on the temperature dependence of soil respiration · pdf fileintroduction release of co2 from...

Functional Ecology 1994 8, 315-323

315

On the temperature dependence of soil respiration

J. LLOYD and J. A. TAYLOR* Environmental Biology Group, Research School of Biological Sciences, Institute of Advanced Studies, Australian National University, GPO Box 475, Canberra ACT 2601 and *Centrefor Resource and Environmental Studies, Institute of Advanced Studies, Australian National University, GPO Box 4, Canberra ACT 2601, Australia

Summary

1. From previously published measurements of soil respiration rate (R) and temperature (T) the goodness of fit of various R vs T relationships was evaluated. 2. Exponential (Q10) and conventional Arrhenius relationships between T and R cannot provide an unbiased estimate of respiration rate. Nor is a simple linear relationship appropriate. 3. The relationship between R and T can, however, be accurately represented by an Arrhenius type equation where the effective activation energy for respiration varies inversely with temperature. An empirical equation is presented which yields an unbiased estimator of respiration rates over a wide range of temperatures. 4. When combined with seasonal estimates of Gross Primary Productivity (GPP) the empirical relationship derived provides representative estimates of the seasonal cycle of net ecosystem productivity and its effects on atmospheric CO2. The predicted seasonal cycle of net ecosystem productivity is very sensitive to the assumed respiration vs temperature relationship. 5. For biomes in areas where soil temperatures are low, soil respiration rate is rela- tively more sensitive to fluctuations in temperature. Nevertheless, more information is required before any predictions can be made about changes in soil carbon pools in response to future temperature changes.

Key-words: Climate change, CO2

Functional Ecology (1994) 8, 315-323

Introduction

Release of CO2 from soils due to production of CO2 by roots and soil organisms and, to a lesser extent, chemical oxidation of carbon compounds is commonly referred to as soil respiration. There have been many studies showing a positive correlation between soil respiration (R) and temperature (T) (for reviews see Singh & Gupta 1977; Reich & Schlesinger 1992) but there is no consensus on the exact form of the relationship. Some workers have used linear regression analysis (e.g. Witkamp 1966; Froment 1972), espe- cially when additional environmental variables such as soil moisture have been included in the analysis (e.g. Gupta & Singh 1981), while others have used the

Q10 relationship, first developed by van't Hoff (1898), and hence assumed an exponential dependence of R on T (see Fig. 5 in Reich & Schlesinger 1992). Power relationships (Kucera & Kirkham 1971) and relationships of the form proposed by Arrhenius (1889) have also been used (Howard & Howard 1979). Only rarely (e.g. Howard & Howard 1979)

have different models been compared. Analysis of residuals to determine the appropriateness of model fits is even rarer.

Our purpose here is to examine published data in order to evaluate the ability of linear, exponential and an Arrhenius type relationship to predict soil respiration rates in the absence of soil moisture limitations. Consequences of the assumed relationship for simple models of the seasonal cycle of soil respiration and net ecosystem productivity are then considered, as are possible changes in soil respiration under conditions of elevated soil temperatures due, for example, to greenhouse gas-induced climate change (Houghton, Jenkins & Ephraums 1990).

Method of analysis

As well as being influenced by temperature, the other abiotic control substantially affecting soil respiration rates is soil moisture. Therefore, also accepting the

This content downloaded on Fri, 8 Mar 2013 14:29:50 PMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/page/info/about/policies/terms.jsp

316 criteria of Reich & Schlesinger (1992) regarding mea- J. Lloyd & surement technique employed (exclusion of data J. A. Taylor obtained with alkali absorbtion techniques if the sur-

face area of absorbent was less than 5% of the surface area of the covered ground), we have extracted data from all published papers known to us where the relationship between soil respiration and either air or soil surface temperature could be evaluated with, for the data used in our analysis at least, there being no evi- dence of soil water limitations on soil respiration rate. These are listed in Table 1. For each study, six to 12 values were obtained from graphs or tables of the annual cycle of R and T or from graphs of the relationship between R and T. In some cases a relationship, as prescribed by the authors, was used with values of R and T being calculated at discrete intervals over the temperature range originally examined. This provided 149 data points covering a wide range of ecosystems and soil temperatures. The relationship between R and Tfor all these points is shown in Fig. 1. This shows an increase in R with T for all studies, but also shows large differences in rates between studies at any given temperature. This is presumably a consequence of differences between studies in the concentration and composition of the carbon in the soil (the effective mass of carbon per unit area, M) available as respiratory substrate. There is also the complication that the air or soil surface temperatures and respiration rates given in Fig. 1 do not take into account variability in vertical temperature gradients (de Vries 1963) and in the distribution of CO2 production rates (de Jong & Schappert 1972) which would be expected to be dependent on plant rooting patterns in the soil below the area sampled. An accurate representation of these heterogeneities, as well as spatial variations in soil structural properties and water status, all of which can affect the rate of CO2 efflux from a soil, requires highly paramaterized simulation models which can

only be solved using numerical methods (Simufnek & Suarez 1993). Thus, bearing the necessary simplifica- tions in mind, we write:

R = kRM eqn 1

where kR is a temperature-dependent rate constant presumed to have the same temperature dependence for all ecosystem and substrate types. Typical units for R are gmol C m-2 s-, for which M would be in gmol C m-2 and kR would have the dimension s-'.

