National and regional net nitrogen balances in Finland in 1990 – 2005

Nitrogen (N) balance has been identified as a principal agri-environmental indicator. In addition to national N balances, calculation of N balances for different agricultural regions is also recommended. In this study, national and regional net N balances for Finland were calculated. The net N balance is the result of deducting the NH3-N losses from manure and fertilisers from the gross N balance. The N balance calculation was based on data for Finnish Rural Centres and calculated per cultivated hectare. The main data inputs for the calculations were agricultural and environmental statistics, coefficients of manure excretion and crop N concentrations. Finnish national net N balance decreased from 90 kg ha–1 in 1990 to 50 kg ha–1 in 2005. The decrease in regional N balances was of the same magnitude. The main reason for the lower N balances was reduced use of mineral N fertilisers. Variation in the N balances was due to yield levels varying according to growing season conditions. The Rural Centres with intensive animal production tended to generate the highest N balances.


Introduction
Nutrient use in agriculture should be sufficient to maintain crop and forage production, but should generate minimal surpluses that pollute water and air."The calculation of nitrogen (N) balances has been identified as a priority agri-environmental indicator by OECD Member countries" (OECD/ EUROSTAT 2003).The information represented by N balances is needed to analyse the interactions between agriculture and the environment and to evaluate the impact of changes in agricultural policy on the environment.
Several methods are used to measure the inputs and outputs and thereby calculate a nutrient balance."Soil surface balance" (Parris 1998) or "gross and net nitrogen balance" (OECD/EUROSTAT 2003) are terms for a calculation method that is used by many OECD countries and international organisations.Basically these methods assess the difference between the annual total quantity of N entering the soil and the annual quantity of N leaving the soil.The gross N balance includes all emissions of N compounds from agriculture into the soil, water and air.The net N balance excludes N emissions into the air, and the N volatilisation and denitrification from fertiliser and manure should be deducted from the gross N balance (OECD/EUROSTAT 2003).
As national authorities calculate national N balances, there is also a demand for regional N balances because some areas can experience nitrate pollution and some areas depletion of N (Parris 1998).Parris (1998) stated that N balance itself indicates only potential for pollution, not actual pollution and suggested that trends in N balance represent a practical and low cost tool for estimating potential environmental effects.
The purpose of this study was to calculate national and regional net N balances in Finland to estimate the potential for N losses to the environment.The net N balance was calculated by excluding ammonia volatilisation but not denitrification losses, which are difficult to estimate, especially those for N 2 .In this study, N balance means net N balance unless otherwise stated.The trends in N balances could also show the influence of the Finnish Agri-Environmental Program started in 1995, although other changes in Finnish agriculture during 1990-2005 have also taken place.An additional objective was to describe available data and methods for N balance calculations.The quality of agricultural data and coefficients are also discussed.

