Storm-triggered transport of coarse woody debris to the oceans
A.J. West1*, C.-W. Lin2, T.-C. Lin3, R.G. Hilton4, M. Tanaka2, C.-T. Chang3, K.-C. Lin5, A.
Galy6, R. Sparkes6, and N. Hovius6
1Earth
Dynamic Systems Research Center, National Cheng Kung University, Taiwan
2Department
of Earth Sciences, National Cheng Kung University, Taiwan
3Department
of Life Sciences, National Taiwan Normal University, Taiwan
4Department
of Geography, University of Durham, UK
5Taiwan
Forestry Research Institute, Taipei, Taiwan
6Department
of Earth Sciences, University of Cambridge, UK
Manuscript in preparation for Limnology and Oceanography
*correspondence: [email protected];
presently at: Department of Earth Sciences, University of Southern California
West et al., Title & Abstract
Abstract
A significant consequence of Typhoon Morakot in August 2009 was the production
of vast volumes of driftwood in Pacific Asia. The flux of this coarse woody debris
(CWD) to the oceans from typhoon-triggered landsliding in Taiwan, where Morakot
made landfall, has been quantified by combining remote sensing (using FORMOSAT2 imagery and aerial photography), analysis of forest biomass, and field
observations. A total of 3.8-8.4 Mtons CWD was transported to the oceans, and this
carried 1.8-4.0 Mt of organic carbon, representing a highly concentrated flux of
carbon and nutrients. Mobilization of coarse woody debris during large storms on
high-relief islands may play a significant role in transfer of terrestrial biomass to the
oceans, with important implications for marine ecology and for feedbacks between
global climate, storm intensity, and carbon cycling.
West et al., Title & Abstract
1
Introduction
2
Dead trees and branches are now known to play key roles in forest ecology
3
and hydrology (Gregory et al., 2003), including regulating biogeochemical cycling of
4
carbon and other nutrients (e.g., Harmon et al., 1986), providing important species-
5
specific habitats (e.g., Flebbe, 1999), and creating structure in stream channels (e.g.,
6
Montgomery et al., 2003). Some woody material can also be exported from forests,
7
but this process has not been quantified in such a way to make it possible to assess
8
the impact both on downstream environments, and on biogeochemical cycles from
9
the regional to global scale. Though coarse woody material, defined as having
10
diameter >10cm, makes up a significant majority of the carbon pool in many forests
11
(e.g. 36-88% in the forests studied by Naiman et al., 1987), most organic material
12
transport is thought to be in the form of particulate organic matter (POM)
13
disseminated in riverine sediment. In part this reflects the physical and chemical
14
breakdown of coarse woody debris (CWD) on the forest floor and in stream
15
channels (e.g., Johnson et al., 2006), but it also reflects a sampling bias. The total
16
transfer of organic carbon in river systems is often assessed by chemical analysis of
17
POM content in river sediment, in parallel with hydrometric measurements (e.g.,
18
Wheatcroft et al. 2010). This inevitably overlooks coarse material. A few studies
19
have specifically considered the transport of coarse POM (typically >1mm), typically
20
by capturing stream-borne material in traps (e.g., Wallace et al., 1995; Johnson et al.,
21
2006; Cordova et al., 2008). However most of this material is leaf, litter, and small
22
wood fragments, and such studies have not captured the transport of the largest
West et al., p.1
23
trunks and branches that make up CWD, because these are not typically mobilized
24
under normal flow conditions.
25
Some work has explored the detailed mechanisms through which CWD is
26
transported, using both observations of the distribution of varying wood fragment
27
sizes in stream systems (e.g. Nakamura and Swanson 1994, and references therein),
28
and a combination of flume experiments with theoretical modelling (e.g., Braudrick
29
and Grant, 2000; Haga et al., 2002). These studies suggest that, while small CWD can
30
be removed by some stream and river systems, significant transport of large CWD is
31
restricted by its large size relative to bank full width of river channels, except where
32
logs are input directly into high-order, wider rivers, for example by logging
33
operations. However, large storms, by generating anomalously high bank-full
34
discharge, should significantly enhance the capacity of stream networks for
35
transporting CWD. This raises concern over the effect of woody debris in
36
exacerbating the damage caused by floods in populated areas (Haehnal and Daly,
37
2004). Nonetheless, it has generally not been possible to directly observe
38
mobilization and transport of CWD in major floods. While post-flood distribution
39
has been occasionally recorded (e.g. Pettit et al., 2005), there is no information
40
about the total amount of wood that is mobilized by very large storms. In particular,
41
the volume of woody debris that is exported from forests and transported to lakes
42
and oceans by rivers is a complete unknown.
