Organic Inputs and Chemical Fertilizer on Carbon Mineralization From Two Ultisols

There are challenges that limit the use of organic inputs for soil fertility management. Amongst them is the limited knowledge of factors that affect rates of decomposition and nutrient release from different organic inputs. A study was conducted on surface soil samples of two Ultisols to determine factors affecting carbon (C) mineralization from selected organic inputs. A loamy sand (LS) from a Kandiustult and a sandy clay loam (SCL) from a Paleustult were used. Fine earth fractions of the soils mixed with organic inputs with and without chemical fertilizer were incubated for 13 weeks and the CO2 evolved was measured. Organic inputs used were biomasses of Cajanus cajan, Tephrosia vogelii, Crotalaria juncea, Mucuna pruriens, a mixture of native grasses and shrubs and composted cattle manure. The latter two inputs are traditionally used by farmers, while the leguminous plants were recommended by scientists. Treatments with chemical fertilizer only, representing the conventional farming practice, and a control with soil alone were included. Addition of organic inputs with or without fertilizer increased total CO2 emissions by 81 to 129% on the LS and by 18 to 34% on the SCL. Adding chemical fertilizer significantly (p < 0.05) increased C mineralization rate constant (k) by 116% on the LS and 48% on the SCL. The mean residence time of organic carbon from treatments grouped by input type followed the order: Control > Traditional > Legumes > Conventional on both soils. In general, the k on the LS was double that on the SCL. The type of organic input, soil texture and application of chemical fertilizer significantly affected C mineralization rates from the soils.


Introduction
Low soil fertility is one of the major constraints to achieving high crop yields, food security, and alleviating poverty in Sub-Saharan Africa (Ajayi, Akinnifesi, Sileshi, & Chakeredza, 2007;Bindraban, Batjes, Leenars, & Bai, 2010;Place, Barrett, Freeman, Ramisch, & Vanlauwe, 2003;Rusinamhodzi, Corbeels, Zingore, Nyamangara, & Giller, 2013). Although conventional farming systems that rely heavily on the use chemical fertilizers have been proposed as a solution to the low soil fertility (Matson, Parton, Power, & Swift, 1997), there are a number of concerns on the suitability of such systems. This is in view of the challenge of balancing the requirement for increased food production with preserving the natural resource base, and coping with threats of climate change (Diacono & Montemuro, 2010;Tilman, Cassman, Matson, Naylor, & Polasky, 2002). The conventional farming systems have further been deemed to be unsuitable for most low-income small-scale farmers in Africa because of the high cost and sometimes unavailability of chemical inputs (Bationo et al., 2012;Lal, 1987).

Description of Treatments Used in the Study
The organic inputs used in the incubation study were part of the treatments used in field experiments by the ORM4Soil Project at Misamfu and Msekera ARSs. Inputs used in the field experiments at each location had been selected following recommendations from local farmers through a participatory rural appraisal to identify traditional and researcher-recommended soil fertility management practices with the potential of improving soil fertility on small-scale farmers' fields. The treatments recommended by researchers at both ARSs were interplants of maize with Cajanus cajan (pigeon peas) and with Tephrosia vogelii (tephrosia). Besides these two interplants, a further interplant of maize with Crotalaria juncea (black sunnhemp) was included at Msekera ARS. All the interplant systems included the use of chemical fertilizers consisting of Compound D fertilizer (10% N: 20% P 2 O 5 : 10% K 2 O) used as a basal dressing fertilizer and urea (46% N) used as a top-dressing fertilizer. The chemical fertilizers were applied at full (200 kg/ha) and half (100 kg/ha) of the recommended rate for small-scale farmers in Zambia.
The traditional soil fertility management practices used was the Fundikila system at Misamfu ARS and composted cattle manure at Msekera ARS. The Fundikila system involves burying native grasses and shrubs under large soil ridges towards the end of the growing season at the end of March to early April. The organic inputs decompose under the ridges, until the ridges are opened-up and flattened before the onset of the rainy season in early December and the crop is then planted. A modification of the traditional Fundikila system which consisted of burying biomass of velvet beans (Mucuna pruriens) grown from December to April instead of the native grasses and shrubs was included at Misamfu ARS. The traditional Fundikila was tested in combination with the full rates of Compound D and urea fertilizers as practiced by farmers, while the modified Fundikila was assessed with full and half rates of the chemical fertilizers. The composted cattle manure was used singly or in combination with half of the recommended rate of Compound D fertilizer. A conventional treatment consisting of a maize monocrop with full rates of Compound D and urea was included in both sites as a control treatment. For the laboratory study all treatments except the control received an equivalent of 4,877 kg maize stover biomass /ha. Table 1 gives a summary of the treatments used on the two soils and the associated treatment codes.

