Effects of biochar and sewage sludge on spinach ( Spinacia oleracea L.) yield and soil NO 3- content in texturally different soils in Glen Valley, Botswana

The effects of biochar and sewage sludge application on spinach ( Spinacia oleracea L.) yield and soil NO 3- content were investigated in typical soils of Botswana (Luvisol, Cambisol) under field conditions. Ten treatments with 3 levels of biochar (0, 2.5, 5 tons ha -1 ) and sewage sludge (0, 6, 12 ton ha -1 ) were applied in 2 subsequent seasons. Significant (p < 0.05) yield increase on the Luvisol occurred if sewage sludge was added at 12 Mg ha -1 with or without biochar. A combination of 6 Mg ha -1 sludge and 5 Mg ha - 1 biochar application resulted in the highest crop yield over 2 seasons. On the Cambisol, only marginal yield increase occurred upon high rates of sole organic amendments and chemical fertilizer, while co-applications decreased yields. Decrease in soil NO 3- content caused yield declines in the second season, while P uptake increased significantly (p < 0.05). Correlations between yields, soil NO 3- and leaf N contents were insignificant (p > 0.05). On the Cambisol, a significant regression model for sludge and soil NO 3- was determined. Therefore, one – time combined application of 6 Mg ha -1 sewage sludge and 5 Mg ha -1 on the Luvisol, and 12 Mg ha -1 sewage sludge are recommended for spinach production on the Luvisol and Cambisol, respectively. In subsequent seasons, crop productivity could be maintained by application of mineral N in order to mitigate over-application of P.


INTRODUCTION
Soil fertilization with sewage sludge is an effective way to recycle nutrients and combat nutrient deficiency in agricultural systems (Sharma et al., 2017). Spinach is one of the most important vegetable crops in Botswana, but good crop yields are constrained by poor soil fertility, especially N and P deficiency. Many studies have reported high spinach yield response to mineral fertilizers and sewage sludge applications (Ngole, 2010;Biemond *Corresponding author. Email: umajaule@gmail.com. Author(s) Wang and Li, 2004;Lefsrud et al., 2007;Stagnari et al., 2007;Rodríguez-Hidalgo et al., 2010;Türkmen* et al., 2004). These studies showed that adequate N availability from sewage sludge is critical for high quality and yields of spinach. When contents of heavy metals, pathogens, and toxic organic compounds in sludge are within the WHO limits, such as the case for the Glen Valley sludge (Ngole, 2010;Mosekiemang and Dikinya, 2012), application rates of sludge to agricultural soils is based on the crop nitrogen (N) demand (Gilmour and Skinner, 1999;Correa et al., 2006).
On dry basis, sewage sludge contains 2 -6% total N which is predominantly organic (Rigby et al., 2016) hence the rate of N mineralization influences potential plantavailable N. In sludge-amended soils, this plant-available N varies between 20 to 63% of organic N in a crop year under field conditions (Magdoff and Amadon, 1980), depending on factors such as sludge application rates, timing, climate, soil properties and moisture dynamics (Weggler-Beaton et al., 2003). Thus, the application rate of sludge should consider the sum of the inorganic N and mineralised organic N in the soil and the added sludge. However, important soil processes such as microbialmediated immobilization, leaching, ammonia volatilization and denitrification can decrease the amount of the plant available N pool (Clough et al., 2013).
At modest sewage sludge application rates (c.a. 10 tons ha -1 ), sludge may not provide adequate N for optimum spinach yields because it has a high N demand. In addition, spinach typically prefers NO 3 because high concentrations of ammonium (NH 4 + ) ions can be toxic and suppress both root development and plant growth (Wang et al., 2009). Thus, the relative concentrations of NO 3 and NH 4 + in sewage sludge, and the nitrification rates determine N uptake and productivity of spinach. But, excessive levels of NO 3 in spinach leaves may be noxious to humans (Citak and Sonmez, 2010), while leaching of NO 3 into groundwater is linked to methemoglobinemia or -blue-baby‖ syndrome in infants, cancer and spontaneous abortions (Spalding and Exner, 1993). Nitrogen availability from sewage sludge in Botswana was sparsely explored by Ngole (2010) under controlled conditions but the results were confounded by lack of mineral fertilizer comparisons.
Biochar-induced changes of soil properties such as soil pH, cation exchange capacity (CEC), moisture dynamics, and microbial activity may in turn significantly influence N transformation reactions (Nelson et al., 2011;Clough et al., 2013;Anderson et al., 2011). Besides the potential direct N supply, biochar effects on soil organic matter decomposition rates could either decrease or increase organic N mineralization from organic amendments, retention of NH 4 + on its surfaces, and therefore, soil NO 3 and NH 4 + ratio. Biochar-induced N deficiency due to net N immobilization have been linked to high biochar C/N and increased microbial activities (Deenik et al., 2010). Other studies showed that such microbial N immobilization and partly due to high biochar cation exchange capacity (CEC) improved N fertilizer use efficiency and plant productivity (Chan et al., 2008;Steiner et al., 2007). Nitrification rates in biocharamended soils can either increase or decrease due to the biochar stimulatory or inhibitory effects, respectively (Nelissen et al., 2012;Clough et al., 2013;Zackrisson et al., 1996), with significant implications for spinach growth and yields as it prefers NO 3 compared to NH 4 + (Wang et al., 2009).
A growing number of studies showed significant synergistic effects of biochar and N fertilizers on N use efficiency and crop yields (Chan et al., 2008;Adekiya et al., 2019;Partey et al., 2014). Contrastingly, Lentz and Ippolito (2012) found no synergistic effects of hard-wood biochar and cattle manure on corn silage yields and nutrient concentrations, except for Mn. Others reported biochar-induced N deficiency due to N immobilization or decreased nitrification rates (Zheng et al., 2013;Cayuela et al., 2013).
The contrasting results suggest that the effects of biochar on N availability is heterogeneous, depending on biochar and soil properties, application rates for both biochar and organic N fertilizers and other experimental conditions. However, information regarding the combined effects of biochar and sewage sludge fertilization on N uptake and yields of spinach is scarce. We hypothesized that synergistic effects of biochar and sewage sludge application would significantly increase NO 3 availability and spinach yields. The objective of this study was to determine the effects of co-application of biochar and sewage sludge on soil NO 3 content, N uptake and spinach yields in typical soils of Botswana.