43
In addition to causing flood-associated damage, such transport could have
44
important regional effects on lacustrine and coastal environments, since woody
West et al., p.2
45
material has distinct biogeochemical and physical roles in these aquatic settings
46
(e.g., Maser et al., 1988). It could also play an important role in the global carbon
47
cycle if it delays the oxidation of organic material by sequestration in deep water or
48
sediment. This would influence the way that changing storm frequency and
49
intensity acts as a feedback in the global climate system (Chambers et al., 2007). The
50
impact of tropical storms on forest vegetation has been shown to be significant, but
51
it is not clear if the net effect is to release carbon to the atmosphere, through
52
reduction of living biomass (Zeng et al., 2009), or to trap carbon through burial of
53
organic material in sediments, facilitated by transfer at high concentrations from the
54
land surface to the oceans (Hilton et al., 2008a). The balance between these effects
55
may determine whether increased storm frequency and intensity acts as a negative
56
or positive feedback in response to changing climate. Woody debris that is carried
57
by flooding rivers may be an important, unrecognized component of the storm-
58
carbon system. This lends particular importance to quantifying storm-driven
59
transport of CWD.
60
In this study we address this shortcoming by determining the total volume of
61
woody debris that was mobilized by Typhoon Morakot and quantifying the flood-
62
driven fluvial export to the ocean. Morakot was a cyclonic storm that delivered near
63
record levels of precipitation to the island of Taiwan, with a maximum local rainfall
64
of 2,965mm recorded over four days beginning on August 7th, 2009 (Taiwan Central
65
Weather
66
http://www.cwb.gov.tw/eng/index.htm). Though only rated as a Category 3 storm,
67
and with lower wind speeds that some other 2009 cyclones, Morakot formed in a
Bureau,
Climate
Statistics:
West et al., p.3
Daily
Time
Series;
68
monsoon trough and delivered anomalous moisture as a result. The resulting
69
flooding caused significant loss of life and property damage (Ling et al., 2009), partly
70
associated with river-bank collapse but largely because of widespread landsliding
71
on steep mountain slopes. The coupling of this landsliding with high water flow in
72
stream and river channels provided optimal conditions for mobilization and
73
transport of CWD. We have combined analysis of remote sensing imagery, data on
74
regional forest biomass, and field observations to determine the amount of CWD
75
mobilized from hillslopes during the storm and subsequently transported to the
76
oceans by rivers in Taiwan. This provides the first quantitative view of the
77
significance of woody debris transport in large tropical storms, made possible by the
78
availability of frequently acquired, high-resolution imagery.
79
80
Methods
81
We adopted two independent approaches to determining the volume of CWD
82
transported by Morakot. The first was to determine the total volume of woody
83
material that was mobilized by landsliding in one large river catchment, the Kaoping
84
(3,320 km2 area), and then assess how much of this was trapped both in landslide
85
deposits and in the river network, in order to determine by difference the volume of
86
material transported to the oceans. The second was to use the amount of wood
87
recovered after the typhoon in the Tsengwen water reservoir (with upstream
88
catchment area of 505.8 km2) to assess the total volume that was transported from
89
the upstream catchment area. Results of the two methods, though independent, can
West et al., p.4
90
be directly compared because they both provide estimates of the volume of wood
91
transported per unit of landslide area in each catchment.
92
FORMOSAT-2 satellite imagery, with a spatial resolution of 8m in multi-
93
spectral mode, was used to identify landslide areas across all of Taiwan (Figure 1)
94
based on a semi-automated routine. Initially, a combined index of NDVI and slope
95
was used to cluster bare areas. These were then screened manually to remove urban
96
areas and riverbeds. Landslides resulting from specifically from Typhoon Morakot
97
were distinguished by overlaying imagery from before and after the event. Pre-
98
event imagery consisted of a mosaic collected between December 1, 2008 and
99
August 4, 2009; post-event imagery was collected between August 21, 2009 and
100
September 10, 2009.
101
Within the Kaoping River basin, the areas affected by Marokot-triggered
102
landslides were gridded at 8m resolution and overlain on a digitized land use map
103
(Taiwan Forestry Bureau, 1999) in order to allocate land use and/or forest type to
104
all areas disturbed by landslides. Biomass harvesting resulting from landsliding was
105
then calculated in two ways. The first method (Figure 1) used the classification
106
system of the Taiwan Forestry Bureau (TFB) land use map together with reported
107
timber accumulation, determined from monitoring of 2,411 permanent forestry
108
plots across 26 different forest types (Taiwan Forestry Bureau, 1995). This method
109
of calculating CWD production has disadvantages. It does not include contribution
110
from bamboo forests to the CWD budget, because this is not included in reported
111
forest timber volume estimates (Taiwan Forestry Bureau, 1995), and it does not
West et al., p.5
112
allow calculation of the total biomass volume (i.e., including non-woody biomass)
113
involved in landsliding. The second method involved separating land uses into 7
114
general categories (Table 1), and classifying the typhoon-triggered landslides into
115
these categories based on land use types described in the TFB inventory (Taiwan
116
Forestry Bureau, 1995). For each category, average total biomass, and average trunk
117
+ branch biomass, was estimated from a compilation of studies on forest ecology in
118
Taiwan (Lin and Ho, 2005). This compilation includes data from 32 detailed plots
119
studies of biomass volume and distribution within forests; these were separated
120
between hardwood (n=8), conifer (n=20), and bamboo (n=4) in order to estimate
121
average biomass totals for each forest type (see Table 1). Natural and plantation
122
forests were not distinguished in this approach, in contrast to the method directly
123
using Taiwan Forestry Bureau data. Biomass in conifer-hardwood mixed forest was
124
assumed to be the equally-weighted average of conifer and hardwood forest totals;
125
mixed hardwood-conifer-bamboo forest were also assumed to have equally-
126
weighted average biomass.