Characterization of Soils and Organic Inputs Used in the Study
Air dried soil samples that had passed through a 2 mm mesh size were tested for their pH in 0.01 M CaCl 2 at a soil to solution ratio of 1:2.5 (McLean, 1973). Soil organic carbon was determined by the potassium dichromate wet oxidation method (Nelson & Sommers, 1982); total nitrogen (N) by the Kjeldahl distillation method (Jones, 2001) following digestion with concentrated H 2 SO 4 ; available phosphorus (P) by the Bray 1 method (Jones, 2001); and exchangeable potassium (K) using the 1 M ammonium acetate method (Chapman, 1965). The particle size distribution was determined using the hydrometer method (Jones, 2001) while the bulk density was determined using the core ring method (Casanova, Tapia, Seguel & Salazar, 2016).
The plant biomasses and composted cow manure were air-dried under a shed for a week and then oven-dried at 65°C for 48 hours (Campbell & Plank, 1998;Dalal, Sahrawat, & Meyers, 1984). The organic materials were then milled using a Thomas-Wiley Laboratory mill (Model 4) and sieved using a 2 mm mesh size sieve. Total nitrogen contents of organic materials were determined using the salicyclic-thiosulphate method described by Amin and Flowers (2004). The total K and P were extracted by dry-ashing and digested in 1M nitric acid (Jones, 2001). Concentrations of K in the extracts were determined on a Perkin Elmer Analyst 400 Atomic Absorption Spectrophotometer while concentrations of P were determined using a JENWAY 6305 UV/Visible spectrophotometer at a wavelength of 882 nm. The organic carbon content of the plant materials was determined by the Walkley and Black method (Okalebo, Gathua, & Woomer, 2002;Schumacher, 2002). The characteristics of the organic inputs are as shown in Table 3.

Set-Up of Incubation Study
The incubation experiment was conducted over 13 weeks to reflect the average growth period of early maturing maize varieties, which according to Oluwaranti, Fakorede and Badu-Apraku (2008) ranges from 80 to 95 days. The 13-week period also corresponds to observed recent duration of the rainy season, when microbial activity is not limited by the availability of water in the field in the agro-ecological region II of Zambia which is the major maize-growing region, in Zambia.
A 100 g of air-dried soil mixed with organic inputs with or without chemical fertilizer was moistened to 60% of total porosity of the soil (Samuel & Ebenezer 2014) and placed in a 1 L plastic jar with a screwable lid. The control treatment consisted of soil alone. A 100 cm 3 beaker containing 5 mL of 0.2M KOH was placed inside the 1 L jar, which was closed air tightly and kept at room temperature. The average room temperature was 22 °C during the incubation period. The amount of mineralized carbon was determined by titrating the alkali solution that absorbed CO 2 with 0.1 M HCl (Kirita, 1971). After every measurement of CO 2 emissions, the soils were moistened to maintain a moisture content corresponding to 60% of the total porosity of the soil. This was done by weighing each jar and adding the required amount of water with a pipette. The experiment was laid out in a Completely Randomized Design (CRD) with four replicates. Carbon dioxide emissions were measured and recorded weekly during the first nine weeks, and fortnightly thereafter.

Determination of Mineralized Carbon
The emission of carbon as a percentage of the total carbon present in the soil was determined using Equation 1 below: Where, C emit is the total amount of carbon emitted during the 13 weeks in mg kg soil -1 ; C appl is the amount of carbon applied as input in mg kg soil -1 ; C soil is the initial soil organic carbon in mg kg soil -1 .
The% CO 2 -C emission based on C applied to the soil is given by Equation 2.
Where, C treat is the cumulative C emitted from the treatment in mg kg soil -1 ; C cont is the cumulative C emitted from the control treatment in mg kg soil -1 ; C appl is the amount of carbon applied as input in mg kg soil -1 .
The carbon mineralization rate constant was determined from the first-order kinetics equation as described in Equation 3 (Stanford & Smith, 1972;Moretti, Bertoncini, & Abreu-Junior, 2017). Where, CO 2 -C cum is the amount of carbon mineralized during the incubation experiment; C to is the potential mineralizable carbon in the sample; k is the carbon mineralization rate constant.
The mean residence time (MRT) of organic carbon, which is defined as the average time that a particular element resides in a given pool at a steady-state (Six & Jastrow 2002) was determined using Equation 4.

