There are many pathways that lead to nutrient loss, especially nitrogen, from composting manure heaps. These include gaseous and leaching losses (Dewes, 1994). There is a need to apply collection and storage management strategies that minimise these losses so that efficient nutrient cycling can be achieved.
In the Central Kenya Highlands, surveys have shown that due to population pressure, per capita land size is diminishing and farming is tending towards more intensive systems where livestock and crop production are integrated (Woomer & Muchena, 1996). One of the main advantages of integrated farming is the opportunity to convert by-products and wastes from one activity into inputs for another (McIntire et al, 1992). Intensive livestock production leads to an increase in use of off-farm fodder and high value feeds, such as animal concentrates, requiring additional financial resources. Manures derived from such diets could be expected to be of high quality if nutrient losses are minimised through better collection and storage strategies. Such strategies are essential if smallholders are to reap the full benefit from the extra resources invested in a manner that is cost effective.
To investigate the effects of livestock and manure management on the quantities of nutrients available for application to soils after composting, a series of experiments was carried out at the Animal Production Farm of the Kenya Agricultural Research Institute (KARI), National Agricultural Research Centre (NARC) Muguga, Kenya. This station in located 25 km west of Nairobi at 1o13' 53.0" S and 36o 38' 1.1" East and an altitude of 2096 m asl. Muguga experiences a bimodal rainfall at peaks between March-May and October-December with annual averages of 986 mm (average of 20 years from 1980). Mean annual maximum temperature ranges between 19.8 and 22.0 oC and minimum ranges between 4.7 and 11.5 oC over the same period.
Staal et al (1997) report that nearly half of farmers in Kiambu used purchased fodder as their main source of feed and that 70% fed concentrates on a regular basis, a value confirmed by a survey in this study. Regular purchase of feed thus represents a major route for the importation of nutrients onto the farm. Overall, the net effect on feedings strategies in these systems will lean towards replacement of the less rumen degradable crude protein contained in fodder with less recalcitrant N in concentrate protein supplements. As intensification becomes more dominant, new feeding practices are simultaneously evolving. The unstable and ever increasing cost of conventional concentrate supplements, which are sometimes unavailable when required, has necessitated the acceptance and use of poultry litter as a cheaper alternative. This is usually mixed with conventional concentrates or used solely to cut down on cost. Since poultry litter contains substantial amounts of urea (a non-protein N), which is absorbed in the rumen, and most of which ends up in the urine, it is important that the urinary N is trapped and efficiently recycled for crop/fodder production. Otherwise it would be most vulnerable to leaching and volatilisation losses. Use of organic materials as bedding in the animal sheds or as intentional additions during composting of cattle manure are methods by which urinary N could theoretically be conserved (Dewes, 1995).
An experiment to examine the effect of feeding diets more rich in nutrients, and conserving urine on manure nutrient content was conducted between the third week of August 1998 and the fourth week of February 1999 allowing two weeks for diet acclimatisation before starting data collection. The study was conducted in two phases; (1) a 61 day collection and accumulation of manure and (2) a composting period lasting 90 days where no additional `new' manure was added. Friesian steers were used for the production of animal excreta (faeces and urine). The experiment comprised four treatments in a 2 x 2 factorial design with concentrate offered at two levels and manures derived from these diets handled in two ways with three replicate steers per treatment.
Manures containing urine were obtained from steers housed so that excreta could be collected separately and measured. Steers for manure collection without urine were housed in a roofed, concrete floored barn where the urine flowed to a drain leaving behind the dung, which was collected and weighed every morning. After weighing and subsampling of faeces and urine, the manures were stored in a roofed, concrete floored barn. Each treatment was replicated three times thereby comprising three steers. The faeces and urine samples were kept under refrigeration at between 2 and 4 oC before the weekly faeces collections was bulked, dried at 65 oC for 72 h and ground to pass through a 2 mm sieve. Total nitrogen was analysed by the modified Kjeldhal oxidation method where salicylic acid is added during digestion so as to include nitrate-N and nitrite-N.