To account for differences in M we have fitted temperature dependencies of R according to:

R = RX f(T) eqn 2

where RX is a data set-dependent parameter giving the respiration rate at some standard temperature and f(T) is some function describing the relationship between kR and T which is common to all 15 data sets.

Large differences between studies in the absolute values of R meant the unweighed least squares was unlikely to provide an unbiased least squares estimator (Fig. 1). Using a Simplex procedure, an error sum of squares (SSE), X(R-R)21JR2, where J is the fitted value of R, was minimized. As much of the variance in Fig. 1 arises as a consequence of variations in absolute respiration rate between the 15 data sets, we express the proportion of variance accounted for by temperature as 1 -MSE/MST where MSE =residual

mean square (SSE/df where df is the degrees of free- dom associated with SSE) for the temperature function being tested and MST is the residual mean square of a model fit using equation 2 but with f(T) = 1.

In order to compress values from all data sets to a common scale, data in Figs 2-6 are presented in units relative to the fitted respiration rate at 10TC. This fitted respiration rate varies with the model being tested. The relative values in Figs 2-6 are also model dependent, and therefore also vary between figures.

Table 1. Data sources used for the analysis in Figs 1-6

Symbol Reference System

E Nakane, Tsubota & Yamamoto (1984) Pinus densiflora (Japan) 0 Nakane (1978) Birch forest (Japan) 0 Yoneda & Karita (1980) Quercus (Japan)

Anderson (1973) Beech forest (UK) * Chapman (1979) Calluna vulgaris Heathland (UK)

Peterson & Billings (1975) Tundra grass and sedges (Alaska, USA) A Kucera & Kirkham (1971) Tall grass prairie (Missouri, USA) A Monteith, Szeicz & Yabuki (1964) Barley (UK) 4 DOrr & Munnich (1987) Temperate grassland (Germany) < Bridge, Mott & Hartigan (1983) Savannah woodland (soil cores: Queensland, Australia) v Richards (1981) Subtropical rain forest (NSW, Australia) V Sivola, Valijokki & Aaltonen (1985) Peatland (Finland) *0. Svensson (1980) Subarctic mire (Sweden) > Reinke, Adriano & McLeod (1981) Pinus palustris (South Carolina, USA) * Hersterberg & Siegenthaler (1991) Grassland (Germany)



317 Temperature and 10

soil respiration E

0Q 8 0~~~~~

0

E 6

4 / A

o~~~~~~~~ .4.<Y A c5 2

*0 0 10 20 30 40 (n Temperature (0C)

Fig. 1. The relationship between soil respiration rate and temperature (either surface soil or ambient air) for the data summarized in Table 1.

Results and discussion

AN EXPONENTIAL DEPENDENCE OF RESPIRATION

RATE UPON TEMPERATURE

Van't Hoff appreciated that many chemical reactions have similar temperature dependencies. In 1898 he wrote 'the ratio of velocities for a given interval of temperature mostly differs little from reaction to reaction' (van't Hoff 1898). In order to 'test this empirical relation for different reactions, since measurements are not always to be had for 10TC intervals' he subse- quently presented 'the equation which that relation between temperature and velocity yields', viz.:

log1ok=a+bT eqn 3

where a and b are constants and k is a rate constant. In our case k = kR. Using natural logarithms one can express equation 3 for the case of respiration rate as:

R=kRM=AeBT eqn 4

where B = bln(10) and A = Mealn(lo). B is related to the common QIO as:

B=ln(Qlo)l1o eqn 5

Despite the fact that it is empirical and has no rational basis, equation 4 has been used extensively in biology, even though it has been known for a consider- able time that the Q10 itself decreases with temperature (Slator 1906; Kanitz 1915). For soil respiration studies Q10 values cited range from 1.3 to 3.3 (Reich & Schlesinger 1992). Much of the variability in the reported Q10 values is probably a consequence of the inherent inaccuracy of the exponential form (see 'Implications for modelling the seasonal cycle'). Nevertheless, we first attempted to fit a Q10 relationship to the data of Fig. 1. This is shown in Fig. 2 (with all data standardized to give a value of 1 for R at

10QC) for which the best-fit curve gives a Q10 of 2.4. The relationship in Fig. 2 accounts for 0 70 of the variation in R as a function of T, but analysis of residuals shows that respiration rates are substantially and systematically underestimated at low temperatures and overestimated at high temperatures. Some workers have previously recognized this problem and have therefore introduced the concept of changing the Q10 itself with temperature (e.g. the third order polynomial Q10 vs temperature relationship of McGuire et al. 1992) but such approaches ignore the basic problem; i.e. the relationship between respiration and temperature is not a simple exponential over the normal range of physiological temperatures. As a quantitative predictor of respiration rates its functional form is inherently wrong.

