Geomorphic impacts of a 100-year flood: Kiwitea Stream,
Manawatu catchment, New Zealand
Ian C. Fuller
Geography Programme, School of People, Environment and Planning, Massey University, Private Bag 11-222, Palmerston North, New Zealand
Received 30 December 2005; received in revised form 4 May 2006; accepted 21 February 2007
Available online 13 May 2007
Abstract
The geomorphic impacts of a 100-year flood are assessed in the Kiwitea Stream (254 km2
), a tributary within the Manawatu
River catchment (New Zealand), using sequential aerial photographs and reach-based morphological sediment budgeting. Channel
expansion and avulsion eroded in excess of one million cubic metres of sediment over 1 km2
of floodplain along a 30-km-long
reach of Kiwitea Stream. Channel transformation was spatially discontinuous and predominantly associated with large-scale bank
erosion in response to a flood over 5 times bigger than the mean annual flood (annual recurrence interval (ARI) 100 years). Total
energy expenditure of this flood in the Kiwitea was 14,900×103
J. The spatial discontinuity of channel transformation relates to
valley floor and channel configurations. High stream powers generated in confined channels at bends produced catastrophic
channel transformation. Where flood flows dissipated overbank, stream powers and the extent of channel transformation were
reduced. Hydrologic, hydraulic and geomorphic variables can be invoked to thus explain the variability of geomorphic impacts
encountered during this event.
 2007 Elsevier B.V. All rights reserved.
Keywords: Geomorphic impact; Flood; River channel change; Erosion; Sediment budget
1. Introduction
1.1. Geomorphic impacts
The impact of floods on channel morphology is
highly variable. Some major floods produce catastrophic
change (e.g. Schumm and Lichty, 1963; Baker, 1977;
Lisle, 1981; Gupta, 1983; Miller, 1995), while others
have little effect (e.g. Costa, 1974; Costa and O'Connor,
1995; Magilligan et al., 1998). Floods of similar
magnitude and frequency may therefore produce dissimilar
morphological response, even within the same
catchment (Costa, 1974; Nolan and Marron, 1985;
Nanson, 1986; Miller, 1990; Magilligan, 1992; Butler
and Malanson, 1993; Pitlick, 1993; Costa and O'Connor,
1995). Wolman and Gerson (1978) suggest the geomorphic
importance of an event is a product of an array of
factors, including magnitude, recurrence interval, processes
occurring during the interval between recurrence
and work performed during this intervening period.
Therefore, given the variety of processes and boundary
conditions, a spectrum of impacts for a given magnitude
event in any one catchment is to be expected.
The role of flooding in fluvial geomorphology has
been persistently controversial (Lewin, 1989) and much
debated since Wolman and Miller (1960) advocated the
view that channels were broadly adjusted to frequent
Available online at www.sciencedirect.com
Geomorphology 98 (2008) 84­95
www.elsevier.com/locate/geomorph
 Tel.: +64 6 3505799x2851; fax: +64 6 3505644.
E-mail address: I.C.Fuller@massey.ac.nz.
0169-555X/$ - see front matter  2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.geomorph.2007.02.026
events. Thus, for example, Harvey et al. (1979) suggest
systems adjust to major flood events (recurrence interval
(RI) up to 2 years), which control channel morphology,
while moderate events (RI 14­30 times a year)
influence adjustments within the overall morphology
created by the major events. However, increasingly, the
role of the extreme event has been recognized as significant
in conditioning channel form (e.g. Reid and
Frostick, 1994). Erskine (1994) suggested catastrophic
floods (N10 times the magnitude of the mean annual
flood) determine channel capacity, while smaller floods
control the form of the channel bed. Similar conclusions
were drawn by Pickup and Warner (1976), and Hack and
Goodlett (1960) suggested that rare, large magnitude
floods could have a dominant impact on some (mountain)
landscapes (Miller, 1995). Erskine (1986) also
described wholesale river metamorphosis during a series
of large floods between 1949 and 1955 in the lower
Macdonald River in NSW, Australia, which persisted
more than 30 years. Such a response was also observed
in the Cimarron River (Kansas) (Schumm and Lichty,
1963). Thus, large floods may either initiate long
periods of river instability and give rise to a flooddominated
channel morphology (Hickin, 1983), or they
may have little impact on a channel (Miller, 1990).
