7
8
5
6
3
4
1
2
9
10
11
12
13
14
27
28
29
30
23
24
25
26
31
32
33
34
19
20
21
22
15
16
17
18
COMPARISON OF BIOGENIC METHANE EMISSIONS FROM
UNMANAGED ESTUARIES, LAKES, OCEANS, RIVERS AND WETLANDS
M.J. Ortiz-Llorente 1, 2 & M. Alvarez-Cobelas 1
1 CSIC-Museo Nacional de Ciencias Naturales, Serrano 115 dpdo., E-28006
Madrid, Spain
2 Author contact (mjllorente@mncn.csic.es)

63
64
65
66
67
68
59
60
61
62
55
56
57
58
47
48
49
50
43
44
45
46
51
52
53
54
39
40
41
42
35
36
37
38
Abstract
A literature review of quantitative data was carried out to conduct a cross-system study on methane emissions relating peak emissions (PE) and annual emissions (AE) in five types of non-managed ecosystems: estuaries, lakes, oceans, streams and wetlands. PE spanned eight orders of magnitude (0.015 μg CH
4
m
-2
h
-1
-300 mg CH
4
m
-2
h
-1
) while
AE spanned seven (0.078-19044 g CH
4
m
-2
yr
-1
). PE and AE were strongly related worldwide ( r
2
= 0.93). There was no relationship between AE and latitude, with highly variable PE across latitudes and climates. The coefficient of variation (CV) was greatest for emissions in oceans and estuaries, while the highest emission rate was recorded in wetlands and lakes. Efflux from coastal areas and estuaries was higher than that from upwelling areas and deep seas. Concerning wetland types, marshes showed the highest
PE with the highest wetland emissions occurring in sites dominated by big helophytes.
Non-stratifying- and eutrophic lakes displayed more emissions than other lake types, but there was no environmental variable that might predict methane emissions from lakes on a worldwide basis. Generally, most ecosystem types followed a seasonal pattern of emissions, with a maximum in summer, except in estuaries which did not show any distinct pattern. Regarding the importance of hot spots within most ecosystems, more spatial variability of CH
4
emissions was observed in lakes than in wetlands and oceans; however, no relationship between emissions and spatial variability was found. A positive relationship, albeit weak, was found between methane flux and either temperature or irradiance in wetlands; a narrow range of both negative and positive values of the water table promoted CH
4
emissions. Previously, little was known about the factors controlling efflux from river and marine environments. Our study suggests that local conditions are important in controlling CH
4
emissions, because the variability explained by the more commonly studied abiotic factors is low worldwide.
This precludes the use of these variables to develop models to predict emissions at regional scales or wider, despite the many attempts made in the past. This makes local assessments of emissions essential, particularly in warm, temperate and tropical areas of the world. Future research aiming to shed light on CH
4
fluxes from estuaries, lakes, oceans, rivers and wetlands must: 1) produce more detailed data on controlling factors;
2) increase efforts to fully characterize spatial and temporal heterogeneity; 3) combine bottom-up (measurements) and top-down (modelling) approaches.
2

69
70
83
84
85
86
79
80
81
82
87
88
75
76
77
78
71
72
73
74
97
98
99
100
101
102
93
94
95
96
89
90
91
92
Keywords : seasonal emissions; global emissions; environmental factors; scale of observation; spatial heterogeneity
1. Introduction
Methane (CH
4
) is a greenhouse gas, whose concentration increased in the atmosphere from 1950 (IPCC 2007) until the 1990s, when emissions decreased; however, there is considerable inter-annual variation (Bousquet et al., 2006). The global warming potential of CH
4
is 25 times greater than that of CO
2
on a 100-year time horizon (IPCC, 2007). Indeed, despite its lower concentration (6 ppb CH
4
y
-1
from 1990 to 2005) compared to CO
2
(1.7 ppm y
-1
from 1990 to 2005), CH
4
contributed ~20% of the global warming effect (IPCC, 2007). The major sources of CH
4
are submerged soils
(natural wetlands, mangroves and rice fields), oceans and fresh water, CH
4
hydrates, fossil fuels, enteric fermentation, animal wastes, domestic sewage, landfills and biomass burning (Dalal and Allen, 2008).
Wetlands are the main type of aquatic ecosystem emitting biogenic CH
4
(Matthews and Fung, 1987). They cover between 4% and 9% of the global surface area (Matthews and Fung, 1987), but store nearly 37% of the global terrestrial carbon (Bolin and
Sukumar, 2000). The role of wetlands in terrestrial carbon cycling is particularly complex, due to the fact that methane production and oxidation are a part of carbon cycling (Roehm, 2005), contributing nearly 40% of global atmospheric methane annually (Fung et al., 1991). Wetlands are the largest single source of atmospheric methane, even when considering all anthropogenic emissions (Christensen, 2011).
Studies on emissions from water-related ecosystems other than wetlands are less frequent. Streams are very poorly studied and ocean areas have been studied in more detail during the last two decades, but many unsolved questions still remain, concerning
CH
4 origin, production, oxidation and advection, and controlling factors (Reeburgh,
2003; 2007). Studies on lake emissions have recently been undertaken (Bastviken et al.,
2011), but they are still more geographically-restricted than those of wetlands.
Methane emissions from wetlands appear to be strongly influenced by environmental variables, such as soil temperature and inundation, at a local scale, and are likely to be influenced by climate change (Bartlett and Harriss, 1993). Many studies have related methane flux changes with the capability of plants to transport gas depending on plant activity and environmental factors. Temperature, water table and irradiance all affect plant cycles, thereby affecting plant-mediated CH
4
emissions and
3

103
104
117
118
119
120
113
114
115
116
121
122
109
110
111
112
105
106
107
108
131
132
133
134
135
136
127
128
129
130
123
124
125
126 even CH
4
emissions from vegetated soils (Kankaala et al., 2005; Käki et al., 2001;
Kutzbach et al., 2004; Mikkelä et al., 1995). In addition, these abiotic factors affect emissions through control of microbial activity; therefore, they have been widely studied and used to estimate global emissions (Roehm, 2005; Walter et al., 1996). In lakes, regional studies suggest that shallow and small lakes release more methane per unit area than deep and large lakes (Bastviken et al., 2004; Juutinen et al., 2009); since there is a covariance between shallowness and eutrophication (Wetzel, 2001), it is not surprising that CH
4
efflux from eutrophic lakes is also higher (Juutinen et al., 2009).
Many global and regional reviews on CH
4
emissions from wetland environments have been carried out (Aselmann and Crutzen, 1989; Bartlett and Harriss, 1993;
Bridgham et al. 2006; Matthews and Fung, 1987; Roehm, 2005; Segers, 1998). Such studies have analyzed yearly CH
4
emissions in wetlands, taking into account environmental variables and locations. Reeburgh (2003) attempted to summarize our knowledge of the global methane budget, also taking into account ocean environments
(Reeburgh, 2007) while Saarnio et al. (2009) carried out a study estimating annual emissions from wetlands and watercourses in Europe. However, it appears that there are no studies which compare emissions among wetlands, lakes, rivers, oceans and estuaries. More specifically, reviews on either maximal or yearly CH
4
emissions have not been carried out that include all these ecosystems. Seasonality and spatial variability of emissions have scarcely been dealt with either (Alford et al., 1997; Purvaja and
Ramesh., 2001).
Due to the importance of CH
4
in terms of global warming, and since there are no recent studies describing how this gas behaves in water-related ecosystems at a global scale, this study compiles all the information available about biogenic methane emissions worldwide.
Furthermore, this study emphasizes the fact that the distribution of existing studies is not uniform. Here the global area for an ecosystem type was compared with the number of studies carried out therein and no significant correlation (p > 0.05) was found. For example, there were only 51 studies found on oceans, with a global area of
363 10
6
km
2
(Bange et al. 1994) whereas there were 317 studies on wetlands, which have a much lower surface area (Aselmann and Crutzen, 1989). This could be because biogenic methane emissions in oceans are lower than those in wetlands; however the mechanisms regulating these emissions in the ocean are largely unknown due to the scarce number of studies carried out.
4

