N2O exchange

Field estimates

According to literature reviews, Papen and Butterbach-Bahl (1999), Borken and Brumme (1997), Schulte-Bisping et al. (2003) and Denier van der Gon and Bleeker (2005) reported that broadleaved forests show a tendency towards re-emitting a higher fraction of the nitrogen input as N2O than coniferous forests due to species-induced differences in litter quality and soil moisture (Butterbach-Bahl et al., 2002c). Papen and Butterbach-Bahl (1999), for example, observed that a beech site in Germany emitted about 10% of the nitrogen input as N2O, whereas a spruce site emitted only 0.5% of the nitrogen input as N2O. Differences between deciduous and coniferous forests are even more striking, because nitrogen deposition increases with surface roughness and is significantly higher in conifer ous than in deciduous forests (Bleeker and Draaijers, 2001).

Generally, N2O emissions increase when going from boreal to temperate forest ecosystems because of the increase in primary production and the increase in nitrogen deposition. For boreal forests, measured fluxes range from 0.18 to 0.27 kg N2O-N/ha/ year (Brumme et al., 2004). Recently, Denier van der Gon et al. (2004) compiled the literature on N2O emissions from forests in Europe (mainly Germany, but also Finland, Sweden, Denmark, UK, Belgium and Austria) and in North America (USA and Canada). Their results show that measured annual N2O fluxes range from 0.02 to 4.5 kg N2O-N/ha/ year (Table 17.2). Some of this site variation can be explained by the effect of forest type on carbon and nitrogen cycling. However, there is usually a huge interannual variability in total annual N2O losses. For example, the annual N2O emissions at the Höglwald Forest ranged from 0.4 to 3.1 kg N2O-N/ha/ year within a 4-year observation period (Butterbach-Bahl et al., 2002b). Evidently, it is difficult to assess long-term average fluxes from short-term measurements, and some of the observed differences between sites and forests types (see Table 17.2) may simply be a result of interannual variability. Table 17.2 also includes information on the nitrogen deposition and the N2O emission fraction,

Table 17.2. Summary of published nitrous oxide (N2O) emission data for deciduous forests and coniferous forests and derived emission factor as function of nitrogen input. (After Denier van der Gon and Bleeker, 2005.)

Average nitrogen

Number of

inputb

N2O emission

Emission

observations3

(kg N/ha/year)

(kg N/ha/year)

factorc

Reference

Deciduous forest

3

22

0.49

0.023

Ambus et al. (2001), Beier et ai

(2001)

1

10

0.23

0.023

Bowden et al. (2000)

8

29

1.67

0.053

Brumme et al. (1999)

2

26

2.98

0.1 11

Butterbach-Bahl et al. (2001)

2

15

0.02

0.001

Corre et al. (1999)

1

20

1.45

0.072

Butterbach-Bahl et al. (1997)

2

46

2.65

0.044

Mogge et al. (1998)

1

16

0.20

0.013

Oura et al. (2001)

2

20

2.66

0.133d

Papen and Butterbach-Bahl

(1999)

6

27

0.65

0.035

Skiba et al. (1998)

3

35

4.03

0.1 15

Zechmeister-Boltenstern et al.

(2002)

Coniferous forest

3

26

0.31

0.016

Borken et al. (2002)

4

26

0.58

0.034

Brumme et al. (1999)

2

35

0.70

0.020

Butterbach-Bahl et al. (2001)

4

28

1 .59

0.073

Butterbach-Bahl et al. (1997)

3

16

0.90

0.056

Butterbach-Bahl et al. (2002a)

1

12

1 .86

0.1 55

Jungkunst et al. (2004)

6

24

0.11

0.005

Klemedtsson et al. (1997)

4

1 05

0.36

0.005

Matson et al. (1992)

2

31

0.39

0.013

Oura et al. (2001)

2

1 00

3.20

0.032

Ineson et al. (1998)

6

30

0.60

0.02 0d

Papen and Butterbach-Bahl

(1999)

14

37

0.37

0.016

Skiba et al. (1998)

aVarious years and plots with different nitrogen treatments and/or tree species within the deciduous forest or coniferous forest class (e.g. an experiment where a beech and alder plot was monitored for two consecutive years would result in n = 4). bOnly indicative, as in some experiments various nitrogen input levels were studied and are averaged. cCalculated from the original data of N2O emission and nitrogen input for individual plots and then averaged. The emission factor gives the fraction of nitrogen input that is re-emitted as N2O.