Study sites
The experiments were established at Glen Valley, Botswana. A Calcic Luvisol (henceforth called Luvisol) and Vertic Cambisol (henceforth called Cambisol) were selected for the study. Some properties of the surface soils (0-15 cm) are shown in Table 1. The Luvisol was classified as a sandy loam textural class (sand -73.3%, clay -16.4% and silt -10.3%) with a bulk density of 1.6 g cm -3 . The Cambisol had sandy clay texture comprising of the following size fractions; sand -48.5%, clay -39% and silt -12.6%), and 1.4 g cm -3 bulk density.

Experimental design
The experiment comprised 10 different treatments with 3 replicates and so each site comprised 30 plots (1.8 m x 1.5 m) arranged in a randomized complete block design (RCBD). Table 2 shows the treatments structure and application rates in each season. The spinach (variety; Fordhook Giant) crop in the first season was planted in March and harvested in April, 2018. The second harvest was done in June, 2018. Similar amendments were applied before the plots were replanted in the second season with spinach seedlings in July, 2018. Harvests 1 and 2 were done in September and October, 2018, respectively. Both sites were disc ploughed to about 30 cm depth before the study. Planting rows were constructed using hand-hoes before organic amendments were incorporated and mixed into soil (15 cm). Transplanting was done one day after irrigation. Mineral fertilizer (2:3:2, 22%) was applied by banding during transplanting at 300 kg ha -1 (Bok et al., 2006). Urea ammonium sulphate (46% N) was top-dressed on CHEM plots at 200 kg ha -1 after 2 weeks of transplanting and after each harvest. The trials were drip-irrigated based on soil moisture conditions for about 2 h per irrigation. In the second season, the plots were cleared of crop residues, but the soil was not ploughed. The planting lines and treatment plots were maintained as during the first season.
Air-dried sewage sludge was collected from the stock piles at the Glen Valley Waste Water Treatment Plant. The sewage sludge was crushed and sieved (2 mm) before analysis and soil application. Biochar was produced from mixed-wood chips via a home-made slow pyrolysis unit (535ºC, 6 h). After cooling, the biochar was airdried, then mixed thoroughly before crushing and sieving (2 mm). The properties of the biochar (BC) and sewage sludge (SS) used for this study are presented in Table 1. The sludge was enriched in N, P, Ca, Mg and Na relative to the biochar, while the K concentration of biochar was over 3-fold that of sludge ( Table 2). The biochar C/N ratio was 64:1, over 3-fold that of the sewage sludge, which indicates a potential for N immobilization during labile biochar C degradation by soil microorganisms. The biochar was characterised by slightly alkaline pH (7.7), high EC (1124 µS/cm), and high contents of available Ca (128 cmol kg -1 ) and K (153 cmol kg -1 ), medium concentrations of Mg (34 cmol kg -1 ) and low contents of Na (3 cmol kg -1 ) ( Table 1).