Van't Hoff appreciated that the exponential relationship was only an empirical convenience to examine temperature effects on different reactions over a limited temperature range (and the Q10 concept still has some utility for that purpose). Indeed, 14 years

OP)12 I i 4

C)~~~~~~~~~Tmeaue(C

CZ ~~~~~~~~~~~~~~~~~~~~~3 a)~ 10

>o 2

_0

8 AM~~~~~~~~~~~~~~~~~~~~~

000 2 3 0 0A0 2 30 4

respiration rate and temperature (equation 6). Also shown is a plot of the residual error term as a function of temperature. Symbols are given in Table 1.



318 CO 12 4 J. Lloyd &

J. A. Taylor a) 10 CZ ~~~~~~~~~~~~~~~~~~~~~2

8 ~~~~~~~~~~~~~A1 0~~~~~~~~~~~~~

A ~~~~~~~~~~1

AA-

a) ~~~~~~~~~~~~~~~~~~~~~~~~~-3 a) 4~~~~~~~~~~~~~~~~~~~~~~~~-

0 10 20 30 40 0 10 20 30 40 Temperature (00)

Fig. 3. Respiration rates (relative to the fitted value at 10 ?C) and the line of best fit for an Arrhenius type relationship between respiration rate and temperature (equation 7). Also shown is a plot of the residual error term as a function of temperature. Symbols are given in Table 1.

before introducing the QIo, it was van't Hoff (1884) who derived the theoretical equation which was the basis of the subsequent rate equation derivation by Arrhenius (1889).

AN ARRHENIUS TYPE DEPENDENCE OF RESPIRATION

RATE UPON TEMPERATURE

In 1884 van't Hoff derived an equation of the form:

d(lnk) E = +C eqn6

dT 9ST2

where T is the absolute temperature (K) and E and c are parameters constant for a particular reaction and 91 is the universal gas constant (8.314Jmol-'K-'). On the basis of experiment, Arrhenius (1889) sug- gested that c=0, which, combined with a considera- tion of energy distributions of reacting molecules enabled him to write:

k =d e-EI(9T) eqn 7

where d is a constant and E is the activation energy. In equation 7 T is the absolute temperature (K). Although deviations from the Arrhenius equation are sometimes observed (e.g. La Mer & Miller 1935; Hulett 1964) it has been found to represent the behaviour of many chemical systems and even some rather complex biological processes (Laidler 1972).

Equation 7 can be rewritten, with the overall reaction rate (kRM) being expressed relative to the rate as some standard temperature, in this case 10TC, as:

/ E 0 l 283 150

R =RIO e(283l5)i 283.15) eqn 8

Using the same procedure as for fitting the exponential Q10 function we have fitted equation 8 to the 15 data sets letting RIO vary between data sets, but fitting

the same activation energy to all data points. This is shown in Fig. 3. The curve in Fig. 3, corresponding to an activation energy of 53 kJ mol-1 provides a better fit to the data (r2 = 0 74) than the exponential function of Fig. 2. Nevertheless, examination of residuals shows that an Arrhenius relationship is also inadequate, similarly underestimating respiration rates at low temperatures and overestimating respiration rates at high temperatures. The theoretically more precise version of the Arrhenius equation given by Eyring (1935) gives a similarly inadequate fit.

The inappropriateness of equation 8, which assumes a constant activation energy is shown in Fig. 4 where the natural logarithm of respiration rate, again standardized to 10 TC is plotted against 1IT (K). For such a plot, the slope is a measure of E and it is

Temperature (0C)

443 299 167 46 25 l

C)~~~~~'

> 1.5

0

> 05

a: -1-

05

r 00030 0 0033 00036 1/T (K-)

Fig. 4. The reciprocal of the respiration rate (relative to the fitted value at 10 C) expressed in relation to the reciprocal of the absolute temperature. The slope of the relationship is a measure of the activation energy.