Statistical Analyses
To determine whether there were significant differences among treatment means an Analysis of Variance (ANOVA) was carried out, and Duncan's Multiple Range Test was used to separate the treatment means when results of the ANOVA indicated the presence of significant differences among treatment means. The analyses were conducted using the SAS software version 9.0 for windows.

Properties of the Soil and Organic Inputs
The properties of the two soils used for the study are presented in Table 2. The soils were very strongly acid and had moderate levels of organic matter. The levels of total N were low on the loamy sand (LS) soil from Misamfu and very low on the sandy clay loam (SCL) soil from Msekera. The soils had higher levels of K than the critical value of 0.2 cmol kg -1 , while the available P was very low on the SCL and very high on the LS according to the interpretation by Hazelton and Murphy (2007). Selected properties of the organic inputs are shown in Table 3. Maize stover and pigeon pea biomass had the highest C content of about 57%, followed by Tephrosia, sunnhemp and velvet beans with about 54% C, while the traditional inputs-cattle manure and indigenous biomass, had the lowest C contents of less than 50%. The velvet bean biomass had the highest contents of total N, P and K while the maize stover had the lowest N, P and K. The legume biomasses had higher N content than the indigenous biomass and maize stover. The C:N ratios of the organic inputs were in the order: velvet beans < sunnhemp < tephrosia < composted manure < pigeon peas < indigenous biomass < maize stover.

CO 2 -C Emissions
The CO 2 -C losses from different treatments on the two soils are shown in Table 4. The total CO 2 -C evolved from the control treatments was significantly lower (p < 0.05) than those from treatments with organic inputs with or without chemical fertilizer in both soils. No statistically significant differences (p < 0.05) in the total CO 2 -C emissions were observed among the other treatments on the SCL. On the LS, the pigeon pea with full rate of chemical fertilizer (PpSF) had significantly higher total CO 2 -C emissions than the control and the conventional treatments (SF). There were no statistically significant differences in the total CO 2 -C emissions among the other jas.ccsenet.org Journal of Agricultural Science Vol. 12, No. 11; treatments on the LS. The effects of the treatments on CO 2 emissions were evident on the LS where significant differences (p < 0.05) in total CO 2 -C emissions and percentage C lost from the inputs were observed. Note. *Treatment means within a column followed by the same subscripts are not significantly different at 0.05 level of significance using Duncan's Multiple Range Test. *All treatment codes are as described in Table 1.

Mineralization Rate Constants (k) and Mean Residence Times (MRT) of the Organic Carbon
The k and MRTs of organic carbon for the different treatments on the two soils are shown in Table 5. The MRT of soil organic carbon ranged from about 5 to 59 seasons on the LS and from 13 to 32 seasons on the SCL. The control treatments had the lowest k on both soils and consequently had the highest MRTs. In contrast, the conventional treatments on both soils had the highest k and the lowest corresponding MRTs. Note. *All treatment codes are as described in Table 1.
The k and MRTs of organic carbon for groups of treatments are shown in Table 6. In both soils, the stability of soil organic carbon in the different treatments was in the order of Control > Traditional > Legumes > Conventional. Treatments with lower k emitted less carbon and had correspondingly longer MRTs.  Note. Treatment means within a column followed by the same subscripts are not significantly different at 0.05 level of significance using Duncan's Multiple Range Test.
* Traditional on LS = IndSF; traditional on SCL = ManS and ManSF½; legumes on LS = VvS, VvSF½, TepSF½, TepSF, PpSF½, PpSF; legumes on SCL = SunnSF½, SunnSF, PpSF½, PpSF, TepSF½ and TepSF. Table 7 shows the k and MRTs of C on aggregated treatments on the two soils with and without application of chemical fertilizer. Applying chemical fertilizers to the soil significantly (p < 0.05) increased the rate of C mineralization and significantly (p < 0.05) reduced the MRT of the organic carbon on both soils.

