16 El

This is a repository copy of Tropical land carbon cycle responses to 2015/16 El Niño as recorded by atmospheric greenhouse gas and remote sensing data. White Rose Research Online URL for this paper: http://eprints.whiterose.ac.uk/135234/ Version: Accepted Version Article: Gloor, E orcid.org/0000-0002-9384-6341, Wilson, C, Chipperfield, MP et al. (12 more authors) (Accepted: 2018) Tropical land carbon cycle responses to 2015/16 El Niño as recorded by atmospheric greenhouse gas and remote sensing data. Philosophical Transactions of the Royal Society B: Biological Sciences. ISSN 0962-8436 (In Press)

Reuse Items deposited in White Rose Research Online are protected by copyright, with all rights reserved unless indicated otherwise. They may be downloaded and/or printed for private study, or other acts as permitted by national copyright laws. The publisher or other rights holders may allow further reproduction and re-use of the full text version. This is indicated by the licence information on the White Rose Research Online record for the item. Takedown If you consider content in White Rose Research Online to be in breach of UK law, please notify us by emailing [email protected] including the URL of the record and the reason for the withdrawal request.

[email protected] https://eprints.whiterose.ac.uk/

Submitted to Phil. Trans. R. Soc. B - Issue

rR

Fo !"

#!

% "

&

$

'

iew

ev

( ) *+ , '& # ) . )% *+ , '& # ) % ' 0 1! , % ' #) *+ , '& # ) % , ) 2 3#3 * & %0) " ) . '- -* / ) )4 " *+ , '& ) " 5 7 ) ! *+ , '& ) "7 ) 5 *+ , '& ) " ( )& * /50 % ) % * /50 ( )& * /50 5 ). *. + , ) * /1 $0 " , ) * /% " % " , ) *+ , '& # )

'( '( '(

" " "

#

, #!

0# 1''

'

, '( -

/ /1

9 !: =

>

#

0 ,

"

%

%

!

; )( !

.

1&1(< " -) 2

)

http://mc.manuscriptcentral.com/issue-ptrsb

-

'5

6

'5 '5

6 6

"

" & !

ly

On

% # 8 # !

* /0 % /

2

$( !

-

Page 1 of 44

1 2 3 4

Emanuel Gloor, Chris Wilson, Martyn P. Chipperfield, Frederic Chevallier, Wolfgang

5

Buermann, Hartmut Boesch, Rob Parker, Peter Somkuti, Luciana Gatti, Caio Correia, Lucas

6

Gatti, Wouter Peters, John Miller, Merritt N. Deeter, Martin Sullivan

7 8

Fo

9

The outstanding tropical land climate characteristic over the past two2three decades is rapid

10

warming, while there are no significant large2scale precipitation trends. Warming is expected

11

to continue but effects on tropical vegetation are unknown. El Niño2related heat peaks may

12

provide a test bed of what to expect. Here we analyse tropical land carbon cycle responses to

13

the 2015/16 El Niño climate anomalies using an atmospheric transport inversion. Based on

14

the global atmospheric CO2 record and long2term behaviour of the global carbon cycle, we

15

find no obvious signs of anomalously large carbon release during the 2015/16 El Niño

16

compared to earlier El Niño events. We find roughly equal net carbon release anomalies from

17

Amazonia and tropical Africa, ~0.5 PgC each, and smaller carbon release anomalies from

18

tropical East2Asia and Southern Africa. Atmospheric CO anomalies reveal fire carbon release

19

from tropical East Asia peaking in October 2015 while Amazonian fires played a smaller

20

role. The positive flux anomalies are consistent with down2regulation of primary productivity

21

during peak negative near2surface water anomaly (Oct 2015 2 Mar 2016) as diagnosed by

22

solar2induced fluorescence. Finally, we find an anomalous positive flux to the atmosphere

23

from tropical Africa early in 2016, coincident with substantial CO release.

iew

ev

rR

ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


24 25



-

0

'

27

%

28

al. 2009) and regulating climate by virtue of their exchanges of carbon, water and energy

29

with the atmosphere. They also play an important role in sustaining endangered fauna and

30

their continued presence is essential for preserving their rich biodiversity. More generally,

31

tropical biomes are home to great cultural diversity and growing economies, and will need to

32

support half the global population by 2050 (State of the Tropics Report, 2017). Thus they

33

have a large impact on livelihoods in these climates. The continued functioning and

34

productivity of vegetation in the tropics is, however, dependent on its response to changing

35

climatic conditions. The dominant climate change signatures across the tropics are rapid

36

warming and an increase of extreme events, severe floods and anomalously dry conditions

37

(Fig. 1, and e.g. Fischer and Knutti 2015).

=

'

?)5@

%

512), storing two2thirds of global plant biomass (e.g. Pan et

ev

rR

Fo

38

iew

39

El Niño events may provide a test bed to examine tropical vegetation responses, likely to be

40

dominated by the response of forests, to these increasingly higher temperatures, paralleled

41

usually by drier than usual conditions. This is because during El Niño events strong positive

42

temperature excursions tend to be spatially correlated with negative precipitation anomalies.

43

These anomalies occur in tropical South2East Asia, tropical South America and to a lesser

44

extent tropical West Africa and Southern Africa (roughly below 10°S, e.g. Dai et al. 2000,

45

Foley et al. 2002).

ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 44

46 47

It has been known since at least the 1970s that El Niño events co2occur with periods of

48

anomalously large atmospheric CO2 growth rates (Bacastow et al. 1976). There is not only a

49

strong correlation between the El Niño index (in essence atmospheric sea surface pressure

50

difference between Darwin and Tahiti in the tropical Pacific) and global atmospheric CO2


Page 3 of 44

)1

1

%

1 &

1

52

temperature anomalies (e.g. Wang et al. 2014). The mechanism causing the correlation with

53

temperature is not entirely clear. One component is increased biomass burning. It has further

54

been argued that water limitation on vegetation performance is important at the local scale,

55

but temperature anomalies are more important at larger scales due to cancelling effects (Jung

56

et al. 2017). The strong correlation between atmospheric growth rate anomalies and El Niño

57

Index suggests that variation of tropical land carbon uptake and release contributes

58

prominently to anomalous atmospheric CO2 growth during positive El Niño phases.

59

Nonetheless, ocean air2sea gas exchange does also play a role. Measurements of this process

60

in the tropical Pacific reveal that during El Niño outgassing in the tropics is reduced, i.e.

61

ocean carbon pool response is in the opposite direction of land carbon pools (Feeley et al.

62

1999). This is because the tropical Pacific thermocline upward tilt towards South America is

63

reduced during El Niño phases of ENSO (El Niño Southern Oscillation) which hinders

64

upwelling of carbon2rich waters along the tropical South American west coast and thus

65

carbon efflux from the sea to the atmosphere is reduced. The decrease over a full El Niño

66

period for the 1997/98 event has been estimated using ocean data to be 0.6±0.1 PgC (Feely et

67

al. 1999, and 2002). Recent estimates of global air2sea gas exchange based on air2sea partial

68

pressure difference measurements and gas exchange parameterization by Feely et al. (2017)

69

suggest a smaller anomaly over the 2015/16 period on the order of 0.120.2 PgC.

iew

ev

rR

Fo

ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


70 71

Similar to the other manuscripts in this volume we attempt to analyse whether, and to what

72

extent, the response of tropical land vegetation during the 2015/16 El Niño event is different

73

from responses during previous similar events, and thus may be a harbinger of future

74

responses not just to climate oscillations but climate variation on top of a rapidly increasing

75

temperature background. There have been, for example, reports indicating that increased



.-

+6

% %

77

=

517). There have also been reports indicating that dry seasons may get

78

drier across the tropics (Murray2Tortarola et al. 2017). It is not clear what the effect of

79

tropical vegetation productivity may be given both stimulating (rising CO2) and limiting (e.g.