Richards (1999) suggests the morphological context
in which the flood takes place is critical to conditioning
the scale of its impacts (cf. Wolman and Gerson, 1978).
Baker (1977) argued that there is a high potential for
catastrophic channel response in small catchments with
highly variable flood magnitudes. Within a broader
context, catchment-scale boundary conditions may
condition the geomorphic effectiveness of floods
(Brooks and Brierley, 1997). Vegetation cover exerts a
fundamental control on hydrology and sediment supply
and may determine the sensitivity of a landscape (or
channel) to flood-induced change, with the possibility of
extreme impacts increasing in cleared catchments
(Erskine and Bell, 1982; Erskine and Warner, 1988).
This paper seeks to quantify the impacts of a single
flood event generated by a "150-year" storm in the
Kiwitea Stream, Manawatu catchment, lower North
Island of New Zealand, which occurred on 15­16
February 2004. The Kiwitea catchment within the
western Manawatu drainage basin has been cleared of
native forest cover within the last 150 years. The
effectiveness of flooding in this system is thus likely to
have increased. The ARI of the flood exceeded 100 years
(Fuller and Heerdegen, 2005) and may thus be classified
as large using Kochel's (1988) definition (ARI N
50 years). This work examines the extent to which the
channel system was modified by this rare event.
1.2. Study area
The western Manawatu catchment drains the southwestern
flanks of the Ruahine Ranges, which here rise to
1643 m, in the southern North Island of New Zealand
(Fig. 1). The Range comprises highly fractured greywacke
(alternating siltstone and sandstone) and forms
part of the North Island's axial ranges (Fig. 1). Uplift and
erosion rates are high: up to 3 mm year-1
and 0.7 mm
year-1
, respectively (Whitehouse and Pearce, 1992).
Dissected hill country immediately to the west of the
ranges forms much of the 254-km2
Kiwitea catchment.
This is located in the eastern margins of the Wanganui
Basin, a major structural depression where up to 4000 m
of marine sediment accumulated above the greywacke
basement during the Plio-Pleistocene (Heerdegen and
Shepherd, 1992). Poorly consolidated sands and gravels
underlie the Kiwitea catchment. In terms of specific
sediment yield, steepland grazed hill country in this
area yields up to 2000­5000 t km2
year- 1
(Hicks and
Shankar, 2003). Land use includes plantation forestry
(pines), varying grades of pasture and scrub.
This physiographic setting places the long (48 km),
narrow (average width 6.5 km) Kiwitea catchment at the
upland fringe of the axial ranges, with a relatively steep
gradient (0.005), gravely bed and highly erodible
boundary conditions. The Kiwitea planform is best
defined as wandering, using Neill's (1973) and Ferguson
and Werritty's (1983) term. This represents a transitional
pattern between multi-thread braided and single-thread
meandering channels; lacking the sinuosity to be
classified as meandering (1.44), or the degree of flow
division to be braided, but combining both mid-channel
bars and some well-developed bends, with extensive
lateral bar forms often present. Wandering rivers are by
nature dynamic (e.g. Ferguson and Werritty, 1983; Fuller
et al., 2003a), although the active channel of the Kiwitea
prior to the flood on 15­16 February 2004 was between
10 and 15 m wide and tree-lined for much of its length
(Philpott, 2005), which compares with the 23-m-wide
meandering channel of 1877 (Anon, 1980). Such channel
constriction increases frequency of sediment transport
(Laronne and Duncan, 1992). The result of this increased
movement of sediment is bed degradation of the Kiwitea
such that the 10-year flood would not overtop its banks
(Anon, 1980). Prior to the February 2004 flood, the
Kiwitea was therefore over-narrow and over-deep, largely
due to riparian plantings in a narrow riparian strip.