The steers in each treatment were balanced for liveweight. This was intended to remove the variability in waste production from the different sized animals as the measurements were expressed as a function of the liveweight. The steers were weighed fortnightly and the fodder and concentrate offers adjusted accordingly. Mean liveweight (LWmean), calculated by averaging steer weights at 0, 15, 30 and 45 days, was used in calculations when expressing parameters in respect to units of liveweight.
Quality parameters of the locally available feeds used in this study, namely: napier grass, commercial concentrate and poultry litter are given in Table 3. Concentrate was purchased from an animal-feed retail store at the cost of Kshs. 850 per 70 kg bag (Kshs. 100 = UK£1). The poultry litter was obtained from a neighbouring smallholder farm. The price of poultry litter in this case was Kshs. 400 per 90 kg bag after sieving through 5 mm mesh openings. Napier grass was obtained daily from the KARI's NARC Muguga estate farms. Wheat straw, added to the excreta at approximately 5g/kg LWmean/day to simulate the addition of bedding, was purchased from a retail store at Kshs. 125 per bale.
The steers were fed on a basal diet of napier grass equivalent to 2% liveweight (LW) in dry matter. Steers receiving high concentrate diets were supplemented with a mixture of 0.5% of LW as dry matter of conventional commercially available dairy meal + 0.5% of LW as dry matter of poultry litter sieved to pass a 5 mm screen. Steers for the low concentrate treatments were provided with 0.5% LW dry matter of dairy meal only. The steers were provided with mineral supplements and water ad libitum. ANOVA was carried out using statistical programme MS Excel version 5.0. All the LSDs were calculated at 5% significance level.
High concentrate diets produced significantly more urine than the low concentrate diets. Daily faecal dry matter (DM) production ranged between 40.5 and 57.0% of total dry matter intake representing between 1.0 and 1.8% LWmean. There was a significant effect of concentrate on the amount of total faecal dry matter production per kilogram liveweight, being higher in the high concentrate diet than in the low concentrate diets. Dependence of faeces and urine on dry matter intake has been reported in a study conducted by Kirchgessner & Kreuzer (1986) who observed that slurry production and dry matter content increased with dry matter or concentrate intake as well as with increasing milk yields.
Table 3. Chemical composition of feeds and straw used in the experiment
There was no significant difference between the urine collection methods nor between high and low concentrate in the amount of total DM added to the heaps over the 61 days accumulation phase. However, manures derived from high concentrate diets showed significantly lower dry matter loss than those derived from low concentrate diets during the composting phase. Total dry matter after 90 days composting was significantly higher with high concentrate than low concentrate diets. Where urine was included, total dry matter after composting was higher than when urine was excluded. One possible explanation for this observation is that addition of urine may have created more anaerobic condition in the heaps than when urine was excluded leading to retardation in the composting process. Similar observations have been reported by Dewes (1996) and Eghball et al (1997).
The N intake ranged from 0.300 to 0.458 g/kg LWmean/day while N excreted ranged from 0.075 to 0.209 g/kg LWmean/day and from 0.033 to 0.055 g/kg LWmean/day in faeces and urine, respectively. Total N excreted (urinary + faecal N) ranged between 36 and 58% of the total N intake. Between 21 and 31% of total N excreted was contained in urine while the rest was excreted in the faeces.
Significant linear relationships were observed between the daily N intake (Nintake) and the daily N excreted in faeces and urine, with the urine better correlated to N intake than the faecal N. The N excreted in the faeces and in the urine could be described by the first order linear equations:
Nfaeces = 0.368Nintake - 0.0112 R2 = 0.534
Nurine = 0.098Nintake + 0.0068 R2 = 0.745
The overall N excreted in faeces plus urine was described by the equation:
Ntotal excreted = 0.446Nintake - 0.0045 R2 = 0.667
Similar relationships have been reported by E. Kebreab (Centre for Dairy Research, Reading University, UK, pers. comm., 1999) and Kirchgessner & Kreuzer (1986) who observed that a linear relationship existed between N intake and faecal N excretion. As the crude protein increased in the diets so did the faecal N excreted. However, these two reports suggest that urine N was better described by a first order exponential fit. Their observation was based on experiments where dietary crude protein levels ranged between 11.5 and 17.0% and N intake ranged between 300 and 600 g/day (Kirchgessner & Kreuzer, 1986; Kebreab, pers. comm., 1999).