165
166
167
168
169
170
161
162
163
164
157
158
159
160
153
154
155
156
149
150
151
152
137
138
143
144
145
146
139
140
141
142
147
148
We conducted a cross-systems review compiling quantitative estimates of CH
4 emissions and their responses to controlling factors across estuaries, lakes, oceans, rivers and wetlands. This cross-system approach has enabled us to summarize research findings across different studies and quantitatively synthesize the ecological information about the process, as well as to explore general patterns among these ecosystems. More specifically, our aims are: (1) to report peak and yearly CH
4
emissions at different latitudes, climates and ecosystem types; (2) to outline how CH
4
emissions vary among different wetland and marine ecosystems; (3) to describe seasonality of emissions in these environments; (4) to consider the influence of spatial variability on emissions; and
(5) to assess the influence of abiotic factors in these environments and how different types of vegetation affect emissions.
2. Methods
2.1 Data collection
This study was carried out by reviewing literature quantifying biogenic CH
4 emissions from estuaries, lakes, oceans, rivers and wetlands. It was restricted to measurements made in unmanaged ecosystems, without any external modification, thus disregarding man-made fertilization (e.g. rice fields). Potential emission rates and CH
4 oxidation rates were excluded from this analysis. Two main types of units appeared in the reviewed articles: weight area -1 time -1 and weight volume -1 time -1 . Only studies using the first unit were chosen, which enabled comparisons of most emission data. The dataset was collected from texts, tables and figures of articles published from 1985 to
December 2008, gathering information on the measurement site, dominant vegetation, maximal and annual CH
4
emission, spatial heterogeneity and seasonality of emissions as well as the factors controlling emissions (Appendices 1-4). Despite the fact that CH
4 emission can result from different processes (diffusion, ebullition, diffusion through aquatic plants), our compilation did not discriminate among them because most published studies did not attempt to make such a distinction; therefore, our reported data must be considered as “total emissions” from a given site, irrespective of the underlying process. Ebullition is negligible in most marine environments (McGinnis et al., 2006) and storage can also be overlooked in shallow lakes, wetlands and rivers due to the lack of thermal stratification.
It was deemed necessary to divide the study in two groups to tackle different time scales that might be important in CH
4
emission studies. The first group comprised those
5

171
172
185
186
187
188
181
182
183
184
189
190
177
178
179
180
173
174
175
176
199
200
201
202
203
204
195
196
197
198
191
192
193
194 conducted over an annual cycle (AE) with at least monthly periodicity of sampling, and was used to study seasonal emissions in different ecosystems. However, most studies on
CH
4
emission were restricted to shorter periods. Therefore, we also compiled data collected over shorter time periods, focusing on emissions occurring in the most favourable month of the year (peak emissions, PE), namely, when environmental factors controlling emission were less limiting. When only monthly data over an entire year were supplied, AE was estimated by the numerical integration of these data (Appendix
3). The relationship between peak and annual emissions in the same sites was also studied. Moreover, the distribution of annual emission data has been studied and divided into two series, each analysed separately. Most comparisons and analyses among ecosystem types would be carried out for PE due to the fact that there were more degrees of freedom in the PE dataset.
Whenever available, data of environmental factors were collected. Thus the relationship was studied between water table, air temperature, soil temperature, PAR in wetlands and lakes. However, in the case of oceans, rivers and estuaries, data were insufficient to search for relationships between CH
4
emissions and environmental factors.
2.2 Data analysis
We studied PE and AE versus latitudes and climate types, the latter following
Köppen´s Climate Classification by FAO-SDRN-Agrometeorology Group (1997). Five ecosystems were studied: estuaries, lakes, oceans, rivers and wetlands. Hopkinson and
Smith (2005) define estuaries as “…those regions at the interface of the terrestrial and oceanic realm where seawater is measurably diluted by freshwater runoff from land.”
The Ramsar Convection (2000) defines wetlands as “the site where water is at or near the soil surface for a significant part of the growing season.” The Merriam-Webster
Dictionary (2011) defines river as “any natural stream of water that flows in a channel with defined banks.” And lakes are defined, as “an inland body of standing water” while the definition of ocean is: “the whole body of salt water that covers nearly three fourths of the surface of the earth”.
There are some reports of emissions by impoundments, which refer to lakes, while emissions from the littoral fringe of lakes are included within wetlands. Ocean environments have also been studied specifically, splitting them into estuaries, coastal shelves, deep seas and upwelling areas. Likewise, wetlands have been further divided
6

205
206
219
220
221
222
215
216
217
218
223
224
211
212
213
214
207
208
209
210
229
230
231
232
225
226
227
228
233
234
235
236
237 into other categories: peatlands, marshes, forested wetlands and coastal wetlands
(estuaries have been considered as different from coastal wetlands when freshwater inputs were suspected to occur), following the outline by Mitsch and Gosselink (2001), who define peatland as “a generic term of any wetland that accumulates partially decayed plant matter”, and marsh as “a frequently or continually inundated wetland characterized by emergent herbaceous vegetation adapted to saturated soil conditions.”
Forested or riparian wetlands are defined as “occurring along rivers and streams, which are occasionally flooded by those bodies of water but are otherwise dry for varying portions of the growing season.” Moreover, we classified wetlands as either ice-free or winter-frozen when the authors mentioned this in their articles. We analyzed the relationship between PE and predominant vegetation, classifying them within the main physiognomic groups (hydrophytes, bryophytes, small helophytes, big helophytes and trees).
Thermal stratification and trophic status are the main drivers of lake ecological functioning (Wetzel, 2001). We compare PE in stratifying vs non-stratifying lakes and oligotrophic vs eutrophic lakes. Data on thermal regimes were obtained from the studies reporting lake emissions, whereas their trophic status was ascertained using yearlyaveraged total phosphorus in the water column of each lake as an index, those lakes having more than 30 μg l -1 were considered as eutrophic (OECD, 1982).
Due to the lack of both Gaussian distribution and homogeneity of variances for most variables, non-parametric tests were undertaken for comparisons (Siegel and
Castellan, 1988). While a Mann-Whitney test was employed to compare medians of
CH
4
emission between two categories, a Kruskal-Wallis test was used to assess differences in CH
4
emission among the categorized aquatic ecosystems when these exceeded two. Latitudinal, climatic, ecosystem types, wetland types and their main vegetation, marine ecosystems and lake trophic status (as total phosphorus in the mixed layer) and stratification types were considered to make comparisons.
The Kavvas and Delleur (1975) method to visualize periodicities in a time series was used with the annual dataset to study seasonality in different aquatic ecosystems and wetlands types. Seasonal dynamics in oceans could not be ascertained because of the lack of data. Seasonal data reported from the Southern Hemisphere were rearranged to enable comparisons with data taken in the Northern Hemisphere. An average over a constant time unit (year) was calculated and formed a new series as the new difference
7