Material and methods
The N balance calculation was based on data from Finnish Rural Centres (Fig. 1).In 1990, there were 20 Rural Centres, of which Nylands svenska lantbrukssällskap was integrated into Uusimaa and Finska Hushållningssällskapet into Farma.During the last three years, 2003-2005, data from Päijät-Häme Rural Centre were not available as they were included with those of Häme Rural Centre.Rural Centres were used instead of other regional districts as N fertiliser data were only available for the Rural Centres.National N balance was calculated on the basis of the regional N balances.The main elements of the N balance calculation and their magnitudes are shown in Table 1.
Data from sales of N fertiliser were obtained from the most important fertiliser suppliers in Finland.Data obtained from Kemira GrowHow Oyj were distributed according to the Rural Centre and sales from other companies were distributed evenly for the entire cultivated area.
The input of manure N from different farm animals was calculated according to manure excretion coefficients (Table 2) used in environmental guidelines for livestock production (Ministry of Environment 1998).The volatilisation of ammonia was calculated according to the coefficients for different farm animals and manure management strategies (Grönroos et al. 1998).
The amount of biological N fixation was calculated from the N content of pea (Pisum sativum L.) production added to N fixed by clover (Trifolium L.) in cultivated grass in organic farming and in seed production.The amount of N fixed by the clover-grass swards was estimated to be 140 kg ha -1 .It is an average value from two years measured in clover-grass swards of organic farms in Finland in the southern Savo region (Väisänen 2000).Associative N fixation was estimated to be 4 kg ha -1 in cereals rye (Secale cereale L.), barley (Hordeum vulgare L.) and oats (Avena sativa L.) and grasses such as timothy (Phleum pratense L.) and meadow fescue (Festuca pratensis Huds.).
Other sources of N entering agricultural soils included seeds and sewage sludge used in agriculture.Sewage sludge comes from wastewater treatment plants and is used as an organic fertiliser or soil conditioner after composting.The amount of sewage sludge used in agriculture was obtained from the VAHTI-database, maintained by the Finnish Environment Institute and N concentration came from the literature (Kulmala and Esala 2000).In our calculations, sewage sludge N was evenly distributed over the cultivated area.Nitrogen input from seeds was calculated according to recommended seeding rates for each crop and seed nutrient content came from the literature (Tuori et al. 1996), and cultivated area of each crop was obtained from agricultural statistics (Information Centre of the Ministry of Agriculture andForestry 1991-2005).
Cultivation areas for the different crops and numbers of different farm animals were obtained directly from the 1990-1991 Yearbook of Farm Statistics (Information Centre of the Ministry of Agriculture andForestry 1990, 1991) and calculated for the Rural Centres by the Information Centre of the Ministry of Agriculture and Forestry for 1992-2005.Crop yields per hectare were taken from national statistics (Information Centre of the Ministry of Agriculture andForestry 1992-2006), using the data from representative Employment and Economic Development Centres or Rural Business Districts in 1992-2005, when data from Rural Centres were not available.Nitrogen contents of crops were calculated from protein concentrations taken from the Finnish tables of feeding recommendations (Tuori et al. 1996).Calculations were done for the time period 1990-2005.
To assess the possible trend in N balances and its components in 1990-2005, a simple linear regression model was used.
where β 0 is the intercept in the year 1990, β 1 is the effect of one year, X is the year and ε is the random error.For both national and regional N balances, linear regression was calculated for fertiliser and manure N input, yield N output and the N balances.All statistical tests were performed at p = 0.05 and a coefficient of determination (R 2 ) was used to describe the accuracy of the predicted trend.

National N balance
The N input decreased from 160 kg ha -1 at the beginning of the 1990s to almost 120 kg ha -1 in 2005 (Table 3).The main reason for this is decrease in the use of mineral N fertilisers.The N output in the harvest ranged from 65 to 80 kg ha -1 .Variation is mainly attributable to the extent of unfavourable climatic conditions, which changed across growing seasons.The net N balance decreased during the calculated time period from 90 to 50 kg ha -1 (Table 3).Gross N balance decreased from 110 kg ha -1 to 60 kg ha -1 .The difference between gross and net N balance decreased from 17 to 14 kg ha -1 .This difference stems mainly from ammonia volatilisa-tion from manure and its decrease is attributable to reduced cattle production.The cultivated agricultural area varied due to set-aside agreements and was lowest in 1991-1994.Thus total gross balance (gross N balance x cultivated area) decreased below the earlier lowest value (in 1993) only after 2000 (Table 3).
The trends in the input of N in mineral fertilisers (-2.2 kg ha -1 yr -1 ) and manure (-0.9 kg ha -1 yr -1 ) were clear and affected the decrease in both gross N balance (-2.5 kg ha -1 yr -1 ) and net N balance (-2.3 kg ha -1 yr -1 , Table 4).Despite the decreased N input, yield N output did not tend to decrease (p=0.156).