127
Not all of the biomass destroyed by landsliding entered the river system. The
128
position of landslides on hillslopes, which varies with triggering mechanism
129
(Densmore and Hovius, 2000), may mean that landslide material is physically
130
separated from the river channel network. We used aerial photographs to assess the
131
connectivity of post-Morakot landslides with debris flow and river channels, by
132
randomly selecting three 5km x 5km plots from a grid of the Kaoping basin, and
133
manually separating landslides with and without visible links for biomass transport
134
into the channel network. We calculated the proportion of landslide area within
West et al., p.6
135
each plot where direct physical connection could allow for transport of material
136
with the large size of CWD from landslides into the river network.
137
Even when landslides are connected to the river network, not all organic
138
material was necessarily evacuated from the landslide areas. The total amount of
139
CWD material on 11 landslide surfaces distributed across the basin was determined
140
by measuring CWD line length in the orthorectified aerial photographs, and by
141
calculating trunk + branch biomass based on allometric scaling relationships.
142
Woody biomass Bw can be described as a function of tree height h based on the
143
equation:
144
ln(Bw) = a + b*ln(h)
145
where a and b are empirically derived coefficients specific to forest type. We used a
146
= 6.14 and b = 4.6 based on coefficients from allometric studies of hardwood forest
147
in Taiwan (Lin et al., 2006). Allometrically-derived biomass was determined for
148
each woody line segment measured in aerial photographs; these were summed to
149
give the total CWD mass on landslide surfaces. Aerial photography does not capture
150
burial of woody debris below the surface, and we used field observations of
151
landslides to assess deep burial of CWD. We identified six deposits at a range of
152
locations in the Kaoping basin where engineering work on roads and bridges left
153
clean cross-sections, and measured the below-ground distribution of woody debris
154
in each deposit.
155
West et al., p.7
156
Results
157
Mapped landslides resulting from Morakot covered an area of 404.3 km2
158
across all of Taiwan. Within the Kaoping river basin, the area of typhoon-triggered
159
landsliding was 129.8 km2, equivalent to a landslide density of 3.9% on an area
160
basis. An additional 38.0 km2 of landsliding was identified within the Kaoping basin
161
but was present prior to Typhoon Morakot. In the catchment upstream of Tsengwen
162
reservoir, the typhoon-triggered landslide area totalled 14.2km2, or 2.8% of the
163
total land surface area. The calculated landslide densities for the Kaoping and
164
Tsengwen reflect concentrated landslide activity but are within the range reported
165
in other settings for landslide erosion following major storms. A global assessment
166
of regions affected by storm-triggered landslide activity prior to 1988 found that
167
0.3-25% of areas surveyed were eroded by landsliding (Crozier et al., 1982).
168
However one distinguishing feature of Marokot was the large total area of Taiwan
169
that was affected, covering a majority of the mountainous area of the island (total
170
land area of 35,980 km2), compared to the largest study area reported by Crozier et
171
al (1982) of 240 km2.
172
Kaoping Woody Debris Budget – Based on the distribution of landslides as a
173
function of timber volume reported by the Taiwan Forestry Bureau (Figure 1), the
174
total yield of timber harvested from landsliding averaged 2.5±1.0x104 m3/km2. For
175
wood density of 613±176 kg/m3 (Zanne et al., 2009), this means a yield of
176
0.015±0.007 Mt/km2, totalling 2.0±0.9 Mt of timber over the entire mapped Kaoping
177
landslide area.
West et al., p.8
178
Based on the land use coding of landslide areas, and estimates of forest type
179
biomass (Table 1), the harvesting of total biomass in the Kaoping basin, including
180
aboveground biomass plus roots and soil litter, was 0.031±0.005 Mt/km2. This
181
implies that the total biomass involved in landsliding amounted to 4.0±0.7 Mt in the
182
Kaoping basin, and 12.5±2.0 Mt across all of the typhoon-triggered landslides in
183
Taiwan. Based on the same approach, the specific mobilization of trunk + large
184
branch biomass was 0.020±0.005 Mt/km2, totalling 2.6±0.6 Mt of material for the
185
Kaoping. This value for trunk + large branch material is within error of the estimate
186
based on the TFB timber volumes. The biomass-based estimate is used in further
187
calculations of the CWD budget for the Kaoping, because it also includes
188
contribution from bamboo forests, which are not included in the timber volume
189
estimates.
