Geographical location of the study area

This study was conducted in the city of Kisangani. Kisangani is the main city of the former Eastern province (today province of Tshopo) in the northeastern part of the Democratic Republic of the Congo (Figure 1). The average geographic coordinates of the area are 0°36’ latitude north, 25°13’ longitude east, with an altitude varying between 376 to 470 m. The region of Kisangani is located in the bioclimatic zone of evergreen tropical rainforest. The zone is characterized by a Af₁-like climate according to the classification of kôppen- Geiger (Upoki, 2001; Thienpondt, 2016; MFAN, 2018). The temperature of the coldest month is above 18°C and the rainfall average of the driest month is above 60 mm. Generally, the temperature revolves around 25.3°C in March and 23.5°C in August with a yearly average of 24.4°C. The relative air humidity varies between 79.1 % in February and 87.3% in July with an annual average of 84.0%. Precipitations are in abundance throughout the year with an annual average of 1782.7 mm with two equatorial maxima around October and April and two solstitial minima around January and July. The annual average number of rainy days is around 155. However, with climatic perturbations observed currently at the global level, the region experiences the occurrence of a small dry season around January (Upoki, 2001;  MFAN,  2018).  The insolation is quite high in the region (Makondambuta, 1997). The yearly average turns around 5.4 h per day, with high intensity between 10 and 14 H especially during the driest months (January). It reaches 1945 h per year which correspond to 45% of total radiation. In contrast, there are months during which it declines up to 36%. Figure 2 depicts the average trend of rainfall and heat prevailing in the region over the trial period in 2013.

Experimental design

The experimental layout was a mono factorial randomized complete block design (RCBD) including three replications with three treatments each. The testing plots had an expanse of 1.20 sq. metres each (1 × 1.20 m) and were separated in between by pathways of 0.6 m, whereas the distance in between blocks  was  1 m wide. On each experimental unit, 110kg of raw materials was piled (Figure 3). A cleared empty distance of 2 m wide was left around the perimeter of the experimental area in order to demarcate it from the surrounding vegetation. The whole experimental design was sheltered with a shed made of palm leaves to protect it from heavy rains. The studied factor was the biomass of local weed plants with three treatments namely, T1: 110 kg of P. maximum biomass, T2: 110 kg T.laxum and T3: 110 kg P.javanicabiomass.

Heap filling and handling

The composting heap filling was implemented gradually: to begin with, three woodcuttings were placed vertically  from  the  bottom  in order to facilitate the ventilation within the heaps being formed. Then, one first layer of 15 kg of wilted biomass (rich in carbon) of the species intended to the experimental unit (treatment) was primarily laid down. This was thereafter followed respectively by 25 kg of fresh biomass of the same species (with high nitrogen content), followed by 10 kg of the mixture pig dejections and rice husk, plus 5 kg of rice husk ash. After a good watering, the above piling steps were repeated in the same way up to reach 110 kg of composting feedstock, followed by a further heavy watering. The whole heap was finally covered with a layer of Panicum maximum straw (± 3 cm) to keep up the heat. To be closer to the real world of small farmers, the first punched biomass turning was given one month after piling the heap and the second two months later and by three months, the compost was ready to be used (FAO, 2003).

Compost biomass weighing and chemical analysis

After three months of composting, the first operation consisted of collecting from the resulting compost, composite samples (9 in total) for the upcoming chemical analyses. Subsequently, compost masses derived from different species were weighed and compared, at first to their initial mass (110 kg) depending on treatment and blocks, as well as among themselves in order to determine which species has provided the highest amount of compost mass. However, it should be noticed that heat measures having prevailed within the different treatments throughout the composting period were not tracked out, given the study considered that all the treatments were subject to the same range of temperature viewing the general typical tropical weather conditions prevailing during the research period. The 9 composite samples were dried in open air (away from the sun) for two weeks in the laboratory of soil sciences of the Faculty Institute of Agronomic Sciences of Yangambi in Kisangani and sifted with a 2-mm sieve squared mesh and put in small bags.

The water and KCl pH were determined using a PH-220s brand pH-meter with a reference electrode, on a soil/water and soil/KCl suspension in 1:2 proportions (w/v). The total organic carbon (TOC) was indirectly determined subsequent to the determination of total organic matter (TOM) from which the content was deducted using a simple ratio between the TOM and the conversion factor (1.724), that is, the percentage of TOC = % TOM / 1,724 (Schumacher, 2002). The total organic matter of samples was determined through weight loosing method (Alongo, 2007). This latter involves oven calcination of 10 g of soil sample at 375 ⁰C during 16 h. The amount of total organic matter was determined by the ratio between the current and the initial weight (% TOM = Wa - Wb/Wa x 100). Besides, the total Kjeldahl nitrogen (TKN) was determined by the standard method (digestion-distillation-titration) using a Kjeldhal digestion flask (Mohammad and Flowers, 2004; Indrayani and Surendra, 2014); whose total nitrogen percentage was deducted as follows:

Where: meq N = 14.10^-3; N₁ = actual Normality of acid (0.01N); V₁ = total volume of diluted digest; V₂ = titre value (volume) of HCl used (100 ml); SampleW = weight of the aliquot (5 g) and V₃ = volume of H₃BO₃ (2%), meq = Milliequivalen.