The difference in urinary N output between that reported by previous authors conducting work in a northern country-context and that observed in this experiment could be explained by the fact that, in this study, the N intake, ranging between 60 and 180 g/day, was far below what previous authors had offered the dairy animals in their investigations. This means that the diet offered might have been just sufficient to provide energy and protein such that the steers were able to utilise most of the consumed N for rumen microbial biomass production and body maintenance with modest amounts excreted in urine. In fact, Mason (1969) observed that high fibre diets such as clover-rye grass or hay and oat straw resulted in significantly higher undigested dietary N in faeces than concentrate supplemented diets in sheep. High fibre diets encourage enhanced rumen microbial activities culminating in rich faecal N excretion of bacterial origin.
Changes in the amounts of N during the manure production cycle are shown in Table 4. The high concentrate diet did not result in significantly higher faecal N excretion but did lead to a significantly higher urinary N excretion. Significantly more N was present in the urine treated manures than in the urine-free manures at the end of the accumulation phase of the experiment.
At the end of composting phase there was significantly higher N in the urine treated manures and in the manures derived from the higher concentrate diets. The interaction between the effect of quantity and quality of diet and presence of urine in the manures was not significant. N losses during the composting phase ranged between 2.0 and 30.1% with significantly higher losses occurring from manures derived from low concentrate diets than from those obtained from high concentrate diets. However, the overall N loss during collection and composting ranged between 14.8 and 43.4% and was not significantly different among the different treatments.
The fact that a high concentrate diet resulted in both higher faecal N excretion (as shown by the regression data in Section 3.1.4, though not by the ANOVA data from Table 4) and higher N concentration in urine is a vital observation in view of the farming practice in the smallholder farming systems of the central Kenya Highlands. The N contained in urine is most prone to losses by volatilisation and leaching. It is crucial to note that studies have shown that 75% of urinary N is in the form of urea. Urea and uric acid from urine are rapidly hydrolysed to ammonia and carbon dioxide gases by enzymes from faecal bacteria (Tveitnes, 1993).
This study suggests that one way of conserving urinary N when making manure composts could be by mixing the urine with faeces and organic materials such as wheat straw to absorb the urine. If, for instance, a steer of 400 kg liveweight was used to make manure for one year, and assuming that it was fed on the types of diets under consideration in this study, then the following scenario of N conservation would be envisaged after composting of the manures:
* high concentrate diet would result in 61.0 kg of N if urine were included (HC+U)
* high concentrate diet would result in 36.0 kg of N if urine were excluded (HC-U)
* low concentrate diet would result in 43.2 kg of N if urine were included (LC+U)
* low concentrate diet would result in 31.2 kg of N if urine were excluded (LC-U)
In the Kenya Highlands, however, wheat or barley straw is an expensive organic material entering the nutrient cycling chain primarily as livestock feed and rarely used as bedding. The benefits of conservation of N observed in this experiment may be outweighed by the cost of obtaining effective amounts of bedding. The rate of straw used in this study, for instance, was high at 1.8 kg/kg LW/yr, which amounts to 720 kg of straw per 400 kg animal per year. If this amount is required to effectively conserve the urinary N then this approach may not be practicable for the smallholder farmer in these farming systems. However, with no treatment lacking straw in this experiment, it is not possible to determine the contribution made to N conservation made by addition of straw.
An experiment was carried out which sheds further light on the effect of manure management practices on nutrient cycling, in which cattle were fed on a low concentrate diet in which maize stover was the fodder rather than napier grass, and in which only the limited available feed refusals, rather than purchased straw, were available as bedding or additions to the compost heaps. Napier grass contained between 14 and 47% (mean 32%) dry matter, whereas maize stover was relatively drier and contained between 59 and 73% (mean 75%) dry matter. The two types of fodder contained different concentrations of nitrogen, ranging between 0.7 and 0.9% (mean 0.8%) in maize stover and 1.0 and 1.2% mean (1.0%) in napier grass suggesting that napier grass was a better diet for crude protein than maize stover.
