; South Africa is classified as a water-stressed country, with an average annual rainfall of around 500 millimeters, which is less than 60 percent of the world average (Rodda et al., 2016).In addition, South Africa's growing water demand is rapidly outstripping its natural availability with a population of 56 million.The country is characterized by a semi-arid climate with rainfall varying from less than 100 millimeters per annum in the west to over 1500 millimeters per annum in the east (Mwendera & Atyosi, 2018;Nel & Sumner, 2007).The country is susceptible and vulnerable to the ravaging effects of climate change, particularly the effects of long-term droughts (Orimoloye, 2022, Muller et al., 2009).For example, several parts of the country experienced severe drought episodes, including the devastating drought incidents of 2017 that affected the Eastern Cape, Northern Cape, and Western Cape (Botai et al., 2020;Botai, Botai, De Wit, Ncongwane, & Adeola, 2017).From 2012, Limpopo province experienced extremely dry conditions that culminated in a meteorological drought between 2014 and 2016.In the Mopani District Municipality of Limpopo Province, the 2014/16 drought had devastating impacts on water levels, crop yields, livestock herds, and rural livelihoods (Nembilwi et al., 2021).Much like the entire South Africa, which is characterized by a semi-arid climate, with an average rainfall of about 465 mm (Du Plessis & Schloms, 2017), the climatic environment of Limpopo Province closely mirrors that of the country, characterized by severe drought and high evaporation rates often resulting in severe water shortage (Davis-Reddy & Vincent, 2017;Zengeni, Kakembo, & Nkongolo, 2016).Communities with access to water reticulation are often left without water for prolonged periods due to a breakdown in the supply infrastructure or a lack of potable supply by service providers (Hoffman & Nkadimeng, 2016).
South Africa has a long history of inequalities due to policy legislation that denied access to basic services such as water allocation to the majority of black South Africans.This inequality denied access to water and the development of infrastructure unfairly disadvantaged the black rural communities.While South Africa transitioned from the apartheid system to democracy, most rural communities still carry the burden of this inequality and injustice (Jegede & Shikwambane, 2021).Most rural communities in Limpopo Province still lack access to potable water for domestic purposes (Hoffman & Nkadimeng, 2016).The province is considered a rural province with over 80% of the population living in rural areas which lack infrastructure development (Malatji, 2020).Water sources are not mutually exclusive, and it may be possible or necessary to have more than one source depending on the intended use.The challenges of water security differ between urbanized and rural settings with most urban dwellings served with piped water reticulation systems while this is absent in most rural areas (Mthethwa, 2018).However, coupled with a changing climate and a growing population, water availability for domestic and agricultural use is impacted and the situation is likely to worsen in the future (Thomas, Twyman, Osbahr, & Hewiston, 2005).
As a result of the physical water shortage, scarcity in access due to lack of adequate infrastructure or failure of institutions to ensure a regular supply, communities have realized the need to "revive this old practice of RWH" (Mwenge- Kahinda, Taigbenu, & Boroto, 2010).Rain Water Harvesting (RWH) has become important for its potential to secure water security (Amos, Rahman, & Gathenya, 2016).For these reasons, RWH practices for water security has been identified as an alternative source of water supply for regions facing water scarcity with varying climatic conditions (Amos, Rahman, & Gathenya, 2016;Barron, 2009;König et al., 2009;Sharma et al., 2009;Traboulsi and Traboulsi, 2017).In literature, an appropriate definition is dependent on the context of RWH.Lani, Yusop, and Syafiuddin (2018) have defined RWH as the direct collection of water from roofs and other purpose-built catchments.According to Yannopoulos, Antoniou, Kaiafa-Saropoulou, and Angelakis (2017), the RWH is the collection, conveyance, and storage of rainwater for future use (domestic, agricultural, and environmental protection), while Haut et al. (2015) argue that the term describes the various methods of using, collecting, and storing rain run-off water for domestic and agricultural purposes.In the South African context, the Department of Water Affairs (2013) describes RWH as a technology more appropriate in rural areas and implies a broader definition than just the collection of run-off from a roof including hillsides and surface runoff.Thus, RWH practice is understood to refer to the collection, storage, and use of runoff, typically from roofs (Arku, Omari, Adu-Okoree, & Abduramane 2015;Roebuck, 2007) and it can contribute to household water security.
In South Africa, RWH has long been a popular practice at small to medium scales, particularly in rural and peri-urban areas (Mwenge-Kahinda, Taigbenu, & Boroto, 2010).Dobrowksy et al. (2014) found that over 55,000 households in southern Africa utilize rainwater as their main source for domestic purposes.RWH also provides water for small-scale home-based productive activities such as vegetable gardening, which could positively contribute to food security for lower socioeconomic groups (Mwenge-Kahinda, Taigbenu, & Boroto, 2010).Despite the many benefits of RWH (Haut et al., 2015), the adoption of the practices has been a blight of debate regarding its efficiency, suitability to different climatic environments and installation costs.There are other factors such as socio-economic considerations that influence the adoption of RWH.It has been reported that RWH is not widespread because the public is uninformed about this practice, lacks knowledge about the investment cost, and doubts the water quality (Damman et al., 2017).Based on the water-related challenges and the current complex realities in most parts of rural Limpopo Province, this study aims to assess the adoption of the RWH practices and their determinants to improve water availability and security at the household level.