238
239
252
253
254
255
248
249
250
251
256
257
244
245
246
247
240
241
242
243
262
263
264
265
258
259
260
261
266
267
268
269
270 between the original series and the mean. For any CH
4
emission time series, a scaled difference, ME t
, was calculated to represent the observed rate, X t
, on time t
(1) ME t
=100*((X t
-X)/X) where X was the yearly mean calculated for that time series. This procedure was repeated for each time series in all aquatic ecosystems of the annual database (Appendix
3).
The range (as difference between the minimum and maximum flux measured) statistic was used to study the spatial variability of emissions, an approach suggested by the lack of both Gaussian distributions and homogeneities of variances in these data.
We chose studies reporting data of CH
4
emission, measured at the same time but in different areas within the same study area (Appendix 4). We only used ranges of CH
4 emissions in wetlands, lakes, estuaries and the ocean because of the lack of stream data.
Moreover, since we suspected that the larger study areas corresponded to the higher variability in methane emission, we used the distance between the most distant sites in the studied areas as a surrogate of the spatial scale.
To study the relationship between peak CH
4
emissions and abiotic factors (air and water temperature, water table and irradiance) we employed the Spearman correlation.
For lakes, average and maximal depth and total phosphorus were also related with AE.
All statistical analyses were performed using STATISTICA 6.1 (Statsoft Inc., 1997).
3. Results
3.1 Emissions and their ranges
Most continental studies were carried out in Canada, USA, Finland and China.
Marine environments were more evenly covered throughout the world but only during
PE (Appendix 1). Methane emissions were more commonly measured in the Northern
Hemisphere, probably due to the abundance of environments related to water and the funding available to create the infrastructure needed to measure methane emissions. A total of 551 estimations were compiled for PE, of which 18 were estuaries, 143 were lakes and impoundments, 51 were oceans, 22 were rivers and 317 were wetlands. As concerning different wetlands, 179 were peatlands, 101 were marshes, 42 were forested wetlands and 8 coastal wetlands (Appendix 1, Table 1). Concerning AE, records were
8

271
272
285
286
287
288
281
282
283
284
289
290
277
278
279
280
273
274
275
276
299
300
301
302
303
304
295
296
297
298
291
292
293
294 taken from 6 estuaries, 55 lakes, 12 rivers and 126 wetlands, but no annual data were available for any oceanic sites (Appendix 3, Table 1).
There were no significant differences in CH
4
emissions between latitude or climate zone classes (Fig. 1). The highest emissions appeared in lakes between 0-20º and 60-
80º, whereas highest emissions for wetland varied between 40 and 60º.
Annual Emissions differed among ecosystem types (Table 1). Efflux from lakes and wetlands was far more variable (CVs: 537 and 489%, respectively) than that from estuaries and rivers. All environments presented contrasting maximal values. The highest values occurred in wetlands, followed by those in lakes, rivers and estuaries.
Mean values showed differences in different ecosystem types with maxima in estuaries and wetlands and minima in lakes and rivers (Table 1).
The scenario was different concerning PE (Table 1). It varied greatly in oceans and estuaries (CV: 630 and 259%, respectively), as compared with those from rivers, wetlands and lakes (CV: 187-198%). Ranges were wider in lakes and wetlands.
Average values were roughly similar in inland waters, with those of the sea being much lower (Table 1).
Oceanic ecosystems differed in their biogenic PE, as judged by a Kruskal-Wallis test ( P < 0.05; Fig. 2a). Coastal shelves (range: 0.0001-64.9 mg CH
4
m -2 h -1 ) appeared to be the most variable, whereas estuaries (mean ± STE 3.29 ± 11.04 mg CH
4
m -2 h -1 ) reached the highest PE averages.
Regarding wetland types (Fig. 2b), marshes presented the highest range of methane release (range: 0.007-280 mg CH
4
m
-2
h
-1
) and the highest mean emission (mean ± SD
20.026 ± 41.98 mg CH
4
m
-2
h
-1
). Wetlands, where big helophytes predominate (Fig 2c), showed the highest range of emissions (range: 0.003-288 mg CH
4
m
-2
h
-1
) and also the highest mean emission (mean ± SD 27.75 ± 45.67 mg CH
4
m
-2
h
-1
).
CH
4
emission in lakes depended upon the lake type. Eutrophic lakes (Fig. 2d) showed more PE variability than oligotrophic ones (range: 0.009-300 mg CH
4
m
-2
h
-1
)
(Mann-Whitney test, P < 0.05). Mean values were higher in eutrophic lakes (mean ± SD
25.53 ± 50.64 mg CH
4
m
-2
h
-1
) than in oligotrophic ones (mean ± SD 6.19 ± 14.13 mg
CH
4
m
-2
h
-1
), as judged by a Mann-Whitney test ( P < 0.05). Variability in CH
4
fluxes also depended on lake stratification (Figure 2d); the range was smallest in stratifyinglakes (range: 0.34-4.93 mg CH
4
m
-2
h
-1
). As a result, mean values were much higher in non stratifying lakes (mean ± SD 15.83 ± 37.87 mg CH
4
m
-2
h
-1
) than in stratifying ones
(mean ± SD 1.24 ± 1.64 mg CH
4
m
-2
h
-1
).
9