Soil sampling and analysis
Soil samples (0 -15 cm) from each plot were collected using the composite sampling procedure at each harvesting stage. Air-dried samples were sieved < 2 mm and analysed in triplicate. Total carbon (TC) of the biochar and sewage sludge was characterized by ashing in muffle furnace at 500°C for 48 h. Exchangeable cations and soil CEC were determined using the ammonium acetate method at pH 7, using a mechanical extractor on a 2.5 g sample (van Reeuwijk, 1993). Exchangeable cations were quantified using a 4210 MP-AES (Agilent Technologies).
The pH of soil and sludge samples was potentiometrically determined in a 1:5 distilled water and 0.01 M CaCl2 solution. Biochar pH was measured on 1 g of sample as described by Wang et al. (2015). The pH values were determined in biochar-to-water ratio of 1:20 (w/v) via an Orion pH meter installed with a glass electrode. Total nitrogen (TN) of soil, sewage sludge, plant and biochar samples was analysed according to the micro-Kjeldahl procedure (van Reeuwijk, 1993). Plant-available phosphorus was determined as described by Ziadi and Tran (2008). Soil bulk density (BD) was determined using 100 cm 3 soil core samplers. Soil particle-size distribution of air-dried samples was measured according to the hydrometer method (van Reeuwijk, 2002). Soil pH and EC were determined at the end of each season while soil bulk density (BD) was measured in the second season only. Soil NO3was quantified according to the Cadmium reduction procedure. Briefly, 3 g thawed soil samples were extracted with 2 M KCl at the soil-to-solution ratio of 1:10 (w/v), while simultaneously determining the moisture factors. The extracts were frozen until they were required for NO3analysis using a Technicon Autoanalyzer II (Technicon Cooperation).

Plant sampling and analysis
The spinach plants were grown for about 60 days from the date of Table 2. Treatments, their application amounts per season and abbreviations.
transplanting. Harvesting was done on plot basis. Randomly selected plants were cut at about 5 cm above the soil surface on each plot. Fresh weights of leaves from each plot were recorded at each harvest stage, before oven drying and sieving (2 mm). For total content of P and bases in sludge and plant, 1.25 g of sample was wet digested in 2.5 ml of sulphuric acid-selenium mixture according to van Reeuwijk (2002). Basic cations were determined in the diluted digests via 4210 MP-AES (Agilent Technologies). In the diluted digests, P was measured spectrophotometrically by the indophenol-blue method (van Reeuwijk, 2002). Total P was measured by the method of Murphy and Riley (1962). Determination of the total content of P, K, S, Mg and Ca in biochar was done according to the modified dry-ashing method (Enders and Lehmann, 2012).

Statistical analysis
A 2-way ANOVA was conducted using SAS version 9.4 (SAS Institute Inc, Cary, NC) with significant differences identified at 5%, unless specified otherwise. Mean differences due to treatments were evaluated using least significant difference (LSD) and were ranked according to Duncan's multiple range tests. The Pearson's correlation procedure was used to analyse relationships between variables at 5% level of significance. The values given at each entry for all the parameters, except soil pH, electrical conductivity and bulk density are the average analyses for each of the two separate harvests in each season.

Effects of amendments on soil NO 3 content
The ANOVA showed that treatments and the interactions between soil type and treatments significantly (p = 0.0001) affected soil NO 3 content. The unamended control on the Luvisol had significantly (p < 0.05) lower NO 3 concentrations compared to the Cambisol in both seasons (Figure 1a and b).