80

the increase in leaf2air water vapour pressure difference) factors. The overarching theme of

81

this manuscript is thus whether the 2015/16 events reveal signs of anomalously increasing

82

vegetation strain.

83

Fo

84

Our analysis is based primarily on a large2scale atmospheric approach which as the main tool

85

uses an inverse model of atmospheric transport (INVICAT, Wilson et al. 2014) to extract

86

information about surface CO2 exchange between land vegetation and atmosphere contained

87

in spatio2temporal variations of atmospheric CO2. Our approach is helped by new data in

88

tropical South America measured by INPE (Instituto Nacional de Pesquisas Espaciais), Sao

89

Jose dos Campos, Brazil. We focus on inter2annual variation of fluxes which should be more

90

robustly estimable than absolute flux magnitudes. To put our results into context we relate the

91

fluxes to climate controls and to distinguish processes, to some extent, we employ solar

92

induced chlorophyll fluorescence and atmospheric carbon monoxide measured from space.

93

We aim to address the following questions: How anomalous is the global CO2 flux anomaly?

94

What are the climate deviations / excesses on land? Where and when do flux anomalies occur

95

and how large are they? How much is due to fire and under what conditions? How much is

96

due to reduction in primary production versus changes in respiration? And finally, are there

97

signs of land vegetation responses outside the usual El Niño patterns given the unprecedented

98

temperatures during the 2015/16 event?

iew

ev

rR

ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 44

99 100

!

"


Page 5 of 44

101

#

102

A possible approach to estimate atmospheric CO2 growth rate anomalies Δ , suggested to our

103

knowledge first by Jones et al. (2001), is as follows

104 Δ (t) =

( )−

∙

( )

105

(1)

106

Here

107

fossil fuel burning and cement manufacture, and

108

the ratio of the annual atmospheric carbon growth rate and fossil fuel emissions. Thus

109

∙

is atmospheric carbon content (in form of CO2), is time,

is global emissions from

is the long2term mean airborne fraction,

Fo

( ) is the expected average increase of atmospheric carbon growth rate for given

rR

110

fossil fuel emissions

in year . Fossil fuel emission estimates used here are from Boden

111

et al. (2017)

112

which are based on energy statistics, and the observed atmospheric CO2 record (the Mauna

113

Loa record)

114

on December 31/January 1 (i.e. mean from July 1 to June 30). As a sensitivity test we have

115

repeated this calculation using

116

conclusions.

iew

ev

~= 0.55. Growth rate is calculated as the difference of annual means centred

118

$

% &

%

$'

ly

117

=0.49 (mean over 190122015) with similar overall

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


#

(

%

119 120

The most robust information provided by atmospheric CO2 concentration records is the

121

global atmospheric inventory and how it changes over time. This reveals, for example, very

122

clearly the well2known rapid increase of atmospheric CO2 over the last decades. In addition

123

to global information the wide2spread surface station observation network, maintained by

124

various groups and in particular NOAA/ESRL (Fig. S1), exhibit spatio2temporal patterns

125

which reflect regional scale variation in CO2 exchange between the land surface and oceans



126

1

0

% &

127

distribution and strength of regional surface fluxes, provided the relationship between fluxes

128

and the concentration patterns they cause can be established. This relationship involves

129

representation of the processes of atmospheric advection and dispersion which can be

130

estimated fairly well using numerical fluid dynamics models of the atmospheric flow

131

('atmospheric transport models') (e.g. Chipperfield et al. 2006). The relation to the actual

132

atmospheric flow is established by using wind and cloud convection transport fields derived

133

from regular observations of the state of the atmosphere for purpose of weather prediction.

134

Flux estimation reduces then to a least2squares minimization problem of the difference

135

between a linear combination of concentration fields resulting from localised fluxes in space

136

and time sampled at the same time and location as the observations, and the actual

137

observations. This problem turns out to be poorly constrained by the number of available in

138

situ measured data and thus a possible approach is to instead optimally combine a set of prior

139

flux 'guesses' %p with the flux estimates which replicate concentration data most closely

140

(Enting et al. 1985), i.e. to minimize

iew

ev

rR

Fo

On

141 ( )=( −

) ∙

∙( −

)+( − ) ∙

142

∙ ( − ) ( )

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 44

143

with respect to %. ) is the a priori flux error covariance matrix,

144

observed atmospheric concentrations and * is the transport2model2calculated matrix, which

145

relates surface fluxes to the atmospheric concentration signal they cause at the sampling sites.

146

This approach solves in essence for small deviations from a prescribed flux model. For this

147

problem an explicit expression for the posterior flux error covariance matrix

148

derived (Tarantola and Valette, 1982):

post

is a vector containing the

=[*t ∙ +21∙* ,)21-21 .

149


post

can be

Page 7 of 44

150

1

%

151

transport model TOMCAT (Chipperfield, 2006). We resolve fluxes monthly and spatially on

152

a grid 5.6o × 5.6o longitude by latitude. The prior flux model includes three components: (i)

153

annually changing fossil fuel emissions, (ii) monthly net land gains or losses which do not

154

change from year to year, based on the CASA (Carnegie Ames Stanford) land biosphere

155

model (average climatology for 200322011). The model estimates primary productivity as the

156

product of solar photo2synthetically active radiation (PAR), land vegetation chlorophyll

157

content (estimated from space NDVI) and a light use efficiency. Respiration is estimated

158

using a carbon cycle model which includes soils (Olsen and Randerson, 2004), and (iii) air2

159

sea fluxes.

ev

rR

160

Fo

161

We have chosen annually repeating and balanced land vegetation 2 atmosphere CO2 flux prior

162

estimates because our interest is in extracting the information on inter2annual variations

163

contained in atmospheric data. For our prior estimates of air2sea fluxes we treat separately the

164

fluxes associated with the pre2industrial carbon cycle (two hemispherical loops with CO2

165

outgassing in the tropics and CO2 uptake at high latitudes) and uptake of carbon induced by

166

the anthropogenic perturbation of atmospheric CO2 (with fluxes steadily increasing and

167

located primarily in the Northern Atlantic and Southern Ocean) (Gloor et al. 2003, Khatiwala

168

et al. 2009). For the former we use the monthly resolved climatology based on air2sea partial

169

pressure differences and an air2sea gas exchange coefficient parameterization, compiled by

170

Takahashi et al. (2009), to which we add a constant and spatially uniform flux such that the

171

fluxes are globally in balance on an annual basis. For the latter we used the spatial air2sea

172

flux pattern of Khatiwala et al. (2009) (their Fig. 1b) which we scale with global net ocean

173

uptake taken from the Global Carbon Project analysis (Global Carbon Project, 2017). This

174

approach leads to improved a posteriori data model fits compared to inversions which use as

iew

ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60




175

2

4

%

2 2

176

measurement differences alone (specifically Takahashi et al. 2009). Minimisation of (%) is

177

done using a quasi2Newtonian method (L2BFGS implemented in M1QN3 minimiser) with

178

gradients calculated with the adjoint of the TOMCAT atmospheric transport model,

179

ATOMCAT (Wilson et al. 2014). We assumed a prior flux uncertainty of 200% per grid cell

180

and we assumed that there is no flux error correlation given the comparably coarse resolution.

181

Atmospheric data are from 81 sites mainly measured by NOAA/ESRL and include in

182

addition the PBL mean (below 2500 m) and the free troposphere mean (above 2500 m) of

183

vertical profile data in the Amazon measured by INPE, Sao Jose dos Campos, Brazil, (e.g.