Effects of Organic Inputs on Carbon Mineralization
The addition of organic inputs with and without chemical fertilizer significantly (p < 0.05) increased CO 2 -C emissions compared to the control on both soils. The increases in C emissions compared to the control ranged from 81 to 129% on the loamy sand, and 18 to 34% on the sandy clay loam. This was as expected as the addition of fresh organic matter normally stimulates microbial activity resulting in increased C mineralization. These results corroborate the assertion of Hossain et al. (2017) that the nature and quantity of organic inputs applied to soils have an influence on the rate of soil carbon emissions. There were significant differences in the proportion of applied C lost from different treatments on both soils. The conventional farming (SF) treatment with maize stover biomass as the main organic input and a full rate of chemical fertilizer had the highest C losses of 17.2% on the LS and 22.5% on the SCL.
Composted cow manure, the traditional organic input used by farmers in the Eastern Province of Zambia had the lowest rate of C lost from organic inputs on the sandy clay loam soil. This could be attributed to the composting process which removed significant amounts of labile carbon from the manure leaving it with more stable carbon compared to the other organic inputs used the study. According to Coleman (1989), besides the C: N ratio, carbon mineralization from organic inputs is also influenced by their lignin and polyphenolic contents. On the loamy sand soil, the treatment with indigenous biomass (IndSF) traditionally used by farmers in the Northern Province of Zambia had the lowest rate of C loss per C applied of 5.4%. These results indicate that on both soils from the two agro-ecological regions, treatments representing traditional farming practices had lower losses of C from the soil, compared to treatments with legumes suggested by researchers as means of shortening fallow periods, and to the conventional farming practices that rely on the use chemical fertilizers.
As discussed above, among the treatments that received organic inputs, the treatments representing traditional farming practices on both soils had the longest MRTs of 16 seasons on the LS and 25 seasons on the SCL, while the shortest MRTs on both soils were those of treatments representing the conventional farming practices (SF) with mean values of 5 seasons on the LS and 13 seasons on the SCL. The longer MRT of organic C in traditional farming practices compared to conventional farming practices and to treatments with leguminous plant biomass is worth noting when considering sustainable management of soil organic matter and carbon sequestration in the soils. The observed shorter MRT of organic C in soils under the conventional farming treatments in this study corroborates the argument of Chirinda et al. (2009) that conventional cropping systems, that rely on chemical fertilizers and focus on short-term management of soil fertility may not be appropriate for improving soil organic matter and sequestering carbon. As indicated by Palm Gachengo, Delve, Cadisch and Giller (2000), a balance between the long-term maintenance of soil organic matter and soil carbon sequestration with the short-term crop nutrient is required for sustainable soil fertility management. The use of leguminous organic inputs with intermediate MRTs could help achieve the goal of balancing short-term soil fertility requirements with the long-term goal of maintaining soil organic matter.

Effect of Applying Chemical Fertilizer on Carbon Mineralization
The aggregated effect of chemical fertilizer application across treatments on C mineralization on the two soils shows that applying chemical fertilizer significantly increased carbon mineralization rates. The increases in CO 2 emissions due to fertilizer application were about 116% on the LS, and 48% on the SCL. There were corresponding significant (p < 0.05) reductions in the MRT of the organic carbon by 73% on the loamy sand soil, and by 33% on the sandy clay loam soil as a result of application of chemical fertilizer. The reductions in the MRT on sandy loam were from 35 seasons without fertilizer to about 10 seasons with chemical fertilizer and from 30 seasons without chemical to about 20 years with chemical fertilizer on the sandy clay loam.
No significant differences in C emissions were observed between the two rates of chemical fertilizers used in the study on the SCL loam. However, some general trends were observed on the two soils regarding treatments with the same type of organic inputs. On the loamy sand higher CO 2 -C emissions were associated with higher rates of chemical fertilizer, while on the SCL soil, higher CO 2 -C emissions generally occurred on treatments with lower rates of chemical fertilizer. The trend observed on the SCL is similar to the findings of Nyamangara, Mtambanengwe and Musvoto (2009) in Zimbabwe, where a reduction in C mineralization rates was observed upon addition of nitrogenous chemical fertilizer. According to Manzoni, Jackson, Trofymow and Porporato (2008) this could be due to the enhanced carbon-utilization efficiency of soil microbes associated with N fertilizer application which results in reduced CO 2 emissions.

Effect of Soil Texture on Carbon Mineralization
Although there were only two soil textural classes evaluated in this study, results in Table 6 show that the carbon mineralization rate was generally higher on the on the LS than on the SCL. The overall mean carbon mineralization rate constant k on the sandy loam of 0.0079 wk -1 was about twice that on the sandy clay loam of 0.0038 wk -1 , with corresponding MRTs of about 10 seasons for the loamy sand and about 20 seasons for the sandy clay loam. Longer MRTs of carbon on the SCL compared to the LS could be attributed to the higher clay content of the SCL of about 27% compared to 10% of the LS. Clay is reported to have a protective effect on soil organic matter which enhances the storage of carbon in soils, a phenomenon Rakhsh, Golchin, Agha and Alandari (2017) refer to as the "carbon storage effect of clay".

Conclusion
Based on the results and foregoing discussion it was concluded that the factors that significantly influenced carbon mineralization in the surface samples of the two Ultisols were the type of organic inputs, the addition of chemical fertilizer and the texture of the soil.