Soil nutrient content in relation to women ’ s agricultural knowledge in the urban gardens of Kisumu , Kenya

Agricultural production in Kenya has been declining since the 1980s, either because soils are mismanaged or because they lack nutrients. In Kisumu, just under 50% of the workers in the urban gardens are female. On average, women spend more hours a day in the gardens than men. To increase yields, women’s knowledge has to be considered in agricultural 20 management. However, women face greater obstacles in land ownership, investment, and farm inputs due to social and cultural constraints as consequence of their gender. This case study aimed to determine the nutrient content in soils of the urban gardens of Kisumu, the agricultural knowledge of the women farmers and how their knowledge influences soil nutrient content through their management. Soils were sampled in Nyalenda, one of Kisumu’s informal settlements where urban gardening is practiced, to determine soil 25 nutrient contents. To determine how agricultural management practices influences total C and N, available N and P, and exchangeable K, Mg and Ca in the soil, two prevailing practices were compared: 1) applying manure only, and: 2) applying manure while intercropping with cowpeas. Interviews and focus group discussions were organized to determine what knowledge the female vegetable farmers possess, and where they acquired their knowledge. SOIL Discuss., https://doi.org/10.5194/soil-2018-24 Manuscript under review for journal SOIL Discussion started: 6 August 2018 c © Author(s) 2018. CC BY 4.0 License.


Introduction
This paper details a case study on the soil nutrient contents in urban gardens in relation to the agricultural knowledge of women vegetable farmers.The study was conducted as part of an interdisciplinary study on soil nutrients and women entrepreneurship in Kenya and Burkina Faso.The women vegetable farmers considered in the present study work in the urban gardens of Kisumu, Kenya.Urban gardening is part of the practice of urban agriculture, which encompasses all agricultural activities practiced within municipal borders (FAO, 2012).Most urban gardens in Sub-Saharan cities are assumed to be marginalized, often polluted, plots of land used for vegetable gardening by the urban poor (Cofie et al., 2003;FAO, 2012).For the urban poor the urban gardens provide employment opportunities and are a source of affordable vegetables.These vegetables are less expensive than those imported from the rural areas due to the lack of transporting costs.
Limited infrastructure makes it difficult for some types of produce to be transported from the rural areas.By growing such vegetables within the municipality the costs are reduced and the lower costs of these vegetables allow the urban gardens to contribute to urban diet diversity and food and nutrition security (FAO, 2012;Gallaher et al., 2013).
In Kenya, 36.5% of the GDP comes from agriculture, however agricultural production in Kenya has been declining since the 1980s (Okalebo, 2009;World Bank, 2016).The primary cause for this decline is believed to be nutrient deficiencies in the soil, specifically nitrogen (N) and phosphorus (P), and in some cases potassium (K) and organic carbon (Org.C) (Cofie et al., 2003;Okalebo, 2009;Tittonell, 2005).The debate regarding these deficiencies is on their cause; some pose that low nutrient content is an inherent soil property (Tittonell, 2005), whereas others suggest it to be a consequence of erosion due to poor soil management and intensive use of the soil (Cofie et al., 2003;Okalebo, 2009).Many of the farmers in Kenya are limited in their choice of management due to the limited size of the farm area.Farmers working with less than 0.5 ha are known as small holder farmers, and they make up 75% of Kenya's agricultural sector (KNBS/AWSC, 2014).The majority of the urban gardens in Kisumu fall within the classification of small holder farm, with farm plots between 0.104 and 0.41 ha (Mireri, 2013).
