Combination of Compost and Mineral Fertilizers as an Option for Enhancing Maize ( Zea mays L.) Yields and Mitigating Greenhouse Gas Emissions from a Nitisol in Ethiopia

: Combined application of organic and mineral fertilizers has been proposed as a measure for sustainable yield intensiﬁcation and mitigation of greenhouse gas (GHG) emissions. However, fertilizer effects strongly depend on the soil type and still no precise information is available for Nitisols in Ethiopia. The study evaluated effects of different ratios of biowaste compost and mineral fertilizers (consisting of nitrogen (N), phosphorus (P), and sulphur (S)) on maize ( Zea mays L. Bako-hybrid) yields in a two-year ﬁeld trial. Soil samples from each treatment of the ﬁeld trial were used to estimate emissions of nitrous oxide (N 2 O), carbon dioxide (CO 2 ), methane (CH 4 ), and microbial activity in a 28-day incubation experiment with two moisture levels (40% and 75% water-ﬁlled pore space, WFPS). The application of fertilizers corresponded to a N supply of about 100 kg ha − 1 , whereby the pure application of mineral fertilizers (100 min) was gradually replaced by compost. Maize yields were increased by 12 to 18% ( p < 0.05) in the combined treatments of compost and mineral fertilizers compared to the 100 min treatment. The cumulative emissions of N 2 O and CO 2 but not CH 4 were affected by the fertilizer treatments and soil moisture levels ( p < 0.05). At 75% WFPS, the N 2 O emissions in the 100 min treatment was with 16.3 g ha − 1 more than twice as high as the treatment with 100% compost (6.4 g ha − 1 ) and also considerably higher than in the 50% compost treatment (9.4 g ha − 1 ). The results suggest that a compost application accounting for 40 to 70% of the N supply in the fertilizer combinations can be suitable to increase maize yields as well as to mitigate GHG emissions from Nitisols in Southwestern Ethiopia.


Introduction
In the context of climate-smart agriculture (CSA), soil management should balance the three CSA pillars of mitigation, adaptation, and productivity [1]. Considering type and amount of fertilizer as well as application time and technique can improve the productivity while reducing nutrient imbalances and nutrient losses from agricultural fields [2,3]. Applying organic fertilizers was shown to have positive yield effects in a broad range of cropping systems [4][5][6][7][8] and also has environmental benefits, as evaluated in a life cycle assessment study [9]. With regard to soil fertility, among others, organic materials were shown to enhance aggregation and stability of the soil and reduce erosion ( [10,11], suppress soil borne diseases [12], store nutrients [13], and improve biological functions [14,15]. Despite the advantages of organic fertilization, various studies agree that the combination of organic and mineral fertilizers can provide even better results concerning CSA than sole to investigate different ratios of compost and urea/NPS applied to a Nitisol regarding crop productivity and GHG emissions. In order to take into account the role of microorganisms in this respect, the activity of the DH was analyzed as well. The concrete objectives of this study were: (I) to quantify GHG emissions of compost and urea/NPS fertilizers as N source, (II) to identify the most suitable ratio of compost and urea/NPS in order to reduce the emissions of GHGs while having positive effects on maize yield, and (III) to evaluate if the ranking of the combinations regarding GHG emissions depends on soil moisture. Considering the state of the art, we hypothesized, that (i) combined N application with compost and urea/NPS to a Nitisol will produce less GHG emissions than the N application with only mineral fertilizers, (ii) the ratio of compost to urea/NPS influences GHG emissions and maize yield, and (iii) the GHG emissions will be higher when the water content in the Nitisol is higher.