Description of the study area
The study was conducted in Capricorn and Mopani District municipalities in Limpopo Province, situated in the northernmost part of South Africa as shown in Figure 1.Limpopo Province includes five district municipalities: Vhembe, Mopani, Greater Sekhukhune, Capricorn, and Waterberg.The population of Limpopo has grown from 4,995,534 in 2001 to 5,404,868 in 2011 and has grown to 5,941 439 in 2022 (Statistics South Africa, 2022).The province occupies an area of about 125,754 km 2 with the highest elevation of 2126 m, topographically dominated by isolated mountains.The Capricorn District is characterized by warm to hot and wet summers, and cold dry winters with colder weather while the Mopani District is known as the subtropical Lowveld region, of hot-rainy summers and warm-dry winters.

Study population and sampling
The study population comprised households within the communities surveyed.A household is defined as a group of persons who normally live and eat their meals together in the same dwelling (Chianu & Tsujii, 2004).Household heads or any adult within the household aged over 18 years and who met the inclusion criteria were invited to participate.The inclusion criteria were that they should have been living in the community for not less than one year, as this was enough time to have set up a water harvesting system and have used it for water supply considering the two rainy seasons prevalent in the area and the giving of consent.Households with no adult (18 years and above) present at the time of the study, or where consent was not given were excluded from the study.The sample size was calculated using the formula for prevalence study (Raosoft Incorporated (2004) with the statistic for a level of confidence (Z) at 95%, population proportion set as 50% is the proportion of people who were familiar with RWH.The margin for error was set at 5%.Both district municipalities have a combined number of 729 391 households (Statistics South Africa, 2022).The sample size was calculated as 384 households.A total of 533 households were invited to participate in the study across nine villages in the area.Of these households, 478 completed the questionnaire, with a response rate of 87.7%.
A structured questionnaire with open and closed questions was developed to collect data on the types of RWH systems, domestic uses and benefits of rainwater, health concerns of rainwater use, and other sources of domestic water.The open-ended questionnaire structure was adopted, and this design was preferred as it allowed respondents to express themselves (Etikan & Bala, 2017).The study employed both convenience and purposive sampling.Convenience samples are the norm within developmental science (Bornstein, Jager, & Putnick, 2013;Jager, Putnick, & Bornstein, 2017).Convenience and purposive sampling methods use the researcher's judgment in selecting the respondents.To minimize bias associated with non-probability sampling, the study used a validated/ standardized questionnaire and avoided complex and vague terminology and questions.

Descriptive statistics
The completed questionnaires were screened for completeness, coded and captured into the Statistical Package for Social Sciences (SPSS) version 20.0 software for analysis.Frequency distribution included creating a frequency table to summarize and organize the data by recording every possible score of the respondents as a column of numbers and the frequency of occurrence of each score.The tables showed the number of frequencies with their percentages.The information was confined to the frequency table and was converted into a form of charts, tables, and graphs.