With over half a million inhabitants Kisumu is Kenya's third largest city.Unemployment in Kisumu is high, in 2013 the unemployment rate in Kisumu was determined to be 40% (Mireri, 2013).Over 60% of Kisumu's inhabitants live in informal settlements (Mireri, 2013;Obade, 2014;UN-Habitat, 2005).It is estimated that up to 60% of the inhabitants of Kisumu practice some form of urban agriculture, including livestock keeping.Agriculture has been practiced on the periphery of the SOIL Discuss., https://doi.org/10.5194/soil-2018-24Manuscript under review for journal SOIL Discussion started: 6 August 2018 c Author(s) 2018.CC BY 4.0 License.city since its founding in 1901, but as the city grew the boundaries between the urban areas and the rural areas have grown vague.Most of the original agricultural areas have fragmented and now fall within the municipal boundary, as such these areas have been reclassified as urban gardens (Anyumba, 1995).Most of the urban gardens are located on the edges of the informal settlements.Mireri (2013) found that approximately 47% of those working in the urban gardens in Kisumu is female, and that on average the women spend more hours a day on the farms than men.Women are culturally expected to take responsibility for family food provision and many of the women farmers work on a subsistence basis (Kabira, 2007;Kameri-Mbote, 2006;Kiriti-Ng'anga, 2015a;Kiriti-Ng'anga, 2015b;Mireri, 2013).Any excess produce is sold by the women to pay for expenses such as their children's school fees (Mireri, 2013;World Bank, 2009).Gender inequality makes it difficult for these women to move beyond subsistence agriculture.Women face greater obstacles than men in regards to land ownership, investment, and farm inputs due to historical, social, cultural and financial constraints as a consequence of their gender (Alunga and William, 2013;Dolan, 2015;Kabira, 2015;Kameri-Mbote, 2006;Kiriti-Ng'anga, 2015a;Kiriti-Ng'anga, 2015b).As a consequence of these obstacles, few modern techniques are applied in the urban gardens.Lack of access to capital and knowledge limits these women to traditional techniques and sensory knowledge passed down within families (FAO, 2012).
Due to a lack of equal access to technologies such as fertilizers, women consistently have yields that are on average 20-30% less than men in developing countries (FAO, 2006).At the same time, there is evidence of gender differentiated access to knowledge.The results of a food security survey held by the African Women's Studies Centre and the Kenyan National Statistics Bureau in 2013 showed that women respond differently to food security issues and consider challenges differently than men.For example, up to 80% of men believe that a small, uneconomical area of land is a hindrance to achieving food security, whereas only 20% of women consider this to be a major hindrance (KNBS/AWSC, 2014).There is further research that suggests that women could potentially produce up to 20% more on the same surface area than men if given equal access to resources (Saito et al., 1994).
Through this interdisciplinary study we aim to determine the soil nutrient status of the soil as reflected by the contents of total soil C and N, available soil N and P and exchangeable soil Na, K, Mg and Ca in the Nyalanda urban gardens.Furthermore, we aim to determine how agricultural management of the women working in the Nyalenda urban gardens determines this soil nutrient status, and how the choice for a particular form of agricultural management is linked to the specific knowledge these women poses as well as to socioeconomic factors.Soil nutrient content and the effects of the two most common agricultural management practices on soil nutrients were assessed through soil sampling and laboratory analysis.Questions regarding knowledge acquisition and sharing were addressed through interviews and single-gender and mixed-gender focus group discussions.These methods are used to triangulate and provide complementary information.SOIL Discuss., https://doi.org/10.5194/soil-2018-24Manuscript under review for journal SOIL Discussion started: 6 August 2018 c Author(s) 2018.CC BY 4.0 License.

Materials and Methods
Kisumu is a city of approximately half a million inhabitants on the northern shores of Lake Victoria.The city is the headquarter of the Kisumu district and the Nyanza Province (Mireri, 2010).Temperature averages at 22 ⁰C year-round and annual rainfall averages between 1000 and 1400 mm.There is a short rain season in November and December and a long rain season lasting from April until June.The city lies on Quaternary sediments and Tertiary Volcanic deposits.Due to the tropical climate, deeply weathered soils can be expected in this area, but the parent material is relatively young and rich in nutrient bearing minerals (Orodho, 2006).
There are several informal settlements in Kisumu, including Nyalenda.Nyalenda lies on the southern edge of the city and is one the largest of the cities six informal settlements, both in number of inhabitants and surface area covered ( UN-Habitat, 2005).Divided over two blocks, A and B, Nyalenda houses nearly 50.000 people within an area of 8.1 km 2 .Existing infrastructure, access to electricity and access to sanitation are limited in the informal settlements (UN-Habitat, 2005).All along the southern edge of Nyalenda there are active vegetable farms adjacent a river and wetland area.One of the groups active in these urban gardens is the Mesopotamia group.The group consists of 14 members, 8 women and 6 men, who cultivate an area of 3-4 ha.Most Mesopotamia members have inherited their land and some rent extra plots within the area; the group is diverse in age and experience.