Season 1
During the first season, the control on the Cambisol had higher NO 3 content (13.1 mg kg -1 ) compared to the same treatment on the Luvisol (9.2 mg kg -1 ). These levels are consistent with commonly reported NO 3 values (10 -25 mg kg -1 ) in agricultural soils (Tisdale et al., 1993). Except for application of low rate of sole sewage sludge (6SS), organic amendments significantly (p < 0.05) increased NO 3 on the Luvisol while the effects were insignificant on the Cambisol during the first season. The highest NO 3 levels during this season were caused by co-application of 6 Mg ha -1 sewage sludge and 5 Mg ha -1 biochar on the Luvisol (19.6 mg kg -1 ) while on the Cambisol, sole sewage sludge application at 12 Mg ha -1 gave the highest NO 3 level (15.1 mg kg -1 ), which also coincided with the highest spinach yields for the respective soils.
Soil NO 3 content insignificantly (p > 0.05) increased with the amount of applied soil amendments on the Luvisol during the first season ( Figure 1a). The same trend was observed for sole sewage sludge on the Cambisol, while increasing biochar amount marginally decreased NO 3 content from 14.1 to 13.2 mg kg -1 over the same period. On the Luvisol, sole sewage sludge application significantly increased NO 3 content relative to the control only when applied at 12 Mg ha -1 , but when combined with both rates of biochar, the lower rate of sewage sludge (6 Mg ha -1 ) resulted in a significant increase in soil NO 3 - (Figure 1a). Co-application of amendments marginally increased soil NO 3 compared to both rates of sole amendments on the Luvisol. With regard to the Cambisol, there were no significant (p > 0.05) treatment effects on soil NO 3 content during the first season, but co-applications decreased soil NO 3 content in comparison to the sole amendments and mineral fertilizer (Figure 1a). Mineral fertilizer (CHEM) did not significantly increase NO 3 above the control on both soil types. Interestingly, all combinations of sewage sludge and biochar on the Luvisol significantly increased NO 3 content relative to mineral fertilizer (Figure 1a), indicating a great potential of these organic amendments to substitute mineral fertilizers Dikinya and Mufwanzala, 2010).

Season 2
During the second season, all the treatments significantly (p < 0.05) increased NO 3 content relative to the control on the Luvisol (Figure 1b). Soil NO 3 content increased with increasing amount of each organic amendment on both soils, but the differences were only significant    (p < 0.05) for sole sewage sludge on the Luvisol. As can be seen in Table 3, the regressions between NO 3 content and sewage sludge amount on both soil types are significant (p < 0.05), and the regression coefficient is higher on the Luvisol. Conversely, biochar had an insignificant, but positive influence on NO 3 content on the Luvisol, while on the Cambisol, increasing biochar amount resulted in a decrease for NO 3 -. Soil NO 3 content generally decreased during the second season for most of the treatments, on both soil types. Notably, substantial decrease was observed for 6 Mg ha -1 sewage sludge plus 5 Mg ha -1 biochar (19.6 to 10.7 mg kg -1 ) on the Luvisol. With the few exceptions where NO 3 content increased (e.g. 12SS, 12SS+5BC on the Luvisol, and CONT, 5BC, 6SS+5BC on the Cambisol; Figure 1a and b), the differences between seasons were marginal.

Effects of amendments on leaf N content
ANOVA indicated that soil type had significant (p = 0.0004) effects on spinach leaf N content, while the effects of treatments, and soil by treatment interactions were insignificant (p > 0.05). Generally, the spinach leaf N content in this study is similar to other studies (2 -5%; Tisdale et al., 1993). In both seasons, treatments maintained statistically similar (p > 0.05) leaf N content compared to the control for both soil types (Table 4).
Increasing the amount of sewage sludge on the Luvisol marginally increased leaf N content during the first season, but the same trend did not occur on the Cambisol. Leaf N content generally increased in the second season for the corresponding treatments on both soils, except for co-application of 6 Mg ha -1 sewage sludge and 5 Mg ha -1 biochar on the Luvisol. Sole biochar at either rate also had no significant effects on leaf N on both soils (Table 4). In general, the effects of factors on leaf N content were marginal as shown by the small regression coefficients for both soil types (Table 3). The effects of biochar were positively related to both NO 3 and plant leaf N contents on the Luvisol, while on the Cambisol, both parameters decreased with increasing biochar rates, as indicated by negative coefficients of the regression equations.