Logistic regression model
To determine the factors that influenced communities' decision to use RWH practices, a logistic model was specified, and its parameters were estimated.The model was chosen over other similar models as it provides a transparent, robust, empirical basis for understanding factors that have a significant influence on the adoption of rainwater harvesting by households.Logistic regression is necessary because the dependent variable is binary.Logistic regression is used to predict a categorical dependent variable based on continuous and categorical independent variables (Gujurati, 1995).The logistic regression is thus necessary to estimate households' probability of adopting the RWH technique.It is also used to determine the percentage of variance in the dependent variable explained by the independents and shows the impact of independent variables on the dependent variable.
The relationship between the probability of Y = 1 and the explanatory variables is determined through the logit function, which is the natural logarithm of odds of Y = 1.This assumes a linear relationship between the log of odds and independent variables.The analysis in this study thus focuses on the probability that households adopt RWH techniques in the study area (Equations 1 and 2).The logistic regression model is based on the probability that Y equals one (p=p 1 ).The value of Y is assumed to depend on the value ofX 1 . . .X k .The logit model representing the relationship of Y and Xis given by: where: p is the predicted probability that Y equals one (dependent variables) β 0 , β k is the estimated parameters X 1 and X k are independent variables.The specific logistic regression model employed in this study is given as follows: where; ADOP = Adoption of Rainwater Harvesting Practices, AGE = Age of the household head, EDUCL= Level of Education, HHLDSZ = Household Size, LVLINC = Level of Income, HRLBR = Hired Labour,WRTSOUR = Water Source and ACINFO = Access to Information.

Ethical considerations
To protect the respondents' interests, well-being and privacy, the researchers obtained informed consent from the participants.Consequently, researchers must also ensure that they do not coerce potential participants into agreeing to participate in studies.Following accepted ethical practice (Sanjari, Bahramnezhad, Fomani, Shoghi, & Cheraghi, 2014;Tripathy, 2013, Grbich, 1999), respondents were assured that their information would not be misused to embarrass or humiliate them and that their participation was voluntary.The respondents were further assured that their identities would not be disclosed in any resulting publication.Ethical clearance was obtained from the North-West University, Faculty of Natural and Agricultural Sciences Ethics Committee (Ethics number: NWU − 01268-21-A9), approved on 23 September 2021.

Socio-economic and demographic characteristics of the respondents
The demographic characteristics of the respondents in the study area depicted a high gender bias with females representing 65% of respondents.The high representation of females was expected, given the rural settings of the study area and the population dynamics of the study area.In these rural settings, the likelihood of more women being at home could reflect the historical background of much of rural South Africa whereby women generally remained in the village while men were engaged in migrant labor on farms, mines and in the urban centers across the country (Vosloo, 2020).A study by Hoffman and Nkadimeng (2016) reported females as decisionmakers, playing an important role as homemakers and are specifically involved in the use and directly impacted water availability at home compared to men.Thus, the high representation of women can be deemed as a reliable and a true reflection of common practice.The age of the respondents surveyed ranged between 18 and 70 years.Most household heads were aged between 46-60 years (42%), followed by the age group 31-45 years at 30%, such that over 72% of the total sampled population was 31 years old and above (Figure 2).The educational level of the respondents in the study area is presented in Figure 3 and shows a high level of academic literacy, the majority (57%) of the respondents have secondary education whilst 16% have some form of tertiary education.Only 2% of the respondents had no formal education.The remaining respondents had technical (16%) and primary education (9%).The literacy or higher levels of education attained by the respondents were expected to positively impact their knowledge of RWH techniques, influencing their ability to practice and adopt the practice.A study by Machete (2011) in the Sekhukhune District of Limpopo Province found that most household respondents who had some form of secondary education (Grade 8-12) were knowledgeable about water supply and access-related issues.According to Tzanakakis, Paranychianakis, and Angelakis (2020), lack of education on water supply, access to water resources as well as increased water demand in communities contribute to low rates of sustainability of water supply systems.The employment status of the respondents as shown in Figure 4 revealed that 42% were unemployed while of the 58% employed 18% were self-employed.The selfemployed respondents mostly engaged in crafts, casual work and selling of primarily produce with a monthly income showing a high disparity ranging from as little as ZAR450 to ZAR3 500.Due to the rural nature of the study area, about 60% of the employed respondents were working in the surrounding farms with an income of under R5 000.With farm work generally being a seasonal employment opportunity, it was noted that the employment rate is likely to increase during the offseason; thus, affecting the respondent's ability to earn an income.The unemployed respondent relied mostly on social grants for income.Due to the economic requirements to set up appropriate RWH systems, employment and income play a decisive role.This affects the household's access to multiple water sources i.e. boreholes and the ability to set up a harvesting system.This is important because RWH is increasingly being adopted to reduce water poverty in rural drought-prone areas to enhance water security, for future use to meet the demands of human consumption and other household activities.
During the survey, it was found that 53% of respondents were semi-skilled, 30% were unskilled whilst it was found that only 17% were skilled as shown in Figure 5.A RWH system can be set up relatively easily with local skills and resources, using simple and easyto-maintain technologies that are inexpensive.However, despite the many benefits, households often find it difficult to implement RWH.The initial installation costs of RWH systems in residential housing are a common barrier to adoption (Lee, Mokhtar, Hanafiah, Halim, & Badusah, 2016).Households with low levels of skills are at a higher risk of non-adoption due to their inability to set up their systems.This will in turn force households to incur costs in setting up a practical system.Factors such as lack of technical skills, income, cost-benefit, and perceived payback have the potential to render the installation of RWH systems uneconomical for many households given its low immediate return on investment (Amos, Rahman, & Gathenya, 2016;Cardos, Blanco, & Duarte, 2020;Lee, Mokhtar, Hanafiah, Halim, & Badusah, 2016).

