Nitrogen and phosphorus output of livestock excreta
Final report, Defra project WT0715NVZ
Report submitted 03 April 2006
1
OutputN&P_ks5
Nitrogen and phosphorus output of livestock excreta – WT0715NVZ
1.0 Introduction.
Farm manures are a major source of nitrogen (N) and phosphorus (P) pollution. Excreta from livestock make a significant contribution to atmospheric ammonia emissions and N and P losses from agricultural land to surface and ground water systems (leading to eutrophication and acidification). The UK Government is committed to reducing nutrient emissions from agriculture, to enable compliance with a number of EU Directives (e.g. Nitrate Directive, National Emissions Ceilings Directive, Water Framework Directive).
Data on manure N outputs/excretion have been published in the UK and elsewhere in Europe and these form the basis for setting environmentally acceptable stocking rates, as part of Nitrate Vulnerable Zone (NVZ) Action Programmes. However, the origins of the data have not always been transparent, and it is not always clear whether the values quoted represent excreted N or output (i.e. adjusted for ammonia and other gaseous losses during housing and storage). The latter are more correctly indicative of N loading to land. Furthermore, many of the data used in estimating N ‘output’ by livestock in the past were derived from feeding or production systems that were less intensive than those in operation today. As a result, data derived from those studies may not be applicable to current systems of production.
The N output figures published as part of the Defra “NVZ Action Programme” and within the Water Code are among the most important and far-reaching, environmental statistics in the UK, as such ‘standards’ control farm livestock stocking rates and may be used within planning and design requirements for manure storage. It is therefore important that such standards are soundly and scientifically based. A recent review of nutrient (N and P) output standards, from a number of EU Member States, highlighted (i) the inconsistency of reporting N output data and (ii) differences from currently used ‘standards’ (Cottrill et al, 2005). This current review of standards now builds on the data accumulated within the European Commission (EC) study and, based on the nutrient balance approach recommended in that review and other data from industry consultation and recent research. The National Ammonia Reduction Strategy Evaluation System (NARSES) mass-flow model will be used to evaluate likely gaseous N losses following excretion (Webb et al, 2004). This will allow the expression of ‘N output standards’ for livestock excreta as (i) ex-animal (i.e. as excreted), (ii) ex-housing and (iii) ex-storage (i.e. as at land spreading).
Phosphorus excretion by farm livestock is a major cause of eutrophication of freshwater systems, and Defra has identified the reduction in P losses from agricultural land as a key objective in improving water quality. There are no current ‘official’ standards for P excretion, other than the guideline values published in the "Organic manures" section of RB209 "Fertiliser recommendations" (Anon, 2000). However, as with N, the EC has indicated a desire to set limits on P loadings to land and has called for recommendations for a common approach to establishing livestock P excretion standards.
2.0 Objectives
The overall study objective was to derive reliable manure N and P output standards for the major categories of farm livestock in England and Wales. The technical and scientific aims included:
1)To review inconsistencies in the expression of current N output standards;
2)To review and update N output standards for farm livestock in the UK;
3)To evaluate the potential impact of ammonia losses during the housing and manure storage phases, on the proposed N output standards;
4)To propose standards of P excretion by farm livestock.
3.0 Methodology and approaches
3.1 Research and industry review
Contact was made with key researchers at centres in the UK and Ireland. Livestock management and commercial feeding practices have been reviewed in consultation with leading livestock advisers and specialist consultants. In addition, data gathered from some commercial pig and poultry producers, including typical inputs (feed consumed) and outputs (products and, in some cases, manure) have been included in the review process. After initial contact with researchers and industry contacts, the availability and quality of suitable data were explored, in some cases, via face-to-face meetings. A number of Defra and Defra-industry funded projects have provided data directly or indirectly relevant to nutrient balances in livestock production systems or manure outputs; these research contracts have included: Defra SP0119, SP0129, WA0632, NT2003, NT2006, NT2402, “AGEWEAN” (SLP LINK IS0212); also the “Pig finishing systems research programme” funded by Defra and BPEX and the SNIFFER research contract UKPIR01 on “Waste streams from intensive livestock production in Scotland and N Ireland” (Sharp and Smith, 2005).