319 C) 12 | I X 4 Temperature and

co ~~~~~~~~~~~~~~~~~~~~3 soil respiration 10 co ~~~~~~~~~~~~~~~~~~~~~2

0

0 -4 Ax d a) 610-- 2--- 0 1 20 3 .>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~U

coc

-2 0 2~~~~0

h ~~~~~~~~~~~-3

0 1 0 2 0 30 40 0 1 0 20 30 40

Temperature (0C)

Fig. 5. Respiration rates (relative to the fitted value at 10'C) and the line of best fit for an Arrhenius relationship with a decreasing activation energy with increasing temperature (equation 9) between respiration rate and temperature. Also shown is a plot of the residual error term as a function of temperature. Symbols are given in Table 1.

clear from Fig. 4 that the relationship is not linear and that over the entire temperature range E is not constant. The polynomial fit, given by the curve in Fig. 4, suggests an increase in activation energy with decreasing temperature from 37-4 kI mol-' at 40 TC to 77.5 kimol-1 at -5 'C.

Soil respiration involves changing populations of many different organisms, each undergoing a complex series of reactions, some of which may have different temperature sensitivities (ap Rees et al. 1988). Further, even for single-step biological reactions it is likely that the activation energies may indeed increase with decreasing temperature (Hulett 1964), perhaps as a consequence of enzyme deactivation at lower temperatures (Sharpe & De Michelle 1977). Kavanau (1951) utilized kinetic theory to derive a semi-empirical formulation that effectively gives a decrease in activation energy with increasing temperature. Kavanau's formulation for soil respiration rate is:

kR =P [1 -i t e-tdtj eqn 9

where x = E01(T- To) and E, no longer has the theoretical significance of an activation energy, To is a temperature somewhere between T and OK and p is a constant. Allowing p to vary between data sets, but forcing common values of E, and To gives the best-fit curve of Fig. 5 for which E0=400 0K and To= 223-8 K. This shows an even better fit than Fig. 3 (r2 = 0-78) but, more importantly, equation 9 gives an unbiased distribution of residuals, being an equally good predictor of respiration rate at all temperatures.

Equation 9 is however unnecessarily complicated as a simple exponential function, viz.:

-E,,

R =Ae (T T) eqn 10

with E0= 308-56 K and To= 227*13 K and where A is

again a data set-dependent variable gives an even better fit (r2= 0-79). As for the case of a constant activation energy, we can express equation 10 in terms of the respiration rate at 10 TC:

E, 1 - 308.56 - _ __ 283 15-T- T-TI 5602 T-227 13 R=R10e JoRe '12 eqnl

Respiration rates predicted using equation 11 are plotted against those used to fit the model in Fig. 6. This shows excellent agreement between the two estimates. Equation 11 provides an accurate unbiased estimate of soil respiration rates across a wide range of ecosystem types and soil temperatures.

IMPLICATIONS FOR MODELLING THE SEASONAL CYCLE

Examination of time trends in the concentration and isotopic composition of atmospheric CO2 indicate

?' 12,,

E

0

E 8~~~~

of C: O

-a 42 0

C,)

0 0 4 8 12 Measured soil respiration rate (kmol CO2 m2 s-1)

Fig. 6. Predicted values of soil respiration rate from equation 11 plotted against the reported values used to fit the model. Symbols are listed in Table 1.



320 6 2-5

J. Lloyd & J. A. Taylor CD0

0

_ '20/

0 -0

-10 0 10 20 30

Temperature (0C)

Fig. 7. The relationship between respiration rate and leaf temperature for Q10 of 1.5, 2.0 and 3 0 and (solid line) the relationship given in equation 12.

systematic intra-annual variations mainly reflecting the changes in the activity of the terrestrial biosphere (Keeling et al. 1989). An accurate modelling of the

seasonality of gas exchange by the terrestrial bio-

sphere is essential for the accurate simulation of

atmospheric CO2 concentrations (Fung, Tucker &

Prentice 1987; Heimann, Keeling & Tucker 1989)

and also provides constraints on the location of net sinks for the uptake of CO2 (Taylor 1989; Taylor &

Lloyd 1992). Ecosystem process models attempting to model the

seasonal cycle of terrestrial gas exchange have tradi-

tionally used a Q10 of around 2 (e.g. Reich et al.

1991), even though respiration studies suggest a mean

Q10 of around 2 5 (Reich & Schlesinger 1992). To model the seasonality of respiration for use in a three- dimensional atmospheric tracer transport model Heimann et al. (1989) used a Q10 of 1.5 with the additional constraint that no respiration occurred at or below -10?C. These different respiration vs temperature relationships are shown, along with the curve of equation 11 in Fig. 7 (in all cases RIO= 1 0). This fig- ure shows that there is no single Q10 value that approximates equation 11. Effectively, equation 11 gives a decrease in Q10 with increasing temperature which, as mentioned previously, is a long-known phe- nomenon (Slator 1906).