Comparison between household's characteristics: Adoption and non-adoption of RWH
The analysis of sample households' characteristics as shown in Table 1 depicts a total of 63.8% of households adopted RWH practices whilst non-adopters were 36.2% of the sample population.The RWH adoption rate in the area is higher than in similar studies within rural communities.Mekuria, Kassegn, and Mekonnen (2020) reported an adoption rate of 53.06% in Kutaber district, South Wollo zone in Ethiopia.Similarly, Munyaneza, Majoro, Hagenimana, and Usabyisa (2016) reported a 57.5% adoption of RWH in a study in Ntarama, Rwanda.The uptake of RWH practices appears to be significantly impacted by various factors of the households in all the reported studies.
A comparison of demographic characteristics between adopters and non-adopters reveals that gender plays a significant role in harvesting water, 64.24% of those who adopted RWH practices were female as compared to 35.42% of males as shown in Table 1.The same is true for non-adopters of RWH practices, female household heads represent a third (33.53%) of those who do not practice RWH.Several studies (Wubetu, 2016;Mekonnen, 2017;Shikur & Beshah, 2013) have reported that gender and level of education have a positive influence on the adoption of RWH.Thus, higher levels of educational attainment are more likely to adopt or practice RWH techniques compared to less educated farmers (Chianu  & Tsujii, 2004).The gender gap proportion in the adoption of RWH technique practices is not unexpected in most rural settings, given the population dynamics in rural areas.Females play an important role regarding water use at home, and thus, are generally responsible for household upkeep and have the responsibility for collecting water (Hutton et al., 2017).
In rural areas such as the study area without reliable water supply, households are forced to seek and use alternative sources.While the household socioeconomic characteristics are important in determining the adoption of RWH, the findings showed 37% of nonadopters relied on groundwater (boreholes) as their main source of water.Further to this, 34% had access to water through a water reticulation system (piped water supply).The perceived reliability of these sources contributes to the lack of adoption of RWH, this is despite its many advantages.
The mean age of the sampled household heads was 45 years with a maximum of 70 and a minimum of 20 years as shown in Table 2.The age profile within those that adopted RWH practices shows a fair distribution with the group aged 46-60 being the most active among the total sampled population of adopters at 39%.Followed by those aged between 31-45 and over 61 at about 29.18% and 21.31%, respectively.Around 80% of those that did not practice any RWH were aged 31-60.Interestingly, on further probing, it was found that this was related more to their beliefs and perception of the quality of the harvested water.Age has a considerable negative effect on the adoption of RWH, i.e. older household heads are generally less inclined to adopt new technologies and practices.This consistency is expected and is in line with the theory of human capital; young members of a household have a greater chance of absorbing and applying new knowledge (Baiyegunhi, 2015).
The family structure in households that responded to the questionnaire survey was characterized by large family sizes with an average of between four males and five females per family.Most respondents in the sample were married (56%), while 39% were not married.The average family size is made of seven individual members.About 94% of households have between one to ten family members including children.In most rural areas, large family size is not uncommon due to family structures and poverty.Overall, most members of the households had low academic levels, over 68% of the respondents had secondary, primary, or no formal schooling.Thus, their household income is influenced by lower levels of education and available jobs.The average household monthly income was R920 (as can be seen in Table 2 above) with most households having no access to formal credit facilities as they cannot offer the financial security required by the financial institutions.This is significant because the level of education and skills have been found to directly determine the ability and means (economic) to set up an effective RWH system.