Experimental Site and Treatments
The study consisted of two experiments-one field experiment to evaluate the maize yield and one incubation experiment to analyze the emission of GHG after application of organic and mineral N sources. The field experiment was performed at the research station of Jimma University College of Agriculture and Veterinary Medicine (JUCAVM) at an altitude of 1710 m above sea level in Southwestern Ethiopia (Eladale; latitude, 7 • 42'N; longitude 36 • 49'E) ( Figure 1). The research site is characterized as humid tropical climate with temperatures between 13 • C and 28 • C ( Figure 2). The annual minimum and maximum rainfall in the area is around 1200 and 2400 mm, respectively, whereby for our experiment, considerably higher rainfall occurred in 2020 than in 2019. The soil texture of the experimental field was silty clay loam with a pH of 4.98, organic carbon content of 2.4%, and total N of 0.22% (Table 1). According to the World Reference Base, the soil was classified as Nitisol, which was characterized as red, well-drained soil with a clay content of more than 30% and a blocky structure. In addition, the site was characterized by low P content, and high iron and aluminum content [42,43]. The soil of this site was also used for the incubation experiment. Table 1. Properties of soil and compost used for the incubation experiment (N = 4, Mean ± standard error).

Parameters Biowaste Compost Soil (5 cm Soil Depth)
Org. C (g kg −1 ) 92.9 ± 0.8 24.0 ± 4.0 N (g kg −1 ) 12.0 ± 0.7 2.2 ± 0.1 S (g kg −1 ) 2.2 ± 0.08 0.5 ± 0.02 Ca (g kg −1 ) 25.1 ± 1.5 2.3 ± 0.2 P * (mg kg −1 ) 718.2 ± 7.5 2.1 ± 0.1 K * (g kg −1 ) 1.9 ± 0.02 0.4 ± 0.1 Mg * (g kg −1 ) 1.3 ± 0.04 0.2 ± 0.01 Cu (mg kg −1 ) 39    In order to assess the impact of the fertilizers on maize yield, GHG emissions and microbial activity in soil, different ratios of mineral, and organic fertilizers were applied to the soil. The fertilizer application based on previous recommendations for N and P supply in the maize cropping systems under similar growing conditions [44,45] and 100 kg N ha −1 and 33.3 kg P ha −1 were defined as the optimum amount of nutrients to be supplied with mineral fertilizers in this experiment (=100% mineral fertilizer). In the other treatments, the nutrient supply with the mineral fertilizers was gradually replaced by a biowaste compost. The maximum amount of compost applied to the field was 7 t ha −1 (dry weight, 1.2% N and 0.072% P) (=100% compost). Compost applications in this range were previously reported to be suitable for maize production in this region [46][47][48]. In total, seven treatments, including control without fertilizers, were established. The nomenclature followed the percentage of mineral fertilizer applied, starting with 100 min (=100% mineral fertilizers), followed by 80 min, 60 min, 50 min, 30 min, and 100 comp (=100% compost) ( Table 2). As a mineral fertilizer commercially available, NPS (19% N-38% P-7% S) and urea (46% N) were applied. A compost based on locally available materials such as residues from vegetable plants, animal manure, and wood ash was prepared following the standard procedure of Tulema et al. [49]. The soil and compost were analyzed regarding nutrient concentration and physical characteristics ( Table 1). The pH of compost and soil were measured using a pH meter (pMX 3000) in 1:2.5 compost/soil: CaCl 2 ratios. The organic C was measured by the Walkley-Black oxidation method and the total N by the micro-Kjeldahl method. The total element concentrations of the compost and soil were measured after microwave digestion (aqua regia) by using inductively coupled plasma optical emission spectroscopy (ICP-OES, Perkin Elmer). The available phosphorus (P), potassium (K), and magnesium (Mg) contents were measured in a spectrophotometer (P) or flame photometer (K, Mg) after extraction with calcium lactate (C 6 H 10 CaO 6 ·5 H 2 O) solution. The cation exchange capacity (CEC) was determined by Chapman [50]. In addition, soil texture was determined using the hydrometer method [51] and bulk density was determined using a core sampler method [52].