Simple simulations of the effects of radiation, temperature and ambient humidity on the seasonal cycle in terrestrial gas exchange are shown in Fig. 8, for which we have simulated respiration using equation 11 and obtained Gross Primary Productivity (GPP) estimates from a whole-tree gas-exchange model which solves leaf temperature budgets and includes stomatal responses to light, temperature and vapour pressure deficit, as well as leaf mesophyll responses to light and temperature (Lloyd et al. 1994; Syversten & Lloyd 1994). We have provided the constraint that over 1 year R = GPP, and hence that the area examined was neither a net sink nor a net source for CO2. We ignored small intra-annual variations in the soil carbon pool which may occur due to the different sea- sonalities of soil respiration, litter fall, etc. Given that

60O40N 135-OOW 39-30N 0O20W 28-OON 82-OOW

0

.a-

U)

0 500

0~

5 00I

0 C-)

_ GSNF . . .

5 -300 - E

3-00

-5:*I00 __________

12 3 45 6 7 89101112 1 2 34 5 67 8 9101112 1 2 3 4 5 67 8 9101112

Month

Fig. 8. Seasonal variation in (0) Gross Primary Productivity (from the model of Lloyd et al. 1994 and Syvertsen & Lloyd

1994), (v) respiration using the relationship in equation 11 and (-) the difference between the two values (net flux). Also

shown is the cumulative integral of the net flux (annual mean = 0), and the growing season net flux (GSNF). These indicate

effects of the modelled fluxes on the ambient partial pressure of CO2.



321 the residence time of soil carbon is on average around Temperature and 32 years (Raich & Schlesinger 1992), this assumption soil respiration is appropriate. We are aware that such a simulation

ignores plant respiration which may have a seasonality modified by factors other than temperature, and perhaps a different temperature sensitivity. We are also ignoring the possibility of soil water deficits affecting either R or GPP. Nevertheless, such a simulation is instructive and we have undertaken it for a boreal type climate (Whitehorse, Canada; 60.4 ?N 135.0 ?W: modelled GPP = 34 mol C m-2

year-'; leaf area basis, leaf area index = 2); a temperate zone climate (Valencia, Spain; 39.3'N0.2'W: modelled GPP = 71 mol C m-2year-1) and a subtropical climate (Lake Alfred, USA; 28.0 ?N 82.0 ?W: modelled GPP = 76 mol C m72year-1). This shows that the seasonal amplitude of respiration can be less or more than for GPP but that the annual cycle of respiration consistently lags GPP by a month or so. This results in maximum net influxes of CO2 into the biosphere in late spring/early summer. Integrating the net flux allows one to view the pattern of the seasonal cycle in atmospheric CO2 that would be observed for the region being examined in the absence of atmospheric mixing. Minimum values are observed in high-latitude regions in mid-summer with the amplitude of the seasonal cycle decreasing with latitude as one approaches the equator. The cumulative net flux of CO2 over the period when the net flux is negative (ecosystem photosynthesis exceeds ecosystem respiration) is commonly referred to as the 'growing season net flux' (GSNF) and is an indication of the magnitude of the net forcing of seasonal variations in atmospheric CO2 (Pearman & Hyson 1980; Fung et al. 1983). In this case the 'growing season' defines the period of time during which the total mass of carbon in the area in question (plants, litter and soil) is increasing. It does not necessarily correspond to the period of an increase in plant biomass. From the lower parts of Fig. 8 the GSNF can be calculated as the difference between the maximum and minimum cumulative ecosystem fluxes. This gives GSNF of around 7 mol C m-2 year-' at the boreal and temperate sites and about 4 5 mol C m-2 year-l in Florida. Thus our modelled GSNF are 0 06 to 0.21 of our modelled GPP. Fung et al. (1987) estimated a global GSNF of 078 pmol C year-', which corresponded to 020 of their modelled global net primary productivity (NPP). Taking a global NPP/GPP ratio of 0 4 (J. Lloyd & G. D. Farquhar, unpublished data), this gives a glob- ally averaged GSNF/GPP ratio of 0 08.

From our models of GPP and R we conclude that seasonal variations in atmospheric CO2 are due to differences in both the timing and the amplitude in the seasonal cycle of GPP and R. This arises mainly as a consequence of surface air temperature lagging day- length and also because of the greater annual variation in incoming radiation as opposed to surface temperature. In warmer areas such as Florida, high leaf tem-

peratures and associated high leaf-to-air vapour pressure deficits in summer months result in stomatal closure and reduced GPP compared to estimates based on radiation and air temperature alone. This dampens the seasonality in GPP, an effect which is even more present in arid climates. This occurs even in the absence of soil water deficits (Syvertsen & Lloyd 1994).