Rainwater harvesting practices
The average water consumption in South Africa is 237 litres/person/day (l/c/d), which is higher than the world average of about 173 l/c/d (Ngobeni & Breitenbach, 2021).However, for households in semi-urban and rural areas who mostly access water services through standpipes, consumption can be as low as 22 l/c/d, regardless of the household size, which is well below the minimum volume of 25 l/c/d recommended by the government through the Free Basic Water Implementation Strategy (2007) for rural populations.Conversely, rural communities with no access to water services to meet their basic needs are obligated to consider alternative sources of water to meet their domestic need.These include the costs, sustainability, and ease of access to these sources.These households are forced to conserve whatever available water and change their consumption patterns for sustenance (Loubser, Chimbanga, & Jacobs, 2021;Oki & Quiocho, 2020).RWH is recognized as one alternative to improve water supply in communities lacking reliable bulk supply systems provided by the service providers (Machado, Oliveira, & Matos, 2022).About 64% of the surveyed households particularly in areas where access to and availability of alternative water sources is problematic practiced RWH whilst 36% did not practice any form of RWH, despite the water challenges faced in the study area.Non-adoption of RWH was driven by the communities' perception and attitude toward harvested water regarding the quality of the water for human consumption.According to Coetzee, Nell, and Bezuidenhout (2016) and Doria (2010), socioeconomic factors have been shown to influence public perceptions of drinking water.Sixty-seven percent of the respondents argued that the water was not fit for consumption without further treatment and thus used it mainly for non-potable domestic purposes including laundry, cleaning the floor, washing vehicles, and watering plants.Further to this, the respondent's perception of water quality only relates to the presence of foreign matter (debris), turbidity and color of the water.Of interest is that the respondents showed a lack of awareness of the difference between "perceptional quality" and 'absolute quality.This is in line with the findings by Anang and Asante (2020) and Dobrowksy et al. (2014) who reported that RWH was mainly used for washing clothes and for cleaning inside and outside the houses due to the perceived inferior quality of the water.
According to Hamilton et al. (2019) and Nnaji and Nnam (2019), well-designed RWH systems provide relatively safe water with good physicochemical quality whilst their primary concern of microbial contamination can be alleviated by proper design and cleaning.Thus, respondents' perception points to a general lack of acceptance, knowledge and understanding of RWH as a solution for water scarcity and its usefulness to provide portable water.According to FAO (2017), due to challenges such as climate change and population increases, it is expected that by 2050, water withdrawal for domestic will increase by 15% around the world.RWH belongs to the promising practices to support sustainable development in sub-Saharan Africa facing climate change (Vohland & Barry, 2009).While the result showed RWH is practised to an extent in the study area, most households have no means to set up and install an effective RWH system i.e. catchment area (e.g. a rooftop), a conveyance system (e.g.gutters and pipes), and a storage facility (e.g.tanks).These households' practices are very rudimentary in nature and opportunistic.In certain instances, water is harvested off the rooftops directly into receiving small containers, usually a bucket or 100-liter drum barrel or any available storage containers as shown in Figure 6.
It was also found that due to the financial costs associated with a well-designed system, the initial costs of RWH systems are prohibitive for most individual households.Therefore, only 24% of the surveyed households had the financial means to install a well-designed system including gutters, downpipes, and a storage facility.The general design of the roof-harvested rainwater system in the study area comprises the following components: roof catchment, gutters, downpipe, and storage tank.This general design common in the study area is similar to that recommended by Mwamila, Han, and Kum (2016) for harvesting rainwater as shown in Figure 7. Thus, most households with intentions to harvest water have a basic system, some use a single gutter without the necessary joints or downpipes, and in other cases, some use old metal sheets folded into a semi-circular structure, then hastily attached, or fixed to their roofs as means of gutters and one or more barrels placed at the gutter end on the ground to collect rainwater.In both instances, there is a high chance of water loss due to wind, contamination, and insufficient storage of the harvested water.While harvested rainwater is both a key domestic and productive resource, its success mostly depends on the system in use.Accordingly, Mwamila, Han, and Kum (2016) argue that low-quality design and poor construction can contribute to several RWH deficiencies.