190
Despite variability in the position of landslide failures, and in the pattern of
191
run-out, pervasive channelization means that, in all three of the 5km x 5km plots
192
that were studied, <1% of landslide area was isolated from the channel system,
193
meaning that this was not a limit on CWD transport. The very high degree of
194
observed connectivity was most likely a consequence of the extremely high runoff
195
generated by the typhoon, which would have decreased the area required to
196
accumulate a flow depth capable of effectively transporting material.
197
Though virtually all landslides were connected to the river channels, trapped
198
woody material was observed on some low-slope surfaces and behind local
199
obstructions within deposits. The distribution measured based on scaling of aerial
West et al., p.9
200
photograph woody fragment line lengths is positively skewed, with a logmean of
201
154 t/km2 and 95% confidence interval of 41-585 t/km2. Over the area of mapped
202
Kaoping landslides, this implies storage of 0.020 Mt (95% CI: 0.005-0.076) CWD on
203
landslide surfaces. In terms of wood storage at depth, while all six deposits
204
examined in the field had variable presence of woody debris on the surface, none
205
had significant observable CWD buried below the top 1m. Though fine-grained
206
organic matter may be more uniformly distributed in landslides (e.g., Hilton et al.,
207
2008b), our observations suggest that CWD is concentrated at the surface, such that
208
the observed surface occurrence reflects total storage in landslides.
209
Of the CWD that was removed from landslides, some remains trapped in the
210
river network, with government surveys reporting 0.718 Mt of CWD in the Kaoping
211
system (Council of Agriculture, Executive Yuan, 2009). Of this, 0.545 Mt was
212
removed during engineering work, preventing further quantitative evaluation in the
213
field.
214
These results make it possible to calculate a total budget for CWD
215
mobilization and transport in the Kaoping basin. Of the 2.6±0.6 Mt CWD produced in
216
Kaoping landsliding, 0.005-0.076 Mt remain stored on landslide deposits, and 0.72
217
Mt was retained in the floodplain. The remaining 1.2-2.5 Mt were transported to the
218
oceans. Over the total landslide area in the Kaoping (129.8 km2) this represents a
219
specific yield to the oceans of 9.3-19.1 kt/km2.
220
Tsengwen Reservoir Woody Debris Budget – Results based on the quantity of
221
wood recovered in Tsengwen reservoir after the storm can be compared with this
West et al., p.10
222
value. A total of 473,000 m3 of CWD was collected by the Taiwan Water Resources
223
Agency immediately following Typhoon Morakot (Council of Agriculture, Executive
224
Yuan, 2009). Again using a wood density of 613±176 kg/m3 (Zanne et al., 2009),
225
this is equivalent to 0.290±0.083 Mt CWD. This was sufficient to cover the reservoir
226
surface in woody debris (Figure 2). All of the wood from the upstream catchment
227
area was trapped by the reservoir, so this total reflects everything that was
228
exported. In the absence of the reservoir, most of this CWD would have been
229
transported to the oceans, though some would have been retained in the floodplains
230
downstream of the reservoir. The budget for the Kaoping implies a wood retention
231
ratio in river channels of 0.2; in other words, 20% of CWD was retained along the
232
river channels. Applying this ratio to the Tsengwen would mean that a total of
233
0.232±0.066 Mt CWD would have been transported to the oceans from this
234
catchment, in the absence of the reservoir. Mapped landslides in the Tsengwen
235
catchment make up 14.3 km2, giving a specific CWD yield of 16.2±4.6 kt/km2.
236
237
Discussion
238
The two independent estimates of CWD yield, one from reservoir
239
accumulation and the other from landslide inventory, are consistent (see Table 2). If
240
this range of yield applies to the landslide area across Taiwan, which is a reasonable
241
assumption given the consistency between the two distinct catchment areas, then a
242
total of 3.8-8.4 Mt CWD was delivered to oceans and reservoirs as a result of
243
Typhoon Morakot. Dammed river basins make up only a small portion of the total
West et al., p.11
244
drainage area of Taiwanese catchments (3404 km2, equivalent 9% of the island
245
surface area; Taiwan Water Resources Agency, 2008), so most of this flux reached
246
the oceans.