The Mesopotamia group had previously been informed by government extension services that their soil might be lacking in N. In response to this apparent lack of N at least 5 group members changed their practices, they started to intercrop the local staple crop Sukuma Wiki, a kale (Brassica oleracea var.Sabellica) with a legume with nitrogen fixating root nodules, cowpeas (Vigna unguiculata L. Walp) (Likoko and Jonkman, 2016).
The four fields selected for soil sampling were all used to grow kales, in two of the fields the kales were intercropped with cowpeas.All four sampled fields are centrally located in the urban gardens, limiting the differential influence the nearby river might have on fields lying closer or farther away from it.The soil on these fields were classified as Vertisols (FAO, 2014), characterized by the presence of heavy clay which shows shrinking and swelling behaviour.All samples were collected in May during the dry season.On each of the four fields 12 samples were collected from the topsoil (0-15 cm) to limit the influence of spatial variability, 48 samples total.All samples were subsequently dried at 70⁰C, sieved at 2 mm and stored for analysis.
The four fields sampled are owned by two female members of the Mesopotamia group, each member owning two of the fields.One of the women grows exclusively kales on her fields and the other woman intercrops the kales with cowpeas on her fields.Both women used manure as fertilizer, however one ploughed it into the soil while the other applied it as a topdressing.The two women that owned the sampled fields, along with the 6 other female members of Mesopotamia, were interviewed to determine how soil nutrient contents were influenced by women's agricultural knowledge.In addition to the interviews two focus group discussions were held with members of the Mesopotamia group.One focus group discussion was held with 6 female participants and another with 11 participants, 6 women and 5 men.A women's only discussion was held because women are more likely to speak their mind when they are not in the company of their male counterparts.The focus groups discussions were based on questions used in the interviews and aimed at determining the extent of agricultural knowledge in the Mesopotamia group as well as their information sources and the relative importance of these to the farmers.Due to the open platform and the presence of multiple participants the focus groups discussions provided more in-depth answers and clarifications, which support the information from the interviews.

Laboratory analyses and data processing
Water extracts of the soil samples were created (ratio 1:2.5) and used to determine pH and EC.These water extracts were then filtered and available P, K, S, Ca and Mg measured using a Perkin Elmer Optima 8000 ICP-OES Spectrometer.
Available NH 4 + , NO x , PO 4 3-and SO 4 2-in the extracts were determined on a Skalar SA-40 continuous-flow analyzer.Total organic and inorganic C in the extracts were measured using a Shimadzu TOC/TN analyzer.
Filtered BaCl 2 extracts were used for the determination of exchangeable Fe, Mn, Mg, Ca, Al, and K with ICP-OES (Schwertfeger and Hendershot, 2009).Extracts were prepared using 100 ml BaCl 2 0,125 M and 4 grams of milled soil sample (<2 mm).CEC was calculated as the sum of the values for exchangeable Ca, Mg, K and Na in cmol c /kg.
Total C and N were determined with 50 mg of soil (<2mm, milled) by using a Elementar Vario EL cube CNS analyzer.Total P, K, Ca, S and Mg were determined by measuring HNO 3 /HCl extracts with ICP-OES; extracts were prepared with 250 mg soil (<2 mm, milled), 6 ml HCl 37% and 2 ml HNO 3 , and underwent microwave destruction (60 min;Tmax 220•C; Pmax 75bar).Total elemental composition of the soil samples was also determined using XRF analysis, using the Thermo Scientific XRF Analyzer Niton; setting: mining Cu/Zn; Standard: NIST 2709a PP 180-649; 160 seconds.
Variance within each field and between fields with different management practices was determined using analysis of variance test.ANOVA was used in case of normal data distribution and Kruskal Wallis with non-normal data distribution (Burt et al., 2009).The strength and direction of the relationship between different parameters was determined using a correlation coefficient, Pearson's R. All statistical analysis was done in Matlab, version R2014b.The measured results and calculated variances where corroborated with the results of the interviews and focus groups discussions..