There is a steep rainfall gradient moving up the
catchment towards the ranges. Mean annual rainfall
varies from 958 mm at Feilding to 1267 mm at
Rangiwahia (locations shown in Fig. 1). Annual,
85I.C. Fuller / Geomorphology 98 (2008) 84­95
86 I.C. Fuller / Geomorphology 98 (2008) 84­95
seasonal and monthly rainfalls throughout the catchment
are subject to variability of up to 20%, with slightly
more rainfall occurring in the winter­spring than summer­autumn
(Anon, 1980).
1.3. The February 2004 storm and flood
Flooding was caused by a storm on 15­16 February
2004, which was one in a sequence of depressions to
affect the North Island. Heavy rain also fell on 1­3, 4­5
and 10­12 February, saturating soils in the region (Parfitt,
personal communication). On 15 February a cold-pool
low became stationary just to the east of the North Island
and intensified. Persistent heavy rain fell over most of the
lower North Island, with rainfall in the Ruahine Ranges
exceeding 200 mm in 24 h (Meteorological Society,
2004). The resulting large area flood was associated with a
long duration rainfall event. Rainfall intensities did not
generally exceed 10 mm h-1
(Fuller and Heerdegen,
2005), but much of the upper catchment had more than
20 h of rainfall at fairly constant intensities (Fuller and
Heerdegen, 2005). Gauges at lower elevations recorded
lesser totals and lower intensities (e.g. a rain gauge near
Feilding at Halcombe Road (100 m) adjacent to the lower
Kiwitea and Oroua recorded 115 mm over 19 h at an
average intensity of 6.1 mm h-1
). The recurrence interval
for the quantities of rainfall recorded in this region over a
24-h period is N150 years (Fuller and Heerdegen, 2005).
Continuous rainfall records at Feilding extend from 1890
(Anon, 1980). Event magnitude in terms of discharge is
more than twice any previous recorded flood event in the
Kiwitea and more than five times the magnitude of the
mean annual flood in this catchment (Fig. 2, Table 1) and
in the adjacent Oroua and Pohangina catchments (Fuller
and Heerdegen, 2005).
2. Methodology
2.1. Impact assessment
To assess the geomorphic impacts of the floods on
the Kiwitea channel and floodplain, aerial photographs
of a 30-km-long reach of the Kiwitea (cf. Fig. 1) were
acquired in February 2004 in the immediate aftermath of
the flooding. These were orthorectified and georeferenced
before being overlaid on February 1999 orthophotos
(2.5 m resolution) using ArcMapTM GIS. The
positional accuracy of these orthophotos is given as 
12.5 m (LINZ, 2005), with the 2.5-m photograph
resolution setting the limit of features discernible. 2004
photography was taken 1 week after the flood, when
river levels had subsided, but not returned to pre-flood
baseflow conditions. Channel changes were identified
from the photography and verified using field visits at
selected sites. On-screen digitizing generated a series of
metric polygons for the entire 30-km reach, identifying
wetted channel, bars, active channel (wetted channel
and bars combined) and areas of inundation (proximal
Fig. 2. Flood hydrographs for the Kiwitea at Spur Road, courtesy Marianne Watson, Horizons Regional Council.
Fig. 1. (a) Location of river catchments referred to in this study: Kiwitea, Oroua and Pohangina, shown in the context of (b) the Manawatu catchment
and (c) the North Island of New Zealand. A sketch section (d) shows the underlying structure and rock type (after Heerdegen and Shepherd, 1992).
The 30-km reach of the lower Kiwitea river studied is highlighted in (a). Letters refer to the river gauging site mentioned in the text: S: Spur Road; and
rainfall recording stations: F: Feilding, R: Rangiwahia.
87I.C. Fuller / Geomorphology 98 (2008) 84­95
and distal to the channel) and floodplain (bank) erosion.
Proximal inundation is defined as that adjacent to the
channel, identifiable by thick drapes of sediment over
the floodplain. Distal inundation is that further away
from the channel, identified by discoloration of paddocks
from water or fine sediment.
During the 1999­2004 period, several floods occurred
(Fig. 2), thus geomorphic impacts identified may
not wholly be attributed to the 15­16 February event,
given that flows during the 1999­2004 period would
have exceeded thresholds for bed sediment transport
(Clausen and Plew, 2004). However, such changes
would be limited to within the active channel.