Effects of amendments on spinach yields
The spinach yield data for the 2 cropping seasons are presented in Figure 2. Treatment effects were significant (p < 0.0001; CV = 33.5%) on yield as indicated by the general ANOVA model. Spinach yields in the control plots were similar between seasons. In the first season, all treatments improved yields relative to CONT on the Luvisol, with the greatest yields resulting from 6 Mg ha -1 sewage sludge plus 5 Mg ha -1 biochar. During the second season, all organic amendments maintained higher yields than control on the Luvisol while CHEM resulted in slightly lower yields relative to the control. On the Cambisol, yields were statistically independent of treatments (p > 0.05).
The changes in crop yield and soil NO 3 between seasons followed contrasting trends on the different soil types. An insignificantly negative correlation (p = 0.43: r 2 = -0.22) between yield and NO 3 was determined on the Luvisol, while the correlations on the Cambisol was positive and significant (p < 0.05: r 2 = 0.57). Decreasing yield during the second season on the Luvisol coincided with decreasing soil NO 3 content (except 12SS, 12SS+5BC), but marginal yields increases on the Cambisol followed an increasing trend of soil NO 3 content.
Considering other plant nutrients, the decline in crop yield in the second season as already highlighted above corresponded with increasing content of both leaf P (Table 4) and available P (Table 5). Other studies (Bhattacharjee et al., 1998;Tisdale et al., 1993;Türkmen* et al., 2004) have reported higher spinach leaf mineral composition than observed in our study; hence, this could be a contributing factor to the decreased yields in the second season.