Finally, the quick perchloric acid (HClO₄ 1N) method (with ethylic alcohol) was used (Rao et al., 1998) to determine the total potassium (K) amount (K₂O % = 100 (0.33996 weight) 25/5).

Statistical analysis

The one-way analysis of variance (ANOVA) was conducted using Genstat 12^th edition to estimate the amount of variability among different samples of compost. Means were separated using Fisher’s LSD test at 5% probability level (Ibanda et al., 2018). Boxplots were plotted with the mean of the three replicates for all parameters measured using R statistical package (Version 3.6.1, 2019-07-05).

Analysis of variance for chemical composition of composts

The results of the mass loss of different biomasses indicated no significant difference by the completion of study, which implies that the mineralization rates were almost similar as shown by theanalysis of variance (Table 1). There was significant difference (p≤0.013) in the composts regarding their pH H₂O and a high significant (p≤0.001) difference of composts in their potassium (K) content. In contrast, the weight, pH KCl, total organic carbon (TOC) nitrogen in percentage and the percentage of potassium showed non-significant differences among them (Table 1).

Mean performance for different compost treatments

The averages of final compost mass produced by the completion of composting process at the 90^th day were almost in contradiction with the supposed initial nitrogen content of species; nitrogen being the main agent for a substantial mineralization of organic matter. In general, the narrow the C/N ratio, the fast and substantial is the mineralization, resulting in low organic matter (compost mass) contribution (Quansah et al., 2001; Musinguzi et al., 2013). In fact, P.javanica(T3) produced the highest mass of compost (76 kg /110kg,  that  is  30.9%  of  initial mass loss), which is a little bit inconsistent with its supposed initial nitrogen content as a legume (Figure 4a). It was followed by T.laxum(T2, 70 kg/110 kg that is 36.3% of mass loss) and P. maximum (T1) with 64.3 kg/110kg that is 41.5% of mass loss. These results contradicted those of other studies by Douglas etal.(1980), who found greater initial mass loss from crop residues with the highest N content. In contrast, the results confirm research conducted by Quansah et al.(2001) in Ghana, who observed the highest initial mass loss in Chromolaenaodorata, which contained the lowest amount of N. Nevertheless, this relative incompatible situation of organic matter production may be due to the immaturity of the two other species (P. maximum and T.laxum) during their collection time (Parr and Papendick, 1978). Thus, if only considering the produced averages of compost mass (Figure 4a), it may be concluded that, of the  three  weed  biomasses  composted,  P. javanica(T3) performed better and may constitute the best compost candidate despite of its initial lowest C: N ratio as a legume.

The results showed that compost prepared from P.javanica(T3), despite its supposed lower initial C/N ratio, demonstrated a high carbon percentage (17.05 %) against 13.97% for T. laxum (T2)and 11.13% for P. maximum (T1) (Figure 4d). This result once again contradicted the hypothesis supported in earlier studies according to which, high nitrogen content results in faster and substantial mineralization producing lower organic matter, which is the carbon matrix (Quansah et al., 2001; Musinguzi et al., 2013). The same assumption of the biomass immaturity collected for the composting process mentioned above may also apply to justify this. In contrast, all the three biomasses proved increased values of compost carbon content. In fact, several studies suggested a soil critical level in carbon content  at  2%  in temperate climate, which is about 3.4% of soil total organic matter (Howard and Howard, 1990; Huber et al., 2008). In tropical soils, carbon thresholds of less than 10 g kg^-1 (less than 0.1 %) without adding organic matter can be too low and can result into disequilibrium in nitrogen supply to plants. However, Musinguzi et al. (2013) suggested that, this critical level is closely related to local physical and climatic conditions undergone by the soil as well as the crop itself. They stated that soil organic carbon (capacity) = f (climate, topography, texture, disturbance, inputs). Moreover, Baize (1993) on his side, categorized soil organic carbon contents of < 0.60% as very low, 0.60 - 1.25% as low and 1.26 -2.50% as medium.

The finding of this study showed that the resulting composts from the three studied species, given the high level of carbon percentage demonstrated can constitute excellent source of soil amendment in the region. In contrast, Musinguzi et al. (2013) confirmed their difficulty to establish the highest soil critical level of carbon content beyond which crops cannot positively respond given that, both the lowest soil carbon content as well as its highest content may negatively impact crops and environment (Figure 4d).