305
306
319
320
321
322
315
316
317
318
323
324
311
312
313
314
307
308
309
310
333
334
335
336
337
338
329
330
331
332
325
326
327
328
3.2 Seasonality and spatial heterogeneity of emissions
Lakes and wetlands usually showed a peak in summer (Fig. 3a), with a maximum in
August and February in the Northern and Southern hemispheres, respectively.
Meanwhile, rivers presented three maxima of CH
4
emission, the first at the beginning of spring and the highest at the end of summer, with a short peak at the beginning of the autumn season. Estuaries did not show any distinct pattern of seasonality (Fig. 3a).
Regarding wetland types (Fig. 3b), seasonal patterns of CH
4
fluxes in frozen and icefree wetlands were similar as both presented a peak in August, higher in winter-frozen wetlands.
The range of spatial variability of CH
4
emission in lakes (0.546-233.6 mg CH
4
m
-2 h
-1
) was higher than those of wetlands (0.02-216.63 mg CH
4
m
-2
h
-1
) and oceans
(0.00003-0.0109 mg CH
4
m
-2
h
-1
), as judged by a Kruskal-Wallis test ( P < 0.05; Fig. 4).
However, no statistically significant relationship was found when we studied how varied methane emissions depended on the distances between the measured sites.
3.3 Factors controlling emissions
A positive relationship ( r 2 = 0.93, P < 0.05) was found between PE and AE (Fig. 5).
Therefore, the higher the PE, the higher the annual emission is. We studied the distribution of annual emission data and divided them in two intervals, above and below
50 g m
-2
yr
-1
, where more of the points are. A positive relationship between PE and AE was found in both intervals (AE < 50 g m
-2
yr
-1 r
2
= 0.92 and AE > 50 g m
-2
yr
-1 r
2
= 0.89; p<0.05). Statistical significant relationships (p< 0.05) appearing between two functions.
Some statistically significant, albeit weak, relationships ( P < 0.05) between PE from wetlands and lakes with environmental factors were observed. Positive relationships existed between CH
4 emissions and irradiance, air and sediment temperature, being the methane efflux-irradiance correlation stronger ( r
2
= 0.44, p < 0.05) than the others tested ( r
2
= 0.11 and 0.20 for air and sediment temperature, respectively; P < 0.05). The highest methane release occurred when water level was between -10 and 15 cm.
However, no statistically significant relationship between methane fluxes and water table was found ( P > 0.05). In lakes, we found no statistically significant relationship between either average or maximal depth and methane emissions ( P > 0.05), both PE and AE. Total phosphorus, used as an index of organic matter production in lakes, was neither related to CH
4
efflux, be it as PE or AE ( P > 0.05).
10

367
368
369
370
371
372
363
364
365
366
359
360
361
362
351
352
353
354
347
348
349
350
355
356
357
358
343
344
345
346
339
340
341
342
4. Discussion
4.1 Emissions and their ranges
We did not find any geographical pattern when we studied the relationship between climate types or latitude and methane emission (Fig. 1), higher emissions occurred in warm temperate climates rather than in tropical or equatorial ones, but also in cold continental sites. The most outstanding feature that emerged from our analyses was the strong variability of methane emissions in these non-managed environments across latitudes and climates.
The highest CH
4
emission in different ecosystem types during PE occurred in lakes, whereas the highest AE were recorded in wetlands (Table 1). Areal rates of PE varied over eight orders of magnitude and AE spanned over seven orders of magnitude (Table
1). However, it is likely that the latter range would widen if some oceanic oligotrophic sites were studied for a year. Ranges and maximum rates reported here were much higher than those described in some recent, albeit regional, reviews (Reeburgh, 2003;
Roehm, 2005; Bridgham et al., 2006). This discrepancy arose from the fact that we undertook a world-wide analysis of the data, thereby increasing environmental variability which usually resulted in a wider range of emissions.
Estuarine and coastal shelves were shown to depict more emissions than upwelling areas and deep seas (Fig. 2a). A part of the estuarine and coastal emission may be augmented by increases in their methane content delivered by rivers (Scranton and
McShane, 1991; Rehder et al., 1998; Semiletov, 1999). Both estuarine and coastal areas are richer in organic matter of continental origin than other marine ecosystems, thereby being better methane-producing places, provided they have more anoxic sites to promote CH
4
production (Ward et al., 1987; Naqvi et al., 2010; Zhang et al. 2008).
Marshes were the wetland type presenting the highest CH
4
flux, followed by peatlands (Fig. 2b). Their warm and waterlogged soil provides ideal conditions for methanogenesis (Roehm, 2005). These data suggest that marshes are a substantial source of atmospheric CH
4
. The deep standing water in marshes could not only provide enough C for CH
4
production by inundating more surface area covered with plant litter, but could also create an anaerobic environment favouring CH
4
production (Ding and
Cai, 2007). Gorham (1991) estimated that peatlands contained approximately one-third of the global terrestrial C pool; thus, peatland is an important source of CH
4
to the
11

373
374
387
388
389
390
383
384
385
386
391
392
379
380
381
382
375
376
377
378
401
402
403
404
405
406
397
398
399
400
393
394
395
396 global C cycle, contributing an estimated 25-30% of global CH
4
released to the atmosphere every year (Aselmann and Crutzen, 1989; Alvarez-Cobelas and Ortiz-
Llorente, unpublished data). However, wetland emissions were not explicitly considered in global climate models and hence, when included in predictions of future climate warming, such models will likely change (IPCC, 2007).
Some studies relate the changing fluxes of CH
4
with the capability of plants to transport the gas depending on plant activity (Käki et al., 2001), photosynthesis (Sutton-
Grier and Megonigal, 2011), decaying plant organic matter accumulating in the sediment (Kankaala et al. 2004) and plant biomass (Käki et al., 2001). Deep rooting helophytes with aerenchymatic cell structure are generally associated with high CH
4 flux rates (Bubier, 1995), supporting our results which show that the highest CH
4 emission occurred in soils whose dominant vegetation either comprised big helophytes or hydrophytes (Fig. 2c); these are the plant species with the highest root-system development in wetlands (Cronk and Fennessy, 2001).
Eutrophic lakes presented higher emissions than oligotrophic ones (Fig. 2d) due to the fact that more organic matter was available for methane production and increased anoxia. However, these relationships are of a qualitative nature since no statistically significant relationship was found between either lake depth or total phosphorus (as a surrogate of eutrophication) and PE or AE. Such relationships have been demonstrated at a regional scale (Juutinen et al., 2009); therefore, it seems such relationships with lake variables were of a very broad nature, losing significance as the geographical scope of the study increased.
In addition, our study showed that shallow lakes appeared to produce more emissions than deep lakes, in contrast with Kiene (1991). The latter stated that sediment methanogenesis was limited in shallow aquatic systems that were well mixed and oxygenated during summer periods. This could be explained taking into account that while CH
4
produced in hypolimnetic sediments could mostly be oxidized to CO
2
by methanotrophic microbes in the water table (Striegl and Michmerhuizen 1998), much of the sediment in non-stratifying lakes was in contact with the upper mixed water layer, in contrast to stratifying lakes. This situation enhanced mineralization over burial, and CH
4 produced in shallow sediments largely escaped oxidation by methanotrophic bacteria and entered the atmosphere (Bastviken et al., 2008).
12