Effects of amendments on NO 3 availability and leaf N content
Soil NO 3 content in the control was significantly lower than the Luvisol compared to the Cambisol (Figure 1a and b), which confirms the lower N availability in this soil, content in the control of the Cambisol could also be explained by greater mineralization of N from soil organic matter, which was higher on this soil type (Table  4).
In addition, the Luvisol and Cambisol had sandy loam and sandy clay textures, respectively. Therefore, the potential movement of soil NO 3 that is mineralized from native organic N or contained in the irrigation water down the profile (below the root zone) is higher on the Luvisol. As a result of these differences in NO 3 content in the control plots, the increase in soil NO 3 due to application of amendments was greater on the Luvisol. Furthermore, organic amendments were more effective in increasing soil NO 3 relative to application of mineral fertilizer on the Luvisol. In particular, significant treatment effects were detected when biochar was co-amended with sewage sludge compared to sole amendments on the Luvisol, whereas the opposite effects were determined on the Cambisol. These findings can be attributed to a number of reasons. Soil-specific effects of biochar on organic matter decomposition (priming), with potentially greater organic N mineralization on the sandy loam textured Luvisol is expected to play a significant role. The higher clay (39%) and native organic matter (2.3%) contents in the Cambisol could result in occlusion of biochar particles (Zackrisson et al., 1996;Wardle et al., 2008;Brodowski et al., 2006), thus restricting the interactions between biochar and sewage sludge particles. The complexation of humus by clay fraction has been linked to reduced soil C and N mineralization (Amlinger et al., 2003). Therefore, greater biochar and sewage sludge interactions on the Luvisol presumably enhanced mineralization of sludge-borne organic N and consequently higher nitrification of NH 4 + into NO 3 -. This effect is well known and can be explained by the supply of nutrients, reduced soil bulk density and improved aeration, optimum soil pH and potential biochar sorption of nitrification inhibitory compounds such as terpenes     (Zackrisson et al., 1996). Furthermore, biochar has greater effects on improvement of water retention on sandy soils than clay soils (Biederman and Harpole, 2013). This effect can result in higher responses in microbial decomposition and mineralization of sludge N on the Luvisol. The expected improvement in water retention could reduce leaching losses of soil NO 3 -, and this effect is likely to be more pronounced on the Luvisol because of its coarse texture, hence coapplication of amendments had more NO 3 content compared to sole sewage sludge treatments (Figure 1a and b).
Soil NO 3 content generally decreased for most treatments in the second season. Significant (p < 0.05) differences were determined for application of sole sewage sludge at 6 Mg ha -1 on the Luvisol, or in combination with 5 Mg ha -1 on both soils. During the first season, both the highest leaf N (Table 5) and yields (Figure 2a) on the Luvisol were determined for combination of 6 Mg ha -1 sewage sludge plus 5 Mg ha -1 biochar. Thus, the significant decrease in NO 3 content in the second season for 6SS+5BC could be a direct result of greater plant N assimilation which accounted for high crop yields during the first season. On the Cambisol, the decrease in soil NO 3 content was also statistically significant for co-application of 2.5 Mg ha -1 biochar and 12 Mg ha -1 sludge ( Figure  1b).
Such decreasing trends in NO 3 content could be attributed to several factors including the variability between agro-climatic conditions between seasons (Figure 3a and b) as shown by the statistically significant (p < 0.05) seasonal effects from the ANOVA. Total rainfall in the first and second seasons as determined from a nearby weather station at Sebele was 160 and 24 mm, respectively. The mean minimum and maximum temperature was 16.2 and 31.3ºC, respectively during the first month (March) after application of organic amendments. On the other hand, these attributes were 4.5 and 21.6ºC in the second season.
Several studies have been conducted which indicate that under warm moist conditions, sewage sludge organic N mineralization is more than in low temperatures (Sierra et al., 2001;Magdoff and Amadon, 1980;Barbarika et al., 1985). N mineralization rates in the first month after sewage sludge application is critical as it precedes the period of high spinach N demand and is usually supplied via top dressing with mineral N fertilizers. Thus, the warmer and humid climatic conditions in the first season potentially contributed to the higher N mineralization of sludge N resulting in high crop yields than in the following season. Furthermore, these results indicate that a significant proportion of the sewage sludge used in this study was organic (Magdoff and Amadon, 1980) which required mineralization before plants N assimilation.
Moreover, it is worth considering that before the trial was established in March 2018, the land was disc ploughed whereas prior to the second season, reduced tillage was applied to retain the planting lines and plots. The differences in mechanical aeration of the soil due to tillage possibly contributed to higher decomposition of sewage sludge and biodegradable biochar C in the first season, leading to accumulation of N into the labile microbial N pool. Although N immobilization can result in reduced yields (Deenik et al., 2010), in a wet season such as the March -June period in this study, it probably mitigated NO 3 leaching potential, which possibly resulted in higher N efficiency. However, from a long-term field trial in Gottingen, Hoffmann et al. (1997) concluded that minimal tillage systems increased Nnet mineralisation compared with conventional tillage systems, mainly due to the greater microbial biomass under reduced tillage. More research is therefore required to determine the factor(s) with overriding effects on sewage sludge N mineralization under Botswana conditions for development of sewage sludge management guidelines on agricultural soils.
There was insignificant correlation (p > 0.05) between soil NO 3 and leaf N contents. While soil NO 3 generally decreased among the majority of treatments in the second season, leaf N increased for all the treatments (Table 5). This disparity can be attributed to the dilution effects of greater yields in the first season. Furthermore, lower crop yields in the second season probably increased leaf N content to sufficiency levels (3%) and above compared to the first season. Table 3 shows that applications of sole organic amendments had greater effects on spinach yields on the Luvisol than on the Cambisol as shown by the higher slope of the regression equations. Moreover, application of biochar on the Luvisol significantly (p < 0.05) increased spinach yield, while the relationship on the Cambisol was insignificant. Crop yields across the two seasons also indicate greater yield response on the sandy loam textured Luvisol, and these responses were better than under mineral fertilizer (Figure 2a and b). Contrastingly, yield responses to organic amendments on the Cambisol were similar to mineral fertilizer. Studies on biochar effects on spinach productivity are lacking, but the results of this study support findings by Boersma et al. (2017). In their study, Eucalyptus polybractea biochar application to a fertile red Ferrosol did not increase yields of several vegetable crops. These results demonstrate the greater prospects of organic amendments in improving the crop productivity of degraded sandy soils which are prevalent in the tropics.

