The study was conducted in Phalombe District under Traditional Authority (TA) Nkhulambe, situated in Southern Malawi (Figure 1). Phalombe is a predominantly rural district characterized by subsistence agriculture, which serves as the primary source of livelihood for the majority of households. The area lies within Malawi’s semi-arid zone, receiving an average annual rainfall of approximately 800–1,000 mm, largely concentrated between November and April (Fiwa et al., 2025). Soils in the district are predominantly sandy loams with moderate fertility, making crop production highly sensitive to input availability and management practices (IUSS Working Group WRB, 2022). Most households in the study area cultivate relatively small parcels of land, typically less than one hectare, and depend predominantly on maize as the principal staple crop. Maize production is often complemented by legumes such as beans and groundnuts, as well as a range of horticultural crops that contribute to household nutrition and provide opportunities for income diversification. Livestock production remains limited in scale, with chickens and goats representing the most common species kept under mixed crop–livestock systems. These animals play an important role in nutrient recycling, income generation, and food security. In recent years, Phalombe has become a focal area for agricultural innovation and adaptive research, particularly in sustainable soil fertility management. Development initiatives and research programs have introduced and promoted the use of organic and locally available fertilizers, including urine-based formulations derived from ecological sanitation (Ecosan) systems. These interventions aim to address declining soil fertility, high costs of synthetic fertilizers, and the need for environmentally friendly alternatives suitable for resource-constrained smallholder farmers. The district thus provides a representative and dynamic setting for examining the socio-economic and institutional factors influencing the adoption of liquid urine fertilizer within smallholder farming systems.

A cross-sectional survey design was employed to examine household-level decisions regarding awareness, adoption, and intensity of urine-based fertilizer use. This design was appropriate because it enabled the collection of quantitative data capturing variations in technology exposure and behavioral outcomes among households during the main agricultural season, which in Phalombe District typically spans from November to April, corresponding with Malawi’s unimodal rainy season that supports the country’s primary rain-fed crop production system (Amosi and Anyah, 2024). The study was conducted among smallholder farmers under Traditional Authority (TA) Nkhulambe in Phalombe District, Southern Malawi. The sampling frame consisted of all households participating in organized farmer groups such as cooperatives, farmer field schools, and village savings and loan associations. These groups provided a well-defined list of active farmers and ensured representativeness of households engaged in community-based agricultural activities.

A stratified random sampling technique was adopted to capture variation across villages while maintaining proportional representation. Villages were first stratified based on population size and accessibility (proximal vs. remote), after which households were randomly selected from each stratum. Respondents were either household heads or principal decision-makers involved in agricultural production, ensuring reliability in responses related to farming practices and input decisions. The total sample size was determined using Cochran (1977) sample size formula for categorical data:

where.

n 0 = initial sample size,

Z = z-value corresponding to the desired confidence level (1.96 for 95% confidence),

p = estimated proportion of households with awareness of urine fertilizer (assumed 0.5 in the absence of prior data to maximize variability), and.

e = desired level of precision (0.06).

Substituting these values:

To account for possible non-responses and incomplete questionnaires, a 10% adjustment was applied, yielding an effective target of approximately 270 households. The final realized sample comprised 251 completed interviews, representing a response rate of 93%, which is acceptable for field-based household surveys in rural contexts. This sample size ensured adequate statistical power for estimating parameters within the triple-hurdle econometric framework and allowed robust disaggregation by socio-economic and institutional variables.

Primary data was collected through structured, face-to-face interviews administered by a team of trained enumerators. A comprehensive questionnaire was designed to capture detailed information on household demographic and socio-economic characteristics, land and livestock ownership, access to agricultural extension services, and farmers’ perceptions of organic and urine-based fertilizers. In addition, the instrument included modules on awareness, adoption history, intensity of use, and perceived constraints associated with urine fertilizer application. To ensure validity and reliability, the questionnaire was pre-tested in a neighboring community with similar socio-economic and agro-ecological characteristics. The pilot exercise allowed for refinement of question wording, sequencing, and flow, ensuring clarity, cultural appropriateness, and consistency in interpretation across respondents. Enumerators received prior training on research ethics, interviewing techniques, and the technical aspects of urine fertilizer use to minimize interviewer bias and enhance data accuracy. The finalized instrument was then deployed at the household level, with each interview conducted in the respondent’s preferred local language to facilitate comprehension and authenticity of responses.