247
Most previous work has assessed the amount of organic material carried by
248
rivers in terms of carbon flux. Based on average carbon content of 48.1±0.8% for
249
Taiwan woody biomass (Lin and Ho, 2005), the Morakot-transported CWD carried
250
1.8-4.0 Mt C. Seo et al. (2008) quantified the riverine yield of woody debris from
251
Japan by looking at records of CWD collected over multi-year periods (varying from
252
1 to 29 years) from 131 reservoirs in Japan. Their results imply a total flux of 7.31
253
ktC.yr-1, an annual flux nearly three orders of magnitude lower than the flux
254
transported by the shorter-duration Morakot flood. In terms of area-normalized
255
yield, the average value for Japanese reservoirs was 0.28 tC.yr-1.km-2, with individual
256
reservoir catchments ranging in average yield from 0.02 tC.yr-1.km-2 to 3.08 tC.yr-
257
1.km-2.
258
represents 50-111 tC.km-2. Compared to the average background value observed in
259
the Japan dataset, the Morakot triggered flux of CWD reflects ~200-400 years worth
260
of transport. Even for a return time of >50 years, such an extreme storm event
261
represents a major contribution to the overall long-term budget.
The Morakot-delivered yield from Taiwan (area 35,980 km2), in contrast,
262
Our results suggest that the major role for CWD may be a critical distinct
263
feature of storm transport. Unfortunately it is not possible to directly estimate the
264
total amount of POC (i.e. the non-coarse, <10cm portion of organic material)
265
transported by Morakot itself. However, Hilton et al. (2008a) used hydrometric
West et al., p.12
266
measurements to estimate the POC flux transported by the Liwu River in eastern
267
Taiwan during Typhoon Mindulle in 2004; their calculated total POC yield of 13
268
tC/km2 is a fraction of the CWD yield estimated from Morakot. The precipitation
269
intensity was lower in Mindulle compared to Morakot, confounding direct
270
comparison, but the contrast suggests that CWD is likely to play an important part in
271
storm-event organic transport. This contrasts with non-storm transport when,
272
based on the analysis of Seo et al. (2009), a maximum of <10% of total organic
273
carbon transport occurs as CWD. We attribute this to the fact that high depths of
274
river flow in extreme storms allow transport of woody material that would not be
275
removed in smaller storms.
276
The Morakot-delivered CWD flux is not only distinguished by the significant
277
role of coarse debris, but also simply by its overall magnitude when compared with
278
global fluxes of POC. The total Morakot-delivered CWD carbon flux represents 10-
279
26% of the yearly flux of particulate carbon from the Amazon River, the largest
280
point source of terrestrial carbon to the oceans (Schlunz and Schneider, 2000). The
281
Morakot yield of CWD (50-111 tC.km-2; see above) is orders of magnitude higher
282
than the 2.6 tC.km-2 transported over the entire year from the Amazon. A key
283
characteristic of the Morakot flux was that it was delivered from a small area over a
284
short amount of time, suggesting that, though major storms such as Morakot are
285
rare, they may have a highly concentrated impact on the reservoir, coastal, and open
286
ocean environments where terrestrial organic material is delivered.
West et al., p.13
287
One of the key impacts of such highly concentrated CWD delivery is physical,
288
with large areas of open water covered by debris. Driftwood generated by Morakot
289
covered reservoirs (e.g. Figure 2) and blocked harbours throughout Pacific Asia in
290
the weeks following the event. This generated a range of secondary natural hazards
291
(including for shipping and drinking water supply). It also would have been
292
accompanied with significant change in the physical conditions (light, temperature)
293
of the shallow water environment, yet the impact of such sudden, short-term
294
disturbance on the marine environment is largely unknown.
295
The concentrated flux of woody debris associated with storms may play a key
296
role in coastal and ocean ecology. Woody debris is known to be critically important
297
both to coastal, pelagic, and benthic ecological communities (e.g., Mortenson, 1938;
298
Gooding and Magnuson, 1967; Hoyoux et al., 2009). Though it is known that the
299
supply of woody material can have important effects on both surface and benthic
300
ecosystems, it is not clear what the role of highly concentrated flux, such as that
301
from a major storm like Morakot, may be. For example, episodic transport during
302
large storms may be important for facilitating waterlogging and sinking of woody
303
material.
304
If significant volumes of woody debris that reach the oceans do sink to the
305
deep ocean, the net transfer of carbon may determine how storms influence the
306
global carbon cycle. In the case of Morakot, the amount of CWD transported to the
307
oceans makes up a significant portion (13-35%) of the total living carbon that was
308
mobilized by landsliding. The remaining biomass is most likely to either decompose
West et al., p.14
309
on land, or be transported as sediment-borne POC after it is physically broken down
310
on the forest floor or in stream channels. While around 80% of terrestrially-derived
311
organic carbon is thought to be decomposed in the oceans (e.g. Hedges et al., 1997;
312
Burdige, 2005), the transport, burial and preservation of CWD in the oceans may
313
differ significantly from POC, with slower oxidation rates and longer transport
314
distances expected for the smaller surface to volume ratio of coarse material
315
(Spanhoff and Meyer, 2004). If significant amounts of carbon are preserved as a
316
result of episodic, storm-driven flooding (e.g., Heezen et al., 1964), such as that
317
associated with Morakot, then regrowth of forest vegetation on landslide surfaces
318
and deposits (e.g., Restrepo et al., 2009) would lead to a net sink of atmospheric CO 2
319
over time scales of forest regeneration (e.g., Hilton et al., 2008b). This facet of storm-
320
driven disturbance has not been accounted for to date (e.g., Chambers et al., 2007)
321
and these impacts remain particularly important to constrain if the frequency of
322
such large storms changes with global climate (IPCC, 2007).