Determination of Maize Yield and Agronomic N Use Efficiency
Maize was cultivated for two growing periods in a randomized complete block design with seven treatments (see Section 2.1) and four replications. The Bako hybrid (BH_661) variety was used, because it is the most commonly used by farmers in the study area. In February 2019 and March 2020, twelve plants per row were planted at 0.75 m inter-row and 0.30 m intra-row spacing with a plot size of 4 m by 2.5 m (10 m 2 ) ( Figure 1). No irrigation was applied during the experiment as the maize crops were sown during the main growing season with sufficient rainfall. Weeding and other agronomical practices were applied manually using labor forces. During maturity (July 2019 and August 2020), the two central rows in each subplot were harvested in order to determine the maize grain yield [53]. The grain samples were oven-dried for 72 h at 70 • C in order to get dry weight.
Beside the yields, agronomic nitrogen use efficiency (ANUE) for each treatment was also calculated, as described by Baligar and Fageria [54].
where GYf is the grain yield of the N fertilized plot (kg), GYu is the grain yield of the unfertilized plot (kg), and Nap is the quantity of N applied with compost or mineral fertilizer (kg).

Incubation Experiment and Greenhouse Gas Measurement
Composite sampling of the topsoil (0-5 cm) of the unfertilized plots was performed assuming farmers usually incorporate fertilizers at the surface of the soil. The soil was homogenized, air-dried, sieved (2-mm pore size), and immediately stored at 4 • C until the beginning of the incubation experiment. Larger (>2 mm) surface aggregates and belowground plant matter were removed beforehand. The laboratory incubation experiment was conducted at the University of Rostock (Germany) with the Nitisol from the field experiment in Ethiopia, applying the same fertilizer treatments as in the field experiment in four replications ( Table 2). Two hundred grams of air-dried soil was filled into a 1000 mL jar, the soil aggregates were evenly compacted to a bulk density of 1.2 g cm −3 (to mimic the natural soil pore spaces), and pre-incubated at 25% WFPS and 25 • C for 15 days. Pre-incubation of soil samples is suggested before starting GHG measurement to settle and standardize the soil microbial community following the disturbance of sampling and sieving [55]. After the pre-incubation, fertilizers were applied and the moisture contents were adjusted to 40% and 75% WFPS in order to mimic the dry and rainy season. The fertilizer addition was adapted to the soil volume in the jars, whereas 100 kg N ha −1 corresponded to 33.3 mg N kg −1 soil. The mineral fertilizers and fresh compost were evenly spread and homogenized with the dry soil. The jars were incubated constantly at 25 • C in the dark in a completely randomized order. Loss of water during incubation was compensated by adding H 2 O demin on a daily basis.
Gas samples were collected each day from the first day to the 13th day. For the first three days, gas samples were collected three times a day and for the remaining ten days, once a day. This approach considered the higher production of GHG immediately after fertilizer application [56]. Gas samples from the headspace of the sealed jars were collected by 60 mL syringes, transferred to evacuated vials, and the gas concentrations of N 2 O, CO 2 , and CH 4 were measured with a gas chromatograph (GC-2014, Shimadzu, Kyoto, Japan) equipped with an electron capture detector for the N 2 O analysis, and a flame ionization detector (FID) for the CO 2 and CH 4 analysis. Jars were opened for 20 min to maintain aeration after every measurement and closed until the next measurement. The loss of moisture was re-adjusted to maintain the chosen moisture content throughout the incubation [35]. Gas fluxes were calculated by assuming a linear increase in gas concentrations inside the incubation bottles over time. The GHG fluxes were calculated area-based by considering the surface of jars filled with soil. We measured the height and diameter of the jar, which was filled with soil (bulk density 1.2 g cm −3 ) and calculated the surface of the jars occupied by the soil. The diameter of the jar was determined by considering the average of the upper and lower surface of the jar. The soil emissions were estimated based on the rate of linear GHGs increase in the container headspace over time from a given amount of soil. Gas fluxes (g ha −1 day −1 ) were calculated by the following equation of Comeau et al. [57] for soil heterotrophic respiration assessment using minimally disturbed soil microcosm cores. The conversion factor of ppm/ppb N 2 O, CO 2 and CH 4 to mg N 2 O, CO 2 and CH 4 was calculated with Equation (2): where Cf = conversion factor of ppm/ppb of N 2 O, CO 2 and CH 4 to mg N 2 O-N, CO 2 -C and CH 4 -C; P = air pressure (kPa); Mm = molar mass of C (12) or N (28); R = gas constant (8.314); T = incubation air temperature (K). Finally, the N 2 O, CO 2 , and CH 4 fluxes were computed on an area basis. The N 2 O-N, CO 2 -C, and CH 4 -C per unit of area were calculated using the following equation: where Flux = linear gas efflux in incubation container on soil area basis (g CO 2 -C m −2 h −1 ); Cf = conversion factor of ppm CO 2 to mg CO 2 -C m −3 ; t = incubation time (hours); C = change in gas concentration during the incubation period; Hs = headspace volume of the incubation jar (m 3 ); 10 −6 = conversion factor from µg to g; Area = area of the microcosm surface (m 2 ). The cumulative GHG emissions were calculated by summing the daily fluxes [58]. The final results were converted from µg where N 2 O EF% = N 2 O emission factor; N 2 O ENI = N 2 O emission in treatments with N input; N 2 O EC = N 2 O emission in the control treatments with no N addition; N input = the amount of N added to the soil. The GWP was determined for fertilizer rate and type using the following equation [60]: where GWP = global warming potential (kg CO 2 eq. ha −1 ); N 2 O = is the amount of N 2 O (kg ha −1 ); CO 2 = the amount of CO 2 (kg ha −1 ); CH 4 = the amount of CH 4 (kg ha −1 ); 298, and 25 = GWP coefficients to convert N 2 O and CH 4 , respectively, to CO 2 equivalents [61].