The analysis focused on three sequential dependent variables that reflect distinct stages in the adoption process of urine fertilizer: awareness, adoption, and intensity of use. Each stage captures a progressively deeper level of engagement with technology and provides a more nuanced understanding of farmers’ decision-making behavior.

Understanding the adoption of agricultural innovations often requires disentangling multiple stages of decision-making, as farmers may differ not only in whether they adopt a technology but also in how intensively they use it. To account for these sequential decisions, this study employed the Triple-Hurdle Model (THM), an econometric framework that extends the conventional double-hurdle specification (Cragg, 1971) by incorporating an additional layer of behavioral complexity. This approach recognizes that awareness, adoption, and intensity of use represent distinct but interrelated processes influenced by different sets of factors.

The first hurdle models awareness, which determines whether a farmer has ever heard of or received information about urine fertilizer. This stage was estimated using a Probit model, where the probability of awareness is expressed as a function of demographic, socio-economic, institutional, and perceptual variables. The second hurdle captures adoption conditional on awareness, reflecting the decision to apply urine fertilizer in practice. This stage was also estimated using a Probit model but included the Inverse Mills Ratio (IMR) derived from the awareness equation to correct for potential selection bias arising from unobserved factors influencing both awareness and adoption decisions.

The third hurdle focuses on intensity of use, which quantifies the frequency or number of applications of urine fertilizer among adopters. Since this variable is censored at zero for non-users, it was estimated using a Tobit model, appropriate for continuous but limited dependent variables. Together, these three equations form a sequential structure that captures the progression from information exposure to behavioral adoption and sustained utilization. Mathematically, the framework can be summarized as follows:

where A i = 1 if A i ∗ > 0 , and 0 otherwise.

where D i = 1 if D i ∗ > 0 , and 0 otherwise.

observed as Y i = Y i ∗ if Y i ∗ > 0 , and 0 otherwise.

Here, X i represents the vector of explanatory variables as defined in Section 2.4; β 1 , β 2 , β 3 denote parameter vectors; ϵ 1 i , ϵ 2 i , ϵ 3 i are independently distributed error terms; and λ i is the Inverse Mills Ratio accounting for sample selection effects.

The THM was chosen for its ability to provide richer behavioral insights compared to single-equation or binary models. It distinguishes between farmers who are unaware of the technology, those who are aware but choose not to adopt, and those who adopt with varying degrees of intensity. This separation not only enhances model accuracy but also allows for more targeted policy implications, since the factors promoting awareness may differ substantially from those influencing actual adoption or intensity of use.

All estimations were conducted using maximum likelihood estimation (MLE) techniques. Model diagnostics included tests for multicollinearity using Variance Inflation Factors (VIF), heteroskedasticity using Breusch–Pagan tests, and significance of the IMR to confirm the presence of selection bias between stages. Robust standard errors were applied to ensure valid inference in the presence of potential heteroskedasticity.

In sum, this analytical framework enables a comprehensive understanding of how socio-economic, institutional, and perceptual factors interact to shape the awareness, adoption, and intensity of urine fertilizer use among smallholder farmers—providing a rigorous foundation for the policy recommendations advanced in later sections.

The estimation followed a sequential procedure consistent with the structure of the Triple-Hurdle Model (THM). Each stage—awareness, adoption, and intensity—was estimated using econometric techniques appropriate for the nature of the dependent variable and its position within the decision process. The model was implemented using maximum likelihood estimation (MLE), allowing simultaneous evaluation of parameter estimates while ensuring statistical efficiency. In the first hurdle, the probability of being aware of urine fertilizer was estimated using a Probit model, where the dependent variable took the value of 1 if the respondent had heard of urine fertilizer and 0 otherwise. The estimation identified factors that increase the likelihood of information exposure, reflecting access to knowledge and institutional linkages. The second hurdle analyzed the determinants of adoption among farmers who were already aware of urine fertilizer. A Probit model with an Inverse Mills Ratio (IMR) correction was employed to account for potential sample selection bias, recognizing that unobserved factors influencing awareness might also affect the adoption decision. A significant IMR term indicated the presence of such bias, justifying the sequential estimation structure. The third hurdle, representing the intensity of use, was estimated using a Tobit regression model because the dependent variable (number of urine fertilizer applications per year) was continuous but censored at zero for non-users. This specification captures variation among adopters while correctly treating non-users as censored observations rather than true zeros. The model thus provides consistent and unbiased estimates of the determinants of usage intensity.