323
The observation that storms are capable of delivering quantitatively
324
significant volumes of coarse organic matter to the oceans leaves open many
325
questions about the fate of this material, and its importance in the marine
326
environment. The prevalence of steep, forested islands in the humid tropics means
327
that transport of CWD driven by very large tropical storms may be an important
328
pathway for transport of organic matter to the oceans that has not been previously
329
recognized, and deserves further attention.
330
Acknowledgments
West et al., p.15
331
Research supported by the Taiwan Ministry of Education, the Taiwan National
332
Science Council, and the U.K. Natural Environment Research Council.
333
West et al., p.16
334
References
335
336
Braudrick, C. A. and G. E. Grant. 2000. When do logs move in rivers? Water Resour.
Res. 36: 571-583.
337
338
Burdige, D. J. 2005. The burial of terrestrial organic carbon in marine sediments: A
re-assessment. Global Biogeochem. Cy. 19: 10.1029/2004GB002368.
339
340
Chambers J. Q., J. I. Fisher, H. Zeng, E. L. Chapman, D. B. Baker, and G. C. Hurtt. 2007.
Hurricane Katrina's carbon footprint on U.S. Gulf Coast forests. Science. 318: 1107.
341
342
343
Cordova, J. M., E. J. Rosi-Marshall, J. L. Tank, and G. A. Lamberti. 2008. Coarse
particulate organic matter transport in low-gradient streams of the Upper Peninsula of
Michigan. J. N. Am. Benthol. Soc. 27:760–771
344
345
Council of Agriculture, Execute Yuan of Taiwan. 2009. Progress Report of the Typhoon
Morakot Reconstruction Committee, October 22nd, Executive Yuan, Taipei, Taiwan.
346
347
Crozier M. J., J. A. McConchie, R. C. Owen, and R. J. Eyles. 1982. Mass Movement
Erosion. Wairarapa, Wellington.
348
349
Densmore, A. L. and N. Hovius. 2000. Topographic fingerprints of bedrock landslides.
Geology. 28: 371.
350
351
Flebbe, P. A. 1999. Trout use of woody debris and habitat in Wine Spring Creek,
North Carolina. Forest Ecology and Management. 114: 367-376.
352
353
Gooding, R. M., and J. J. Magnuson. 1967. Ecological significance of drifting objects to
pelagic fishes. Pacific Science. 21:486–497.
354
355
Haehnel, Robert B. and Steven F. Daly. 2004. Maximum Impact Force of Woody
Debris on Floodplain Structures. Journal of Hydraulic Engineering, 130:112-120.
356
357
358
359
Harmon M. E., Franklin J. F., Swanson F. J., Sollins P., Gregory S. V., Lattin J. D.,
Anderson N. H., Cline S. P., Aumen N. G., Sedell J. R., Lienkamper G. W., Cromack K. Jr,
Cummins K. W. 1986. Ecology of coarse woody debris in temperate ecosystems. Adv.
Ecol. Res. 15: 133–302.
360
361
Hedges, J. I., R. G. Keil, and R. Benner. 1997. What happens to terrestrial organic
matter in the ocean? Org. Geochem. 27: 195-212.
362
363
Heezen, B. C., R. J. Menzies, E. D. Schneider, W. M. Ewing, and N. C. L. Granelli. 1964.
Congo submarine canyon. Bull. Am. Assoc. Petrol. Geologists 48: 1126-1149.
364
365
366
Hilton, R. G., A. Galy, N. Hovius, M.-C. Chen, M.-J. Horng, and H. Chen. 2008a. Tropicalcyclone-driven erosion of the terrestrial biosphere from mountains. Nature Geosci.
1: 759-762.
West et al., p.17
367
368
Hilton, R. G., A. Galy, and N. Hovius. 2008b Riverine particulate organic carbon from an
active mountain belt: importance of landslides. Global Biogeochem. Cy. 22: GB1017.
369
370
371
Hoyoux, C. M. Zbinden, S. Samadi, F. Gaill, and P. Compère. 2009. Wood-based diet
and gut microflora of a galatheid crab associated with Pacific deep-sea wood falls.