A number of methods exist for quantifying nutrient excretion by farm livestock. These include:
- Direct measurements with livestock
- Direct measurements with manure
- Metabolic/mechanistic models
- Input-output measurements
Direct measurements with livestock may provide the most accurate measure of nutrient excreted, but require either total collection of faeces and urine or reliable markers for spot sampling. This is an expensive and time-consuming method, and the values obtained can only be applied to similar types of livestock (breed, age, sex, growth rate etc) and diets.
Estimates of nutrients excreted in manure by direct measurements and analysis of the manure may be achieved at less cost (in terms of the number of samples and analyses required). However, the amounts of manure produced are difficult to quantify, and obtaining representative samples for analysis can be particularly difficult. This approach also suffers from the fact that the results obtained are only really applicable to the particular factors and conditions prevailing during the period of observations and sampling. As discussed elsewhere, N losses via gaseous ammonia emission occur very rapidly following excretion and need to be considered in relation to the point of assessment and to what extent NH3 may impact on the measurements or estimates of manure N output.
Mechanistic/metabolic models aim to describe in detail the processes that go on within the body of the animal. An integral component of these models is a prediction of the partitioning of nutrients into excreta. These models are, however, complex and because they tend to work on an individual animal basis are generally not suited to whole farm application.
The approach used for estimating nutrient excretion by farm livestock was to assume that the amount of nutrients excreted in faeces and urine is the total consumed minus the nutrient content of products (milk, meat, eggs, liveweight gain etc). This applies for both N and P and may be represented as:
N or P excreted = N or P intake – N or P in products
In this study, the approach was applied at the level of the individual animal, with an adjustment made according to the length of the production cycle and for non-use of the stock accommodation, to provide an annual output factor per “animal place”. The latter is necessary to allow for non-productive time needed for cleaning and re-stocking the housing.
Detail of the information required is provided in the section outlining the results obtained for each of the livestock categories. Nutrition specialists provided typical input and performance data on which to base the calculations and, where possible, industry data was also considered. In this way, representation of appropriate classes of livestock types and productivity levels was achieved, e.g. milk production in dairy cows and liveweight range.
3.2 Validation
Empirical assessments of manure output on a number of commercial poultry units provide some scope for validation of the nutrient balance estimates, alongside the more detailed research data from Defra funded and other research sources outlined above.
3.3 Accounting for ammonia emissions
The mass-flow model, NARSES (Webb et al., 2004) has been used to estimate the likely losses of N via ammonia emission and denitrification, following excretion. In particular, losses during housing and manure storage are of special interest for the major livestock classes. This will allow the expression of ‘N output standards’ for livestock excreta as (i) ex-animal (i.e. as excreted), (ii) ex-housing and (iii) ex-storage (i.e. as at land spreading). The estimated N outputs after gaseous N emissions provide some guidance with typical manure N content at land spreading (i.e. total N and readily available N content of manure), providing a validation check in comparison to the “standard data” (Chambers, 2005).
4.0 Nitrogen balance in livestock
4.1 Dairy cattle
Although a number of factors influence the amount of N excreted by dairy cows, the one having the greatest effect is N intake, which is a function of the amount of feed consumed and its’ N content. Many feeding studies have demonstrated the close correlation between N excreted in manure and N consumed. This is illustrated in Figure 4.1, which combines data from 25 studies in Europe and North America involving dairy cows fed a wide range of diets (Cottrill, unpublished).
Figure 4.1 The relationship between N excreted in manure and N consumed by dairy cows.
As discussed above, a number of methodologies have been proposed for estimating N excretion of N by dairy cows. The mass balance approach is widely used, but in order to be useful it requires reliable information on both input (N intake) and output (N products) factors:
N manure = N intake – N products(Equation D1)
N intake is calculated as feed intake (as kg of dry matter, DM) x N content of the feed, while N products includes the N in milk, liveweight gain of the cow and the N in the calf. Of all the variables, feed intake is the most difficult to estimate. The amount of feed consumed by lactating dairy cows can vary considerably during lactation, being influenced by both animal and diet-related factors. Of these, the size (weight) of cow and milk yield (or stage of lactation) have the greatest effect. Other factors such as the quality of feed available, feed constituents and method of feeding are also known to influence feed intake.