Regional N balances
The N surplus of the regional N balances also decreased in all but one Rural Centre during the period of calculation (Table 5).In the Rural Centre of Österbotten, N surplus was fairly constant in the 1990s.On closer inspection the use of N in fertilisers was 20-30 kg ha -1 lower than in the other Rural Centres, mainly due to a high percentage of organic soils (25%) in Österbotten.In addition, clay soils are almost absent in the area of Österbotten Rural Centre, and sandy soils are the dominant soil type.Compared with sandy and organic soils, clay soils are associated with 10-20 kg ha -1 and 20-40 kg ha -1 , respectively, higher N recommendations for cereals in the area (Viljavuuspalvelu 2000).
During the early 1990s N surpluses in the intensive livestock regions were clearly higher than N surpluses in the cereal production regions (Table 5, Fig. 2).These N surpluses were reduced towards 2000.In 1990-2005, the annual decrease in N balance in intensive livestock regions was -2.4 --3.5 kg ha -1 yr -1 .The decrease was only slightly lower, -1.5 --2.9 kg ha -1 yr -1 , in regions that concentrate more on cereal production.
Figure 2 shows the main components of N balance from two Rural Centres.Cereal production dominates in Uusimaa and milk production in Pohjois-Savo.Grassland occupies 20% of the cultivated area in Uusimaa and 60% in Pohjois-Savo.As for the Rural Centres in Figure 2, the decrease in N balance in most Rural Centres is based on reduc- ) in Finland in 1990Finland in -2005Finland in . 1990Finland in 1991Finland in 1992Finland in 1993Finland in 1994Finland in 1995Finland in 1996Finland in 1997Finland in 1998Finland in 1999Finland in 2000Finland in 2001Finland in 2002Finland in 2003Finland in 2004  Ammonia emission -0.7 -0.7 -0.6 -0.6 -0.6 -0.6 -0.6 -0.5 -0.5 -0.5 -0.5 -0.5 -0.5 -0.5 -0.2.9 2.9 5.9 6.5  1922 1936 1964 2000 1965 2006 1990 1969 1992 2023 1993 93.9 Total gross balance on the cultivated area (1000 t) tion of both mineral fertiliser and manure N input.
Only in the Rural Centres of Österbotten and Lappi was the N input in mineral fertiliser not clearly decreased.Manure N input and animal densities did not decrease in Farma + Finska Hushållningssällskapet and Österbotten Rural Centres.
Local weather conditions cause variation in N surpluses among the Rural Centres.For example, high N surpluses in Uusimaa and Farma in 1999 were caused by low rainfall during the growing season.Cereal yields are usually more vulnerable to unfavourable weather conditions than grass yields, as can be seen from the yield N variation in Uusimaa and Pohjois-Savo Rural Centres (Fig. 2).In 1990-2005, there were three Rural Centres (Pirkanmaa, Etelä-Pohjanmaa and Oulu) where a decreasing trend explains approximately 30% of yield N reduction.The effect of 0.6-0.8kg ha -1 a -1 can partially be related to the substitution of grassland with cereals.

N balance surplus
The decrease in net N balance from 90 kg ha -1 to 50 kg ha -1 implies more efficient use of N in agriculture and should result in reduced N leach-  1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 1990 1992 1994 1996 1998 2000 2002 2004 2006 N ing.However, as the cultivated area was smaller in 1991-1994, due to set-aside, the total N balance decreased beyond that only after 2000.This calculation led to a similar N balance as reported by Antikainen et al. (2005) for 1995-1999.There the nitrogen balance of Finnish agricultural soils was 62 kg ha -1 between 1995 and 1999, which agrees with the average value of 63 kg ha -1 from our calculation for those years.Antikainen et al. (2005) divided the N surplus in soil into leaching of 15 kg ha -1 (Vuorenmaa et al. 2002) and denitrification of 18 kg ha -1 (Finnish Ministry of the Environment 2002), and thus the remaining N surplus in soil was 29 kg ha -1 .Measurements of N leaching provided annual figures of 10-20 kg ha -1 (Salo and Turtola 2006), which agree with the estimate of Vuorenmaa et al. (2002).Measurements of nitrous oxide emissions indicated annual losses of 2-8 kg ha -1 from mineral soils (Syväsalo et al. 2004) and 4-25 kg ha -1 from peat soils (Regina et al. 2004).Regarding nitrous oxide emissions, it is difficult to estimate total denitrification as N 2 :N 2 O can range from 0.1 to 5.7 (Mathieu et al. 2006).
As organic carbon in Finnish arable soils decreased according to results of a field survey (Mäkelä-Kurtto and Sippola 2002) and a field experiment (Esala and Larpes 1984) by 0.3% in 10 years, it is unlikely that soil organic matter can retain N.This suggests that N leaching, denitrification and volatilisation losses are higher than currently verified by measurements.Because denitrification as N 2 gas is the process for which there are practically no measurements in Finland, it can be considered the most likely loss pathway.While some researchers estimate that denitrification explains 50-90% of N surplus (e.g.Kroeze et al. 2003), others estimate that only about 10% of the soluble N entering the ecosystem might be lost via denitrification (Janzen et al. 2003).