To ensure robustness and reliability of results, a series of diagnostic and validation tests were conducted:

The estimation strategy therefore ensured that each behavioral stage—awareness, adoption, and intensity—was empirically validated and statistically robust. This sequential approach not only enhanced explanatory power but also enabled the derivation of nuanced insights into the distinct socio-economic, institutional, and perceptual factors influencing each stage of urine fertilizer adoption among smallholder farmers.

A total of 251 smallholder farming households were surveyed across villages under Traditional Authority (TA) Nkhulambe in Phalombe District, Southern Malawi. Table 1 summarizes the key demographic, socio-economic, and institutional characteristics of the sample population. The results depict a community characterized by modest resource endowments, limited farm sizes, and strong participation in organized farmer groups. The average age of household heads was approximately 41.7 years (SD = 12.4), indicating that most respondents were within the productive age bracket actively engaged in farming. The gender distribution was slightly male dominated, with 61% of respondents being men. Educational attainment averaged 7.1 years of formal schooling (SD = 3.9), reflecting a moderate literacy level conducive to participation in agricultural extension and innovation programs.

Households were relatively large, with an average size of 5.4 members (SD = 2.2), and operated small landholdings averaging 1.9 acres (SD = 1.5). These figures are typical of Malawi’s smallholder farming systems, where land scarcity and population density constrain production potential. The mean monthly household income was approximately MK 47,800 (SD = 28,600), signifying a predominantly low-income, subsistence-based economy. Livestock ownership was common but limited in scale. Respondents reported an average of 6.8 chickens and 4.1 goats per household, indicating that small livestock serve both as a financial buffer and a potential source of organic nutrient recycling. Importantly, all respondents belonged to at least one farmer organization, such as cooperatives, farmer field schools, or savings and loan associations, highlighting the district’s strong culture of collective action. Additionally, 52% of farmers reported having interacted with an agricultural extension officer at least once in the preceding 12 months, underscoring moderate but uneven institutional reach.

Regarding urine fertilizer, 46% of farmers were aware of its potential agricultural use, while 32% had applied it on their fields. Among adopters, the average number of applications in the previous year was 4.7 times (SD = 3.1). These figures suggest that although awareness levels are growing, actual uptake remains limited, and intensity of use is modest. The gap between awareness and adoption implies the presence of behavioral, economic, or perceptual barriers—issues that are further explored through econometric analysis in subsequent sections.

Collectively, these descriptive statistics portray a farming community characterized by low asset accumulation, strong institutional embeddedness, and growing exposure to innovative soil fertility management practices. The heterogeneity in education, extension contact, and livestock ownership provides a useful empirical foundation for explaining the variation in awareness, adoption, and intensity of urine fertilizer use examined through the triple-hurdle model.

Results from the first stage of the triple-hurdle model (Table 2) show that education, extension contact, and poultry ownership were significant positive predictors of awareness of urine fertilizer among smallholder farmers. In contrast, age and odour concerns were negatively associated with awareness. Each additional year of education increased the odds of awareness by 12% (p < 0.01), while each unit increase in extension contact raised the odds by 27% (p < 0.01). Ownership of chickens also had a positive effect on awareness, with the odds increasing by 4% per additional bird (p < 0.05). Conversely, age was associated with a slight reduction in awareness (odds ratio = 0.98; p < 0.05), and farmers who expressed odour concerns were 44% less likely to be aware of urine fertilizer (p < 0.05). These findings indicate that awareness is influenced primarily by educational attainment, access to extension services, and household resource endowment, while perceptual and demographic factors contribute to variation in information exposure among farmers.

The second stage of the triple-hurdle model (Table 3), estimated conditionally on awareness, revealed that extension contact and farm size were significant positive predictors of urine fertilizer adoption, whereas cost concerns and odour concerns exerted significant negative effects. Farmers with more frequent extension contact had 41% higher odds of adopting urine fertilizer (p < 0.01). Each additional acre of farmland increased the likelihood of adoption by 21% (p < 0.05). In contrast, farmers who perceived urine fertilizer as costly were 53% less likely to adopt (p < 0.05), and those expressing odour concerns were 62% less likely to adopt (p < 0.01). The Inverse Mills Ratio (IMR) derived from the awareness model was negative and statistically significant (coefficient = −1.08, p = 0.02), confirming the presence of selection bias between the awareness and adoption stages and reinforcing the appropriateness of the triple-hurdle framework for this analysis.