A number of attempts have been made to predict the dry matter intake (DMI) of dairy cows. In the report to the European Commission (ERM, 1999), dry matter intakes of 0.85 kg feed DM/kg FCM[1] for both small and large breeds of cows were assumed. This figure was based on estimates of DMI and milk production in Denmark and The Netherlands, from observations reported in 1995/96 and 1991/92, respectively. The report of the European Communities (2002), which formed the basis of the ERM report, included live weight and milk yield as determinants of production[2], although the origins of the coefficients are not stated.
Several equations for predicting DMI have been developed in North America for use in ration formulation by farmers and nutritionists, although only a few have been subject to peer review. Some of these require detailed dietary information that makes them inappropriate for this exercise. However, data on the live weight of cows and their milk yield are usually available, and a number of equations have been published (e.g. Chase and Sniffen, 1985[3]; NRC, 2001[4]) in which intake is predicted using these variables. For a 600 kg cow with lactation yield of 6900 kg FCM, the equations of ERM, Chase and Sniffen and NRC predict average daily dry matter intakes of 17.6, 18 and 19.9 kg, respectively.
All of these equations require information on the weight of the cows. This can very considerably, depending principally on the breed and age of the animal. In estimates of N excretion (below) a typical weight of 600 kg has been assumed. In practice, estimates of N excretion using these equations are not particularly sensitive to the weight of the cow. In the estimates of N excretion using the ERM mass balance model (Table 4.1), a 10% difference in liveweight resulted in a 3% difference in N excretion.
The N content of the diet can be very variable, depending on the feeds used and the stage of lactation (or milk yield) of the cow. Most dairy farmers would aim to provide a minimum of 28 g N/kg DM (17.5% crude protein/kg DM), and this was proposed as a default by ERM (1999). In practice, however, the average N content over the year (including the period when cows are not lactating) is likely to be less than this, and a value of 2.65% (16.5% crude protein) is used in the calculations below to estimate N excretion (Table 4.1).
The N content of milk is usually estimated as milk protein production divided by 6.38. The weight of calves at birth varies, but an average (for large breed cows) of 45 kg is proposed, with an N content of 30 g/kg. Cows usually gain some liveweight during the course of the lactation; an average 25 kg/lactation, with an N content of 25 g/kg live weight is proposed, although ERM (1999) suggest that cows may gain as much as 40 kg/lactation. In practice, the contribution of N from the calf and liveweight of the cow have a minimal impact on the total N in animal products.
During the period when cows are not lactating, DM intakes and diet N contents are lower. Although not explicitly stated, the ERM estimates appear to include the dry period since the report refers to estimates of N excretion per cow per year. Based on data from ARINI[5], AgriSearch (2005) have proposed that N excretion of cows (kg/day) in the dry period may be estimated from the metabolic liveweight of the cow[6], and the use of this equation for predicting N intake by non-lactating cows is recommended.
Using the mass balance approach, N excretion of a 600 kg live weight dairy cow with an annual milk yield of 6,973 kg/lactation[7] are summarised below. For comparison, N excretion by a small breed of cow (e.g. of the Jersey or Guernsey breed) is also given.
Table 4.1 Calculation of manure N output using the ERM (1999) balance model and UK production data
Parameter / Average / SmallMilk yield / kg FCM / 6,973 / 4500
Adult cow live weight / kg / 600 / 500
Total feed intake / kg DM / 5,790 / 4,259
Diet N content / g/kg DM / 26.5 / 26.5
Total N intake / kg/cow/year / 153 / 113
N retention in milk / kg/cow/year / 35 / 22.5
N retention in calf / kg / 0.65 / 0.65
N retention in liveweight gain / kg / 0.63 / 0.63
Total N retained / kg/year / 36 / 24
N excreted / kg/year / 117 / 89
A recent study in Italy (Xiccato et al., 2005) collected feed and production data from 104 dairy farms and estimated N excretion. Winter diets consisted predominantly of maize silage, and as a result dietary N concentration averaged 24.5 g/kg DM. Average milk yield was 8366 kg/cow/year. Using the balance approach, they estimated 116 kg N excreted per cow per year. The milk yield in this study was greater than that assumed in Table 4.1; the marginally lower N excretion in this study, compared to the figure above, is a result of the lower N diets fed on the Italian farms.