407
408
421
422
423
424
417
418
419
420
425
426
413
414
415
416
409
410
411
412
435
436
437
438
439
440
431
432
433
434
427
428
429
430
4.2 Seasonality and spatial heterogeneity of emissions
Seasonal CH
4
emissions in lakes and wetlands (Fig. 3a) showed a step-function response to soil temperature, emissions being zero or near zero during winter and much higher in summer with rapid shifts between both seasons. While lakes presented a sharp increase in summer and lower values throughout the year, CH
4
release by wetlands remained steadier throughout the year (Fig. 3a). The same results were reported by Dise et al. (1993), Kang and Freeman (2002) and Xing et al. (2005). The three riverine peaks shown here were also reported by Alford et al. (1997). The spring peak was the result of labile organic material accumulating over winter months when low temperatures reduced microbial activity. The summer peak appeared to match the physiological activity of plants when the release of organic material by root systems was increased.
Autumn maxima were attributed to the dramatic influx of organic carbon-rich leaf litter to the system, which was readily converted into organic substrates by CH
4
producing bacteria. These results are in agreement with those by Bergström et al. (2007), who demonstrated that lake littoral plant cover and the distribution of littoral plant species must be taken into account when estimating regional CH
4
budgets.
Emission differences between winter ice-free and frozen-wetlands (Fig. 3b) could be explained by the fact that ice cover did not enable bacterial activity to begin, and that methane bubbles were trapped within the ice, so the increased emission in non-frozen wetlands started earlier. When high temperatures promoted ice-cover thaw, bacteria could resume their activity and methane was released to the atmosphere.
The preliminary spatial study demonstrates that there is no relationship between
CH
4
emission variability and the spatial scale. It seems the larger the ecosystem, the smoother the variability of methane emissions, with areas of high and low emissions cancelling each other out. Thus, successful scale-up is difficult. Hot spots of methane emission, however, appear to be the rule in spatially heterogeneous environments
(Bastviken et al., 2010; Bianchi et al., 1996; Kock et al., 2008; Linke et al., 2010;
Moore et al., 1994).
A positive relationship between PE and AE was found (Fig. 5) in lakes and wetlands
(Alvarez-Cobelas and Ortiz-Llorente, unpublished data). This could enable optimization of the duration of experiments and related costs, thus diminishing work effort and saving money.
13

441
442
455
456
457
458
451
452
453
454
459
460
447
448
449
450
443
444
445
446
469
470
471
472
473
474
465
466
467
468
461
462
463
464
4.3 Factors controlling emissions
Relationships tested between environmental variables and CH
4
emissions, however, have poor explanative power. We studied these relationships in wetlands and lakes, first individually and then together. Both attempts gave the same results. A statistically significant, albeit weak, positive relationship emerged between irradiance and CH
4 emission, which has also been reported in other studies (Mikkelä et al., 1995; Käki et al., 2001). Concerning other abiotic factors, we found a complex relationship between the depth of water table and CH
4
emission because its maximum values were recorded within a short range of negative and positive values.
The water table position determines the relative extent of oxic and anoxic horizons within soils and, consequently, the ratio between CH
4
production and CH
4
oxidation, the fundamental microbial processes underlying the CH
4
cycle (Christensen et al.,
2001). The effect of the water table position on CH
4
emissions can be compared with an on-off switch. When the depth of water table falls well below the soil surface, microbial
CH
4
oxidation is increased and CH
4
emission is hence reduced, as occurs when water table level is increased above the sediment. Methane emissions in water-related ecosystems are partly controlled by water table, a parameter closely related with climate. Rainfall decrease, temperature increase and an elevated extraction of water from groundwater led to decreased water table and increased methane emissions from lakes and wetlands. Ringeval et al. (2010) stated that over the period 1993-2000, the variability in wetland areas strongly influenced the seasonal and interannual variability of CH
4
emissions. On a seasonal timescale, the variable flooded area mainly plays a role in controlling variable fluxes at boreal latitudes (~30 %). The difference with other studies (Shannon and White 1994) is the strength of these relationships. While these authors have found very strong correlations between water table and methane emissions, world-wide data have shown a much lower statistical relationship.
The factors controlling CH
4
emission in lakes and wetlands are difficult to determine on a worldwide basis. Locally, they have been related to water depth, dissolved organic carbon, total phosphorus and anoxic volume fraction (Bastviken et al.,
2004; Juutinen et al., 2009). However, no such relationships hold when studying them on a worldwide basis. Not only do CH
4 emissions from lake environments appear to be more spatially heterogeneous than initially suspected (Bastviken et al., 2010) but also, temporal variations in efflux occur ranging from diel (Bastviken et al., 2010) to interannual changes (Bartlett and Harris, 1993; Pulliam, 1993; Shannon and White,
14

475
476
489
490
491
492
485
486
487
488
493
494
481
482
483
484
477
478
479
480
503
504
505
506
507
508
499
500
501
502
495
496
497
498
1994; Hughes et al., 1999). Thus, controlling factors are likely to change depending upon the spatial scale involved. This makes us suspect that the relationship between methane emissions and either water table or sediment temperature are only valid at small scales, for individual study areas. By contrast, when we study an ecosystem type
(lake or wetland) at the global scale these relationships disappear, and a more complex relationship may emerge among them.
The latitudinal variability is also striking (Fig. 1), thus preventing meaningful prediction of the factors controlling emissions worldwide. It seems that temperature, organic matter and anoxic conditions may be some of them, but there are not enough data to support this hypothesis on the worldwide scale.
It is more difficult to study the control of biogenic CH
4
efflux in the sea. Most reported studies do not address it causally, maybe because the multivariate nature of controls and the expense of studying them all preclude a mechanistic approach. The picture is clearer in estuaries and closed basins, where CH
4
and organic matter inputs transported by incoming streams and anoxic conditions prevail at the sediment-water interface (Scranton and McShane, 1991; Zhang et al., 2008; Naqvi et al., 2010). Deep seas and upwelling areas are quite different because there are long, advective effects
(Brüchert et al., 2009) and short-range effects, such as oxidation of methane (Reeburgh,
2007), vertical transport associated to settling fecal pellets (Karl and Tillbrook, 1994), mixing of different water masses (Rehder and Suess, 2001) and oxygenic methane production due to methylphosphorus derivatives arising from decomposition of algal blooms in phosphate-stressed waters (Damm et al., 2008; Karl et al., 2008). It is therefore not surprising that methane fluxes from the ocean are so poorly understood and that the ocean methane paradox (Reeburgh, 2007) needs further investigation.
Strong seasonality is also evident for marine environments (Zhang et al., 2008; Yang et al., 2010), but to obtain a good resolution of seasonal changes of methane efflux from the sea is very expensive. Observations of biogenic methane efflux from ocean areas with frequent sampling for a year or more, can hardly be accomplished by oceanographic vessels and would be better undertaken using remote sensing
(Frankenberg et al., 2005) and surface drifters (Boutin et al., 2008).
4.4 Conclusions and recommendations
The global distribution of studies in methane emissions is not homogeneous.
Countries like the USA, Brazil, Canada, Finland and China account for almost 70% of
15

