Elevated CO2 does not increase eucalypt forest productivity on a low-phosphorus soil

TL;DR: In this paper, a large-scale experiment on a mature, phosphorous-limited eucalypt forest showed that aboveground productivity was not significantly stimulated by elevated CO2, despite a sustained 19% increase in leaf photosynthesis.

Abstract: Experimental evidence from a mature, phosphorous-limited, eucalypt forest finds that aboveground productivity was not significantly stimulated by elevated CO2. Findings suggest that this effect may be limited across much of the tropics. Rising atmospheric CO2 stimulates photosynthesis and productivity of forests, offsetting CO2 emissions1,2. Elevated CO2 experiments in temperate planted forests yielded ∼23% increases in productivity3 over the initial years. Whether similar CO2 stimulation occurs in mature evergreen broadleaved forests on low-phosphorus (P) soils is unknown, largely due to lack of experimental evidence4. This knowledge gap creates major uncertainties in future climate projections5,6 as a large part of the tropics is P-limited. Here, we increased atmospheric CO2 concentration in a mature broadleaved evergreen eucalypt forest for three years, in the first large-scale experiment on a P-limited site. We show that tree growth and other aboveground productivity components did not significantly increase in response to elevated CO2 in three years, despite a sustained 19% increase in leaf photosynthesis. Moreover, tree growth in ambient CO2 was strongly P-limited and increased by ∼35% with added phosphorus. The findings suggest that P availability may potentially constrain CO2-enhanced productivity in P-limited forests; hence, future atmospheric CO2 trajectories may be higher than predicted by some models. As a result, coupled climate–carbon models should incorporate both nitrogen and phosphorus limitations to vegetation productivity7 in estimating future carbon sinks.

Summary

Methods

• Methods, including statements of data availability and any associated references, are available in the online version of this paper.
• Diameter class (cm) These effects are shown for leaf net photosynthesis (a, left side) and aboveground net primary production, ANPP from 2013 to 2015 (b, right side).
• The mixed-model repeated-measures analysis for photosynthesis was done using data shown in Fig. 1a ), with the time term indicating sampling date across three years.
• For ANPP, the time term is 'year', the first to third year of the full eCO2 treatment.
• In both analyses, a mixed-model repeated-measures analysis was done using a fixed treatment (CO2) and a random plot effect, and Type III sums of squares computed using restricted maximum likelihood estimates for F-tests.

Methods (online)

• Six large circular plots (0.05 ha each) were established in 2010 in a mature eucalypt woodland on an alluvial spodosol in western Sydney, Australia.
• The location receives 800 mm of precipitation per annum on average and has a mean annual temperature of 17.5°C (www.bom.gov.au).
• For the leaf component, the productivity was computed as the sum of annual litterfall whilst for twigs the authors assume strictly annual turnover across the three years.
• The authors analysed the photosynthesis data 35 using a mixed-model repeatedmeasures analysis of variance in R v3.3.1 using the 'lme4' function within the 'nlme' package, with CO2 treatment as a fixed factor and plot as a random factor nested within CO2 treatment.
• All data were checked for normality using the Q-Q plots and Levene's test, and residuals from model fitting were checked for evidence of heteroscedasticity.

Full text

Open Research Online
The Open University’s repository of research publications
and other research outputs
Elevated CO
2
does not increase eucalypt forest
productivity on a low-phosphorus soil
Journal Item
How to cite:
Ellsworth, David S.; Anderson, Ian C.; Crous, Kristine Y.; Cooke, Julia; Drake, John E.; Gherlenda, Andrew
N.; Gimeno, Teresa E.; Macdonald, Catriona A.; Medlyn, Belinda E.; Powell, Jeff R.; Tjoelker, Mark G. and Reich,
Peter B. (2017). Elevated CO2 does not increase eucalypt forest productivity on a low-phosphorus soil. Nature
Climate Change, 7(4) pp. 279–282.
For guidance on citations see FAQs.
c
 2017 Macmillan Publishers Limited
https://creativecommons.org/licenses/by-nc-nd/4.0/
Version: Accepted Manuscript
Link(s) to article on publisher’s website:
http://dx.doi.org/doi:10.1038/nclimate3235
Copyright and Moral Rights for the articles on this site are retained by the individual authors and/or other copyright
owners. For more information on Open Research Online’s data policy on reuse of materials please consult the policies
page.
oro.open.ac.uk

1
Elevated CO
2
does not increase eucalypt forest productivity on a low-
phosphorus soil
David S. Ellsworth
1

, Ian C. Anderson
1
, Kristine Y. Crous
1
, Julia Cooke
2
, John E.
Drake
1
, Andrew N. Gherlenda
1
, Teresa E. Gimeno
3
, Catriona A. Macdonald
1
, Belinda
E. Medlyn
1
, Jeff R. Powell
1
, Mark G. Tjoelker
1
, Peter B. Reich
1,4
1
Hawkesbury Institute for the Environment, Western Sydney University, Penrith, NSW 2751,
Australia,
2
School of Environment, Earth and Ecosystem Science, The Open University,
Milton Keynes MK7 6AA, UK,
3
ISPA, Bordeaux Science Agro, INRA, 33140 Villenave
d'Ornon, France,
4
Department of Forest Resources, University of Minnesota, St. Paul, MN,
55108 USA.