Results
The soil of the Mesopotamia group was analyzed on the nutrient content, and by FAO standards generally fell within the ranking 'high' (FAO, 2006).Table 1 provides the average values for the main soil parameters.The pH of the soil in the sampled fields ranged from very slightly acid to very slightly alkaline, with an overall average of 7.25 ( was high overall with an average value of 33.5 cmol c kg -1 , likely as a consequence of the high clay content of the soil (Table 1).Similarly, with an average of 36.6 g kg -1 the total soil carbon was also high.The laboratory analyses showed relatively high amounts of water soluble and exchangeable cations, however there is a significant difference in nutrient content depending on the management practice.
The women of the Mesopotamia group possess limited technical knowledge and are aware of the effects of the management practices that they apply only in terms of the visible effects of these practices.They know that plants need nutrients from the soil and that they can add nutrients to the soil by applying fertilizers, mineral or organic.The women spoke of a variety of agricultural management practices during the meetings, including: crop rotation, fallow periods, fertilization with manure, compost and artificial fertilizer, intercropping, and mulching.Not all of the mentioned practices were familiar to all of the women, and some of the women found that they were limited in choice of management practice due to their socio-economic circumstances.For example, for many women fallow periods are not an option as their lands are simply too small.A consequence of the lack of fallow periods is more pressure on the land, which can lead to increased erosion and may result in An alternative practice to fallow periods practiced by two of the women at the Nyalenda site is the seasonal rotation of their crops, which gives the soil time to recover as different crops have different nutrient requirements.Intercropping is done by some (n=4), mainly with cowpeas, to improve soil nutrients.However, the intercropping technique is not always applied correctly; the plants are harvested and not plowed into the soil.Plowing the cowpeas into the soil is needed for the nutrients accumulated to become available for other crops through the enrichment of the soil with organic matter (Okalebo, 2009).
The women also use intercropping to prevent soil erosion while the main crops, often kale, is still growing.Intercropping also provides a source of income while the farmer waits for the kale to mature as cowpeas mature faster.Soil analysis shows that the application of intercropping has a significant effect on the soil nutrient content (Figures 1, 2, Table 1).
Specifically, in case of manure application combined with intercropping the pH leaned towards being very slightly acid, whereas in case of only manure application the pH leaned towards very slightly alkaline.The CEC was nearly 10 cmol c kg -1 higher in fields under only manuring than in the fields where there is also intercropping.Similarly, total soil carbon is nearly SOIL Discuss., https://doi.org/10.5194/soil-2018-24Manuscript under review for journal SOIL Discussion started: 6 August 2018 c Author(s) 2018.CC BY 4.0 License.20 g kg -1 higher in the fields where only manure was applied in comparison with the fields where there was also intercropping (Table 1; Fig. 2a).The contents of the macronutrients N, P, K, Ca and Mg were almost all higher under the field management type manure application only, as compared to manuring combined with intercropping (Fig. 1; Fig. 2).
Figure 1a, 1b and 1c show the amounts of water soluble and exchangeable Mg, Ca and K as part of the total amount of the cation present in the soil, clearly demonstrating that the levels are higher under the practice of applying manure only.Figure 1d, 1e and 1f show the proportion of the total amount of Mg, Ca and K in the soil that is water soluble or exchangeable.
Notable is that while the absolute amounts are higher under manuring only, under the practice of manure application combined with intercropping, often a larger proportion of the nutrients was water soluble or exchangeable (Fig. 1).
Specifically, the average exchangeable fraction was higher for Mg and Ca under intercropping + manuring, and the average water soluble fraction was higher for Ca and K under intercropping + manuring.The Kruskal Wallis and ANOVAs tests showed that all the described difference between the fields and between the management practices were significant for these characteristics at a confidence interval of 95%.