184

Gatti et al. 2014) (Fig. S1 and S2). Observational data has an uncertainty of 1 ppm, plus an

185

estimate of representation error derived by averaging the absolute prior concentration

186

variation between the model grid cell containing the measurement location and the

187

surrounding grid cells. This leads to overall observational uncertainties of between 1 and 6

188

ppm, depending on the measurement location. To assess the influence of the Amazon vertical

189

profile data we have also performed an inversion without these data. The effect of including

190

the data is to reduce the magnitude of flux anomalies while the timing and location of

191

anomalies are not affected much (section 3).

iew

ev

rR

Fo

ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 44

192 193

./ (

%(

#

194

One cause of plant water stress is negative deviations (or anomalies) of the abundance of soil

195

water (or soil water content) from the climatological mean representative for a region. A

196

proxy for soil water content over large spatial scales can be measured from satellites because

197

large2scale land surface water content anomalies cause Earth gravity anomalies. Such gravity

198

anomalies are being estimated from space by the twin satellite mission GRACE (Gravity

199

Recovery and Climate Experiment, Tapley et al., 2004) with the satellites following each


Page 9 of 44

55

%

51

B

1

1

202

again once the front satellite has passed the anomaly and the rear satellite is approaching the

203

anomaly. To confirm the realism of the gravity anomaly data measured from space we

204

compared gravity anomaly anomalies with precipitation anomalies measured by TRMM

205

(Tropical Rainfall Measuring Mission, Huffman et al. (2007), Fig. S4). The signatures of the

206

two data types are very consistent (taking into account that gravity anomalies are to first

207

order equal to cumulative precipitation anomalies). To calculate monthly gravity anomaly

208

anomalies we subtracted monthly means calculated using the full 200222016 record from the

209

continuous record of monthly means.

210 211

%% &

(

%

iew

212

0

ev

rR

Fo

213

We use daytime CO air column inventories estimated from MOPITT radiometer data on the

214

TERRA satellite (Deeter et al. 2013) to estimate carbon emissions from fires. To do so we

215

first estimate carbon monoxide fluxes from monthly CO air column anomalies. We then

216

convert the carbon monoxide fluxes to carbon fluxes, assuming they are from fires, by

217

multiplying the carbon monoxide fluxes with a biomass burning emission ratio of (1/74)

218

(ppm CO2)/(ppb CO) (Gatti et al. 2010) although this ratio may vary with the type of fire.

219

The MOPITT CO record we are using is version 6 (version 6 (L3V95.2.3), Deeter 2013)

220

which covers March 2000 to December 2016. This version uses both thermal infrared (TIR)

221

and near infrared (NIR) radiances and so, compared to the other two MOPITT products (TIR2

222

only and NIR2only), it provides the maximum sensitivity to surface2level CO. Nonetheless,

223

because of the non2uniform weighting function of the retrievals, column content estimates

224

may contain a bias (an underestimate of column CO if signals are concentrated to the lower

ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60




)

> =

')> 0

2

. / 3

0 9

4

8 1

4 =

5 1

8

+

!"

+ #

%$≈ 0.85

1

%

DE − #!

!&

− ∇ ∙ (() ∙ *) 8 0

4

≈ 9 23

≈ 0.1 23 lifetime of CO in the atmosphere, #

%$#!

232

atmosphere, 0

233

to oxidation of non2methane hydrocarbons,

234

to the total air volume above a fixed region to obtain a relation between a CO flux

235

perturbation 5 at the earth surface and the 5() anomalies it causes:

236

5 =

237

normal vector to the vertical walls :; orthogonal to the surface region of interest, and df is

238

an infinitesimal area element of :;. The contribution of in2 and outflow into and out of the

239

air volume above the region is negligible if region boundaries are chosen such that Δ() ≈ 0.

Fo

air flow velocity vector. We apply the equation

+

6

− 7 8 5() ∙ < , > df ≃

6

+

6 !&

where,

is an outwards directed unit

iew

!&

ev

6

1

%

21 34

ly

On

240 241

CO volume source due

rR

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 44

242

Photosynthesis is associated with fluorescence. A small fraction of solar radiation trapped by

243

chlorophyll escapes instead of being used to fix CO2. This fraction is reemitted into the

244

atmosphere from the leaf at larger wavelengths (in the range of 670 and 800 nm, e.g. Meroni

245

et al. 2009) compared to the originally trapped radiation. Fluorescence has been shown to be

246

related to productivity (Yang et al. 2015, van der Tol et al. 2014) and thus we use it here as a

247

proxy for productivity. We specifically use here SIF retrieved from GOSAT (Greenhouse

248

Gases Observation Satellite, Kasuya and Hamazaki, 2005) measurements at 772 nm using the


Page 11 of 44

249

2%

9

%

& %

= 511). The bias

250

correction2procedure, which is an essential part of the post2retrieval processing, was

251

performed using the European Space Agency Climate Change Initiative land cover maps

252

(Bontemps et al. 2013). GOSAT measurements over permanently non2vegetated areas (where

253

zero fluorescence can be assumed) were identified using these maps in order to derive

254

radiance2dependent calibration curves on a monthly basis. Based on these monthly curves, a

255

2D2spline interpolation was used along time and radiance dimensions to obtain the bias

256

correction term for any given GOSAT sounding. This ensures that the time2dependence of

257

the instrument2related bias is taken into account. The retrievals are available for the period

258

April 2009 to September 2016.

rR

Fo

259

ev

260

Monthly compilations of SIF retrievals exhibit missing values. To obtain sufficient data

261

coverage we therefore calculated quarterly (32monthly) means and similarly calculated

262

anomalies for three2month periods. Despite lumping three months together there are still

263

pixels with no retrievals. To calculate quarterly anomalies we therefore kept track of number

264

of existing retrievals on an individual pixel basis. For comparison of anomalies for specific

265

areas and 32monthly periods we calculated region mean anomalies including only those

266

pixels for which retrievals exist.

iew

ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


267 268

. +

269

.

* #

270

( (

5

271

The largest annual global atmospheric CO2 increase rates recorded with modern analytical

272

tools (i.e. since 1959) occurred in 2015 and 2016 with values of 2.94 and 2.85 ppm,

273

respectively

(NOAA/ESRL,

Boulder,


Colorado,

USA;


.!

E77

7

7

7

7

P P

4> 1

275

1998 (2.81 ppm). While of concern per se, to detect changes of El Niño land

276

vegetation responses at the global scale, the fossil fuel contribution to the growth rate needs

277

to be separated from other flux contributions, in particular nonlinearly increasing fossil fuel

278

emissions. As explained in Section 2.1 to achieve this we assume a constant fossil fuel

279

airborne fraction which we subtract from the atmospheric carbon inventory growth rate

280

(Figure 2). The anomalies in 2015 and 2016 were positive and when summed up were

281

approximately 1.7 PgC, with the 2016 anomaly being approximately twice as large as the

282

2015 anomaly. Including the reduction in CO2 outgassing from tropical oceans during

283

positive El Niño episodes based on the air2sea flux estimates summarized in Feely et al.

284

(2017), then the total flux anomaly of global land carbon to the atmosphere is ~ 1.92 2.1 PgC

285

over the two years (201522016).

iew

286

ev

rR

Fo

287

A comparison with the 1997/98 El Niño anomaly reveals that the 2015/16 anomaly was not

288

extraordinarily large, certainly of a smaller magnitude than the 1997/98 anomaly. This

289

conclusion does not depend much on the period chosen to estimate the airborne fraction

290

(Section 2.1). From a vegetation process response point of view, the 1997/98 anomaly is,

291

however, somewhat unusual in that it includes a strong direct human2impact large2scale peat

292

drainage component which in 1997/98 led to 'catastrophic' peatland / peat forest fires and

293

carbon release (Page et al., 2002). Thus, part of the 1997/98 positive anomaly is unrelated to

294

climate2induced variation in productivity and respiration of living vegetation or soil

295

respiration in a strict sense. A noticeable indirectly El Niño2related aspect revealed by growth

296

rate anomalies (Fig. 2) is the negative (land carbon uptake) anomalies from roughly 2008

297

onwards.

ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

298


Page 12 of 44

Page 13 of 44

33 55 51

# B

4 =

>

1

+6

7

4 2

2

302

usual conditions, we briefly summarize here measures of vegetation stress related to climate.

303

The first and primary measure is plant water stress caused by negative deviations (or

304

anomalies) of the abundance of soil water (or soil water content) from the climatological

305

mean representative for a region. We use here monthly gravity anomaly anomalies measured

306

by GRACE (Section 2.3) as a proxy for soil water stress. Although these anomalies include

307

both below and aboveground water anomalies our use here as a vegetation water stress

308

indicator is supported by anti2correlation between annual pan2tropical land / tropical South

309

American land gravity anomaly anomalies and global atmospheric CO2 growth rate

310

anomalies g (see Section 2.1) shown in Fig. 3 (Pearson r=20.69 and 20.72, and p=0.0046 and

311

0.0025, respectively).

iew

ev

rR

312

Fo

313

The main features of vegetation water deficits during the 2015/16 El Niño period according

314

to both gravity anomaly as well as precipitation data (Fig. 3, Fig. S4) are (i) in the Amazon

315

basin, an east2to2west spreading and steadily increasing area with large water deficit, with

316

this process starting at the beginning of 2015. The deficit reaches its peak and covers the

317

entire basin by the first quarter of 2016, with water deficit remaining high throughout the

318

basin until the final quarter of 2016; the most pronounced negative precipitation anomaly

319

occurred during the final quarter of 2015 all across the basin; (ii) in Africa one notable signal

320

is a considerable water deficit developing south of roughly 10°S during the first three

321

quarters of 2016. The anomaly is not as strong as for Amazonia. The most pronounced

322

precipitation anomaly for the region southward of 10°S occurred during the final quarter of

323

2015, i.e. a bit earlier than in the other two regions. A second notable feature is excessively

ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60




!

%

325

>

%

=

9

:

O

:

516, while according to gravity anomalies and precipitation

326

estimated by TRMM there were no very clear indications of drought conditions but this will

327

need further investigation; (iii) tropical South2East Asia experienced strong negative

328

precipitation anomalies and associated water stress during the second half of 2015. These

329

three regions experienced strongly elevated temperatures nearly synchronously with the

330

substantially drier than usual conditions, with peak temperatures all exceeding existing

331

historical records (Figs. 4 & S4).

Fo

332 333

Among the three continents the climate anomalies for the Amazon seem to be the strongest

334

with temperature and precipitation anomalies centred around the last three months of 2015

335

and first three months of 2016, and with the effects of precipitation anomalies on soil

336

moisture lasting over nearly all of 2016, reflecting the time it takes for water deficits to

337

propagate through the catchment (Fig. S3).

339

Overall the observed climate anomalies are similar to the canonical El Niño patterns as

340

described e.g. by Dai and Wigley (2000), with the tropical Asian precipitation anomaly being

341

somewhat weaker. The possible exception is the Congo basin, which was excessively hot

342

during the first quarter of 2016.

ly

On

338

iew

ev

rR

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 44

343 344

$

% &

%

$'

(

%

345 346

What do atmospheric inversion results suggest? Motivated by the relation between inter2

347

annual variation of the global atmospheric CO2 growth rate and gravity anomaly anomalies

348

on tropical land we compare land CO2 flux anomalies with gravity anomaly anomalies for


Page 15 of 44

!3

'

:

%

)5

$: 4

=

)> 0

4

1

)1

0

352

synchronicity of positive flux anomalies (fluxes to the atmosphere) with negative gravity

353

anomaly anomalies and vice versa (Pearson r=20.42 for monthly means, p < 1023), which is

354

consistent with the global record (Fig. 2). This result demonstrates the inversion’s ability to

355

detect and attribute expected flux anomalies from the atmosphere data. Splitting up the flux

356

estimates by those regions with notable climate anomalies we find the following (Fig. 6).

357

According to our calculations four regions released significant amounts of carbon during the

358

201522016 period: Tropical South America, Tropical Africa, Southern Africa and Tropical

359

East Asia. The losses from tropical South America and tropical Africa are similar in

360

magnitude while losses from Tropical East Asia and Southern Africa are smaller (Table 1).

361

The timing of peak carbon losses differs between the regions, with peak in October 2015 for

362

tropical East Asia, February 2016 in South Africa, February and March 2016 for tropical

363

Africa November2December 2015 and March2April 2016 for tropical South America.

iew

ev

rR

Fo

364 365

Table 1 Tropical land carbon flux anomalies Period

ly

Region

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Net carbon flux to Biomass atmosphere

carbon flux

(PgC)

(PgC)

Tropical S America

Sep 2015 2 Jun 2016

0.5 ± 0.3

0.0520.1

Tropical Africa

Nov 2015 2 Jul 2016

0.6 ± 0.3

0.0820.16

Southern Africa

Jan 2016 2 May 0.2 ± 0.1 2016


burning


0

'

:

'

5152 Dec 2015 0.2 ± 0.1

0.320.4

366 367 368

!

369

CO2 estimates based just on atmospheric CO2 concentration data and inversion of

370

atmospheric transport provide net fluxes but cannot discern between the different underlying

371

processes, such as biomass burning, or changes in vegetation productivity and respiration

372

processes (e.g. by living trees /vegetation and/or dead organic matter in soils). Here in

373

addition we analyse information from atmospheric total column carbon monoxide (CO)

374

retrieved from MOPITT (Measurements of Pollution in the Troposphere) radiometer for the

375

period 200022017 on the NASA TERRA polar orbiting satellite, as an indicator of release of

376

carbon via fire (Fig. 7) (Deeter et al. 2016) and solar2induced fluorescence (SIF) retrieved

377

from GOSAT radiance measurement as an indicator of land vegetation productivity and

378

covering the period April 2009 to September 2016 (Fig. 8).

% &

380

Monthly CO air column anomalies, Δ %

515 and to lesser extent

392

subsequent months. When spatially integrated over tropical South America the decrease

393

during October2December 2015 is ~20%. To obtain a rough estimate of the associated

394

decrease in carbon uptake we use an estimate of tropical South American vegetation annual

395

productivity (gross primary productivity, GPP) estimated by Jung et al. (2011) based on CO2

396

flux measurement between atmosphere and forest canopies. The annual productivity of

397

tropical South American vegetation according to Jung et al. (2011) is ~18 PgC. According to

398

SIF we obtain a 32month reduction in productivity in this region of 20% and thus obtain a

399

reduction of carbon uptake of ~0.9 PgC during this quarterly period. Given limited evaluation

400

of the SIF2GPP relationship in the tropics this estimate needs to be taken with some caution.

401

For the other quarterly periods fluorescence anomalies are lower and signs of change less

402

coherent across large regions.

iew

403

ev

rR

Fo

404

!

405

From a climate perspective the outstanding development in the tropics on land over the past

406

decades is rapid warming. The 2015/16 El Niño adds a positive temperature anomaly on top

407

of

408

experiment of tropical forest vegetation subject to high temperatures in the future. Because of

409

the elevated background temperatures vegetation responses might be more severe compared

410

to responses observed during previous El Niño events. Based on just the global atmospheric

411

CO2 record we do not find any obvious sign of anomalously large carbon release during the

412

2015/16 El Niño compared to El Niño events in the past. This does not exclude compensating

413

effects at continental to regional scales. At these scales there is a strong spatial correlation

414

between positive temperature peaks and negative soil water anomalies diagnosed via gravity

ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


this already rapidly warming 'background'. It thus provides a natural sensitivity



!15

'

1

4

%

%

416

exchange anomalies which is indeed consistent with correlation of anomalies on land with the

417

global atmospheric CO2 record. By far largest negative near surface water content anomalies

418

occurred in the Amazon Basin during the final quarter of 2015 and the first quarter of 2016.