In Ireland, a stochastic budgetary simulation model of a dairy farm[8] has been developed “to allow investigation of the effects of varying biological, technical, and physical processes on farm profitability” (Shalloo et al., 2004). This model, which estimates grass and silage intakes, uses a mass balance approach to calculate N output per cow per annum. For a milk yield of 4,676 kg/cow/annum, the model estimates N intake of 121.1 kg and N utilisation - in milk, liveweight gain and calf - of 27.1 kg, giving an estimated N excretion of 94 kg. The ERM (1999) model would predict N excretion of 98 kg/cow/annum.
As discussed above, information on the dry matter intake of dairy cows is frequently unreliable or not available. An alternative approach is to use empirical relationships based on measurements made with dairy cows.
Using data from feeding studies at the ARINI, researchers there have developed predictions of N output based on animal production data (Yan et al., 2006). In common with other researchers (see Figure 1) a close correlation between the amounts of N excreted in manure and N intake and (R2 = 0.89; Equation D2) was observed. However, they also observed a strong correlation between total N intake and milk yield (R2 = 0.77; Equation D3), which allowed N intake to be accurately predicted. Adding liveweight (as metabolic live weight) as a secondary predictor of N intake improved the relationship (R2 = 0.79; Equation D4).
Manure N (g/day) = 0.713 x N intake + 4(Equation D2)
N intake (kg/305 days) = 0.0130 x MY + 33.8 (Equation D3)
N intake (kg/305 days) = 0.0108 x MY + 0.4951 x liveweight0.75 – 12.7 (Equation D4)
All dairy producers will have detailed records of the amount of milk produced during the course of lactation, and a reasonable estimate of the liveweight of their cows. Using Equation D4 therefore provides an estimate of N intake that can then be applied to Equation D2 to predict N excretion during the course of lactation.
To obtain an accurate estimate of total N excretion during the full lactation cycle, it is necessary to include estimates of N balance during the dry period. Based on studies with non-lactating cattle, the ARINI model includes a prediction of manure N output during the dry period:
Manure N (kg/day) = 0.00124 x liveweight0.75 (Equation D5)
Using these equations, it is possible to estimate likely N excretion values for UK dairy cows based on average milk yields (kg/cow), using equations D4 and D5. Assuming an average milk yield of 6770 litres (6973 kg), a dry period of 60 days and liveweight of 600 kg, estimated N excretion is 98 kg/cow/year. This is somewhat lower than the balance approach (Table 4.1) suggests.
The ARINI equations (Yan et al., 2006) were derived from 4 long-term dairy cow feeding studies. In two of them cows were fed indoors for the whole lactation, while two involved cows fed indoors during the winter and grazed grass during the summer. Dietary N contents in some treatments were low, but in others were typical of commercial practice. Thus the database had a very wide range of dietary N concentrations.
As discussed above, the area of greatest uncertainty associated with calculating manure N output of dairy cows is estimating feed intake. Figure 4.2 illustrates the relationship between milk yield and DM intake obtained in the four ARINI studies.
Figure 4.2 The relationship between milk yield and DM intake (T. Yan, personal communication).
As would be expected, daily DM intake increased with increasing milk yield. However, at the same milk yield there was a large variation of DM intake. For example, at milk yield of 6000 kg/305 days, DM intake ranged from just over 10 to over 16 kg/d. This resulted from both animal factors and dietary quality (ME concentration, CP concentration and quality, etc.). This may help to explain the variations between predicted results using different models. Generally, equations developed from a large dataset which contains a large range in quantity and quality of dietary and animal factors, can produce a more accurate average prediction.