Discussion
Results of the sample analysis of the soil of the urban gardens in Nyalenda showed a pattern consistent with the soil typology.The soils in the Nyalenda urban gardens can be classified as Vertisols, locally known as black cotton soils, which are soils characterized by a high clay content that shows shrink and swell patterns (FAO, 2006;FAO, 2014).Vertisols are generally fertile and productive soils, high in Ca, K and Mg, but often poor in N and P (FAO, 2006, FAO, 2014).Soil analysis has shown the rating for exchangeable Ca and K was very high and the rating for Mg was high according to FAO classification (Table 1; Fig. 1) (FAO, 2014).In comparison, Tittonell et al., (2005) analyzed topsoil conditions in 3 smallholder farms in western Kenya.Their analyses showed N values almost 1 g kg -1 lower, similar P values, and K values a tenth of the values found in Nyalenda.The high amounts of nutrients in the Nyalenda soils is most likely because of the parent material, which consists of river and lake sediments with inputs from the African rift valley, and the limited age of the soil material.
Earlier research on soil nutrients and possible solutions for soil fertility problems in western Kenya concluded that socioeconomic factors determine how likely it is that scientific findings are taken up by farmers (Gicheru, 2012;Okalebo, 2005).When asked to name and rank their primary sources of information all farmers of the Mesopotamia group, male and female, indicated a preference for inherited knowledge, followed by trainings and demonstrations."Next to the information we got from our ancestors, we get information from the look of things, when you come and see somebody farming and you ask what they are doing and how it is going.So by observing is also how we get information" (FGD 1, participant I5).There was some discussion on the differences between demonstrations and observations, and some of the farmers consider them equal in importance.Information from training and demonstrations is however often lost because techniques or elements thereof are forgotten over time and/or materials needed are unavailable or too expensive.Using and adapting techniques by observing other farmers is more common.Information from television or internet has less impact because these farmers lack access to these media.Exhibitions are considered good, but the expense to go and visit them is often considered to be too high.
Limited access to sources of information means that most of the Mesopotamia farmers possess limited technical knowledge regarding soil processes, however they are aware of soil processes and their consequences in practical terms from sensory knowledge and daily experience.For example, the women are aware of the need to rest the soil with fallow periods or crop rotation and that mulching improves soil structure.The majority of the women in the Mesopotamia group possess knowledge regarding agricultural management practices and the effects of these practices on soil in these basic terms.As this knowledge is mainly passed down from previous generations or disseminates through observation of other farmers in the group or community, this knowledge does not travel far (Alunga and William, 2004;Kabira, 2007).
The women farmers of Mesopotamia report that they prefer to share information with other women.The women are wary of sharing information with men, as men might feel offended by 'being taught' by women.The women believe it is easier for other women to understand their knowledge because of their shared backgrounds and responsibilities and they indicate that they often continue beyond scheduled meetings to further discuss issues and solutions."Women are easier to work with because they are the people who take responsibility in the houses and can solve this." (FGD 2, Participant I5).On occasion they will choose not to share information with another, if for example they believe that the other woman does not have the resources to apply the technique.
In Nyalenda, a context where poverty is widespread, agricultural management and decisions are heavily influenced by social-economic constraints.These constraints can work against sustainable farming practices.For example, the farmers explained that some had taken to intercropping with cowpeas originally because they were told that they needed to increase N in the soil by government extension workers.The analysis of the soil samples shows that the soil nutrient content is significantly lower when the kales are intercropped with cowpeas ( Soil samples from fields with intercropping show lower amounts of soil nutrients on average (Figs 1a, 1b, 1c, 2), however a comparison on the availability ratios of some of the nutrients showed that they have a higher availability or are more readily exchangeable under intercropping (Fig. 1d, 1e, 1f).The cause for this difference is not clear, but may be the influence of the presence of a legume species.The presence of the rhizobacteria on the root nodules of the legume can promote the availability or exchangeability of nutrients beyond nitrogen by immobilizing nutrients and preventing them from leaching from the soil (Lavakush et al., 2014;Vejan et al., 2016).Furthermore, the lower amount of nutrients in the intercropped fields may be due to the different approaches that the farmers have to applying manure.Farmer I2, who applies intercropping, applies the manure as a topdressing only.Farmer I7, who does not practice intercropping, ploughs the manure into the field.Ploughing the manure into the field preserves the N and promotes the biological breakdown of the manure, which increases the availability of the nutrients therein (Baligar, 2001).The decrease in soil nutrients in intercropped fields (Fig. 1) was most likely caused by the farmers not ploughing the cowpeas into the soil in combination with the different manuring practice (Okalebo, 2009).