509
510
523
524
525
526
519
520
521
522
527
528
515
516
517
518
511
512
513
514
537
538
539
540
541
542
533
534
535
536
529
530
531
532 studies worldwide. India, Russia, Sweden and Taiwan make up for a further 10%. A database including the real area of a specific ecosystem type around the world does not exist, except some examples as reported in Saarnio et al. (2009) for Europe.
Furthermore, the estimation of areas using satellite techniques has numerous errors. It would be extremely difficult to estimate the percentage corresponding to the global coverage area of these ecosystems. Our study highlights how the existing studies concentrate on specific zones.
Moreover, data on the factors controlling methane emissions are commonly restricted to wetlands and lakes, making it difficult to study how methane emissions vary worldwide in each ecosystem type. This study suggests that local conditions are very important in controlling CH
4
emissions, because the variability explained by the more commonly studied abiotic factors is rather low, as the review of wetland CH
4 emission by Segers (1998) has already pointed out (see his Table 1b). Regarding lakes, better knowledge of factors controlling emissions has been acquired when studying them on a local or, at most, regional basis (Juutinen et al., 2009) while knowledge of these factors in seas and rivers is more elusive. This precludes the use of these variables to either model or predict emissions at regional and on even wider scales, despite many attempts in the past, making local assessments of emissions very valuable still (e.g.
Roehm, 2005; Bridgham et al., 2006; Bastviken et al., 2011). To complicate the situation further, the occurrence of emission hot spots and strong temporal variability of methane fluxes also contribute to poor estimates at the ecosystem scale (Bubier et al.,
2005). Meanwhile, other neglected variables should be considered; for example, despite claims by Segers (1998), methanotrophic biomass is seldom quantified in aquatic environments but it might explain a high fraction of CH
4
flux variability. However, it must be stated that likely controlling variables are not frequently reported, and this is a drawback when carrying out global reviews of the topic, like the one presented here.
Not only are factors controlling methane efflux still unresolved on a global scale, but for many areas of the world our knowledge of methane emissions from waterrelated ecosystems is still very inaccurate, particularly in oceans and inland waters of semiarid, arid and tropical regions.
Reeburgh (2007) has recently reviewed novel techniques for studying methane metabolism in the ocean, which can also be used in freshwaters. Notwithstanding, there is still a lot of intercomparison work on methane flux methodologies to be done.
Particularly, the technique most widely used for studying CH
4
efflux in the sea is that of
16

559
560
561
562
555
556
557
558
563
564
543
544
549
550
551
552
545
546
547
548
553
554 the water-air gradient, which relies upon accurate determination of the gas piston velocity. However, this mass balance method poses theoretical problems (Wanninkhof et al. 2009) and also wide discrepancies in results depending upon the exchange equation used to calculate that velocity (e.g. Zhang et al., 2008).
Furthermore, studies must be carried out highlighting short-range spatial heterogeneity including areas of the Earth that are virtually unknown as yet. Hence, bottom-up (incubation chambers, surface drifters, remote sensors, eddy covariance towers) and top-down (inverse modelling) measurements must be combined to enhance the global estimation of methane emissions to the atmosphere. The significance of methane as a greenhouse gas (IPCC, 2007) makes these improvements essential if a better and more accurate picture of global emission is to be found, which is a precondition to reducing methane emissions worldwide.
Acknowledgements.
The research has been supported by the Spanish Ministry of
Environment, projects 81/2005 and 1/2008 of the Spanish Network of National Parks.
Clara Blanco, who was the librarian of the late Centre for Environmental Sciences
(CSIC), was very helpful in compiling most references used in this review. Juan Carlos
Rodríguez-Murillo is also acknowledged for his insightful reviews of an earlier draft of this study and Interglobe Language Links for his review of English language. Referees suggested improvements to an earlier draft.
References
565
566
Alford, D.P., Delaune, R.D., and Lindau, C.W., 1997. Methane from Mississippi river deltaic plain wetlands, Biogeochemistry, 37, 227-236.
567
568
569
Aselmann, I. and Crutzen, P.J., 1989. Global distribution of natural freshwater wetland and rice paddies, their net primary productivity, seasonality and possible methane emissions, J. Atmos. Chem., 8, 304-358.
570
571
572
Bange, H.W., Bartell, U.H., Rapsomanikis, S., and Andreae, M.O., 1994. Methane in the Baltic and North Seas and a reassessment of the marine emissions of methane, Global Biogeochem. Cy., 8, 465-480.
573
574
Bartlett, K.B. and Harris, R.C., 1993. Review and assessment of methane emissions from wetlands, Chemosphere, 26, 261-320.
17

575
576
577
Bastviken, D.,Cole, J., Pace, M., and Tranvik, L., 2004. Methane emissions from lakes:
Dependence of lake characteristics, two regional assessments, and a global estimate, Global Biogeochem. Cy., 18, GB 4009, doi: 10.1029/2004GB002238.
578
579
580
Bastviken, D., Cole, J.J., Pace, M.L., and van de Bogert, M.C., 2008. Fates of methane from different lake habitats: connecting whole-lake budgets and CH4 emissions,
J. Geophys. Res., 113, G02024, doi: 10.1029/2007JC000608.
581
582
583
584
Bastviken, D., Santoro, A.L., Marotto, H., Pinho, L.Q., Caherios, D.F., Crill, P., and
Enrich-Prast, A., 2010. Methane emissions from Pantanal, South America, during the low water season: Toward more comprehensive sampling, Environ.
Sci. Technol., 44, 5450-5455.
585
586
587
Bastviken, D., Tranvik, L.J., Downing, J.A., Crill, P.M., and Enrich-Prast, A., 2011.
Freshwater methane emissions offset the continental carbon sink, Science, 331,
50.
588
589
590
Bergström, I., Mäkela, S., Kankaala, P., and Kortelinen, P., 2007. Methane efflux from littoral vegetation stands of southern boreal lakes: An upscale regional estimate,
Atmos. Environm., 41, 339-351.
591
592
593
594
Bianchi, T.S., Freer, M.E., and Wetzel, R.G., 1996. Temporal and spatial variability, and the role of dissolved organic carbon (DOC) in methane fluxes from the
Sabine River Floodplain (Southeast Texas, U.S.A.), Arch. Hydrobiol., 136, 261-
287.
595
596
597
598
Bolin, B. and Sukumar, R., 2000. Global perspective. In Land Use, Land-Use Change and Forestry. R.T. Watson, I.R. Noble, B. Bolin, N.H. Ravindranath, D.J.
Verardo, and D.J. Dokken (eds) A Special report of the IPCC, Cambridge
University Press, Cambridge, UK, pp. 23-51.
599
600
601
602
603
Bousquet, P., Ciais, P., Miller, J.B., Duglokencky, E.J, Hauglustaine, D.A., Prigent, C.,
Van der Werf, G.R., Pylin, P., Brunke, E.-G., Carouge, C., Lafenfelds, R.L.,
Lathière, J., Papa, F., Ramonet, M., Schmidt, M., Steele, L.P., Tyler, S.C., and
White, J., 2006. Contribution of antrophogenic and natural sources to atmospheric methane variability, Nature, 443, 439-443.
18