Dehydrogenase Enzyme Activity
Dehydrogenase enzyme activity (DHA) was determined following the modified method based on [62]. During this procedure, 0.8% triphenyl-tetrazolium-chloride (TTC) was added to 1 g of soil and incubated for 24 h at 37 • C. As a result of DHA, TTC was reduced to triphenyl-formazan (TPF) by most microorganisms. TPF was extracted with acetone after incubation and measured with the spectral photometer (Specord 40, Analytik Jena, Germany). The activity was expressed as 1 g TPF per g soil released within 24 h (1 g TPF g −1 24 h −1 ). Soil samples were taken three times during the incubation period and analysed for DHA. The first sample was taken immediately after the incorporation of different fertilizers. The second sample was taken after seven days of incubation.

Statistical Analysis
The normality of residuals was assessed by the Kolmogorov-Smirnov normality test [63], and it was shown that our data was approximately normally distributed. Oneway analysis of variance (ANOVA) was used to determine the effect of different fertilizer types on GHG emissions, N 2 O EF, GWP, and DHA. The interaction effect of moisture content and fertilizer types was analyzed by a two-way ANOVA. The mean values were determined by using the Tukey multiple-comparison test by using SPSS (22.0 version). Pearson correlation analysis was used to determine the relationship between C inputs and emissions of N 2 O, CO 2 , CH 4 , and N 2 O EF.

Maize Yield and Agronomic Nitrogen Use Efficiency
The maize grain yields were measured in two consecutive years in an on-station experiment ( Table 3). The maize yield depended on the experimental year as well as on the fertilizer treatments. Averaged across the fertilizer treatments, the yields were lower in the second year, which is linked to unexpected rainfall and windy weather conditions. Relatively high yields were found for the combined fertilizer treatments. This was especially true for the 60 min treatment with significantly (p < 0.05) higher yields (9.9 Mg ha −1 ) than the control without fertilizers or the single fertilizer applications in both years. For example, averaged across both years, the 60 min treatment had 9.8 Mg ha −1 , which was 18% higher than the 8.3 Mg ha −1 in the 100 min treatment. The combined treatment with only 80 min was not found to be more effective than the 100 min treatment. No differences were found between the 100comp and 100 min treatments. In accordance to the yields, a combined application of compost and mineral fertilizers increased the ANUE of maize, and for the 60 min and 50 min treatment, about three times higher values than in 100 min treatment were measured (18.2 and 13.9 vs. 5.5 kg grain per kg N applied).