The objectives of this part of the work were to study the effect of combining cattle excreta (faeces and urine) and organic materials originating from the bedding and feed refusals by different strategies in order to conserve nutrients upon composting.
3.2.1 Experimental details
Twenty steers were used for faeces, urine and feed refusal production in the collection phase of the experiment with each treatment comprising four animals (replicates). The steers were approximately balanced for weight for each of the treatments so that the total weight of the animals at the beginning of the study were similar. The five methods of manure collection were:
* faeces + urine + feed refusals mixed on the floor of the cow shed by the animal (S)
* faeces + urine + feed refusals mixed manually (F+U+FR)
* faeces + feed refusals mixed manually but without urine (F+FR)
* faeces and urine only, mixed manually (F+U)
* faeces only, without urine or feed refusals (F)
The steers were fed on maize stover obtained from the KARI-Muguga estate farm, at 2.5% initial liveweight (that is, the animal weights at the beginning of the study, LWi) dry matter, and 2 kg dairy meal concentrate, purchased locally, split into two 1 kg rations fed in the morning and afternoon, and provided with minerals and with water ad libitum. Steers used for treatments F+U+FR and F+U were enclosed in metabolism units where faeces and urine could be collected separately. Steers for treatment S were enclosed in cubical sheds with feed refusal (maize stover) bedding on a concrete floor while for treatments F+FR and F steers were kept as in treatment S but not provided with bedding, and raw manure was collected daily and recombined appropriately. Treatment F resulted in a high moisture content product as a result of inevitable contamination with urine on the floor of the cow shed.
The amount of feed given to the animals and what they rejected was recorded daily. A composite sample of the feed was obtained daily and at the end of each week this was bulked and ground for laboratory nutrient analysis as described in Section 3.1. No other measurements were taken for treatment S until the end of the accumulation phase. For the other treatments, additional measurements included mass of faeces and the mass of urine produced by the animal daily, and the amount of feed refusal that went into the composting heap. These three components, that is, faeces, urine and feed refusal, were daily recombined appropriately according to treatments for 60 days. In order to estimate the amount of faeces and urine produced for the treatments whose animals were not kept in the metabolism units, pro-rata measurements based on food intake were calculated. At the beginning of the composting phase, the amount of manure going into the replicate compost heaps was weighed. After 84 days composting, and before taking the manures to the field the heaps were weighed. Nutrient mass balances were calculated for N to ascertain losses occurring during accumulation and composting. Analyses of materials was carried out as described in Section 3.1.
Accumulation and composting of the manures were carried out on the concrete floor of a roofed cow shed. The manure heaps were stored in 1 m3 chicken wire cages with 10 mm openings mounted on steel frames. On the inside, a finer plastic netting of 2 mm openings was used that would retain collected waste, prevent the manures coming into contact with the metal frames, and yet still allow free air circulation. The concrete floor below the cages was lined with non-porous plastic sheet, extending to about 20 cm high up the sides of the cages, to minimise leaching from the heaps.
There was no significant treatment dry matter differences in the amount of feed intake, faecal production or in the total amount of waste produced during the 60 days accumulation period. This is as expected as the animals were approximately balanced for weight by treatments and the data is expressed as per kg LW. It was observed that of the 2.4-3.3% (mean, 2.8%) of initial body weight of dry matter consumed as maize stover and concentrates, the steers produced 0.73-1.1% (mean, 0.8%) of initial body weight dry matter as faeces.
Loss of dry matter during composting was highest with the methods containing feed refusals and significantly lower in F and F+U, probably reflecting a greater rate of aerobic composting in the larger, more loosely packed heaps with refusals. Overall dry matter loss between collection and the end of composting was 29-51% for all methods except F+U which was very low (15%). F+U formed a wet heap that rapidly became capped with a dry layer, probably reducing the rate of further decomposition. Highest dry matter return to the soil after composting was with S (0.46 kg/kg LWI). However, this would have been even higher with F+U if the refusals were added directly to the soil without composting (F+U manure (0.4 kg) + feed refusals (0.23 kg) = 0.63 kg/kg LWI) as well as with F alone (F manure (0.29 kg) + feed refusals (0.31 kg) = 0.60 kg/kg LWI). Lowest return of dry matter to the soil was with F+U+FR and F+FR, both giving 0.35 kg/kg LWI.