e-mail: D.Ellsworth@westernsydney.edu.au

2
Rising atmospheric CO
2
stimulates photosynthesis and productivity of forests, offsetting 1
CO
2
emissions
1,2
. Elevated CO
2
experiments in temperate planted forests yielded ~23% 2
increases in productivity
3
over the initial years. Whether similar CO
2
stimulation occurs 3
in mature evergreen broadleaved forests on low-phosphorus (P) soils is unknown, 4
largely due to lack of experimental evidence
4
. This knowledge gap creates major 5
uncertainties in future climate projections
5,6
as a large part of the tropics is P-limited. 6
Here, we increased atmospheric CO
2
concentration in a mature broadleaved evergreen 7
eucalypt forest for three years, in the first large-scale experiment on a P-limited site. We 8
show that tree growth and other aboveground productivity components did not 9
significantly increase in response to elevated CO
2
in three years, despite a sustained 10
19% increase in leaf photosynthesis. Moreover, tree growth in ambient CO
2
was 11
strongly P-limited and increased by ~35% with added phosphorus. The findings suggest 12
that P availability may potentially constrain CO
2
-enhanced productivity in P-limited 13
forests; hence, future atmospheric CO
2
trajectories may be higher than predicted by 14
some models. As a result, coupled climate-carbon models should incorporate both 15
nitrogen and phosphorus limitations to vegetation productivity
7
in estimating future 16
carbon sinks. 17
18
Limited understanding of the size of the CO
2
-induced fertilisation effect on forest carbon 19
sinks remains among the largest quantitative uncertainties in terms of terrestrial feedbacks to 20
the carbon (C) cycle-climate system
6,8,9
. Coupled climate-C cycle models project a 24-80% 21
increase of net primary productivity (NPP) for forests in the next 50 years with rising 22
atmospheric CO
2
concentration, with substantial atmospheric CO
2
responses expected for 23
forests in the tropics
4,10
. These model projections are partly based on elevated CO
2
(eCO
2
) 24
experiments in young temperate planted forests, which have yielded on average ~23% 25

3
increases in production
3
over several years with 200 µmol mol
-1
increases in atmospheric 26
CO
2
concentrations
4,11
. Due to the lack of experimental evidence, we presently do not know 27
how large the eCO
2
fertilisation response is for mature forests that grow on soils where 28
phosphorus (P) is limiting productivity
4,10
, as is the case for many evergreen broadleaved 29
forests. This knowledge gap creates major uncertainties in future climate projections
9
because 30
evergreen broadleaved forests comprise over a third of global forest area, and dominate the 31
atmospheric CO
2
sink at lower latitudes
5,6
. Many eCO
2
experiments have taken place in 32
young tree plantations
3
on relatively P-rich soils, but unlike aggrading forests, mature forests 33
are more likely near nutritional equilibrium with their underlying soils. Hence mature forests 34
may be more appropriate for understanding in situ nutrient limitations to productivity and C 35
storage with rising atmospheric CO
2
. Without clear understanding of this nutrient feedback to 36
the C cycle in evergreen broadleaved forests, we cannot accurately estimate the trajectory of 37
future atmospheric CO
2
, thus limiting our ability to estimate climate change mitigation by 38
such forests and constrain internationally-allowable CO
2
emissions
9,12
. 39
40
Soil nutrient limitation may restrict eCO
2
-induced biomass enhancement and related C 41
storage processes
11
, but it is unclear if the type of nutrient limitation is important. Studies in a 42
temperate grassland and a forest ecosystem under contrasting CO
2
and N supply suggest a 43
large initial stimulation in productivity, often followed by reduced CO
2
stimulation when N is 44
limiting
13,14
. Limited P supply might affect tree growth and ecosystem C sequestration 45
processes differently than the N-supply limitation
15
that has thus far been demonstrated in 46
eCO
2
experiments on N-poor soils. In heavily weathered soils common in tropical and 47
subtropical regions, P is typically bound to Fe and Al oxides, hydroxides and secondary 48
minerals and not available to plants. One possibility is that increased plant carbohydrate 49
availability from eCO
2
leads to increased plant investment in the secretion of organic acids 50