Daily Greenhouse Gas Emissions
The emission of GHGs was estimated in an incubation experiment with different soil moistures. Generally, GHG emissions were lower in dry soil (40%WFPS) than in wet soil (75% WFPS). High GHG emissions were measured on the second and third day of incubation. After the sixth day, the emission clearly decreased and remained at a similar level until the end of the experiment.
The N 2 O fluxes varied depending on the treatments, although a treatment effect was not found on each day of the experiment ( Figure 3, Tables A1 and A2). Relatively high fluxes were observed on the second day for the 60 min (3.17 g N 2 O-N ha −1 day −1 ) and the 100 min (2.71 g N 2 O-N ha −1 days− 1 ) treatments in wet soil. On day three to five, the treatment with 100% mineral fertilizer stood out with about three to five times higher N 2 O emissions than the control (0.44 to 0.81 g N 2 O-N ha −1 day −1 ) and the 100comp treatment (0.54 to 0.91 kg N 2 O-N ha −1 day −1 ) (p < 0.05). Under dry conditions at 40% WFPS, the differences between the fertilizer treatments were less pronounced, though significant at several days of measurement with high values found again for the 100 min treatment. Lowest N 2 O emissions throughout the measurement time were usually observed in the control and 100comp treatments under both moisture conditions.  Similar to N2O, we usually observed greater daily emissions of CO2 from amended soil than from the control soil in the first days of measurement (Figure 4, Tables A3 and A4). The peaks were observed on day two and three for both moisture levels. The fertilizer treatments showed different patterns depending on the soil moisture. For 75% WFPS, the 100 min treatment showed high values which were significantly higher than the control and the 100comp treatment (p < 0.05) and tendentially higher than all other fertilizer treatments on day two with 2.27 kg CO2-C ha −1 day −1 and three with 2.20 kg CO2-C ha −1 day −1 . Under dry conditions, the 40comp treatment emitted more CO2 than the control and the 100comp treatment (p < 0.05) and tendentially more than all other fertilizer treatments on days two with 1.17 kg CO2-C ha −1 day −1 and three with 1.21 kg CO2-C ha −1 day −1 . With running incubation time, as for the wet conditions, again the 100 min was found to release relatively high amounts of CO2. : Control (no input); 100 min: 100% mineral fertilizer N (100 kg N ha −1 ) and P (33.3 kg P ha −1 ), 80 min: 80% mineral fertilizer + 1.4 t ha −1 compost; 60 min: 60% mineral fertilizer + 2.8 t ha −1 compost; 50 min: 50% mineral fertilizer + 3.5 t ha −1 compost; 30 min: 30% mineral fertilizer + 4.9 t ha −1 compost, and 100comp: 100% compost (7 t ha −1 compost). * indicates significant differences among the treatments (Tukey HSD test, p < 0.05). Error bars indicate the standard error of the mean (n = 3). Similar to N 2 O, we usually observed greater daily emissions of CO 2 from amended soil than from the control soil in the first days of measurement (Figure 4, Tables A3 and A4). The peaks were observed on day two and three for both moisture levels. The fertilizer treatments showed different patterns depending on the soil moisture. For 75% WFPS, the 100 min treatment showed high values which were significantly higher than the control and the 100comp treatment (p < 0.05) and tendentially higher than all other fertilizer treatments on day two with 2.27 kg CO 2 -C ha −1 day −1 and three with 2.20 kg CO 2 -C ha −1 day −1 . Under dry conditions, the 40comp treatment emitted more CO 2 than the control and the 100comp treatment (p < 0.05) and tendentially more than all other fertilizer treatments on days two with 1.17 kg CO 2 -C ha −1 day −1 and three with 1.21 kg CO 2 -C ha −1 day −1 . With running incubation time, as for the wet conditions, again the 100 min was found to release relatively high amounts of CO 2 .