2.2. Maximum stream power
Peak stream power was calculated using the
narrowest flood channel width in the lower Kiwitea
(55 m) where channel morphology remained stable.
This method used Baker and Costa's (1987) equation:
x  gQS=w 1
where x is stream power per unit width;  is specific
weight of water (9800 N m-3
for clear water); Q is
discharge (m3
s-1
); S is energy slope; and w is water
surface width.
Table 1
Flood flows in the Kiwitea, Pohangina and Oroua contextualized (based on Fuller and Heerdegen, 2005)
Gauging site
[area km2
]
16 February 2004
flood (95% CI)
Average recurrence
interval (years)
Previous maximum
flood (m3
s-1
)
Date of
previous
maximum
Mean annual
flood (Q2.33)
Ratio
100 years:2.33 years
flood
Years of
record
Spur Road
[224]
358 m3
s-1
(98) 100 166 m3
s- 1
02.09.1988 72 m3
s-1
5.03 29
GEV distribution (two-parameter); based on EV2 and including the 15­16 February flood event. Details of these distributions are available in
Fuller and Heerdegen (2005).
Fig. 3. Sub-reach of the Kiwitea used to derive estimation of sediment erosion budgets. Each bank erosion unit is numbered sequentially from upstream
to downstream. Volumes of erosion are given in Table 3, derived using Eq. (1). The location used to measure maximum stream power is labelled: .
88 I.C. Fuller / Geomorphology 98 (2008) 84­95
Values of discharge were taken from the stream
gauge at Spur Road (cf. Fig. 1), no significant tributaries
enter the river between the reach used to calculate
stream power (Fig. 3) and this gauging station; thus,
despite the spatial offset (6 km), the values of
discharge are considered to be far more reliable than
those which could be reconstructed from channel crosssections.
The stream gradient at this site (0.005) is taken
as a surrogate for energy slope. The maximum value of
stream power is given in Table 2. Energy expended per
unit area was derived by averaging the maximum stream
power expended during the course of the 32-h duration
flood and multiplying this by the total number of
seconds to provide a value of energy expenditure per
unit area (J) (cf. Costa and O'Connor, 1995). The average
maximum stream power assumes a constant channel
width of 55 m during the flood, which may not be
unreasonable given the narrowness and stability of the
channel at this location (cf. Fig. 3), although this may
lead to an underestimation of power towards the start
and end of the event; thus, the estimation of total energy
expended is conservative. As gradient is also assumed
to be constant, in this calculation stream power is a
function of changing discharge as gauged during the
course of the event. Gauge records provide a measure of
discharge every half hour.
2.3. Sediment erosion
Morphological budgeting provides a first approximation
of the volume of material eroded during the
flood in a sub-reach of the lower Kiwitea (Fig. 3) and
based on a profile × planform approach (Brewer and
Passmore, 2002; Fuller et al., 2002, 2003b).
EV 
dAxs
L
 A
 
2
where EV=erosional volume; Axs =change in crosssectional
area (pre- to post-flood); L=length of erosion
unit along cross-section; and A=planform area of
erosion unit: this area is derived from digitized polygons
mapping discrete areas of bank erosion using the aerial
photography described above.
Estimation of material eroded (only) was based on a
reconstruction of pre-flood topography along surveyed
cross-sections (cf. Fig. 4). The position of the pre-flood
banks along each cross-section is derived from the 1999
LINZ orthophoto. The height of the pre-flood bank was
extrapolated visually from the surveyed cross-section to
the edge of the 1999 channel (cf. Figs. 4 and 5). Crosssections
were surveyed using a Topcon GT501 electronic
total station in coarse mode (precision 5 mm),
having been fixed in position using RTK-GPS (Trimble
R8 receiver and rover) with a horizontal precision of
25 mm and vertical precision of 50 mm. The precision of
the ground survey greatly exceeded the precision of the
orthophoto (2.5 m resolution). As such, volumes derived
(Table 3) represent a first approximation of volumetric
loss rather than exact values.