Analysis of the final stage of the triple-hurdle model (Table 4) shows that the frequency of urine fertilizer application among adopters was strongly influenced by institutional contact, farm resource capacity, and perception-related factors. The coefficients indicate that extension contact, farm size, and poultry ownership were positively associated with intensity of use, whereas age and odour concerns exhibited negative effects. An additional extension interaction corresponded to an average increase of 0.9 urine fertilizer applications per year (p < 0.01). Similarly, a one-acre increase in farm size raised application frequency by approximately 1.1 times annually (p < 0.05). Poultry ownership contributed positively, with each additional chicken linked to a 0.1 increase in application frequency (p < 0.05). In contrast, older farmers tended to apply the fertilizer less frequently (coefficient = −0.05; p < 0.05), and those reporting odour concerns applied it 1.4 times fewer per year (p < 0.05). Overall, the intensity analysis confirms that institutional engagement and on-farm resource availability enhance the sustained use of urine fertilizer, while perceptional constraints and demographic factors continue to moderate utilization patterns among adopters.

Consistent with adoption theory, education and access to extension services significantly improved farmers’ awareness and subsequent adoption of urine fertilizer. Farmers with more years of formal schooling were better positioned to process technical information and perceive the agronomic value of alternative soil fertility technologies (Tapsoba et al., 2020; Martignoni et al., 2022). The low adoption of urine fertilizer could be linked to the observed moderate literacy level of an average 7.1 years of academic exposure by the population. Similarly, frequent extension contact increased the odds of awareness by 27% and adoption by 41%, aligning with earlier studies that emphasize extension as a critical vehicle for technology dissemination in low-input agricultural systems (Smith et al., 2022; Petersen-Rockney et al., 2021). These findings reinforce the argument that social learning and demonstration-based communication are essential to overcome behavioral inertia and misconceptions surrounding the use of human-derived agricultural inputs (Verhagen and Scott, 2019). According to the social exchange theory (Thibaut and Kelley, 1959), an impactful change in agricultural yield guided by extension personnel will fundamentally influence the social behavior of the community by embracing the proposed technology as they weigh the personal benefits against the costs. This will enhance sustainable growth in agricultural production and subsequently improve livelihood of the population. In observation of male dominance within the community at 61%, it would be recommendable to ensure inclusivity in technology awareness drives to reduce possible adoption bottlenecks.

Farm size and livestock ownership were positively associated with adoption and intensity of urine fertilizer use. Larger farms imply greater land heterogeneity and the need for nutrient management innovations (Dimkpa et al., 2023). Ownership of chickens, which serves as an indicator of household asset wealth and capacity for nutrient recycling, was also associated with a higher intensity of urine fertilizer use. This relationship suggests that households with greater livestock assets are better positioned to integrate organic nutrient sources into their farming systems, thereby enhancing the efficiency and sustainability of resource use. This finding corroborates the emerging understanding that integrated crop-livestock systems enhance the uptake of circular bioeconomy technologies, as nutrient flow between livestock and crop systems strengthens the feasibility and regularity of urine-based fertilization (Mathobela et al., 2024). Farmers with broader productive assets may therefore view urine fertilizer as complementary rather than substitutive to existing organic manure practices.

Despite the clear agronomic and economic advantages of urine fertilizer, cultural and sensory perceptions, particularly concerns related to odour, were found to significantly diminish awareness, adoption, and intensity of use. Similar patterns have been observed in other ecological sanitation settings, where psychological discomfort and deeply rooted cultural taboos have constrained public acceptance and hindered widespread utilization (Nkusi, 2018; Mariwah, 2017). These findings highlight the critical importance of addressing perceptual and cultural barriers through targeted education, sensitization, and participatory engagement to enhance the social acceptability of ecological sanitation technologies. Odour concerns lowered the probability of awareness by nearly half and reduced adoption odds by 62%. This indicates the importance of perception management and knowledge co-creation through participatory demonstrations that normalize human-waste recycling practices (Leblanc et al., 2019). Furthermore, these findings resonate with the broader discourse on the social construction of technology, where cultural narratives shape perceived risk and appropriateness of innovations (Nhamo et al., 2024). Addressing these barriers requires not only technical optimization to minimize odour but also deliberate social marketing that reframes urine fertilizer as a symbol of ecological responsibility and economic pragmatism.