dResults presented in Papen and Butterbach-Bahl (1999) were corrected by excluding the extreme 1996 emissions. Results are thus based on the data for 1 995 and 1 997 for beech site (value of 0.133 instead of 0.222) and for 1994, 1995 and 1997 for the spruce site (value of 0.020 instead of 0.028).

aVarious years and plots with different nitrogen treatments and/or tree species within the deciduous forest or coniferous forest class (e.g. an experiment where a beech and alder plot was monitored for two consecutive years would result in n = 4). bOnly indicative, as in some experiments various nitrogen input levels were studied and are averaged. cCalculated from the original data of N2O emission and nitrogen input for individual plots and then averaged. The emission factor gives the fraction of nitrogen input that is re-emitted as N2O.

dResults presented in Papen and Butterbach-Bahl (1999) were corrected by excluding the extreme 1996 emissions. Results are thus based on the data for 1 995 and 1 997 for beech site (value of 0.133 instead of 0.222) and for 1994, 1995 and 1997 for the spruce site (value of 0.020 instead of 0.028).

assuming that nitrogen deposition is linearly related to N2O emission, and this aspect is further discussed in Section 17.4.

Process-based model approaches

Process-oriented models simulate the biosphere-atmosphere exchange of N2O-N

based on individual production, consumption and emission processes, and their intera ctions. Nitrification rate (d[NHJ]/di) is either calculated from a mechanistic description of the growth and development of nitrifying populations or as a function of substrate concentration [NH+], oxygen partial pressure (pO2), temperature (T) and pH, according to:

where k1 is the first-order nitrification coefficient under optimal conditions and f([NH+]), f(pO2), f(T) and f(pH) are dimensionless reduction functions for [NH+], molecular oxygen, temperature and pH, respectively. Often, nitrifying activity is related to [NH+] via a Michaelis-Menten type of relationship, i.e. f([NH+]) = [NH+]/{k2 + [NH+]}. In this case, [NH+] limits nitrifying activity (cf. first-order process) at low concentrations and does not limit nitrifying activity (zero-order) at high concentrations. Constant k2 is the Michaelis-Menten half-saturation constant or the concentration at which f([NH+]) = 0.5. It should be noted that the meaning of k1 changes to 'potential nitrification activity' when a Michaelis-Menten type of relationship is used for substrate dependence. For the dependence of nitrification rate on f(pO2), f(T) and f(pH), various descriptions have been formulated, from simple to complex, but discussion of these is beyond the scope of this chapter (e.g. Shaffer et a!., 2001; Heinen, 2005).

Like nitrification, denitrification rate is either based on: (i) a mechanistic description of the growth and development of denitrifying populations; (ii) fractional or structural models; or (iii) substrate dependence using first-order kinetics. In the latter case, deni-trification activity (d[NO-]/dt) is described as a function of the estimated potential denitrification rate, which is often proportional to organic matter decomposition rate, corrected for nitrate concentration ([NO-]), pO2, temperature (T), and pH according to:

where kD is the estimated potential denitri-fication rate or organic matter decomposition rate, and f([NO-]), f(pO2), f(T) and f(pH) are dimensionless reduction functions for [NO-], molecular O2, temperature and pH, respectively. Again, the description of these reduction functions is beyond the scope of this chapter (e.g. Shaffer et al., 2001).

A well-known process-based model is the PnET-N-DNDC model (Li et al., 2000), which integrates the interactions among pri mary drivers (climate, soil characteristics, forest type and management), soil environmental factors (e.g. temperature, moisture, pH, redox potential and substrate concentration gradients) and various biogeochem-ical reactions. These interactions control transformation and transport of carbon and nitrogen in the ecosystem and thereby the emission of N2O from forest soils into the atmosphere. Extensive validation of the model for different forest sites worldwide has shown that the model is capable of predicting the emission of N2O (Stange et al., 2000; Butterbach-Bahl et al., 2001).