419

Negative anomalies also occurred during a brief period in tropical East Asia centred on

420

October 2015 and Southern Africa during the first two quarters of 2016. We estimated

421

continental scale CO2 flux estimates based on atmospheric concentration data from surface

422

station networks complemented by vertical profile data in the Amazon. The results of these

423

calculations should be taken with some caution, because uncertainty in model transport can

424

lead to biased flux results (e.g. Gurney et al., 2002). Despite biases, inter2annual variability

425

may still be robust because transport modelling biases will affect all years in a similar way,

426

meaning that correlations with environmental variables can still be reliable. In our study, high

427

correspondence between tropical South American flux anomalies and negative precipitation

428

anomalies gives some confidence in the results, as well as the covariation in time with

429

climate anomalies and atmospheric CO anomalies. We find roughly equal net flux anomalies

430

from Amazon and tropical Africa of around 0.5 PgC each, and somewhat smaller positive

431

flux anomalies from tropical East Asia and Southern Africa. According to atmospheric CO

432

anomalies our analysis attributes anomalous carbon release from tropical East Asia to fires

433

peaking in October 2015, while consistent with fluorescence data from space, biomass

434

burning played a smaller role in the Amazon and with the flux anomaly reasonably consistent

435

with down2regulation of primary productivity during peak negative water anomaly (final

436

quarter of 2015 and first quarter of 2016). The one feature in our results which seems

437

somewhat unexpected, as this is not usually a region considered to be affected significantly

438

by El Niño, is the anomalous flux from tropical Africa coincident with substantial CO release

439

from the Congo Basin, during the first quarter of 2016. Our estimate of CO2 released by fires

iew

ev

rR

Fo

ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 44


Page 19 of 44

!!5

:

!!1

4

4

:

1

%

1 &1

%

$:

1

442

may have caused an anomalous decrease in productivity, SIF data do not give strong support

443

to this mechanism. Thus, in addition to changes in productivity, enhanced heterotrophic

444

respiration may have contributed also to this signal.

445 446

Finally, we examine how our results summarized together with main controls in Table 2

447

compare with the recent analyses of Liu et al. (2017) based primarily on satellite data. For the

448

comparison it is important to realise that the Liu et al. study calculated anomalies with

449

reference to flux estimates from the year 2011, a La Niña year. It is well established that

450

during La Niña years global CO2 growth rate anomalies are strongly negative. Thus Liu et

451

al.'s point of reference is quite different from ours. Taking this into account our results are

452

quite similar with exception of Africa. At pan2tropical level Liu et al. (2017) estimate a

453

difference of flux from land to atmosphere of 2.5 PgC for the period May 2015 2 April 2016

454

compared to Jan 2011 2 Dec 2011. Their specific choice for the 201522016 period is likely

455

motivated by maximum positive anomalies. If we use a similar criteria and thus use the

456

period July 2015 2 June 2016 we find a difference of 2.4 PgC. With regards to tropical Africa,

457

in contrast to Liu et al. (2017), we find a substantial carbon loss from tropical Africa at the

458

same time as the very strong heat peak in the Congo Basin (the beginning of 2016) when a

459

clear CO anomaly also occurred. Our rough biomass burning estimate cannot explain this

460

result on its own 2 thus some down2regulation of tropical forest productivity or enhanced

461

respiration would be needed to explain it. In comparison to Liu et al. (2017) our inverse

462

calculations also attribute less carbon release from Southern Africa during the 201522016 El

463

Niño period.

iew

ev

rR

Fo

ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


464



!-) !-!-. !-/ !-3 !.5 !.1 472

Fo

473 474

Table 2 Chronology and magnitude of carbon flux anomalies ('Cflx') (sign convention based

475

on a land vegetation perspective, i.e. anomalous carbon loss to the atmosphere has a negative

476

sign while anomalous uptake has a positive sign), climate ('H2O': soil water status, 'T'

477

temperature) and process diagnostics: carbon monoxide ('CO') and solar induced fluorescence

478

('SIF'). Symbols indicate existence of positive (+) and negative anomalies (2) and the number

479

of symbols the strength of the anomalies.

iew

ev

rR

480 481 Tropical S America H2O

T

Cflx

JAS

2

++

OND

222

++

22

JFM

222

+

222

AMJ

22

+

JAS

2

Tropical Africa CO

SIF

H2O

T

Cflx

ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 44

Southern Africa CO

SIF

H2O

T

Cflx

CO

SIF

Tropical East Asia H2O

T

Cflx

CO

2015 JFM AMJ

+

222

22

+

22

++

22

2 +

2016

OND

?

2

+++

222

++

22

22

++

22

NA

22

NA

2

482 483


2

+++

SIF

Page 21 of 44

!/! !/)

6

!/!/.

# &

"

!//

1

%

0

N':

% 1 %

+8::7 '$(

+8::78:$7 '$( ',

E77111 &

1

7

7

8

+ $

489

+ 7 55)/5-71, NE/K01353X/1 and NERC MOYA. MPC, CW, HB, PS acknowledge

490

support from NCEO (NERC National Centre for Earth Observation). LVG acknowledges

491

support from FAPESP and CNPQ grants. WP was partly funded by ERC2CoG grant ASICA

492

(649087).

496

iew

ev

494 495

rR

493

Fo

497

EG, CW and MPC conceived the study, CW did the atmospheric transport inverse

498

calculations with contributions from MPC, CF and EG. PS, RP, HB provided solar

499

fluorescence retrievals from GOSAT, LG, CC, WP, CC led / contributed to the Amazon

500

greenhouse data collection and laboratory analysis, MS contributed to the climate analysis

501

and MND provided CO air column retrievals from MOPITT. EG wrote the manuscript. All

502

co2authors commented on and contributed to the science and writing of the manuscript.

ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


503 504 505 506 507 508 509



)10

'

511 512

Table S1 Domains used for spatial integration of air column CO anomalies. Region

Latitude

Longitude

Tropical South America

20.5 S,...,19.5 N

80.5 W,..., 20.5W

Tropical Africa

20.5 S,...,19.5 N

Tropical Asia

20.5S,...,19.5 N

20.5 W,...39.5E 49.5 E, ..., 149.5 E

513

Fo

514 515 516

rR

+ %

ev

517 518

Bacastow, R. B. (1976) Modulation of atmospheric carbon dioxide by the Southern

519

Oscillation, Nature, 1162118.

521

Boden, T.A., G. Marland, and R.J. Andres (2017) Global, Regional, and National Fossil2Fuel

522

CO2 Emissions. Carbon Dioxide Information Analysis Center, Oak Ridge National

523

Laboratory,

524

10.3334/CDIAC/00001_V2017.

U.S.

Department

of

Energy,

Oak

ly

On

520

iew

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 44

Ridge,

Tenn.,

U.S.A.

doi

525 526

Bontemps, S., P. Defourny, J. Radoux, E. Van Bogaert, C. Lamarche, F. Achard, P. Mayaux,

527

M. Boettcher, C. Brockmann, and G. Kirches (2013) Consistent global land cover maps for

528

climate modelling communities: current achievements of the ESA’s land cover CCI, in:

529

Proceedings of the ESA Living Planet Symposium, 9–13 September 2013, Edinburgh, UK,

530

9–13.