During the interviews it became clear that instead of ploughing the cowpeas into the fields, the farmers are harvesting the cowpeas for sale.Harvesting the cowpeas means a greater uptake of nutrients from the soil and no additional organic material is added to the soil.Selling the cowpeas has become the primary motivation for intercropping as it gives the farmers a source of income in the period before the kales are mature and the advantage of doing so is more readily apparent to farmers than a potential increase of N in the soil.This shows that the lack of effect of intercropping on soil N contents in the examined soils is most likely not the cause of a lack of women vegetable farmers' knowledge of proper application of intercropping.Rather it appears to be a conscious choice related to a shift in the aim to be achieved by intercropping, i.e.
gaining a secondary crop to be harvested and sold rather than increasing the yields or quality of the primary crop.

Conclusion
The results of the soil analysis showed that the soil in the Nyalanda urban gardens is rich in macronutrients.Further analysis indicated that, while seemingly small, the impact of different agricultural management practices on soil nutrients is significant.Growing cowpeas and ploughing them into the soil should increase soil N content, however the farmer that applied intercropping sold the cowpeas on the market rather than ploughing them into the soil.This practice resulted in the observed decrease in nutrient contents, but provided the farmer with income at a time when the kales were still maturing.
The FGDs with the Mesopotamia group showed that there is knowledge of a wide range of agricultural management practices present.However, the interviews with the individual women member of the group showed that the knowledge on these practices is unequally distributed and that while they may be known to a technique they do not always possess technical knowledge on the management practice.We conclude that the incomplete knowledge of these farmers is a consequence of the way they acquire and rank knowledge.During the FGDs a clear preference was given by all farmers to knowledge gained from family members.Observation of other farmers and trainings by outside groups were also appreciated, but considered less important.Knowledge from the trainings was often forgotten or materials needed were unavailable or unaffordable, making the training virtually ineffective.
The ineffectiveness of trainings showed that these should be adapted to take the socio-economic circumstances of the trainees into account.Furthermore, the gender differences in ability and access should similarly be taken into account in order to improve the effectiveness of a given training or agricultural recommendations.The case study showed that women are influenced by their socio-economic and cultural status when making decisions in agricultural management and that these decisions may differ from those of men in the same or similar circumstances.While the women of the Nyalenda group are willing to share their knowledge with other women, they are more wary of sharing with men.This wariness on the part of women contributes to the presence of gender differentiated knowledge and hampers the spread of knowledge in general.
The eight women varied in age and experience, capturing a broad spectrum of views and knowledge.The semi-structured interviews used open SOIL Discuss., https://doi.org/10.5194/soil-2018-24Manuscript under review for journal SOIL Discussion started: 6 August 2018 c Author(s) 2018.CC BY 4.0 License.questions to determine what knowledge women farmers had about the effects of fertilizers on crops and soil, where they received this information, and to what degree and with whom they shared this knowledge.

Figure 1 .
Figure 1.1a, 1b and 1c: Bars total length show the total amount of Mg, Ca and K in mg kg -1 soil under management 'manure' and 'intercropping and manure', darker sections of the bar diagram depict the portion of the total that is exchangeable and plant available/water-soluble. 1d, 1e and 1f: Water soluble and exchangeable Mg, Ca and K as percentage of total.
SOIL Discuss., https://doi.org/10.5194/soil-2018-24Manuscript under review for journal SOIL Discussion started: 6 August 2018 c Author(s) 2018.CC BY 4.0 License.Gender-aware research has shown that women possess important knowledge regarding agricultural management, distinctly different from the knowledge of men in agriculture(Saito et al., 1994; KNBS/AWSC, 2014).When farmers do not use scientific findings it is often regarded as a sign of unwillingness, lack of understanding, or ignorance.This view is particularly damaging for the collaborative interactions between different institutions and farmers, and the success of any potential innovations in agriculture that are adaptive, affordable and applicable to the context.