Effects of amendments on spinach yield
Mean yield data across the two seasons (data not shown) indicates significant synergistic effects of biochar and sewage sludge on the Luvisol for most of the treatments but no significant complimentary effects on the Cambisol. That there was significant (p < 0.05) yield increase from combined application of high rates of amendments on the Luvisol in the second season, while yield declined for the majority of the treatments (Figure  2a), is evidence of the synergistic effects of the amendments on the relatively infertile soil. Contrasting results were reported in a comparative short-term pot study in Zimbabwe by Gwenzi et al. (2016). Working on a clayey soil, the authors reported synergistic effects of sewage sludge and its biochar amended at 15 Mg ha -1 in increasing maize biomass yields, while sole biochar application without mineral fertilizer was less effective in increasing biomass yields. In our study, sole biochar application on the clay soil had more positive influence on yields than combination of biochar and sewage sludge. The difference in performance of amendments with their results for a soil type with similar texture to the Cambisol can be attributed to the variability in experimental conditions .
Crop yields were statistically independent from soil NO 3 content (p > 0.05), which presumably was caused by high coefficient of variation of the yield data (CV = 33.5%). Nonetheless, the decrease in yields on the Luvisol closely followed the same trend as soil NO 3 and leaf P contents between seasons (Figure 1a and Table  5), while leaf N increased. These data demonstrate that soil NO 3 content was the limiting factor for yields in this study because available P increased above the critical range of 45 -50 ppm (Ziadi and Tran, 2008) between seasons. Comparison of yield data between the two soil types indicates that there are more beneficial effects of application of organic amendments on the Luvisol than on the Cambisol.
Distinctly, 12SS+5BC consistently increased soil NO 3 and plant N contents, and yields (p < 0.05) between seasons on the Luvisol (Figures 1 and 2). The average yields for 12SS+5BC (26.9) was less than that for 6SS+5BC (33.6 ton ha -1 ), hence there is no added benefit of increasing the amount of sewage sludge combined with 5 Mg ha -1 biochar from 6 to 12 ton ha -1 . Also, available P increased well above the critical level to 177 mg kg -1 under 12SS+5BC (Table 3) on the Luvisol. Although P is not toxic and less mobile compared to NO 3 in the environment, its potential loss into the nearby Notwane River could result in eutrophication and degradation of the aquatic life.
Biotic and abiotic oxidation of biochar surfaces increases the CEC (Liang et al., 2006;Glaser et al., 2000;Wiedner et al., 2015), and this presumably retained significant levels of ammonium (NH 4 + ), thus suppressing nitrification (Clough et al., 2013;Nelson et al., 2011). This effect is hypothetically greater on the Cambisol due to its high clay content (Amlinger et al., 2003). Spinach typically prefers NO 3 to NH 4 + (Wang et al., 2009). Thus, in maintaining a relatively small pool of NO 3 -, formation and assimilation by spinach is suppressed, which could account for the yield declines in the second season for most treatments on both soils, specifically the biocharamended treatments, because elevated levels of NH 4 + can be toxic to aerobic plants and suppress both root development and plant growth (Wang et al., 2009;Deenik et al., 2010). Therefore, the decline in the yields under co-applications (Figure 3b) might be attributed to NH 4 + toxicity in biochar treated plots, since the other nutrients were in adequate supply (Table 3). Biochar addition of 5 Mg ha -1 on the Luvisol significantly reduced spinach yields (mean = 23 Mg ha -1 ) compared to application of biochar at 2.5 ton ha -1 (mean = 28 Mg ha -1 ), emphasizing the possibility of greater NH 4 + accumulation under higher sole biochar application.
High yields for a combination of intermediate sewage sludge (6 Mg ha   -1 ) and biochar applications (5 Mg ha -1 ) during the first season depleted soil NO 3 -, leading to decreased yields the following season. This hypothesis and the likely low mineralization rates of sludge organic N during second season due to cooler temperatures played a significant role in decreasing yields relative to the first season. Since available P was in adequate supply, crop yields on the Luvisol could be sustained by additional mineral N fertiliser instead of annual addition of sewage sludge to prevent excessive levels of P in the environment. This is true across all the treatments because P reached the critical level on this soil type for most of the treatments.

Conclusion
Overall, the results showed that co-application of biocharsewage sludge had statistically similar effects on soil NO 3 content on both soil types. Further, the improvement in N bioavailability and yields was greater on the Luvisol, while on the Cambisol, the effects of organic amendments were similar to mineral fertilizer. The decline in spinach yields in the second season was linked to the decrease in soil NO 3 because biocharsewage sludge addition supplied adequate P for spinach growth and leaf P reached sufficiency levels in the second season. Available P, SOC, CEC, soil bulk density, exchangeable bases all improved due to organic amendments. However, leaf micronutrients levels were comparatively lower than those reported for spinach in other studies, which should be the subject of future research. Therefore, Glen Valley farmers can reduce their fertilizer costs by using 5 Mg ha -1 of biochar on both soils, or combined application of 6 Mg ha -1 sewage sludge plus 5 Mg ha -1 biochar on the Luvisol. However, to prevent excess application of P, one-time application of organic amendments followed by mineral N fertilizer is necessary to maintain crop yields.

CONFLICTS OF INTERESTS
The authors have not declared any conflicts of interests.