Cumulative Greenhouse Gas Emissions, Global Warming Potential, and Nitrous Oxide Emission Factor
Over the 28 days of incubations time, the cumulative N 2 O and CO 2 but not CH 4 emissions were affected by the fertilizer treatments and moisture levels (p < 0.05) ( Table 4). In both moisture levels, the application of 100% mineral fertilizers resulted in higher (p < 0.05) N 2 O emissions than the application of 100% compost under wet (156% more) and dry (31% more) conditions. The different ratios of compost and mineral fertilizers rarely resulted in significant differences of N 2 O emissions, but tendentially more N 2 O was emitted when the ratio of mineral fertilizers increased. Similar statements can be made for CO 2 , with low emissions in the control and 100comp treatment.
The N 2 O emission factor (N 2 O EF) depends mathematically on the N 2 O emission and consequently; as for the N 2 O emissions, the N 2 O EF values were found to be higher (p < 0.05) in the 100 min treatment than in the 100comp treatment at both moisture levels ( Table 4). At 75% WFPS, the N 2 O EF was in the 100 min treatment more than ten times higher as in the 100comp treatment (0.28 vs. 3.85%).
Clear differences between the mineral and the compost treatments were also found for the global warming potential (GWP) ( Table 4), which is mathematically based on the emissions of the three GHGs. At 75% WFPS, the GWP in the 100 min treatment was with 15.0 kg CO 2 eq. ha −1 higher than all other treatments, except the 60 min treatment. With increasing ratios of mineral fertilizer, there is a trend of increasing GWP values under both soil moisture conditions.

Dehydrogenase Enzyme Activity
The dehydrogenase (DH) activity hardly varied between the two sampling dates on day 1 and day 7 ( Table 6). The control without any amendments had with about 65 µg TPF g −1 24 h −1 usually lower DH activities than the treatments with fertilizer application. The ratio of organic to mineral fertilizers was not decisive for the activity of the DH. In contrast to the other characteristics, the soil moisture was also not relevant for the activity of the DH (p < 0.05).

Higher Maize Yields and Agronomic Nitrogen Use Efficiency in the Combined Fertilizer Treatments
The results of our study showed that higher yields and ANUE were found in the combined application of compost (compost N: 40-70%) and mineral fertilizers (mineral fertilizer N: 30-60%) compared to other treatments. Positive yield effects of combined applications of organic and mineral fertilizers were also found in other studies under varying growing conditions [29,[64][65][66], and often this was attributed to an improved soil structure [67,68], intensification of biological processes in soil [68], higher water storage capacity [64,65], and higher cation exchange capacity [69] (see also Introduction).
The rainfall pattern in the second year was less suitable for plant productions than in the first year, which resulted in lower yields, even if the differences were not particularly great (8.26 vs. 9.39 Mg ha −1 , averaged across all treatments). Even under less-favorable conditions, higher yields and ANUE were found when mineral fertilizers were combined with compost. The results suggest that under extreme weather conditions and stronger yield depressions, which will probably occur more frequently in Ethiopia in the future [70], compost application can contribute to maintaining yields which has been demonstrated for agricultural and horticultural crops [8,9,68].