3. Results and discussion
3.1. Channel response
The geomorphic impact of large (sensu; Kochel,
1988) floods is variable. Sometimes impacts are major,
while at other times only minor changes may occur (e.g.
Magilligan, 1992; Costa and O'Connor, 1995). The
changes observed along the lower 30 km of the Kiwitea
are summarized in Table 2, and the impact of this
100 year event was categorized as severe (Miller, 1990)
to catastrophic (Magilligan, 1992). The dimensions of
the wetted channel enlarged by 171% and in some
reaches active channel width increased by over 500%,
providing accommodation space for substantial deposition
of sand and gravel within the greatly widened active
channel. Bar area thus increased by 600% in the lower
30 km. This represents substantial modification of the
channel and adjacent floodplain. Given that true baseflow
conditions did not prevail during the 2004 photo
acquisition, some of the wetted channel increase could
be attributed to higher discharge, but in turn this means
the 600% increase in bar area is a conservative estimate.
Table 2
Summary of channel changes observed along 30-km valley lengths of the Kiwitea, Oroua and Pohangina (after Fuller and Heerdegen, 2005)
Peak stream
power a
(W m-2
)
Total energy
expendedb
(J×103
)
Bar
1999
(km2
)
Bar
2004
(km2
)
Bar area
increase
(%)
Wetted
channel 1999
(km2
)
Wetted
channel 2004
(km2
)
Wetted
channel
increase (%)
Bank
erosion
(km2
)
Total
inundation
(km2
)
Overbank
sediment
drape (km2
)
319 14,928 0.2 1.4 600 0.24 0.65 171 1.1 4.4 2.8
a
Minimum flood channel width used: 55 m.
b
Assumes maximum stream power.
89I.C. Fuller / Geomorphology 98 (2008) 84­95
Fig. 4. Channel changes on the Kiwitea c. 11 km north of Feilding at NZMST23 349 132. Selected channel cross-sections, with reconstructed preflood
profiles, are shown to illustrate the scale of change. Minimal overbank inundation indicates much of the flood flow remained contained within
the stream channel. Severe erosion occurred on the outside of bends where stream powers were highest (cf. Miller, 1995).
90 I.C. Fuller / Geomorphology 98 (2008) 84­95
Channel widening in response to moderate and large
floods is common. For example, Sloan et al. (2001) cite
a maximum 80% increase in width along the Eel River
in California, and Nolan and Marron (1985) described
width increases of between 7% and 105% elsewhere in
California. Similarly, significant channel change occurred
in the upper Hunter valley in New South Wales,
Australia, in response to a series of large floods acting as
effective geomorphic agents within a catchment cleared
of native vegetation, including the erosion of 250 ha of
floodplain, dramatic increase in channel width and
reduction in sinuosity (Erskine and Bell, 1982; Erskine
and Warner, 1988). However, the scale of the widening
quantified in reaches in the Kiwitea is extraordinary.
Fig. 4 illustrates the dramatic widening of the Kiwitea
channel that typically occurred in discrete reaches along
the lower 30 km of the valley. This is contrasted with the
less dramatic (relatively) channel changes observed a
short distance (3 km) downstream (Fig. 5), where the
impact of the flood on the channel has been mitigated by
substantial overbank flow, draping thick overbank fines
across a broad area of floodplain. Lower banks in this
reach permitted overbank flow at lower discharges,
which limited the energy available for erosion in the
Fig. 5. Channel changes on the Kiwitea c. 8 km north of Feilding at NZMST23 330 106. Dissipation of flood flows overbank causing substantial
inundation reduced stream powers and lessened geomorphic impacts of the flooding at this site.
91I.C. Fuller / Geomorphology 98 (2008) 84­95
active channel. While erosion still occurred in the channel,
its effects are less dramatic than the site upstream
(Fig. 4). Stream powers in this lower reach, given a flood
flow width of 250 m, were 70 W m-2
; while in the upper
reach (Fig. 4) widths of 55 m produced stream powers in
excess of 300 W m- 2
. This is significant, as Magilligan
(1992) proposed a threshold of stream power for
catastrophic change at 300 W m-2
, a figure which
subsequently was supported by Lapointe et al. (1998).