Uncertainty in N balance calculations
Manure excretion coefficients are usually, as in this calculation, fixed values that are not adjusted for changes in feeding regimes for milk and meat production.Furthermore, there can be considerable differences in excretion coefficients used among different countries (van Eerdt and Fong 1998), which can complicate comparisons among countries if the coefficients are not reliable.The variation in N excretion coefficients can be seen from Table 1, where the default values for OECD and Finnish coefficients for environmental authorities (Ministry of Environment 1998) and for greenhouse gas emission calculations (Statistics Finland 2006) are shown.Considering the coefficients for Finland, values calculated for greenhouse gas emission would probably be the most reliable as they are checked regularly on the basis of recommended animal feeding.In future studies the expertise of animal nutrition should be used in environmental nutrient balance studies when calculating the estimates for manure and nutrient excretion.
Concerning other N inputs, the N fertiliser use data that was based on sales statistics can differ from the amount actually applied to crops over a given year (Parris 1998).Biological N fixation is rarely studied in Finland and amounts of fixed N probably vary considerably among fields.
Estimation of ammonia volatilisation from manure is based on various coefficients that are dependent on manure storage and treatment.Manure storage and field application methods on farms are poorly documented in national statistics.Volatilised ammonia is readily absorbed by vegetation and soil and thus most volatilised ammonia can be redeposited close to the site of emission (Pitcairn et al. 1998).An alternative method for calculating ammonia volatilisation was suggested by Janzen et al. (2003), who assumed that 30% of soluble manure-N is volatilised and 30% out of that is later deposited on other than agricultural land.This results in 9% output of soluble manure-N from the agricultural system.Probably the OECD recommendation to use gross N balance derives from the difficulties in estimating ammonia volatilisation, which is an important element in net N balance.
Crop yield statistics are seldom absolute, especially in the case of grass production and grazing.Annual variation in N content of grains can also introduce error into the balances.Results from an annual survey of the Finnish Food Safety Authority (Salo et al. 2007) suggested that variation of N content in cereals was 0.3-0.5 percentage points over years and regions.Regional calculations could be improved by using these data.

Comparison among countries and regions
In the OECD Nitrogen Balance Database (OECD 2001) the highest national gross N balances in 1985-1997, 100-300 kg ha -1 , were for countries with intensive animal production (Netherlands, Belgium and Denmark) and intensive agriculture concentrated on small cultivated areas (Japan and South Korea).The majority of European countries are similar to Finland, with gross annual N balances of 50 to 100 kg ha -1 .Countries with large areas of extensive agriculture, such as Canada, have gross N balances as low as 17 kg ha -1 (Janzen et al. 2003).As the annual decrease in N balance in Finland was more than 2 kg ha -1 in 1990-2005 and net N balance reached 50 kg ha -1 , the decrease will most likely cease in the coming years.
In general, countries with high livestock densities and intensive agricultural production systems have the highest N surpluses.The overall trend in national N balance surpluses over the last decade is downwards or constant for most OECD countries (Parris 1998).
While an annual national N balance provides an impression of the performance of the agricultural sector as related to its use and management of N, there is usually significant spatial variation in N balances, largely attributable to variation in cropping and livestock production patterns and systems, soil types, topography, climatic conditions and farm management practices (Parris 1998).Lord et al. (2002) calculated a "farm gate" N balance of 140 kg ha -1 for agricultural grassland and 51 kg ha -1 for arable land in the United Kingdom.However, nitrate concentrations in rivers were generally greater in arable areas, which shows that N leaching is also dependent on land use, soil type and climate (Lord et al. 2002).Distribution of N surplus over fields within a farm affects N leaching because the relationship between field surplus and N leaching (Watson andFoy 2001, Salo andTurtola 2006) is one of increase after a certain threshold value is reached (van Beek et al. 2003, Korsaeth andEltun 2000).
In Finland, the main production sector of the northern and eastern Rural Centres is milk production and these Centres had high N balances at the beginning of the 1990s.During the study period, the livestock density and manure N input decreased in most of these Rural Centres and this was associated with a clear decrease in mineral fertiliser input.The decreased mineral N fertiliser input is a combination of decreased N use for grassland and a shift towards cereal cultivation with lower N application rates compared with grassland.In some of these Rural Centres, the yield N uptake also slightly decreased, most probably due to the lower N uptake of cereals than of grasses.In southern Rural Centres dominated by cereal production, the

Fig. 2 .
Fig. 2. The main components of nitrogen (N) balance in the Rural Centres of Uusimaa and Pohjois-Savo.

Table 1 .
The balance sheet for the nitrogen balance calculation.

Table 3 .
Gross and net nitrogen (N) soil surface balances (kg ha -1

Table 4 .
Linear regression equations for the national nitrogen (N) balances and their components (kg ha -1 ).