The temperature sensitivity assumed for respiration has a large effect on the predicted seasonal cycle. For example, in cold climates, even a Q10 of 2 5 is not high enough to simulate equation 11. On the other hand, compared to equation 11 a Q10 of 1.5 more than doubles the modelled GSNF in cold climates but has much less effect at high temperatures (not shown). The difference at low temperatures would be less apparent if respiration was constrained to equal zero at all temperatures less than -10TC (Heimann et al. 1989), but as can be seen from Fig. 7, equation 11 automatically goes to very low values at such temperatures and requires no empirical adjustments of that kind.

Fung et al. (1987) used a different approach to modelling the seasonality of respiration by deriving straight line equations from previously published respiration and temperature studies, using different equations for four different biome groups (temperate/boreal needle-leaved vegetation; temperate/boreal broad-leaved vegetation; tropical/subtropical vegetation and grasslands). However, Fig. 5 suggests a marked curvilinearity in the respiration vs temperature response, and this can also be seen in Fig. 3b of Fung et al. (1987). Further, for tropical/subtropical woody vegetation and grasslands Fung et al. (1987) had to scale the temperature sensitivity according to the mean monthly maximum surface temperature. The relationship demonstrated here makes no distinc- tion between biomes, nor does it require a modifica- tion of the temperature sensitivity of respiration with temperature itself.

EFFECTS OF HIGHER GLOBAL TEMPERATURES ON

SOIL RESPIRATION AND SOIL CARBON POOLS

The likely future warming of the globe due to anthro- pogenic release of greenhouse gases (Houghton et al. 1990) should have a impact on kR, but it is unlikely that the effective soil pool (M) will remain unchanged. Therefore simple equations such as equation 11 cannot provide accurate predictions of changes in soil respiration in response to increased temperature. Even models which take variations in M into account (Jenkinson, Adams & Wild 1991) assume that the rate of carbon -input into the soil (from litter fall and root turnover) is constant and will neither increase nor decrease with future higher temperatures and elevated CO2 levels. This is unlikely to be the case (Taylor & Lloyd 1992) and hence, in the absence of accurate information on the temperature



322 A I . . J. Lloyd & 0 J. A. Taylor 0.2 -

0

35-

0.1 5 . 25

C,

U) 1.5

-10 0 10 20 30 Temperature (0C)

Fig. 9. The relative sensitivity of soil respiration to changes in soil temperature (RTS) (equation 12). Also shown is the RTS for Q10 of 1.5, 2 5 and 3 5.

and CO2 sensitivity of Net Primary Productivity (NPP) we limit the current analysis to the temperature sensitivity of kR itself. There is the additional complication that kR does not continue to increase with temperatures much above 35TC (Bunt & Rovira 1955). Therefore, over the temperature range that equation 11 has been calibrated, by keeping M constant we obtain the relative sensitivity of kR to temperature change. That equation, which can be safely applied to the temperature range -10TC to 30TC is:

1 dkr Eo n 12 kr dT (T-T0)2

lIkRx dkRIdT, the relative temperature sensitivity (RTS) is plotted as a function of temperature in Fig. 9. This shows that at 0C a 1 TC increase in temperature would cause a 22% increase in kR whereas at 25 TC the increase would only be about 5%. Thus, in the absence of moisture limitations, ecosystems associated with low soil temperatures such as tundra and boreal forest have the greatest relative sensitivity of soil respiration rate to changes in soil temperature. Nevertheless, we cannot draw any conclusions about the sensitivity of soil carbon pools to temperature change as we do not know whether the relative temperature sensitivity of NPP and hence carbon input into soils also shows such a marked non-uniformity.

Conclusions

The temperature sensitivity of soil respiration is best- described by an empirical relationship that effectively gives an increase in activation energy with decreasing temperature. When combined with seasonal estimates of GPP this relationship provides a realistic simulation of net ecosystem productivity and its effect on atmospheric CO2 concentrations. Valid conclusions regarding the effects of possible changes in temperatures on soil carbon pools cannot however yet be

made, as detailed knowledge of the temperature sensitivity of carbon input into the soil (via NPP) is also required.

Acknowledgements

We thank H. Marshall for meticulous extraction of data from the original sources.

References

Anderson, J.M. (1973) Carbon dioxide evolution from two temperate deciduous woodland soils. Journal of Applied Ecology, 10, 361-378.

Arrhenius, S. (1889) Uber die Reaktionsgeschwindigkeit bei der Inversion von Rohrzucker durch Sauren. Zeitschrift fur Physik Chemique, 4, 226-248.

Bridge, B.J., Mott, J.J. & Hartigan, R.J. (1983) The forma- tion of degraded areas in dry savanna woodlands of northern Australia. Australian Journal of Soil Research, 21, 91-104.