Butterbach-Bahl et al. (2004) made regional inventories of N2O emissions from forest soils for Saxony, Germany, using a coupled PnET-N-DNDC-GIS database system. Mean N2O emissions for German forest soils were estimated at 1.4 kg N2O-N/ha/ year. The PnET-N-DNDC model has also been used to make an inventory of N2O-N emissions from European forest soils, covering the whole of Europe excluding Cypress, former Yugoslavia and Albania, by coupling it to a Geographic information system (GIS) with a geographic resolution of 2527 grids of 50 x 50 km representing a forested area of 141 million hectares. The estimated total annual N2O emissions were 81.6 Kt N2O-N with an uncertainty range of 50.7-97.0 Kt N2O-N for 2000 (Kesik et al., 2005). Average N2O emissions from forest soils across Europe were estimated at 0.58 kg N2O-N/ha/ year, which is less than half of the estimate for forests in Saxony.

Empirical approaches

On the basis of field measurements, Brumme et al. (1999) suggested that forest soils can be stratified according to their N2O emission characteristics into seasonal, background and event emission pattern (SEP, BEP and EEP, respectively) types. Broadleaved forests with thick oxygen horizons on acid soils can be characterized as a large and seasonal source of N2O (classified as SEP), which is assumed to be primarily controlled by temperature and moisture. Broadleaved forests, with thin oxygen horizons, and all coniferous forests have been identified as a low and non-seasonal source of N2O (classified as BEP) since the oxygen horizon hardly inhibits diffusion. Estimated mean annual N2O emissions from German forest soils for these forest types are 2.05 and 0.37 kg N/ha/year for broadleaved forests with thick and thin oxygen horizons, respectively and 0.17 kg N/ha/year for coniferous forests (Schulte-Bisping et al., 2003). EEP occurs primarily during freeze-thaw processes in SEP and BEP types under beech, spruce, alder and oak forests (Brumme et al., 1999; Papen and Butterbach-Bahl, 1999; Teepe et al., 2000). Additional emissions by freeze-thaw processes may occasionally contribute to the BEP by up to 70%, whereas the contribution is much smaller in forests with SEP (up to 40%), but its overall impact is likely to be low (Brumme et al., 1999). Schulte-Bisping et al. (2003) used the SEP/ BEP stratification to calculate the regional emission strength of forest soils in Germany, although they used the pH of the top soil and not the thickness of the oxygen horizon as diagnostic criterion for separating forests in SEP- and BEP-type forests. Using the stratification approach and climate data for a period of 30 years (1961-1990), they calculated the average source strength of forest soils in Germany to be 0.32 kg N/ha/year, which is four times smaller than the estimate derived by Butterbach-Bahl et al. (2004) for Saxony with the PnET-N-DNDC model, but close to the European average estimate. However, Schulte-Bisping et al. (2003) did not include effects of freezing and thawing, as well as nitrogen deposition, suggesting that the estimates may not be too different from each other. Additional factors explaining the difference between the two estimates are forest typology and upscaling procedures.

tent, Schulte-Bisping et al. (2001) estimated the CH4 uptake by forest soils at 1.84 kg (CH4)/ha/year. Brumme et al. (2004) used a stepwise regression technique to explore the relationship between CH4 uptake and biome, annual rainfall, mean annual temperature, soil texture and latitude. The analysis showed that biome and soil texture class accounted for 18.1% of the variation in the global data-set of annual fluxes. The mean uptake rate was 6.6 kg CH4/ha/year for coarse textured soils, 2.3 kg CH4/ha/year for medium and fine-textured soils and the average for all soil types was 3 kg CH4/ha/ year (Brumme et al., 2004).

An overview of published CH4 flux rates between soils and the atmosphere for forest ecosystems as a function of climate zone, forest type, soil type and texture is given in Table 17.3. Lindner et al. (2004) estimated the CH4 budget of European forest soils by linking these data to a GIS with spatial information about the forest area distribution in Europe (Schuck et al., 2002), the European soils database of the European Soils Bureau (http://eusoils.jrc.it/) and the climatic zonation mentioned in Table 17.3. The total forest area in this analysis was 320 million hectares. According to these calculations, European forests have a mean net CH4 uptake of 0.2 kg CH4/ha/year, but with a large uncertainty range. Schulte-Bisping et al. (2001) and Brumme et al. (2004) estimated a higher net uptake of CH4 by European forests. However, they did not consider water-saturated soils, which can be a rather large source of CH4 (Roulet et al., 1992). The current global soil sink for atmospheric CH4 is estimated to be 30-40 Tg/year (Mosier et al., 1994; Houweling et al., 1999), with much of this sink occurring in well-aerated forest soils.

Was this article helpful?

0 0

Post a comment