Marine Biology 156: 2421-2439.
372
373
374
Intergovernmental Panel on Climate Change (IPCC). 2007. Climate Change 2007:
The Physical Science Basis—Summary for Policymakers, Cambridge Univ. Press,
Cambridge, UK.
375
376
377
Johnson, M. S., J. Lehmann, E. Carlos Selva, M. Abdo, S. Riha, and E. Guimaraes-Couto.
2006. Organic carbon fluxes within and streamwater exports from headwater
catchments in the southern Amazon. Hydrol. Process. 20: 2599-2614.
378
379
380
381
Lin, K. C. and S. L. Ho. 2005. Carbon sequestration in forests of different ages
estimated from biomass, p 97-108. In Proceedings of Conference on Contribution of
Forest Management to Carbon Dioxide Sequestration, Taiwan Forest Bureau and
National Taiwan University, Taipei, Taiwan. [in Chinese]
382
383
384
385
Lin, K. C., C.-T. Duh, C.-C. Hsu, and C.-M. Huang. 2006. Aboveground Carbon
Accumulation of Subtropical Natural Broadleaf Forests in the Liukuei Experimental
Forest of Southern Taiwan. Jour. Exp. For. Nat. Taiwan Univ. 20: 153-164. [in Chinese
with English abstract]
386
387
388
389
390
Ling, H. I., F.-Y. Menq, M.-H. Wu, A. Y.-C. Hunag, and C.-W. Hsu. 2009. A
reconnaissance report on Typhoon Morakot. Report of the NSF-sponsored
Geoengineering Extreme Events Reconnaissance Team, GEER Association Report No.
GEER-018. Available through
http://research.eerc.berkeley.edu/projects/GEER/Post_EQ_Reports.html
391
392
Maser, C. and J. E. Sedell. 1994. From the Forest to the Sea: The Ecology of Wood in
Streams, Rivers, Estuaries, and Oceans, St. Lucie Press, Florida.
393
394
395
396
Montgomery, D. R., B. D. Collins, T. B. Abbe and J. M. Buffington. 2003. Geomorphic
effects of wood in rivers, p 1-10. In K. L. S. Gregory and A. Boyer [eds.], The Ecology
and Management of Wood in World Rivers, American Fisheries Society Symposium
vol. 37, American Fisheries Society, Bethesda, MA.
397
398
Mortenson, T. 1938. On the vegetation diet of some deep-sea echinoids.
Annotationes Zool. Japon. 17: 225-228.
399
400
401
Naiman, R. J., J. M. Melillo, M. A. Lock, T. E. Ford, and S. R. Reice. 1987. Longitudinal
Patterns of Ecosystem Processes and Community Structure in a Subarctic River
Continuum. Ecology 68: 1139-1156.
402
403
Nakamura, F., and F. J. Swanson. 1994. Distribution of coarse woody debris in a
mountain stream, western Cascades Range, Oregon. Can. J. For. Res. 24: 2395-2403.
West et al., p.18
404
405
406
Pettit, N. E., R. J. Naiman, K. H. Rogers, and J. E. Little. 2005. Post-flooding
distribution and characteristics of large woody debris piles along the semi-arid
Sabie River, South Africa. River Research and Applications 21: 27-38.
407
408
409
Restrepo, C., L. Walker, A. Shiels, R. Bussman, L. Claessens, S. Fisch, P. Lozano, G.
Negi, L. Paolini, G. Poveda, C. Ramos-Scharron, M. Richter, and E. Velazquez. 2009.
Landsliding and its multi-scale influence on mountainscapes. BioScience. 59: 1-14
410
411
Schlunz, B. and R. R. Schneider. 2000. Transport of terrestrial organic carbon to the
oceans by rivers: re-estimating flux- and burial rates. Int. J. Earth Sci. 88: 599.
412
413
414
415
Seo, J.-I., F. Nakamura, D. Nakano, H. Ichiyanagi, and K. W. Chun. 2008. Factors
controlling the fluvial export of large woody debris, and its contribution to organic
carbon
budgets
at
watershed
scales.
Water
Resour.
Res.
44:
doi:10.1029/2007WR006453.
416
417
Spanhoff, B. and E. I. Meyer. 2004. Breakdown rates of wood in streams. J. N. Am.
Benthol. Soc. 23: 189.
418
419
420
Taiwan Forestry Bureau. 1995. 3rd Taiwan Forestry Resource and Land Use
Investigation and Report, Executive Yuan, Taipei, Taiwan. Also available at
http://www.forest.gov.tw
421
422
423
Taiwan Water Resources Agency. 2009. Hydrological Yearbook of Taiwan, Republic
of China. Ministry of Economic Affairs, Taipei. Available at
http://gweb.wra.gov.tw/wrweb/
424
425
426
Wallace, J. B., M. R. Whiles, S. Eggert, T. F Cuffney, G. J. Lugthart, and K. Ching.