The pH values were all near neutral (7) as well for pH H₂O (Figure 4b) and for pH KCl (Figure 4c), ranging from 6.63 to 6.99 for pH H₂O and 6.37 to 6.39 for pH KCl; but compost from P. maximum (T1) and T.laxum (T2) which were shown to be significantly (p≤0.013) different with high pH than P. javanica would be preferred when focusing on improving soil related acidity.

Landon (1991) categorized pH values as follows: > 8.5 = very high, 7.0 - 8.5 = high, 5.5 - 7.0 = medium and < 5.5 = low. Therefore, in view of these results, all biomasses in the study had demonstrated a medium pH. Therefore, the moderate pH value near the neutral demonstrated throughout all the three biomasses clearly supported the appropriateness of the three species for compost production in order to amend soils of the region whose pH are mainly acidic with an average water pH and KCl (1N) 4.9 and 3.7 respectively (Kombele, 2002).

Despite the non-significant variation found in nitrogen content among treatments, the mean performance for nitrogen content showed that P. javanicacompost(T3) had the highest N rate with an average of 0.44%. It was followed by the compost made of P. maximum (T1)(0.39 %) and finally that of T. laxum (0.35%) (T2) (Figure 4e). The high nitrogen content in P.javanica (which is a legume) compost was not surprising given the supposed high initial nitrogen rate in legumes in general. However, as it is known, nitrogen is a key nutrient for crop growth and is always qualified as limiting factor. Furthermore, soil nitrogen concentration has impacts on other nutrient content such as C. The critical threshold of N concentrations in tropical soil is suggested to range between 20 to 25 mg g^-1 (0.2 to 0.25%) below which net N immobilization from soil would  be  expected  (Janssen, 1993; Quansah et al., 2001). Considering Janssen’s baseline, all composts from the three species which demonstrated high nitrogen content may be qualified of good quality in terms of nitrogen content. Consequently the three species could help in increasing nitrogen levels of most soils in the region of Kisangani for enhancing crop production.

The mean performance of percentage of potassium showed that the highest rate of potassium was found in compost from P. maximum (T1) (41.36%). T.laxum (T2) and P.javanica (T3) produced 20.39 and 15.86% of potassium respectively (Figure 4f), confirming the high significant (p≤0.001) differences found among them in Table 1. Result of potassium content related to P.maximum compost was relatively different from those found by Quansah et al. (2001) in Ghana. They found a quite low percentage 0.71 mg g^-1 of potassium in compost made from P.maximam which is equivalent to about 0.071% of K. According to Kaddar et al. (1984), the critical concentrations of exchangeable soil K in the tropics at which deficiencies occur range up to 0.45 meq K/100 g of soil. However, this level may differ depending on soils and crop type. Thus, although the three different composts obtained from the biomasses studied differed significantly in terms of their characteristics in potassium content; still, the level of the nutrient they demonstrated was quite satisfactory compared to the lowest threshold required for potassium in tropical soils.Therefore, the three species can constitute good source for compost to control potassium related issues in tropical soils but with preference for P. maximum.

This study, which is one of the pilot researches aimed at transforming the baffling problems related to the abundant weed biomasses especially that of P. maximum, T. laxum and P.javanica, often cause of crop yield loss. The goal was to transform them into advantage to enhance soil fertility and increase crop productivity. The study demonstrated that these common weeds in the region of Kisangani contain underutilized resources in the form of nutrients for soil fertility management. The total organic carbon, nitrogen and potassium were quite at high level to restore soil fertility for a good agricultural productivity for small farmers in a region where organic matter such as poultry manure and cow dung are scarce and the access to inorganic fertilizers is limited due to high cost. The study revealed also that, the pH values of all the composts produced from the three weed species were closer to neutral on average, able to fix the recurrent issue related to soil acidity in the region, which constitutes one of the leading roots to the low soil productivity and transient fertility. Finally, the research showed that the species used were able  to   produce   fairly  good  compost  mass  and  in  a relatively short time, despite prevailing climatic conditions supposed to be responsible of substantial mineralization of organic matter and which could  lead to the loss of most key elements for crops through natural pathway. However, there might be limitations as to the quantities of weed biomass a farmer may obtain from his field. So, the suggestion is that farmers keep on their fields, for instance on a small portion of field, in situ weed biomass intended to compost production instead of destroying them using fire as it is the common practice in the region. Moreover, it is recommended that the pH H₂O and pH KCl, the total organic carbon, nitrogen and potassium contentsof weed species be measured prior to the experiment.

The authors have not declared any conflict of interests.