4
from roots
16
or the investment in P-acquisition by mycorrhizal symbionts. This would thereby 51
reduce P-limitation to broadleaved evergreen forest productivity
17
by increasing plant access 52
to scarce soil P. Consistent with this idea, there is evidence that recent rising CO
2
may have 53
driven a substantial portion of the observed historical increase in tropical forest carbon 54
stocks
18
though future increases remain in question. 55
56
Although there is considerable variation in soil fertility across the world, tree growth in 57
highly weathered tropical and sub-tropical soils may be limited by P availability in addition 58
to, or rather than, N availability
19,20
. Hence nutrient availability and the type of nutrient 59
limitation may both be important in regulating forest CO
2
fertilisation responses in those 60
regions
7,17
. There is still little agreement on how to appropriately represent P limitations to 61
productivity in Earth systems models
7,21
, and there has been no direct experimental test of the 62
CO
2
fertilisation effect in P-limited forests (Supplementary Fig. 1). 63
64
To help fill this gap, we established a free-air CO
2
enrichment experiment on six circular 25m 65
diameter plots in mature Eucalyptus forest (EucFACE) on a low P soil near Sydney, Australia 66
(23 m elevation; 33° 37' 4" S, 150° 44' 25" E) (Supplementary Fig. 2). The main canopy 67
species, Eucalyptus tereticornis, has a distribution through tropical and temperate zones. 68
EucFACE has unique characteristics compared to prior forest elevated CO
2
experiments: the 69
presence of mature broadleaved evergreen trees in natural unmanaged forest, and nutrient-70
poor soil with a demonstrated P limitation to tree growth
22
. A gradual CO
2
enrichment began 71
in Sept 2012 at 30 µmol mol
-1
above ambient CO
2
concentration, and slowly ramped up to 72
the full-strength eCO
2
treatment of 150 µmol mol
-1
above ambient CO
2
concentration
23
, 73
which began on 6 Feb 2013. This full CO
2
treatment was maintained throughout the 74
following three years (Feb. 2013-Feb. 2016) that are the focus of this report. We 75

Frequently asked questions

Q1. What are the contributions in "Elevated co2 does not increase eucalypt forest productivity on a low-phosphorus soil" ?

Elevated CO2 does not increase eucalypt forest productivity on a low-phosphorus soil this paper. 

Q2. How was the CO2 effect size calculated?

379 Confidence intervals for the CO2 effect size estimate were computed in R (http://cran.r-380 project.org) using the function ‘confint’, which applies quantile functions for the t-381 distribution after model-fitting. 

Q3. How many trees were measured across the ambient and elevated plots?

Of a total of 146 trees measured across the ambient and 353 elevated plots, 49 suppressed trees, 6 co-dominant trees with trunk defects, and 4 trees 354 showing shrinkage possibly preceding mortality were omitted from the mixed-model 355 analysis. 

Q4. What is the mean of the ANPP in a mature Eucalyptus?

Stemwood 281 production is determined as the annual biomass increment, and foliage+fine twig production 282 are measured as annual biomass turnover collected monthly in permanent litter baskets. 

Q5. What was the productivity of the leaf 367 component?

For the leaf 367 component, the productivity was computed as the sum of annual litterfall whilst for twigs the authors 368 assume strictly annual turnover across the three years. 

Q6. What is the mean temperature in the warmest month?

Mean maximum temperature in the warmest month is 30°C and mean 319 minimum temperature in the coldest month is 3.6°C, with monthly mean temperatures always 320 > 10°C. 

Q7. What was the effect of the CO2 on the tree crown?

A smaller set of 339measurements on shaded foliage within the tree crowns was used to confirm results from the 340 upper-crown measurements in terms of the CO2-enhancement effect on photosynthesis, thus 341 the entire crown can be expected to behave similarly. 

Q8. How was the data checked for normality?

All data were checked for normality 386 using the Q-Q plots and Levene’s test, and residuals from model fitting were checked for 387 evidence of heteroscedasticity. 

Q9. What is the mean response ratio for photosynthesis in a mature Eucalyptus?

The CO2 fertilisation response ratio for photosynthesis over time, 268 with grey areas representing two-sided 95% confidence intervals for the CO2 fertilisation 269 response ratio for each of the measurement timepoints. 

Q10. What is the ANPP of the Eucalyptus?

Total ANPP is represented by the 279 combination of stemwood biomass production (stippled), fine twig and bark production 280(striped), seed and capsule production (hatched), and leaf production (solid). 

Q11. How long did the eCO2 study last?

The P-addition treatments were 331 maintained through the duration of the study, resulting in 4 years of P-fertilisation concurrent 332 with the 3-year eCO2 study. 

Q12. How was the diameter of each tree measured?

The diameter 345 of each tree was measured at 1.3 m height at approximately monthly intervals starting 346 February 2011, 2 years prior to commencement of the full CO2 treatment. 

Q13. How many trees were measured across all years?

The authors thus used a total of N=87 trees measured across all years and without stem 356 defects, suppression or shrinkage in the mixed-model analyses. 

Q14. how many datasets are available in a research data australia repository?

The datasets generated during and/or analysed during the current study are 390 available in a Research Data Australia repository (http://doi.org/10.4225/35/57ec5d4a2b78e).