Page 23 of 44

) 1 532

Chipperfield, M. P. (2006) New version of the TOMCAT/SLIMCAT offline chemical

533

transport model: Intercomparison of stratospheric tracer experiments, Q. J. Roy. Meteor.

534

Soc., 132, 1179–1203, doi:10.1256/qj.05.51

535 536

Dai A. and T. M. L. Wigley (2000) Global Patterns of ENSO2induced Precipitation,

537

Geophys. Res. Lett., 27(9), 128321286.

538

Fo

539

Deeter, M. N., S. Martínez2Alonso, L. V. Gatti, M. Gloor, J. B. Miller, L. G. Domingues, and

540

C. S. C. Correia (2016) Validation and analysis of MOPITT CO observations of the Amazon

541

Basin, Atmos. Meas. Tech., 9, 3999–4012.

ev

rR

542 543

Deeter, M. N. (2013) MOPITT (Measurements of Pollution in the Troposphere) Version 6

544

Product User's Guide.

546

Fan, Y., and H. van den Dool (2008) A global monthly land surface air temperature analysis

547

for 1948–present, J. Geophys. Res., 113, D01103, doi:10.1029/2007JD008470.

ly

On

545

iew

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


548 549

Feely, R. A., R. Wanninkhof, T. Takahashi & P. Tans (1999) Influence of El Niño on the

550

equatorial Pacific contribution to atmospheric CO2 accumulation, Nature, 398, 5972601.

551 552

Feely, R. A., et al. (2002) Seasonal and interannual variability of CO2 in the equatorial

553

Pacific. Deep2Sea Res., 49, 2443–2469.

554



555

$ : $

&

(

W;

" $

* : 0 6

= 517)

556

BAMS state of climate Chapter 3, Global Oceans, Global ocean carbon cycle, Bulletin

557

Americ. Met. Soc., S892S92.

558 559

Fernandes, K., L. Verchot, W. Baethgen, V. Gutierrez2Velez, M. Pinedo2Vasquez, and C.

560

Martius (2017) Heightened fire probability in Indonesia in non2drought conditions: the effect

561

of increasing temperatures, Environ. Res. Lett. 12, 054002.

562

Fo

563

Fischer, E. M., and R. Knutti (2015) Anthropogenic contribution to global occurrence of

564

heavy2precipitation and high2temperature extremes, Nature Climate Change, 5, 5602564,

565

doi:10.1038/nclimate2617.

ev

rR

566 567

Foley, J. A., A. Botta, and M. T. Coe (2002) El Niño–Southern oscillation and the climate,

568

ecosystems and rivers of Amazonia, Glob. Biogeochem. Cycl., 16 (4), 1132,

569

doi:10.1029/2002GB001872.

iew

On

570 571

Frankenberg, C., J. B. Fisher, J. Worden, G. Badgley, S. S. Saatchi, J.2E. Lee, G. C. Toon, A.

572

Butz, M. Jung, A. Kuze, and T. Yokot (2011) New global observations of the terrestrial

573

carbon cycle from GOSAT: Patterns of plant fluorescence with gross primary

574

productivity,

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 44

, 38, L17706, doi:10.1029/2011GL048738.

575 576

Gatti, L. V., M. Gloor, J. B. Miller, C. E. Doughty, Y. Malhi, L. G. Domingues, L.S. Basso,

577

A. Martinewski, C. S. C. Correia, V. F. Borges, S. Freitas, R. Braz, L. O. Anderson, H.

578

Rocha, J. Grace, O. L. Phillips, and J. Lloyd (2014) Drought sensitivity of Amazonian carbon


Page 25 of 44

).3 580

%

%

+

)5-

.-2/5

E10.1038/nature12957

581 582

Gatti, L. V., J. B. Miller, M.T.S. D’Amelio, E. Gloor, A. Martinewski, L. S. Basso, S. Wofsy,

583

and P. Tans (2010) Vertical profiles of CO2 above eastern Amazonia suggest a near neutral

584

carbon balance between 2000 and 2009, Tellus, 62, 5812594, DOI: 10.1111/j.16002

585

0889.2010.00484.x

586

Fo

587

Gloor M., N. Gruber, J. L. Sarmiento, C. Sabine, R. Feely, and C. Rödenbeck (2003) A first

588

estimate of present and preindustrial air2sea CO2 flux patterns based on ocean interior carbon

589

measurements and models, Geophys. Res. Lett., 30, No. 1 10.1029/2002GL015594.

ev

rR

590

iew

591

Gurney, K. R., R. M. Law, A. S. Denning, P. J. Rayner, D. Baker, P. Bousquet, L. Bruhwiler,

592

Y.2H. Chen, P. Ciais, S. Fan, I. Y. Fung, M. Gloor, M. Heimann, K. Higuchi, J. John, T.

593

Maki, S. Maksyutov, K. Masarie, P. Peylin, M. Prather, B. C. Pak, J. Randerson, J.

594

Sarmiento, S. Taguchi, T. Takahashi, and C.2W. Yuen (2002) Towards more robust estimates

595

of CO2 fluxes: control results from the TransCom3 inversion intercomparison, Nature, 415,

596

6262630.

ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


597 598

Huffman, G.J., R.F. Adler, D.T. Bolvin, G. Gu, E.J. Nelkin, K.P. Bowman, Y. Hong, E.F.

599

Stocker, and B. Wolff (2007) The TRMM Multi2satellite Precipitation Analysis: Quasi2

600

Global, Multi2Year, Combined2Sensor Precipitation Estimates at Fine Scale. J. Hydrometeor.,

601

8, 38255.

602



603

*

,

4

' : '

= 551) The Carbon Cycle Response to

604

ENSO: A Coupled Climate–Carbon Cycle Model Study, J. Clim., 411324129.

605 606

Jung, M., M. Reichstein, H. A. Margolis, A. Cescatti, A. D. Richardson, M. Altaf Arain, A.

607

Arneth, C. Bernhofer, D. Bonal, J. Chen, D. Gianelle, N. Gobron, G. Kiely, W. Kutsch, G.

608

Lasslop,1 B. E. Law, A. Lindroth, L. Merbold,L. Montagnani, E. J. Moors, D. Papale, M.

609

Sottocornola, F. Vaccari and C. Williams (2011) Global patterns of land‐atmosphere fluxes

610

of carbon dioxide, latent heat, and sensible heat derived from eddy covariance, satellite, and

611

meteorological observations, J. Geophys. Res., 116, G00J07, doi:10.1029/2010JG001566

Fo

612

rR

613

Jung, M., M. Reichstein, R. Schwalm, and C. Huntingford (2017) Compensatory water

614

effects link yearly global land CO2 sink changes to temperature, Nature, 541, 516 – 520.

ev

615

iew

616

Khatiwala, S., F. Primeau, and T. Hall (2009) Reconstruction of the history of anthropogenic

617

CO2 concentrations in the ocean, Nature, 462, 3462349.

On

618 619

Kasuya, M., and T. Hamazaki (2005) Overview of the Greenhouse Gases Observing Satellite

620

(GOSAT), Proceedings of WSANE 2005 (Workshop for Space, Aeronautical and

621

Navigational Electronics 2005), KARI, Daejeon, Korea, March 325, 2005, 1452149.

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 44

622 623

Global

Carbon

Project

(2017)

624

[www.globalcarbonproject.org/carbonbudget]

Carbon

budget

and

trends

2017.

625 626

Liu, J., K. W. Bowman, D. S. Schimel, N. C. Parazoo, Z. Jiang, M. Lee, A. A. Bloom, D.