Bunt, J.S. & Rovira, A.D. (1955) Effect of temperature and heat treatment on soil metabolism. Journal of Soil Sci- ence, 55, 129-136.

Chapman, S.B. (1979) Some interrelationships between soil and root respiration in lowland Calluna heathland in southern England. Journal of Ecology, 67, 1-20.

DOrr, H. & Miinnich, K.O. (1987) Annual variation is soil respiration from selected areas of the temperate zone. Tellus, 39B, 114-121.

Eyring, H. (1935) The activated complex and the absolute rate of chemical reactions. Chemical Reviews, 17, 65-77.

Froment, A. (1972) Soil respiration in a mixed oak forest. Oikos, 23, 273-277.

Fung, I.Y., Prentice, K., Matthews, E., Lerner, J. & Russell, G. (1983) Three-dimensional tracer model study of atmospheric CO2: response to seasonal exchanges with the terrestrial biosphere. Journal of Geophysical Research, 88, 1281-1294.

Fung, I.Y., Tucker, C.J. & Prentice, K.C. (1987) Applica- tion of advanced very-high resolution radiometer vegetation index to study atmosphere-biosphere exchange of CO2- Journal of Geophysical Research, 92, 2999-3015.

Gupta, S.R. & Singh, J.S. (1981) Soil respiration in a tropical grassland. Soil Biology and Biochemistry, 13, 261-268.

Heimann, M., Keeling, C.D. & Tucker, C.J. (1989) A three- dimensional model of atmospheric CO2 transport based on observed winds: 3. Seasonal cycle and synoptic timescale variations. Aspects of Climatic Variability in the Pacific and the Western Americas, Geophysical Monograph, no. 55 (ed. D. H. Peterson), pp. 277-303. American Geophysical Union, Washington D.C.

Hersterberg, R. & Seigenthaler, U. (1991) Production and stable isotopic composition of CO2 in a soil near Bern, Switzerland. Tellus, 43B, 197-205.

van't Hoff, J.H. (1884) Etudes de dynamique chemique. Frederik Muller & Co., Amsterdam.

van't Hoff, J.H. (1898) Lectures on Theoretical and Physi- cal Chemistry. Part I. Chemical Dynamics (translated by R. A. Lehfeldt), pp. 224-229. Edward Arnold, London.

Howard, J.A. & Howard, D.M. (1979) Respiration of decomposing litter in relation to temperature and moisture. Oikos, 33, 457-465.

Houghton, J.T., Jenkins, C.J. & Ephraums, J.J. (eds) (1990) Climate Change, The IPCC Scientific Assessment. Cam- bridge University Press, Cambridge.



323 Temperature and soil respiration

Hulett, J.R. (1964) Deviations from the Arrhenius equation. Quarterly Reviews, 18, 227-242.

Jenkinson, D.S., Adams, D.E. & Wild, A. (1991) Global warming and soil organic matter. Nature, 351, 304-306.

de Jong, E. & Schappert, H.J.V. (1972) Calculation of soil respiration and activity from CO2 profiles in the soil. Soil Science, 113, 328-333.

Kanitz, A. (1915) Temperatur und Lebensvorgange, pp. 151-165. Gebruder Borntraeger, Berlin.

Kavanau, J.L. (j1951) Enzyme kinetics and the rate of biological processes. Journal of General Physiology, 34, 193-209.

Keeling, C.D., Bacastow, R.B., Carter, A.F., Piper, S.C., Whorf, T.P., Heimann, M., Mook, W.G. & Roeloffzen, H. (1989) A three-dimensional model of observed atmospheric CO2 transport based on observed winds: 1. Ana- lysis of observational data. Aspects of Climatic Variability in the Pacific and the Western Americas, Geo- physical Monograph, no. 55 (ed. D. H. Peterson), pp. 165-236. American Geophysical Union, Washington D.C.

Kucera, C. & Kirkham, D. (1971) Soil respiration studies in tallgrass prairie in Missouri. Ecology, 52, 912-915.

Laidler, K.J. (1972) Unconventional applications of the Arrhenius law. Journal of Chemical Education, 5, 343-345.

La Mer, V.K. & Miller, M.L. (1935) Temperature dependence of the energy of activation in the dealdolization of diacetone alcohol. Journal of the American Chemical Society, 57, 2674-2680.

Lloyd, J., Wong, S.C., Styles, J., Turnbull, C., McConchie, C.A. & Batten, D. (1994) Measuring and modelling whole tree gas exchange. Australian Journal of Plant Physiology, (submitted).