This pattern of alternating degrees of channel change is
repeated along the entire 30 km valley reach.
Over the course of the lower 30 km of the Kiwitea,
impacts of the flood are thus considerable. Erosion of
adjacent floodplain totalled 1.1 km2
(Table 2). In terms
of volume of sediment eroded, a 3-km sub-reach of the
lower Kiwitea (Fig. 3) generated 266,833 m3
(Table 3).
This sub-reach is representative of the extents of erosion
observed in the remaining 27 km of the river studied, in
that both low terraces and active floodplain have been
eroded. The mean multiplier of change in cross-section
area (pre- to post-flood) divided by length of erosion unit
for this 3-km reach (1.23) (in effect a surrogate for mean
bank height in the reach) thus gives a tentative means of
estimating sediment loss from the entire 30-km reach.
Multiplication of this multiplier by the total area of valley
floor eroded gives a volume of 1,384,083 m3
.
However, application of this mean multiplier to the
total area of erosion in the 3-km sub-reach yields a figure
of 215,227 m3
, suggesting an underestimation of total
volume of around 25%.
3.2. Conditioning factors
Channel morphology exerts a substantial influence on
flood power at any given point in a river (Graf, 1983), and
spatially varying boundary conditions affect the location
and extent of flood impacts (Miller, 1995). Thus,
geomorphic impacts in humid systems are normally
maximized in steep, narrow channels, compared with
broad, low-gradient valleys (Miller, 1995), as here stream
power is maximized (Meyer, 2001). The 30-km reach
assessed in this article cannot, however, be defined as a
steep, narrow, mountain valley. Nevertheless, terrace
bluffs confine reaches in the Kiwitea and generate higher
flood powers than less confined reaches in the same river,
or elsewhere where the floodplain is broader (cf.
Magilligan, 1992; Fuller and Heerdegen, 2005). Furthermore,
the bed of the Kiwitea had degraded to the extent
that the 10-year flood would not overtop its banks (Anon,
1980). Similar affects of constriction (enhancing flood
impacts) were observed by Butler and Malanson (1993).
Nanson (1986) also identifies such narrowing of channels
as being critical in concentrating erosional energy to the
extent that high flows greatly exceed erosional thresholds,
causing a catastrophic channel and floodplain stripping.
The combination of high stream powers in some overnarrow
channels has produced severe and in places
catastrophic (sensu; Magilligan, 1992) erosion and
channel transformation in the Kiwitea.
However, channel confinement in a narrow slot is not
necessarily required to maximize stream power. Indeed,
Fuller (in press) indicates that the valley floor width
index (cf. Grant and Swanson, 1995) is not a good
predictor of the spatial variability of erosion in the
Kiwitea. Using flood flow modelling, Miller (1995)
suggests that the shear stresses on the floodplain along
the outside of a bend are comparable with the shear
stresses generated in narrow, canyon-like reaches. Thus,
local channel and floodplain configuration can maximize
the available stream power. In reaches of the
Kiwitea where the flood flow was confined between
terrace bluffs at the outside of bends, the full force of the
floodwaters was contained, and stream powers probably
exceeded 300 W m-2
, considered by Magilligan (1992)
and Miller (1990) as the threshold for catastrophic
channel change. It is therefore not surprising that
such reaches have undergone "major morphologic
adjustments: major erosion, deposition or channel realignment"
(Magilligan, 1992, p. 384). However, Miller
(1990) acknowledges that stream power is in fact a poor
predictor of the extent of erosion due to the array of
factors that influence a reach's susceptibility to change
(e.g. specific valley floor configuration, slope, planform,
Table 3
Erosion volumes calculated from a 3-km sub-reach of the Kiwitea
defined in Fig. 3
Erosion
unit
Area
(m2
)
Change in crosssectional
area (m2
)
Length of
unit (m)
Volume
(m3
)
1 12881 165 93 22853.4
2 3971 62 35 7034.3
3 15674 149.8 104.36 22498.7
4 16734 72.78 24 50745.9
5 5198 46 40 5977.7
6 5235 46 40 6020.3
7 3451 30 22 4705.9
8 2431 24 24 2431.0
9 5155 58.5 39 40645.5
10 27097 24 24 27724.0
11 27724 46.4 57.2 22489.4
12 21072 79.2 44 37929.6
13 13460 64 77 11187.5
14 7269 3.4 12 2059.6
15 5089 3.4 12 1441.9
16 1938 1 1.78 1088.8
Total 266833.4
Refer to Fig. 3.