627

Wunch, C. Frankenberg, Y. Sun, C. W. O’Dell, K. R. Gurney, D. Menemenlis, M. Gierach,


Page 27 of 44

628

,

:

= 517) Contrasting carbon cycle responses of the tropical

629

continents to the 2015–2016 El Niño. Science, 358, 191

630 631

Meroni, M. Rossini, L. Guanter, L. Alonso, U. Rascher, R. Colombo, and J. Moreno (2009)

632

Remote sensing of solar2induced chlorophyll fluorescence: Review of methods and

633

applications, Rem. Sens. Env., 113, 2037–2051.

634 635

Mitchell, T. D., and P. D. Jones (2005) An improved method of constructing a database of

636

monthly climate observations and associated high2resolution grids, Int. J. Climatology, 25,

637

6932712, Doi: 10.1002/joc.1181.

ev

rR

638

Fo

639

Murray2Tortarolo, G., P. Friedlingstein, S. Sitch, S. I. Seneviratne, I. Fletcher, B. Mueller, P.

640

Greve, A. Anav, Y. Liu, A. Ahlström, C. Huntingford, S. Levis, P. Levy, M. Lomas, B.

641

Poulter, N. Viovy, S. Zaehle, and N. Zeng (2016) The dry season intensity as a key driver of

642

NPP trends, Geophys. Res. Lett., 43, 2632–2639, doi:10.1002/2016GL068240

iew

On

643 644

Olsen, S., and J. Randerson (2004) Differences between surface and column atmospheric

645

CO2 and implications for carbon cycle research, J. Geophys. Res., 109, D02301,

646

doi:10.1029/2003JD003968.

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


647 648

Page, S. E., F. Siegert, J. O. Rieley, H.2D. V. Boehm, A. Jayak and S. Limink (2002) The

649

amount of carbon released from peat and forest fires in Indonesia during 1997, Nature, 61265.

650 651

Pan, Y., Birdsey, R. A., Fang, J., Houghton, R., Kauppi, P. E., Kurz, W. A., Phillips, O. L.,

652

Shvidenko, A., Lewis, S. L., Canadell, J. G., Ciais, P., Jackson, R. B., Pacala, S. W.,



653 654

: ,

' $

: '

'

#

, : = 511) A large and

persistent carbon sink in the world’s forests, Science, 333, 988–993.

655 656

Scheffers B. R., Joppa L. N., Pimm S. L., and W.F. Laurance (2012) What we know and

657

don't know about Earth's missing biodiversity, Trends Ecol. Evol. 27, 501–10.

658 659

State of the Tropics (2017) Sustainable Infrastructure for the Tropics. James Cook

660

University, Townsville, Australia.

661

Fo

662

Takahashi, T., S. C. Sutherland, R. Wanninkhof, et al. (2009) Climatological mean and

663

decadal change in surface ocean pCO2, and net sea–air CO2 flux over the global oceans.

664

Deep2Sea Res, Pt II 56, 554–77.

iew

665

ev

rR

666

Tarantola A. and B. Valette (1982) Generalized nonlinear inverse problems solved using the

667

least squares criterion, Rev. Geophys. Space Phys., 20, 2192232.

On

668 669

Tapley, B., Flechtner, F., Bettadpur, S., and M. Watkins (2004) The Gravity Recovery and

670

Climate Experiment Mision overview and early results, Geophys. Res. Lett., 31, doi:

671

10.1029/2004GL019920.

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 44

672 673

Van der Tol, C., J. A. Berry, P. K. E. Campbell, and U. Rascher (2014) Models of

674

fluorescence and photosynthesis for interpreting measurements of solar2induced chlorophyll

675

fluorescence, J. Geophys. Res. Biogeosci., 119, 2312–2327, doi:10.1002/2014JG002713.

676 677

Wang, X., S. Piao, P. Ciais, P. Friedlingstein, R. B. Myneni, P. Cox, M. Heimann, J. Miller,


Page 29 of 44

678

'

0

# M

679

sensitivity

680

doi:10.1038/nature12915.

to

tropical

:

= 514) A two2fold increase of carbon cycle

temperature

variations,

Nature,

506,

2122215,

681 682

Wilson, C., M. P. Chipperfield, M. Gloor, and F. Chevallier (2014) Development of a

683

variational flux inversion system (INVICAT v1.0) using the TOMCAT chemical transport

684

model (2014) Geosci. Model Dev., 7, 2485–2500.

685

Fo

686

Yang, X., J. Tang, J. F. Mustard, J.2E. Lee, M. Rossini, J. Joiner, J. W. Munger, A. Kornfeld,

687

and A. D. Richardson (2015) Solar2induced chlorophyll fluorescence that correlates with

688

canopy photosynthesis on diurnal and seasonal scales in a temperate deciduous forest,

689

Geophys. Res. Lett., 42, 2977–2987, doi:10.1002/ 2015GL063201.

ev

rR

690

iew

691 692

On

693 694

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


695 696 697 698 699 700 701 702

3



.5 704

1 Climate trends for tropical and subtropical forest biome based on CRU (Climate

705

Research Unit) TS 3.24 climatology (Mitchell and Jones, 2005). Stippling denotes

706

statistically significant trends.

707 708

Fig. 2 Global atmospheric CO2 growth rate anomalies ∆g estimated using Eq. 2 and El Niño

709

3.4 Index (obtained from http://www.cpc.ncep.noaa.gov).

710

Fo

711

Fig. 3 Tropical land gravity anomaly anomalies measured by the GRACE satellites and

712

global CO2 growth rate anomalies ∆g estimated using Eq. 2.

713

ev

rR

714

Fig. 4 Land gravity anomaly anomalies and monthly means of daily minimum and maximum

715

temperatures, respectively, for tropical land regions. Temperature data are from Climate

716

Prediction Center (CPC), Global Land Surface Air Temperature Analysis (Fan et al. 2008).

718

Fig. 5 Tropical South America to atmosphere CO2 flux anomalies estimated with inverse

719

modelling of atmospheric transport and atmospheric CO2 concentration observations, and

720

tropical South American gravity anomaly anomalies estimated by GRACE satellite mission.

ly

On

717

iew

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

721 722

Fig. 6 Time2series of land2to2atmosphere carbon flux estimates (low2pass filtered) for tropical

723

land regions. The portion for which Amazon vertical profile data have been available and

724

included in the atmospheric transport inversion calculations is coloured in dark black.

725

Stippled red are estimates which do not include tropical South American data.

726


Page 30 of 44

Page 31 of 44

727 728

.0

%

4

515/16 of total air column carbon

monoxide measured from space (MOPITT, Deeter et al. 2013).

729 730

Fig. 8 Solar induced fluorescence anomalies (measured from GOSAT satellite (Kasuya and

731

Hamazaki 2005) and based on retrievals at 772 nm).

732 733

Fig. S1 Locations where CO2 mixing ratio data are being measured which are used to

734

estimate CO2 surface fluxes using atmospheric transport inversions.

Fo

735 736

Fig. S2 Availability of atmospheric CO2 data used for atmospheric transport inversions.

rR

737 738

ev

Fig. S3 Gravity anomaly anomalies as measured by GRACE satellites for 2015 and 2016.

739

iew

740

Fig. S4 Precipitation anomalies calculated from TRMM (Tropical Rainfall Measuring

741

Mission) (Huffman et al. 2007) version 7 and monthly daily maximum and minimum

742

temperature from NOAA Climate Prediction Center (CPC) climatology (Fan et al. 2008).

On

743

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


744

Fig. S5 Example (October 1, 2015) of CO retrieval weighting kernels for total column CO

745

estimation using MOPITT data for 32by23 degree latitude by longitude regions in Amazonia,

746

Africa, Indonesia, and the Indian Ocean respectively. The weighting kernels are normalized

747

such that they represent the response of the retrieved total column to a perturbation in CO

748

partial column applied at each level in the retrieved profile. The ideal value would be 1 at all

749

levels.