604
605
606
Boutin, J., Merlivat, L., Hénocq, C., Martin, N., and Sallée, J.B., 2008. Air-sea CO
2
flux variability in frontal regions of the Southern ocean from carbon interface ocean atmosphere drifters, Limnol. Oceanogr. 53(5, part 2), 2062-2079.
607
608
609
610
Bridgham, S.D., Megonigal, J.P., Keller, J.K., Bliss, N.B., and Trettin, C., 2006. The carbon balance of North American wetlands, Wetlands, 26, 889-916.
Brüchert, V., Currie, B., and Peard, K.R., 2009. Hydrogen sulphide and methane emissions on the central Namibian shelf, Prog. Oceanogr., 83, 169-179.
611
612
Bubier, J. L., 1995. The relationship of vegetation to methane emission and hydrochemical gradients in northern peatlands, J. Ecol., 83, 403-420.
613
614
615
Bubier, J.L., Moore, T., Savage, K., and Crill, P., 2005. A comparison of methane flux in a boreal landscape between a dry and a wet year, Global Biogeochem. Cy.,
19, GB1023, doi: 10.1029/2004GB002351.
616
617
618
Christensen, T. R., Lloyd, D., Svenson, B., Martikainen, P., Harding, R., Oskarsson, H.,
Friborgh, T., Sogaard, H., and Panikov, N., 2001. Biogenic controls on trace gas fluxes in northern wetlands, IGBP- Global Change Newsletter, 51, 9-15.
619
620
Christensen, T. R, 2011. Wetlands. In Methane and climate change, Reay, D., Smith, P., and van Amstel, A. (eds.), Earthscan Ltd. London, UK.P. 27-42.
621
622
Cronk, J. K. and Fennessy, M.S., 2001. Wetland Plants: Biology and Ecology, Lewis,
Boca Raton, Fla.
623
624
Dalal, R.C. and Allen, D.E., 2008. Greenhouse gas fluxes from natural ecosystems,
Aust. J. Bot., 56, 369-407.
625
626
627
Damm, E., Kiene, R.P., Schwarz, J., Falck, E., and Dieckmann, G., 2008. Methane cycling in Arctic shelf water and its relationship with phytoplankton biomass and DMSP, Mar. Chem. 109, 45-59.
628
629
Ding, W. and Cai, Z., 2007. Methane emissions from natural Wetlands in China:
Summary of years 1995-2004 studies, Pedosphere 17, 475-486.
630
631
632
Dise, N.B., Gorham, E., and Verry, E.S., 1993. Environmental factors controlling methane emissions from peatlands in Northern Minnesota, J. Geophys. Res., 98,
10583-10594.
19

633
634
635
FAO. 1997. Köeppen Climate Zones, FAO – Environment and Natural Resources
Service (FAO – SDRN Agrometeoroly Group), Global Climate maps series. http://www.fao.org/sd/eidirect/CLIMATE/Elsp0001.html.
636
637
638
Frankenberg, C., Meirink, J.F., van Weele, M., Platt, U., and Wagner, T., 2005.
Assessing methane emission from global space-borne observation, Science, 308,
1010–1014.
639
640
641
Fung, I., John, J., Lerner, J., Matthews, E., Prather, M., Steele, L.P., and Fraser, P.J.,
1991. Three-Dimensional Model synthesis of the global methane cycle, J.
Geophys. Res., 96, 13033-13065.
642
643
Gorham, E., 1991. Northern peatland: role in the carbon cycle and probable responses to climatic warming, Ecol. Appl. 1:182-195.
644
645
646
647
Hopkinson, C.S.Jr and Smith, E.M. 2005. Estuarine respiration: an overview of bentic, pelagic, and whole system respiration, In Respiration in aquatic ecosystems,
P.A. del Giorgio and P.J. le B. Williams (eds),, Oxford University Press, Oxford, pp. 83-102.
648
649
650
Hughes, S., Dowrick, D.J., Freeman, C., Hudson, S.A., and Reynolds, B., 1999.
Methane emissions from a Gully mire in mid-Wales, UK, under consecutive summer water table drawdown, Environ. Sci. Technol., 33, 362-365.
651
652
653
654
655
IPCC: Climate change, 2007: The physical science basis, Contribution of working group I to the fourth assessment report of the intergovernmental panel on climate change, edited by Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M.,
Averyt, K.B., Tignor, M., and Miller, H.L., Cambridge University Press,
Cambridge, UK and New York, USA, 996 pp.
656
657
658
Juutinen, S., Rantakari, M., Kortelainen, P., Huttunen, J.T., Larmola, T., Alm, J.,
Silvola, J., and Martikainen, P.J., 2009. Methane dynamics in different boreal lake types, Biogeosciences, 6, 209-223.
659
660
661
Käki, T., Ojala, A., and Kankaala, P., 2001. Diel variation in methane emissions from stands of Phragmites australis (Cav.) Trin. ex Steud. and Typha latifolia L. in a boreal lake, Aquat. Bot., 71, 259-271.
20

662
663
664
Kang, H. and Freeman, C., 2002. The influence of hydrochemistry on methane emissions from two contrasting northern wetlands, Water Air Soil Pollut., 141,
263-272.
665
666
667
Kankaala, P., Ojala, A., and Käki, T., 2004. Temporal and spatial variation in methane emissions from a flooded transgression shore of a boreal lake, Biogeochemistry,
68, 297-311.
668
669
670
671
Kankaala, P., Käki, T., Mäkelä, S., Ojala, A., Pajunen, H., and Arvola, L., 2005.
Methane efflux in relation to plant biomass and sediment characteristics in stands of three common emergent macrophytes in boreal mesotrophic lakes,
Glob. Change Biol., 11, 145-153.
672
673
Karl, D.M. and Tilbrook, B.D., 1994. Production and transport of methane in oceanic particulate organic matter, Nature,368, 732-734.
674
675
Karl, D.M., Beversdorf, L., Björkman, K.M., Church, M.J., Martinez, A., and DeLong,
E.F., 2008. Aerobic production of methane in the sea, Nat. Geosci., 1, 473-478.
676
677
Kavvas, M.L and Delleur, J.W., 1975. Removal of periodicities by differencing and monthly mean substraction, J. Hydrol., 26, 335–353.
678
679
680
681
Kiene, R.P., 1991. Production and consumption of methane in aquatic system. In:
Rogers, J.E. and Whitman, E. (Eds) Microbial production and consumption of greenhouse gases: methane, nitrogen oxides and halomethanes, America Society of Microbiology. Washington D.C. pp 111-149.
682
683
Kock, A., Gebhardt, S., and Bange, H.W., 2008. Methane emissions from the upwelling area off Mauritania (NW Africa), Biogeosciences, 5, 1119-1125.
684
685
686
Kutzbach, L., Wagner, D., and Pfeiffer, E.M., 2004. Effect of microrelief and vegetation on methane emission from wet polygonal tundra, Lena delta, Northern Siberia,
Biogeochemistry, 69, 341-362.
687
688
689
Linke, P., Sommer, S., Rovelli, L., and McGinnis, D.F, 2010. Physical limitations of dissolved methane fluxes: The role of bottom-boundary layer processes, Mar.
Geol., 272, 209-222.
21