Empirical results of the logit model
A set of explanatory variables as shown in Table 3 were designed to define the main socio-economic characteristics of the local community.The study employed 12 variables to describe how household socio-economic characteristics influence and affect the adoption or nonadoption of RWH.

Gender
The gender of the respondents had a statistically significant positive influence on the adoption of RWH as a source of alternative water in the households.The findings show that the probability of female respondents adopting RWH is three times higher than their male counterparts.The finding fits the expected outcome due to the higher representation of females the rural settings, whereby women generally remain in the home and are generally responsible for managing water at the household level.Studies such as Doss and Morris (2001) and Gilbert, Sakala, and Benson (2002) have reported on the gender disparities among respondents in studies of the adoption of new technologies.Furthermore, the results show a continuing trend between RWH adoption and gender, there is a higher rate of female-headed households found in non-adopters.This finding supports the widely held assumption that adoption may be higher among females because of their proximity to the practical  daily burden of water collection and management (Hutton & Chase, 2016;Olajuyigbe, 2010;Howard, Bartram, Water, & World Health Organization, 2003).

Age
This study found that age has a statistically significant negative influence on and is strongly related to the probability of adopting RWH practices at a 5% level of significance.Thus, for the older respondent, the probability of adoption increased at a ratio of 0.778.Older household heads are observed to be risk-averse and reluctant to adopt alternative water management practices that require skills, labor and finances (Mekuria, Kassegn, & Mekonnen, 2020).This is true, especially in areas where there is no support or subsidies to support the initial setup of a RWH system.Furthermore, the descriptive statistics revealed a significant difference in the mean age of RWH adopters (48 years) compared to RWH nonadopters (39), thus younger household heads are less likely to adopt RWH, this can be attributed to access to finance, the necessity of alternative source of water sources.According to Mekuria, Kassegn, and Mekonnen (2020), the age of the household head was reported to negatively influence the adoption of RWH technology whilst Murwirapachena (2021) has reported that the age of respondents is a statistically significant variable influencing the adoption of RWH.

Level of education
The level of educational attainment is positively linked with an increased probability for the adoption of RWH as expected.Thus, household heads who have attained a level of education are likely to adopt RWH practices as compared to those with no or less educational attainment.The probability of educated respondents adopting RWH is 1:01.The higher probability is due to the ability to acquire information such as climate science, waterrelated challenges in their community and the potential benefit of adopting RWH practices.Higher education attainment has been shown by Ahmed, Onwonga, Mburu, and Elhadi (2013) to increase the probability of adoption of RWH at 1% significance.This finding is in line with those of Chianu and Tsujii (2004), and Ahmed, Onwonga, Mburu, and Elhadi (2013).Interestingly, the findings of this study contradict previous findings by Baguma and Loiskandl (2010) and Staddon, Rogers, Warriner, Ward, and Powell (2018) who reported no link between the level of educational attainment and adoption of RWH in rural and central Uganda.The authors found that subsidy provision was statistically significant for the adoption of RWH technologies in rural Uganda.In the latter study, Staddon, Rogers, Warriner, Ward, and Powell (2018) noted the potential participation of NGOs had the potential to render formal education attainment less significant concerning RWH adoption due to the ability of such organizations to expose households of varied education levels to this information, mobilizing.Contra to this study, it was found that only the government was involved in building capacity for RWH by providing storage tanks in the Giyani District Municipality with minimal NGO or other stakeholder participation.