Submittedtrend to Phil.1981 Trans. R. B - Issue Mean annual temperature − Soc. 2016 ( oC/ decade)

.... .. .. .. .. .... ..... .. . .... ............. .. . 1 ....... ........... ............ ............. 2 ................ . . ................... ................. .............. .. 3 .............. .............. ... .... .. ..... .. ... 4 .. ... .. ...... .. ........ . ....... 5

Fo

6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

.... ........ . ... .. .. ..... ..... ...... . ..... .. ...... ........... . ............ .. .... ...... ..... .... ...... .. ....... ...... . .. ..... ... .. . ... ..... ... . ... . ... .. .. ..

rR −0.3

evi −0.1

ew 0.1

Page 32 of 44

....... .............. ...... ...... ...... . .. . ... ... ..... .... ......... .. .... .... .. ... ... .... . . .. ...... ... ...... ... .. ... . .. ..... . ... ..... ...... ..... ..... ... .. ... ..... ...... ....... ..... ..... .. ..... . .. ..... ... ..... . .. . 0.3

>0.5

On

Mean annual precipitation trend 1981 − 2016 .(mm/ decade) ..

.

. .. ... .. .... .. .. . .. . .. ..... ... ... ... .. .. . .. . .. .. . ..

..... ...

. .. .. .. . .... . ... ... ... . ... . ... . .. . ... ... ... ..

ly

.... . .. .... . . .. .. .. . ... .. .. ... .. ... ...

http://mc.manuscriptcentral.com/issue-ptrsb −275 −175

−75

25

125

225

..... .. ... .. . ..

Rev

iew

−2

−1

0

1

For

On

ly

Atm. carbon growth rate − AF*(Fossil Fuel emissions) ElNino 3.4 Index (scaled)

−3

(PgC yr−1)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26


2

Page 33 of 44

1960

1970

1980

1990

Year


2000

2010


Page 34 of 44

2 −2 −1 0 1 (PgC yr−1)

Fo

rR

2005

ev

2010

2015

ie

w

On

Tropical South America Anomalies 2

ly

−2 −1 0 1 (PgC yr−1)

Gravity anom anomaly (cm) −10 −5 0 5

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38

Gravity anom anomaly (cm) −4 −2 0 1 2

Pantropical Anomalies

Grav Anom Anom AGR − AF*FF, scaled 2005

2010 Year http://mc.manuscriptcentral.com/issue-ptrsb

2015


Tropical S America 24

Min Day T (C)

23

32

22

31

21

30

20

29 28

−10

0

5 10

Max Day T (C)

33

Tropical South America (20S − 10N)

2010

2015

20

28

21

29

22

30

Min Day T (C)

23

Central Africa (5S,5N)

19

27

Max Day T (C)

1990

On

2005

2010

2015

2010

2015

2010

2015

20

Min Day T (C)

Southern Africa

1995

15

2000

2005

16

26

18

28

20

30

Min Day T (C)

22

32

Tropical East Asia

24

Max Day T (C)

2000

ly

1990

Tropical South East Asia (10S − 10N, 90W − 150W)

1995

10

5 10 0

30

iew Max Day T (C)

Southern Africa (40S − 10S)

2005

26

ev

25

−10

0

rR

2000

20

5 10

Fo

1995

31

North and Central Africa (10S − 30N)

−10

Grav anom anom (cm)

Grav anom anom (cm)

1990

Grav anom anom (cm)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46

Grav anom anom (cm)

Page 35 of 44

http://mc.manuscriptcentral.com/issue-ptrsb 2002

2004

2006

2008

2010

2012

2014

2016

2018

1990

1995

2000

2005


Page 36 of 44

5

ly

1997

1999

2001

2003

−15

Land to Atmosphere C flux Gravity Anomaly Anomaly 1995

(cm)

0

2 −1 −2

On

−5

iew

−10

Rev

0

1

For

−3

(PgC yr−1)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

10


2005

2007

2009


2011

2013

2015

2017

Tropical Land

2 0 1 −1 −3 1 −3

−1

iew ly On

1

Southern Africa

−1

(PgC yr−1)

−3

Tropical Africa

ev

(PgC yr−1)

rR

Tropical South−East Asia

(PgC yr−1)

(PgC yr−1)


Fo

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41


−3

(PgC yr−1)

Page 37 of 44


1997

1999

2001

2003

2005

2007

2009

2011

2013

2015

2017

view

nuscriptcentral.co toPage Phil. 38 Trans. of 44 R. S

0

0

50

−10

−10

0

−100

−30

−30

−50

10

20

JAS

250

300

350

100

Fo

150

200

250

300

ie

w

250

300

100

350

150

200

100

20 10 0 −10

Longitude

250

300

350

250

300

350

ly


300

On

150

200 Longitude

150

250

AMJ

JAS

100

350

OND

−30

−10

0

10

20 10 0 −10

200

300

0

350

−30

150

250

−10

ev

JFM 2016 anomalies

100

200

−30

−30

−10

0

rR

100

150

20

200

10

150

20

100

−30

Latitude Latitude

Latitude

Latitude

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44

AMJ

20

(%)

Submitted to100 Phil. Trans. R. Soc. B - Issue 10

20

JFM 2015 anomalies

10

Page 39 of 44

350


iew

ev

rR

Fo ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Page 40 of 44

Page 41 of 44

abp alt aoc asc ask azr bal bhd bkt bme bmw brw bsc cba cgo chr cib cmo cpt crz drp dsi eic gmi goz hba hpb hsu hun ice izo kco key kum kzd kzm lef llb lln lmp mbc mex mhd mid mkn mlo nat nmb nwr oxk pal pao poc psa pta rpb scs sdz sey sgp shm smo spo stm sum syo tap thd tik ush uta uum wis wlg wpc zep alf rba tef tab

iew

ev

rR

Fo ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1996

1998

2000

2002

2004

2006

2008

2010

2012

2014

number of observations per month 0

1

2

3

4

5

6

7

8


9

10

11

2016


iew

ev

rR

Fo ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Page 42 of 44

Page 43 of 44


2016

2015 JFM

cm

JFM

cm

10

0

0

40 20

0

10

20

0

−20 −40 0

50

100

150

200

250

300

−20

−20

−20

Latitude

40

350

−40 0

50

100

150

250

300

350

200

250

300

350

200

250

300

350

200

250

300

350

10

10

100

150

200

250

−20

0

300

0

350

ev

−20 0

50

100

150

200

250

300

350

iew 0

150

On 50

100

150

OND

10 0

−20

−20

0

100

ly

OND 10

50

JAS

−20

0

10

JAS

rR

10

50

Fo

0

0 −20

Latitude

0

Latitude

200

AMJ

AMJ

Latitude

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46

0

50

100

150

200

250

300

350

0

50

100

150

Longitude

Longitude



30.0

31.0

31.0

32.0

32.0

S E Asia S E Asia

30.0

2000

2005

2010

29.0

Fo

rR 2015

1990

ev

iew

19.5

20.5

15 16 17 18 19 20

29.0

1995

21.5

1990

1990

1995

2000

2005

2010

2015

1995

2000

2005

2010

2015

1995

2000

2005

2010

2015

1995

2000

2005

2010

2015

On

ly

0 −10

−5

−2

cm/mo

0

2

5

4

6

1990

−6

Day max. Temp Day(C) Max Temp (C)

Tropical S America Tropical S America

Precip. Anomaly Day min. Temp cm/mo (cm/mo) (C)Temp (C) Day Min

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41

Page 44 of 44

1990

1995

2000 Year

2005

2010

2015

1990


Year