On average, 28 and 18% of the N input as feed was recovered in faeces and urine, respectively (Table 5). Of the total N excreted, faecal N contributed between 47 and 76% (mean, 61%) while urinary N ranged between 24 and 53% (mean, 39%). Greatest loss in nitrogen during the accumulation phase was observed in heaps that had high moisture contents from the addition of urine (S, F+U+FR and F+U). These lost 39, 37 and 37 %, respectively, of the N collected during the accumulation phase, whereas the urine-free F and F+FR lost only 24 and 1% N, respectively during the accumulation phase.
During the composting phase, N losses were 39, 37, 33%, respectively, for S, F+U+FR and F+FR, but only 22 and 14%, respectively for F and F+U, suggesting that manures with maize stover refusals manually added had the greatest N losses. Overall, N losses ranged between 34 and 63% with F+FR resulting in the lowest loss. S and F+U+FR showed the highest overall N losses of 63 and 61%, respectively.
The use of bedding as a means to conserve N in manure is a traditional practice in mixed farming in many climates and agricultural systems. However, in this experiment, the addition of feed refusals was either insufficient or ineffective in conserving N added to the manure heap as urine in the S and F+U+FR treatments, since in both of these cases very large losses of N occurred compared with F+FR despite the presence of the feed refusals. In fact, the combination of F+FR very effectively conserved the N in the faeces and feed refusals. In the S and F+U+FR treatments, not only was the additional N due to urine completely lost but, with these treatments, even lower total N in the final compost (5.5 and 5.6 g N/kg LWi, respectively) was obtained than with the urine-free F+FR (6.5 g N/kg). In the F+FR treatment, 4.7 g N/kg LWi from urine was also available for direct return to the soil, possibly as a liquid slurry feed to perennial fodder crops, resulting in a 11.2 g N/kg LWi return to the soil.
It is not clear to what extent this urine N would contribute to crop production if applied as a liquid, since loss of N from urine applied to the soil can be both rapid and extensive (Powell et al, 1998). Reports have indicated between 10-80% N loss from fresh manure application depending on the weather conditions, state of manure (slurry or solid) and the method of application (Klausner & Guest, 1981; Kemppainnen, 1989; Smith & Chambers, 1993). To arrest these nitrogen losses, proper strategies for handling and applying manure to the fields need to be employed such as stabilising the N with organic materials so that it can mineralise and be available when the crop needs it (Beauchamp, 1986; Myers et al, 1994).
However, in the F+FR manure scenario, any additional gain from the direct use of urine would be a net benefit above the alternative of adding the urine to the manure heap. A further scenario, of composting the faeces alone and adding urine and feed refusals to the soil would also theoretically add considerable N to the soil (7.2 g/kg LWi), as would the direct addition of all materials to the soil daily. However, with these approaches, the more rapid loss of N from volatile sources, leaving high C:N residues might lead to higher levels of soil N immobilisation than with the well rotted, mature composts.
Similarly, although F+U gave a greater N accumulation after composting than F, F+U did not conserve significantly more or less N after composting than F+FR did, because of the high loss of nitrogen with this treatment, with the latter treatment offering the possibility of using the urine directly, as discussed above.
The treatments F+U+FR and F+FR can broadly be compared with low concentrate diets with or without urine added to the manure heap (LC+U and LC-U) in Section 3.1. In this experiment F+FR retained most N at the end of the composting period while in the study in Section 3.1 LC+U retained more N than LC-U. This difference may indicate that the wheat straw used in Section 3.1 is better at conserving urinary N than the maize stover obtained as feed refusals in this experiment. Not only was the wheat straw added at a 50% higher rate on a dry matter basis, but the finer wheat straw appeared to absorb urinary N more effectively than the more coarse maize stover stems. It can tentatively be suggested that urinary N can be partially conserved if composting with adequate and absorbent organic material (e.g. LC+U) or by the somewhat anaerobic F+U combination, However, under aerobic conditions with only maize stover feed refusals available as an organic addition, it is difficult to conserve urinary N. These tentative conclusions needs verifying since the two experiments were conducted with different conditions and feed sources.