1995. Long-term dynamics of coarse particulate organic matter in three Appalachian
Mountain streams. J. N. Am. Benthol. Soc. 14: 217-232.
427
428
429
Wheatcroft, R. A., M.A Goñi, J. A. Hatten, G. B. Pasternack, J. A. and Warrick J.A.
2010. The role of effective discharge in the ocean delivery of particulate organic carbon
by small, mountainous river systems. Limnology and Oceanography 55: 161–171.
430
431
432
Zanne, A.E., G. Lopez-Gonzalez, D. A. Coomes, J. Ilic, S. Jansen, S. L. Lewis, R. B.
Miller, N. G. Swenson, M. C. Wiemann, and J. Chave. 2009. Global wood density
database, Dryad. Also available at http://hdl.handle.net/10255/dryad.235
433
434
435
Zeng H., J. Q. Chambers, R. I. Negrón-Juárez, G. C. Hurtt, D. B. Baker, M. D. Powell.
2009. Impacts of tropical cyclones on U.S. forest tree mortality and carbon flux from
1851 to 2000. Proc Natl Acad Sci U S A. 106: 7888-92.
436
West et al., p.19
Figure Captions
Figure 1. Map of Taiwan, showing landslide areas identified as resulting from
Typhoon Morakot, with color scale indicating the volume of timber accumulation
based on the reported volumes for each forest type (Taiwan Forestry Bureau, 1995).
Inset shows FORMOSAT-2 imagery for identified region, illustrating the landsliding
that stripped vegetation from hillslopes and supplied CWD directly to the river
system. Black dot shows the location of Alishan, where the long-term precipitation
record shown in Figure 3 was collected.
Figure 2. A water reservoir in Taiwan in the days following Typhoon Morakot,
showing the accumulation of coarse woody debris. Calculations of the total flux of
this debris highlight the importance of such storms for transferring terrestrial
biomass to the oceans, and suggest that coarse material may play a significant role
that is not accounted for in measurement of river sediment (photo courtesy C.-B.
Lin).
Figure 3. Rainfall record for Alishan, Taiwan (Taiwan Central Weather Bureau,
Climate Statistics: Daily Time Series; http://www.cwb.gov.tw/eng/index.htm), for
2009, showing heavy rainfall associated with Typhoon Morakot. Landslides
triggered by the typhoon were assessed by comparing mosaics of FORMOSAT
imagery collected before and after the event. Inset shows the 2009 precipitation in
comparison with the record from the previous 10 years. The extreme magnitude of
Morakot confounds direct calculation of its recurrence interval.
West et al., Figures & Tables, p.1
Figure 1.
West et al., Figures & Tables, p.2
Figure 2.
West et al., Figures & Tables, p.3
Figure 3.
West et al., Figures & Tables, p.4
TABLE 1. Biomass mobilized during landsliding in the Kaoping catchment
Landslide Area
Total forest
biomass3
Mobilized total
biomass
Forest trunk +
branch
biomass
Forest type
(km2)
(Mg.ha-1)
(Mt)
(Mg.ha-1)
(Mt)
Conifer
17.3
265±54
0.4±0.1
179±50
0.31±0.09
Hardwood
73.2
387±891
2.8±0.6
240±80
1.76±0.58
Coniferhardwood
forest
15.0
326±104
0.49±0.2
210±94
0.31±0.14
Bamboo forest
3.2
117±96
0.04±0.03
101±96
0.03±0.03
Mixed forest
10.2
222±1422
0.2±0.1
155±135
0.16±0.14
Grassland &
shrub
3.0
04
0
0
0
Other
(unvegetated)
7.9
0
0
0
0
Mobilized trunk
+ branch
biomass
Calculated as equal mixture of conifer and hardwood; 2Calculated as equal mixture of hardwood,
conifer, and bamboo; 3Total aboveground biomass plus soil litter and roots; 4Total biomass assumed
to be insignificant compared to the forest biomass.
1
West et al., Figures & Tables, p.5
Table 2. Summary of calculated woody debris fluxes and yields
Catchment
Kaoping Catchment
Data Source
Timber
Volume
Biomass
Tsengwen
Reservoir
Catchment
Mobilization of woody debris
from landslides
2.0±0.9 Mt
2.6±0.6 Mt
0.29±0.08 Mt
Yield* of woody debris into
river system
7.9-22.3
kt/km2
14.8-24.6
kt/km2
20±5.7 kt/km2
Yield* of woody debris to the
oceans
6.3-17.8
kt/km2
9.3-19.1
kt/km2
16.2±4.6 kt/km2 **
*yield calculated relative to typhoon-triggered landslide area in each catchment
**reflects total which would have been transported in the absence of the reservoir
West et al., Figures & Tables, p.6