Perceptions
The odds ratio of the respondent's perception was found to be statistically significant at a 5% significance level.The respondent's negative perception regarding RWH was found to be generally associated with the quality of the water, social status, and the cost of the initial construction, however, in certain instances, traditional and cultural beliefs are found to be exerting considerable influence.The results also revealed that the households' perception of water quality has the potential to lead to water insecurity and over-reliance on unreliable sources to meet their domestic needs.Thus, it would be expected that households with a positive perception and attitude to RWH are likely to adopt the practice.In similar studies, Sheik, Rehman, and Yates ( 2003) and D 'Emden, Llewellyn, and Burton (2008) showed that household perceptions do influence their adoption decisions to RWH practices.The results show that households with a positive perception of RWH are four times more likely to adopt RWH than a respondent with a negative perception.Thus, perception significantly influences the respondent's willingness to adopt RWH.

Household size
The results revealed a positive relationship between household size and the adoption of RWH.The positive effect of the probability of adopting RWH was found to be statistically significant at 5%.Household size is often considered a proxy for labor availability.This means households with large family sizes are more likely to adopt water harvesting since they would have sufficient family labor for RWH activities, more so if the practice is rudimentary.With the increase in household size, there is a need to increase or secure a sustainable alternative water source.Larger households are expected to be more engaged in securing water than their smaller counterparts.This is in line with findings by Mekonnen (2017) who reported that family size positively influences the adoption of RWH practices.

Household income
Due to the nature of the sample population, it was expected that household income would have a statistically significant positive effect on the adoption of RWH.Households with higher family incomes are more likely to adopt RWH.The odds ratio for income is 3.08, the higher the household income, the more increased the probability of adoption.This might be a result of the appetite for risk-aversion and waterproofing their homes.A higher level of household income implies a greater incentive for investment in technology and the ability to bear the risk associated with its adoption.Such variables may also influence expected usefulness indirectly, through a relationship with risk.The findings suggest that lower-income non-adopters of RWH practices are less water-secure.

Access to information
Access to information on RWH practices and demonstrations has a statistically positive effect on the adoption of RWH practice.The odds ratio for the adoption of RWH is increased by 1.67 with access to information.A household with perceived knowledge about the requirements (either financially or physically) and benefits of RWH is more likely to adopt the practice.Overall, households with access to information are better placed to make decisions regarding the adoption of RWH (Mtyelwa, Yusuf, & Popoola, 2022).

Conclusion
Given the concern of water security in rural communities as a result of a changing climate, alternative water sources assessment is of significance to ensure water availability for domestic use.This study investigated the socioeconomic factors and other determinants of RWH practices in rural communities of Limpopo Province in South Africa.The findings show that a significant number of households practised RWH to augment available water for domestic use, albeit few households used the harvested water for portable use.Due to the nature of the practices, it was found that the reliability of the systems for RWH was low.Overall, the study has established that RWH plays an important role in rural communities where they are being practised.The communities' perception regarding the quality of rainwater is a hindrance toward the full use of rainwater.Perception was found to be statistically significant in determining the adoption of RWH.While studies have shown that with a welldesigned system, the water quality is generally of an acceptable standard and despite the benefits associated with RWH, communities are still reluctant.Additionally, factors such as household income, level of education, income, and household size were statistically significant determinants of RWH adoption.The findings of the study show pathways in understanding the adoption of RWH in rural communities and highlight the vital role of education, knowledge, and income in rural settlements with limited or no access to a water supply.There is therefore a need to educate rural communities on the benefits of harvesting rainwater for domestic use.Thereby ensuring water security for the most vulnerable and marginalized communities in developing nations.It is also recognized that RWH is a seasonal practice, possible mainly during October to March in the study region.
Other options for water access are necessary during the half-year dry season, whilst also considering projected climate change impacts on surface and groundwater resources.

Figure 1 .
Figure 1.Map of the study area showing Capricorn and Mopani District Municipalities.

Figure 2 .
Figure 2. The age of respondents in the study area.

Figure 3 .
Figure 3.The educational level of respondents in the study area.

Figure 4 .
Figure 4. Employment status of the respondents.

Figure 5 .
Figure 5. Technical skills of the respondents in the study area.

Table 1 .
Descriptive statistics comparison between RWH adopters and non-adopters.

Table 2 .
Explanatory variables used in the analysis and empirical model statistics.

Table 3 .
Empirical results of the logistic model for factors influencing RWH adoption.
WATER SCIENCE