Another advantage of compost application is the supply of plant nutrients. The mineral fertilizer application in this experiment only consisted of N, P, and S, while composts contain all plant nutrients, albeit in differing concentrations, depending on the original material [65]. And although the site was not described as being deficient in nutrientsapart from the low P content-the application of various nutrients could have supported plant growth. However, despite of all these positive impacts of the compost application described, the treatment with 100% compost application was (at least tendentially) agronomically less suitable than the fertilizer mixtures with 40 to 70% of the N provided by compost. This can be explained by the availability of mineral N. The majority of N in composts is bound in stable organic compounds [71], and assumed 35% of N released in the year of application [48]. This can hamper maize growth, especially in periods of high N demands during the plant development [72]. Our results showed that shares of 40 to 70% N from compost in the fertilizer combinations are most suitable for maize growth under these growing conditions. The C:N ratio in these combinations were 2.7, 3.5, and 5.1, respectively. The addition of only 20% compost with a C:N ratio of 1.2 was obviously not enough to benefit from the organic matter supply.
The experimental field in our study was well managed in previous years, including adequate fertilizer management. The total content of N (about 2.2 g kg −1 , which corresponds to about 5000 kg N ha −1 in the upper 30 cm of soil) as well as the content of organic C (about 24 g kg −1 ) in the soil were rather favourable and in the range of other Nitisol sites with proper soil management [73,74]. A fallow was applied at this site one year before our study started. These facts can explain the relatively high maize grain yield with the control treatment without fertilizers. In contrast to the fertilizer mixtures, we observed a non-responsiveness of maize yields to the application of sole mineral fertilizer (100 min) in both experimental years. This is partly related to the fertility of the soil, as also shown in a study by Negassa et al. [75]. However, we believe that the non-responsiveness of maize yields to mineral fertilizers in this study was also attributed to low availability of P. The bio-available soil P content was with 2 mg kg −1 very low at the beginning of the study, which can be reasoned with the acidic soil conditions (pH= 4.9) and high iron content, which usually reduce the availability of P [42,48,76]. As organic matter in soil can reduce P fixation, it contributes to a better availability of P for crops [77], which can explain the positive effect of compost in the mixtures. Unfavorable soil or weather conditions were cited in 68% of the surveyed agricultural fields in sub-Saharan Africa as a reason why mineral fertilizer use did not increase maize yield [78]. This indicates that the multiple interacting factors affecting crop yields are difficult to quantify in general, and that a careful evaluation of fertilizer practices for each cropping site is necessary to ensure returns on fertilizer investments.
As described above, the organic material in the compost is stabilized during the composting process. Although it was not tested in our experiment, fresh organic materials such as farmyard manure may have different effects, because of the faster decomposition of organic matter and cycling of nutrients [79]. For areas with same soil type, recent results showed an advantage of compost over farmyard manure [29,65].
Higher yields in the combined treatments were related to higher NUE, which is of great importance in Ethiopian agriculture. The results of the two growing seasons showed that the 30 min, 50 min, and 60 min treatments had with 12.7 to 18.2 kg grain per kg N about three times higher NUE than the other treatments. Thus, the results indicate that combined fertilizer application having 40% to 70% of the total N from compost can be a suitable measure to stabilize maize yields and increase nutrient efficiency in the study area.