92 I.C. Fuller / Geomorphology 98 (2008) 84­95
geometry, roughness, local flow obstructions), and
Meyer (2001) suggested that major channel alteration
occurred in an alluvial reach at stream powers in the
range 50­200 W m- 2
. Nevertheless, in reaches where
flow overtopped lower banks (cf. Fig. 5), unit stream
power was limited and did not exceed, or approach, the
threshold for major morphological adjustment. This
limited the geomorphic impact of the flood on these
reaches. Thus, the bend shown in Fig. 5, which has been
largely untouched by the flood in the Kiwitea, remains
intact because lower banks upstream permitted substantial
extra-channel flow, in effect by-passing this reach of
channel.
3.3. Comparison with adjacent catchments
Similar amounts of rainfall fell on the adjacent Oroua
and Pohangina catchments (Fuller and Heerdegen,
2005), but channel responses differ substantially (see
Fuller, in press). Floods that were generated in these
catchments had ARIs of 116 and 38 years respectively.
The Pohangina flood could not therefore be classified
as "large" (Kochel, 1988), but in terms of magnitude
frequency, the Oroua flood was bigger than that in the
Kiwitea. Despite the equivalence of flood magnitude and
duration, changes in the Oroua were confined to
relatively minor and localized channel widening. In
this system, the river is not confined in a narrow, overdeepened
channel but is relatively wide and shallow,
with a 35-m-wide wetted channel and a 65-m-wide
channel fairway (Horizons, 2002), which dissipates
flood flow. Extensive inundation also occurred in the
Oroua, which further limited the extent of erosion in this
river, with just over half of the total bank erosion area
measured in the smaller Kiwitea (Fuller and Heerdegen,
2005). Stream powers in the Oroua were lower, at a
maximum of 106 W m-2
(Fuller and Heerdegen, 2005),
which falls some way below Magilligan's (1992) threshold
for catastrophic change. In the Pohangina, the
impacts of the smaller magnitude flood were largely
limited to avulsion within the active channel belt and
reworking of vegetated bars within the channel fairway,
typical of processes observed in similar wandering rivers
(e.g. Ferguson and Werritty, 1983; Fuller et al., 2002,
2003a, 2005). The area of bank erosion was approximately
half that which occurred in the Oroua.
The magnitude of the February 2004 flood event
relative to the mean annual flood was far higher in the
Kiwitea than adjacent catchments (Fuller and Heerdegen,
2005). The Kiwitea channel was therefore comparatively
poorly adjusted to accommodate the flood discharge of
15­16 February. The pre-February 2004 channel was on
average just 10 to 15 m wide (Philpott, 2005), which
compares with the 23-m-wide meandering channel of
1877 (Anon, 1980). The Kiwitea channel had over time
adjusted to the smaller mean annual flood, which was
several orders of magnitude smaller than the 100-year
event of 2004. Richards (1999 p.15) suggests that "the
geomorphological significance of an event depends on
the circumstances encountered by that event--because
the effect of the event is conditioned by pre-existing
morphology". The impact of the February flood event in
the Kiwitea has therefore been enhanced by the pre-flood
(narrowed) channel morphology. The same magnitude of
event, had it occurred in the now much widened channel
would not have had the same impacts. Gardner (1977)
attributes minimal impacts of a 1 in 500 years flood event
in Ontario to the valley floors being "well-adjusted to
handling infrequent, high magnitude flows, particularly
in the absence of man-made modifications and structures
on the floodplain." (Gardner, 1977, p. 2300).