McGuire, A.D., Melillo, J.M., Joyce, L.A., Kicklighter, D.W., Grace, A.L., Moore III, B. & Vorosmarty, C.J. (1992) Interactions between carbon and nitrogen dynamics in estimating net primary productivity for potential vegetation in North America. Global Biogeochemical Cycles, 6, 101-124.

Monteith, J.L., Szeicz, G. & Yabuki, K. (1964) Crop photosynthesis and the flux of carbon dioxide below the canopy. Journal of Applied Ecology, 1, 321-337.

Nakane, K. (1980) Comparative studies of cycling if soil organic carbon in three primeval moist forests. Japanese Journal of Ecology, 30, 155-172.

Nakane, K., Tsubota, H. & Yamamoto, M. (1984) Cycling of soil carbon in a Japanese red pine forest I. Before a clear-felling. Botanical Magazine (Tokyo), 97, 39-60.

Pearman, G. & Hyson, P. (1980) Activities of the global biosphere as reflected in atmospheric CO2 records. Journal of Geophysical Research, 85, 4457-4467.

Peterson, KA.' & Billings, W.D. (1975) Carbon dioxide flux from tundra soils and vegetation as related to temperature at Barrow, Alaska. American Midland Natural- ist, 94, 88-98.

ap Rees, T., Burrell, M., Entwhistle, T.G., Hammond, J.B.W., Kirk, D. & Kruger, N. (1988) Effects of low temperature on the respiratory metabolism of carbohydrates

by plants. Plants and Temperature (eds S. P. Long & F. I. Woodward), pp. 377-393. The Company of Biologists Limited, Cambridge.

Reich, J.W., Rastetter, E.B., Melillo, D.W., Kicklighter, D.W., Steudler, P.A., Peterson, B.J., Grace, A.L., Moore III, B. & Vorosmarty, C.J. (1991) Potential net primary production in South America. Ecological Applications, 1, 399-429.

Reich, J.W. & Schlesinger, W.H. (1992) The global carbon dioxide flux in soil respiration and its relationship to climate. Tellus, 44B, 81-99.

Reinke, D.C., Adriano, D.C. & McLeod, K.W. (1981) Effects of litter alteration on carbon dioxide evolution from a South Carolina Pine forest floor. Soil Science Society ofAmerica Journal, 45, 620-623.

Richards, B.N. (1981) Forest floor dynamics. Production in Perpetuity, pp. 145-157. Proceedings of the Forest Nutrition Workshop, Canberra, Australia. CSIRO Divi- sion of Forest Research, Canberra.

Sharpe, P.J. & De Michelle, D.W. (1977) Reaction kinetics of poikotherm development. Journal of Theoretical Biol- ogy, 64, 649-670.

Silvola, K., Vdlijokki, J. & Aaltonen, H. (1985) Effect of draining and fertilization on soil respiration at three ame- liorated peatland sites. Acta Forestalia Fennica, 191, 1-32.

Simunk, J. & Suarez, D.L. (1993) Modeling of carbon dioxide transport and production in soil 1. Model development. Water Resources Research, 29, 487-497.

Singh, J.S. & Gupta, S.R. (1977) Plant decomposition and soil respiration in terrestrial ecosystems. Botanical Review, 43, 449-528.

Slator, A. (1906) Studies in fermentation. I. The chemical dynamics of alcoholic fermentation by yeast. Journal of the Chemical Society, 89, 136-142.

Svensson, B.H. (1980) Carbon dioxide and methane fluxes from the ombotrophic parts of a subarctic mire. Ecologi- cal Bulletin, 30, 235-350.

Syvertsen, J. & Lloyd, J. (1994) Citrus. CRC Handbook of Environmental Physiology of Fruit Crops (eds P. Anders- son & B. Schaeffer), (in press).

Taylor, J.A. (1989) A stochastic Lagrangian atmospheric transport model to determine global CO2 sources and sinks-a preliminary discussion. Tellus, 41B, 272-285.

Taylor, J.A. & Lloyd, J. (1992) Sources and sinks of CO2. Australian Journal of Botany, 40, 407-418.

de Vries, D.A. (1963) The thermal properties of soils. Physics of Plant Environment (ed. R. W. van Wijk), pp. 210-235. North Holland, Amsterdam.

Yoneda, T. & Karita, H. (1978) Soil respiration. Biological Production in a Warm-Temperate Evergreen Oak Forest of Japan (eds T. Kira, Y. Ono & T. Hosokawa), pp. 239-249. University of Tokyo, Tokyo.

Witkamp, M. (1966) Decomposition of leaf litter in relation to environment, microflora and microbial respiration. Ecology, 47, 194-201.

Received 15 April 1993; revised 22 October 1993; accepted 28 September 1993



on the temperature dependence of soil respiration · pdf fileintroduction release of co2 from...

Documents