UK losses of soil carbon – due to climate change?

Pat BellamyNatural Resources DepartmentCranfield UniversityUK

Outline

• Brief overview of Carbon losses from all soils across England and Wales 1978-2003 published in Nature2005

• Potential causes of carbon loss• Changes that have occurred in UK over the twenty

years between samplings• How we are investigating what is driving the

change in soil carbon• Preliminary results based on simple models• Work in progress

The National Soil Inventory of England & Wales

• Whole of England & Wales sampled at each intersect of a 5 km x 5 km grid – 1979-1983.

• Soil profiles described and topsoil (0-15cm) samples taken

• For each sample organic carbon, pH, metal concentrations and nutrients measured.

• Site properties such as land use, slope, aspect recorded.

The National Soil Inventory was madeto obtain an unbiased estimate of the distribution of the soils of England and Wales and of the chemistry of the topsoil (0–15 cm depth)

20m

20m

Sampling at each site

Soil organic carbon content c.1980

Resampling of NSI

• Proportion of NSI sites were re-sampled: 1994-1996 (cropland and managed

grassland) 2002-2003 (forestry, moorland,

extensive grazing land)• Sampling scheme designed to detect changes of organic

carbon ± 2 g kg-1

Resampling of the NSI

Protocols verified for:

• Original surveyor• Field sampling – methods

and tools• Site location• Laboratory analyses

Results from resampling of NSI

Rat

e of

cha

nge

(g k

g-1 y

r-1)

-6

-4

-2

0

2

ArableBog

Conife

rous w

ood

Decidu

ous w

ood

Rotatio

nal g

rass

Lowlan

d hea

th

Perman

ent g

rass

Rough

graz

ingScru

b

Upland

gras

s

Upland

heath

/ moo

r

© Bellamy et al (2005)

Annual change in carbon – grouped by land use

Results from NSI across all land uses

Corg (g kg-1)0 100 200 300 400 500

Rat

e of

cha

nge

(g k

g-1 y

r-1)

-8

-6

-4

-2

0

2All land uses

Eqn (1)

Eqn (1): Rate of change = 0.6 - 0.0187 × Corg

Annual change in carbon (g/kg/yr)

© Bellamy et al (2005)

Changes in soil carbon 1978-2003

Annual change in carbon g/kg/yr

•Carbon lost from soils at a mean rate of 0.6 gm of organic carbon per kg of soil per year.•For soils with more than 100 gm per kg carbon rate of loss greater than 2 gm per kg per year

Estimated annual soil C loss:England & Wales

≈ 4.4 million tonnesUK ≈ 4.4 x UK / E&W C stock

≈ 13 million tonnes

For comparison:UK industrial CO2 emission

≈ 140 million tonnesReduction since 1990

≈ 13 million tonnes

Annual changes in soil carbon stocks

Organic Carbon (%)No Data 0 - 1 1 - 2 2 - 5 5 - 10 10 - 25 25 - 35 > 35

Potential causes of carbon losses

Land use/managementExtension of agriculture- e.g. grassland and woodland

conversion to cropland

Intensification of land use- e.g. drainage, increased nutrient

loading, pesticide use

Non-agricultural land uses- e.g. afforestation on wet land,

managed burning

Environmental changeAtmospheric- e.g. decreased acid rain,

increased nitrogen deposition

Climatic - e.g. temperature and

precipitation

Mean monthly UK temp and rainfall 1978-2003

Month of year1 2 3 4 5 6 7 8 9 10 11 12

Tem

pera

ture

(o C)

0

5

10

15

20

Rai

nfal

l (cm

)

6

8

10

12

14

16Tmax

Tmin

rain

(a) Annual pattern

0 5 10 15 20

∆ m

onth

ly T

max

(o C)

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

Year after 19780 5 10 15 20

∆ m

onth

ly T

min

(o C)

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

0 5 10 15 20∆

mon

thly

pre

cip

(cm

)

-4

-2

0

2

4JanFebMarAprMayJunJulAugSepOctNovDec

Jan

(b) Changes since 1978

Changes in soil pH between NSI samplings

Original pHAnnual change

in pH(over 12-25 yr)

How do we investigate what is driving thechange in soil carbon?

• We do not have detailed information on many of the potential drivers of soil carbon change at the NSI sites

• An alternative approach is to use simple models of soil carbon turnover to evaluate possible explanations for the observed trends based on changes in land management or environmental change including climate change

• We can reject explanations for the changes that require unrealistic parameters to deliver the measured magnitudes of change

Soil carbon dynamics

AtmosphereCO2

immobilization

Plant carbon

Leached carbon

Soil C accumulates ifimmobilization > mobilization

Lost ifmobilization > immobilization

mobilization

Soil carbon

Simple carbon model

Rate of change = Input – k x organic carbon(k = decomposition rate constant)

• Assume Input & k changed some time before first sampling –i.e. soils are adjusting to earlier change - Input & k independent of OC, dependant on land use only

• Assume change in decomposition rate, k, (proportional to organic carbon) – i.e. decomposition rate changes in response to climate change

• Assume change in Input (proportional to carbon content) – i.e. Input changes in response to environmental change

The model fitted to data for each land use class

Simple carbonmodel:results

0 2 4 6 8 10 12 14 16 18-1.2-1.0-0.8-0.6-0.4-0.20.00.20.40.60.81.0

observedpredicted

0 2 4 6 8 10 12 14 16 18

Rat

e of

cha

nge

(kg

C m

-2 y

-1)

-1.2-1.0-0.8-0.6-0.4-0.20.00.20.40.60.81.0

Mean soil carbon content (kg C m-2)

0 2 4 6 8 10 12 14 16 18-1.2-1.0-0.8-0.6-0.4-0.20.00.20.40.60.81.0

Arable

Permanent grass

Non-agricultural

Arable and rotational grassland

Permanent grassland

Non-agricultural

Simple carbon model: results

• The model parameter estimates were assessed against known effects of land use/management change, climate change and environmental change

Simple carbon model: results

• Estimates of k and Input obtained for each land use

• If k changed proportional to OC the large changes required suggest climate change cannot be sole driver

• Large changes in Input required suggest environmental changes not a major driver

Initial conclusions

Results so far indicate the dominant driver of soil carbon losses was changes in land use/management, but climate change was responsible for part of the losses, and is likely to cause increased losses in future

Further work in progress

• More sophisticated models of carbon change are being developed which will represent the whole range of soils in England and Wales

• These models are being validated using the NSI data• These will allow more precise estimation of the

contribution of climate change to the change in soil carbon observed in the NSI data

Acknowledgements

• Funding for the National Soil Inventory of England and Wales came from Defra (Department for Environment Food and Rural Affairs UK)

• I would like to thank past and present colleagues at the National Soil Resources Institute at Cranfield University including Ian Bradley, Guy Kirk, Peter Loveland, Richard Andrews, Ilkka Leinonen and Bob Jones, also Murray Lark at Rothamsted Research.

Carbon losses from all soils across England and Wales 1978-2003 (2005) Pat H. Bellamy, Peter J. Loveland, R. Ian Bradley, R. Murray Lark & Guy J.D. Kirk Nature 437 pp245 - 248