Although source separated urine is generally considered to pose lower microbiological risk than fecal matter, its relative safety cannot be assumed under real world smallholder farming conditions. Recent studies emphasize that the risk profile of urine derived fertilizers is highly contingent on system design, user compliance, and post collection handling practices, all of which are prone to variability in low resource settings (Roxburgh et al., 2020; Yalin et al., 2023). Cross contamination between urine and fecal matter may occur through improper toilet use, malfunctioning or poorly maintained urine diversion infrastructure, shared collection containers, and unhygienic handling during storage, transportation, or field application (Qadir et al., 2020; Fatta-Kassinos et al., 2023). In the present study, no empirical data were collected on urine treatment methods, storage duration, pH stabilization, or adherence to hygienic handling protocols. Consequently, the analysis does not evaluate compliance with World Health Organization and Food and Agriculture Organization guidelines for the safe reuse of human derived nutrients (WHO, 2022; FAO, 2023).

While urine diversion offers a promising pathway for safe nutrient recovery and aligns with circular bioeconomy principles, its integration into agricultural systems must be explicitly guided by a precautionary risk–benefit framework. Such an approach balances agronomic benefits with necessary public health and environmental safeguards (Yalin et al., 2023; WHO, 2022). The promotion of urine-derived fertilizers should therefore not proceed in isolation but must be accompanied by rigorous and context-specific risk assessment protocols. Potential trade-offs that require careful management include pathogen transmission, residual pharmaceuticals, antimicrobial resistance genes, and heavy metal accumulation in soils and food chains. These risks are particularly pronounced in systems characterized by inadequate treatment, inconsistent user compliance, or limited monitoring infrastructure (Fatta-Kassinos et al., 2023; Hamilton et al., 2020; Yalin et al., 2023). Consequently, efforts to scale urine fertilizer use must be coupled with enforceable safety standards, participatory farmer training on safe handling and storage, and periodic biosafety audits. These measures are essential to mitigate unintended consequences and maintain public trust (FAO, 2023; Qadir et al., 2020). Adopting this integrated perspective ensures that the pursuit of nutrient recycling and Sustainable Development Goal (SDG) attainment (2, 6, and 13) does not compromise food safety or ecosystem health. In doing so, it aligns circular sanitation-agriculture innovations with the broader objectives of sustainable agriculture and One Health.

The absence of standardized treatment and monitoring information represents an important limitation and warrants careful interpretation of the findings. While observed adoption patterns provide valuable insights into social acceptance, perceived utility, and institutional drivers of uptake, they do not constitute evidence of biosafety compliance or effective risk mitigation. This distinction is critical, as inadequate treatment or improper handling of urine-based fertilizers may compromise soil safety, crop quality, and public health, even in systems designed around urine diversion principles (Hamilton et al., 2020; Yalin et al., 2023). These findings reinforce the need for future research and policy frameworks to integrate socio economic adoption analysis with rigorous biosafety assessment to ensure that the promotion of urine reuse aligns with food safety standards and One Health objectives.

A unique finding is the universal membership of respondents in organized farmer groups, suggesting a strong institutional foundation for scaling innovations. However, this did not automatically translate into higher adoption levels, implying that collective structures alone are insufficient without effective facilitation and knowledge brokerage. Group membership may enhance information exposure but not necessarily influence behavioral change unless supported by continuous mentorship and resource mobilization (Verhagen and Scott, 2019; Klassen and Vreysen, 2020). This observation reinforces the view that successful dissemination of sustainable technologies depends on functional institutions that bridge scientific knowledge and local practice, enabling experimentation, peer influence, and adaptive learning (Cookey et al., 2022).