690
691
692
Matthews, E. and Fung, I., 1987. Methane emission from natural wetlands: global distribution, area, and environmental characteristics of sources, Global
Biogeochem. Cy., 1, 61-86.
693
694
695
McGinnis, D.F., Greinert, J., Artemov, Y., Beaubien, S.E., and Wüest, A., 2006. Fate of rising methane bubbles in stratified waters: How much methane reaches the atmosphere?, J. Geophys. Res., 111, C09007, doi: 1029/2005JC003183.
696 Merriam-Webster.com
. 2011. http://www.merriam-webster.com
697
698
699
Mikkelä, C., Sundh, I., Svensson, B., and Nilsson, M., 1995. Diurnal variation in the methane emission in relation to the water table, soil temperature, climate and vegetation cover in a Swedish acid mire, Biogeochemistry, 28, 93-114.
700
701
Mitsch, W. J., and Gosselink, J.G., 2001. Wetlands.2
nd
edition. John Wiley and Sons,
New York.
702
703
Moore, T.R., Heyes, A., and Roulet, N.T., 1994. Methane emissions from wetlands, southern Hudson Bay lowland, J. Geophys. Res., 99, 1455-1467.
704
705
706
Naqvi, S.W.A., Bange, H.W., Farías, L., Monteiro, P.M.S., Scranton, M.I., and Zhang,
J., 2010. Marine hypoxia/anoxia as a source of CH
4
and N
2
O, Biogeosciences, 7,
2159-2190.
707 OECD, 1982. Eutrophication of waters: Monitoring, assessment and control, Paris .
708
709
Pulliam, W.M., 1993. Carbon dioxide and methane exports from a south eastern floodplain swamp, Ecol. Monogr., 63, 29-53.
710
711
Purvaja, R. and Ramesh, R., 2001. Natural and anthropogenic methane emission from coastal wetlands of South India, Environm. Managemn., 27, 547-557.
712
713
714
Reeburgh, W.S., 2003. Global methane biogeochemistry, In Treatise on Geochemistry,
The Atmosphere, H.D. Holland and K.K. Turekian (eds.), Elsevier/Pergamon,
Amsterdam, pp. 65-89.
715 Reeburgh, W.S., 2007. Oceanic methane biogeochemistry, Chem. Rev. ,107, 486-513.
22

716
717
Rehder, G., Keir, R.S., Suess, E., and Pohlmann, T., 1998. The multiple sources and patterns of methane in North Sea waters, Aquat. Geochem., 4, 403-427.
718
719
Rehder, G. and Suess, E., 2001. Methane and pCO
2 in the Kuroshio and the South China
Sea during maximum summer surface temperatures, Mar. Chem. 75, 89-108.
720
721
722
723
Ringeval, B., de Noblet-Ducondré, N., Ciais, P., Bousquet, P., Prient, C., Papa, F., and
Rossow, W.B., 2010. An attempt to quantify the impact in wetland extent on methane emissions on the seasonal and interannual time scales, Global
Biogeochem. Cy., 24, 1-12.
724
725
726
Roehm, L. C., 2005. Respiration in wetland ecosystems, In Respiration in aquatic ecosystems, P.A. del Giorgio and P.J. le B. Williams (eds),, Oxford University
Press, Oxford, pp. 83-102.
727
728
Saarnio, S., Winiwarter, W., and Leitao, J., 2009. Methane release from wetlands and watercourse in Europe, Atmos. Environ., 43, 1421-1429.
729
730
Scranton, M.I. and McShane, K., 1991. Methane fluxes in the southern North Sea: The role of European rivers, Cont. Shelf Res., 11, 37-52.
731
732
Segers, R., 1998. Methane production and methane consumption: a review of processes underlying wetland methane fluxes, Biogeochemistry, 41, 23-51.
733
734
Semiletov, I.P., 1999. Aquatic sources and sinks of CO
2
and CH
4
in the Polar regions, J.
Atmos. Sci., 56, 286-306.
735
736
Shannon, R. D. and White, J.R., 1994. A three-year study of controls on methane emissions from two Michigan peatlands, Biogeochemistry, 27, 35-60.
737
738
Siegel, S. and Castellan, N.J., 1988. Non parametric Statistics for the behavioral
Sciences. Mc-Graw-Hill Books, New York.
739
740
741
Striegl, R.G. and Michmerhuizen, C.M., 1998. Hydrologic influence on methane and carbon dioxide dynamics at two north-central Minnesota lakes, Limnol.
Oceanogr. 43, 1519-1529.
23

742
743
744
Sutton-Grier A. and Megonigal, P., 2011. The power of “green” energy: plant trait influences on microbial competition and greenhouse gas production, Nature, doi:
10.1038/npre.2011.5942.1.
745
746
747
Walter, B. P., Heimann, M., Shannon, R.D., and White, J.R., 1996. A process-based model to derive methane emissions from natural wetlands, Geophys. Res. Lett.,
23, 3731-3734.
748
749
750
Wanninkhof, R., Asher, W.E., Sweeney, C., and McGillis, W.R., 2009. Advance in quantifying air-sea gas exchange and environmental forcing, Annu. Rev. Marine
Sci., 1, 213-244.
751
752
753
754
755
Ward, B.B., Kilpatrick, K.A., Novelli, P.C., and Scranton, M.I., 1987. Methane oxidation and methane fluxes in the ocean surface layer and deep anoxic waters,
Nature, 327, 226–229.
Wetzel, R.G., 2001. Limnology: Lake and River Ecosystems, 3 rd
edition, Academic
Press, San Diego.
756
757
758
Xing, Y., Xie, P., Yang, H., Ni, L., Wang, Y., and Rong, K., 2005. Methane and carbon dioxide fluxes from a shallow hypereutrophic subtropical Lake in China, Atmos.
Environm., 39, 5532-5540.
759
760
761
Yang, J., Zhang, G.L., Zeng, L.X., Zhang, F., and Zhao, J., 2010. Seasonal variation of fluxes and distributions of dissolved methane in the North Yellow Sea, Cont.
Shelf Res., 30, 187-192.
762
763
764
Zhang, G., Zang, J., Liu, S., Ren, J., Xu, J., and Zhang, F., 2008. Methane in the
Changjiang (Yangtze River) estuary and its adjacent marine area: riverine input, sediment release and atmospheric fluxes, Biogeochemistry, 91, 71-84.
765
24