Mitigation of GHG Emissions by Compost Application
Fertilizer types and rates had a significant effect on N 2 O and CO 2 emissions and GWP from the Nitisol soil in the incubation experiment, although their influences varied in dependence on the soil moisture. The 100 min treatment resulted in higher N 2 O and CO 2 emission than the control or 100comp, especially under wet soil conditions (75% WFPS).
High amounts of available N usually intensify the denitrification process and the N 2 O emissions [32,55,80] (see also introduction). In our study N 2 O emissions were reduced when the mineral fertilizers were combined with compost. This can be explained by the replacement of the mineral N by organic N, and consequently by an initial microbial immobilization of N [80,81] and/or slow release of N from the organic part in the ratios. Furthermore, compost application can increase the abundance of denitrifying microorganisms and thus favoring the complete denitrification and production of dinitrogen gas instead of N 2 O [14]. The availability of N also plays a role in CO 2 emissions. Due to a reduction of available N, the microbial activity and decomposition of native soil organic matter usually decreases [82].
Besides the availability of N, the interactive effects of N and C supply also influence the emission of GHGs [83]. High microbial activity due to C supply with organic fertilizers can result in an intensification of microbial-induced processes [84]. However, negative correlations were observed between C input and N 2 O and CO 2 emissions in our incubation experiment. The increased C:N ratio with increasing portion of compost in the mixtures could be the reason, which resulted finally in a limitation of N for microbial activities despite the higher C stocks. This is also supported by the microbial activity (DHA), which was not increased with increasing rates of compost in comparison to 100 min.
The reduction of CO 2 and N 2 O emissions after the application of organic material was also highlighted in other studies with other amendments, such as crop residues [35,80] or manure [85]. Amendments with high C:N ratio like straw (up to 100:1) resulted in low N 2 O emission and is also an option to replenish soil organic matter [80] but may hamper the N nutrition of crops [86].
The majority of the gases were emitted during the first days after incubation, which was also shown in other incubation studies, as for Ferralsol [35] and Vertisol [80]. The results indicate the risk of high GHG emissions in a relatively short period after fertilizer application during the main crop growing season in Ethiopia when the soil is relatively wet. For sites with a high availability of N, the addition of organic material with a high C:N ratio like crops residues could be a good means under these conditions to reduce N 2 O emissions [35,80].
Generally, for soils rich in C and N, higher CO 2 and N 2 O emissions can be expected. The C and N content of our soil (C, 24 g kg −1 and N, 2.2 g kg −1 ; see Table 1) were in the range of other Nitisols in Ethiopia [42] but higher than the majority of other soils in Ethiopia and other East African regions [87]. The results of other incubation experiments can also be interpreted in this context, with very low N 2 O emissions from Ferralsol with little or no N input [35] and high N 2 O emissions in fertilizer treatments with N application of >200 kg ha in Vertisol [80].
The fertilizer types and rates did not affect CH 4 emissions in either moisture level and no correlations between C or N-input and CH 4 emission were found in our incubation experiment. The emission of CH 4 comes primarily from fields under flooded conditions [24,88] with higher water content than in our experiment. In this context, Brembong et al. [89] described soils with normal WFPS as very effective CH 4 sinks.
Results of management strategies from other studies have to be considered with caution, as GHGs emissions vary depending on the physicochemical properties of soils [32,83]. For instance, a higher clay content of soil is usually related to higher water retention and higher emission of CH 4 and N 2 O, which can explain that relatively high GHG emission were often found for Vertilsols due to their tendency to become waterlogged [80]. Nyamadzawo et al. [32] reported comparably low N 2 O emissions for Lixisol (about 0.5 kg ha −1 ) and Inceptisol of (about 1.5 kg ha -1 ) for different fertilizer treatments during a cropping season in Zimbabwe, which was attributed to the soil texture with high content of sand and low water retention. Another important soil characteristic regarding GHG emissions is soil pH. A low pH value is not suitable for most microorganisms involved in CH 4 and CO 2 metabolism [90], and from acid soils like Nitisols (the pH of soil in our study was 4.9) potentially lower CH 4 and CO 2 emissions can be expected than from neutral soils.
The effect of moisture was especially important for N 2 O emissions in the 100 min treatment, which were much higher under wet (75% WFPS) than under dry soil conditions. In wet conditions, anaerobic bacteria use NO 3 − as an electron acceptor during microbial oxidation and release N 2 O through the process of denitrification [19]. The CO 2 emissions were also generally higher under wet conditions and the effects of the treatment were more pronounced than under dry conditions. The proportion of the pores filled with water and soil aeration affect CO 2 emissions [91,92] and CO 2 emissions from soil can increase linearly with the soil water content until saturation point, after which the emissions decrease again. For most soils, the saturation point for CO 2 emission is >70% [36,93]. In this study, no effects of soil moisture were found on DH activity. Probably the range of moisture was still relatively suitable for microbial activity and with about 70 to 80 µg TPF g −1 TS 24 h −1 the activity of DH was relatively high (e.g., in comparison to Stagnic Cambisol [84]). Clear inhibitions of microbial activities can be found for very low water contents of air-dried soils [24].
Compost was shown to be a suitable amendment considering GHGs and maize yield in our study. However, like mineral fertilizers, compost is also limited in Ethiopia, especially in the area where organic resources are used for another purpose such as fuel, food for animals, or construction material [65,94,95]. As different ratios of compost and mineral fertilizers in the mixtures were found to be suitable to reduce N 2 O and CO 2 emissions and increase maize yield (see Section 4.1), upon the availability of resources, the proportions of these types of fertilizers can be set flexibly in a certain range around 50:50. Beside the evaluation of the fertilizer effect, the ANUE might also be a good indicator to predict GHG emissions [80] and accordingly, the ANUE was found to be highest in the 30 min, 50 min, and 60 min treatments (see Section 4.1).

Conclusions
The results of this study showed that the combined application of compost and mineral fertilizer can be an option for enhancing maize yields and mitigating GHG emissions from Nitisols in Southwestern Ethiopia. Utilization of compost as fertilizer can be especially suitable during the wet season and might be an option to mitigate negative yield effects of extreme weather conditions, which will probably occur more frequently due to climate change in Ethiopia. To verify the results of the GHG emissions from the incubation experiment, further investigations should take place at the field level.