The negative and significant Inverse Mills Ratio (IMR) indicates an inverse relationship between awareness and adoption, suggesting that the unobserved factors influencing awareness, such as exposure through external projects or promotional campaigns, do not automatically translate into actual behavioral change. This finding shows that awareness alone is insufficient to drive adoption unless it is accompanied by clear economic incentives, practical feasibility, and technical support mechanisms that enable farmers to implement the innovation effectively. Cost perception was a major barrier, reducing adoption by 53%, even though urine fertilizer is materially low-cost. This apparent contradiction suggests that the cost concern is likely a proxy for transaction costs, such as labour for collection, storage infrastructure, and transportation (Qadir et al., 2020). Policies should thus focus on reducing these indirect costs through local infrastructure support and communal urine collection systems integrated into sanitation services. Moreover, the alignment of this initiative with Malawi’s climate-smart and circular economy strategies can catalyze funding and institutional backing for upscaling (Nhamo et al., 2024).

This study provides empirical evidence supporting the dual benefits of integrating ecological sanitation and sustainable agriculture. By closing nutrient loops, urine fertilizer contributes to both environmental protection and food security, echoing global calls for circular resource use (Yalin et al., 2023; Qadir et al., 2020). The findings affirm that ecological sanitation can move beyond household-level waste management toward a systemic strategy for nutrient recovery and reuse, aligning with Sustainable Development Goals (SDGs) 2, 6, and 13. For Malawi and similar agrarian economies, urine fertilizer adoption can contribute to nitrogen self-sufficiency, reduce import dependency, and promote climate adaptation through soil organic matter improvement.

While the triple hurdle model employed in this study provides robust insights into the sequential nature of awareness, adoption, and intensity of urine fertilizer use, future research should extend this work through longitudinal and experimental designs to track behavioral change over time and establish causal relationships. Integrating biophysical assessments such as soil fertility indicators, crop yield responses, and economic returns would further strengthen understanding of the long term agronomic and livelihood impacts of urine-based fertilization. Equally important is a more explicit examination of gendered dynamics, particularly the roles and responsibilities of women in sanitation management, nutrient recovery, and fertilizer application, as these dimensions are central to the effectiveness and sustainability of ecological sanitation initiatives (Martignoni et al., 2022). In addition, greater attention should be paid to cultural norms, beliefs, and taboos, which continue to shape risk perceptions and strongly influence farmers’ willingness to adopt urine derived fertilizers.

A major limitation of the present study is the absence of direct biosafety and public health assessments of urine fertilizer. Future research should therefore incorporate microbiological, parasitological, and chemical analyses to evaluate pathogen survival, pharmaceutical residues, and antimicrobial resistance risks under locally relevant storage, treatment, and application conditions. Longitudinal field trials that explicitly link adoption behavior with soil quality, crop contamination, and food safety outcomes are particularly necessary to ensure that the promotion of urine fertilizer is aligned with public health protection, regulatory standards, and broader One Health objectives.

Furthermore, while this study provides robust evidence on the socio-economic and institutional factors shaping awareness, adoption, and intensity of urine fertilizer use, it does not evaluate the microbiological, parasitological, or chemical safety of urine-derived inputs. No laboratory-based assessments were conducted to examine the presence of pathogens, pharmaceutical residues, heavy metals, or antimicrobial resistance genes. As such, the observed adoption patterns should not be interpreted as validation of agronomic safety, food safety, or public health suitability. Therefore, future research must integrate direct biosafety and public health assessments to ensure that the promotion of urine fertilizer aligns with established safety standards. Specifically, studies should include: (1) pathogen reduction efficiency, assessed through microbiological and parasitological analyses to evaluate the survival of pathogens such as E. coli and helminth eggs under locally relevant storage and treatment conditions; (2) pharmaceutical residue analysis, including screening for antibiotics, hormones, and other bioactive compounds to assess potential risks of antimicrobial resistance and ecological disruption; and (3) the validation of safe application protocols through field trials to test and optimize locally feasible treatment methods, for example storage duration, pH elevation, and dilution ratios, alongside application techniques that minimize risks to farmers, consumers, and the environment.

Although source-separated urine is often considered relatively low-risk compared to fecal matter, potential hazards remain, particularly in contexts where storage duration, treatment conditions, and handling practices are inconsistent. Emerging evidence indicates that human-derived fertilizers may contain residual pharmaceuticals and antimicrobial compounds that warrant careful risk assessment before large-scale agricultural reuse. The absence of such analyses represents a key limitation of this study and underscores the need for integrated socio-economic and biophysical research frameworks when promoting ecological sanitation-based nutrient recycling systems.