4. Conclusions
The Kiwitea Stream eroded 1.4 million m3
of valley
floor as the narrow and over-deepened channel responded
to the largest recorded flood, which was five
times bigger than the mean annual flood with an ARI of
100 years. Geomorphic impacts were, however, spatially
discontinuous and highly reach specific. In some
reaches, channel change was catastrophic while in others
minimal changes occurred. The variability was conditioned
by thresholds of flood power, in conjunction with
the local channel configuration and planform geometry.
Localized confinement of flood flows in the Kiwitea
enhanced stream powers and thus reinforced the propensity
towards major morphological adjustment. By
contrast, in adjacent reaches and catchments, flood flows
dissipated across the floodplain, limiting the unit stream
power available for significant erosion and lessening the
geomorphic impact of the 2004 flood.
Acknowledgments
I would like to thank Erin Hutchinson for provision of
cross-section data. Horizons Regional Council staff
(Palmerston North), especially Marianne Watson and
Joseph McGehan, are thanked for contributing hydrological
data and towards Fig. 1. Richard Heerdegenis thanked
for facilitating communication with Horizons and hydrological
analyses. David Livingston (Research Volunteer,
Geography Programme, Massey University, spring 2004)
is thanked for processing aerial photograph overlays,
made available and possible through the Centre for
93I.C. Fuller / Geomorphology 98 (2008) 84­95
Precision Agriculture, Massey University, whose cooperation
is also gratefully acknowledged. Lawrie Cairns is
thanked for having the foresight to fly the aerial
photography in the immediate aftermath of the flooding
in February 2004, making this assessment of flood
impacts possible. Helpful reviews of this manuscript
were received by Vishwas Kale, Basil Gomez and Varyl
Thorndycraft.
References
Anon., 1980. The Kiwitea Scheme. Manawatu Catchment Board and
Regional Water Board unpublished report.
Baker, V.R., 1977. Stream channel response to floods with examples from
central Texas.GeologicalSocietyof America Bulletin 88, 1057­1071.
Baker, V.R., Costa, J.E., 1987. Flood power. In: Mayer, L., Nash, D.
(Eds.), Catastrophic Flooding. Allen & Unwin, Boston, pp. 1­21.
Brewer, P.A., Passmore, D.G., 2002. Sediment budgeting techniques in
gravel bed rivers. In: Jones, S., Frostick, L.E. (Eds.), Sediment Flux to
Basins: Causes, Controls and Consequences, vol. 191. Geological
Society Special Publication, pp. 115­131.
Brooks, A.P., Brierley, G.J., 1997. Geomorphic responses of lower Bega
River to catchment disturbance, 1851­1926. Geomorphology 18,
291­304.
Butler, D.R., Malanson, G.P., 1993. An unusual early-winter flood and its
varying geomorphic impact along a sub-alpine river in the RockyMountains
of Montana, USA. Zeitschrift für Geomorphologie 37,
145­155.
Clausen, B., Plew, D., 2004. How high are bed-moving flows in New
Zealand rivers? Journal of Hydrology (New Zealand) 43, 19­37.
Costa, J.E., 1974. Response and recovery of a piedmont watershed
from tropical storm Agnes, June 1972. Water Resources Research
10, 106­112.
Costa, J.E., O'Connor, J.E., 1995. Geomorphically effective floods. In:
Costa, J.E., Miller, A.J., Potter, K.W., Wilcock, P.R. (Eds.), Natural
and Anthropogenic Influences in Fluvial Geomorphology. Geophysical
Monograph, vol. 89. American Geophysical Union, pp. 45­56.
Erskine, W.D., 1986. River metamorphosis and environmental change
in the Macdonald Valley, New South Wales, since 1949. Australian
Geographical Studies 24, 88­107.
Erskine, W.D., 1994. Sand slugs generated by catastrophic floods on the
Goulburn River, New South Wales. In: Olive, L.J., Loughran, R.J.,
Kesby, J.A. (Eds.), Variability in Stream Erosion and Sediment
Transport. IAHS Publication, vol. 224, pp. 143­151.
Erskine, W., Bell, F.C., 1982. Rainfall, floods and river channel changes
in the upper Hunter. Australian Geographical Studies 20, 183­196.
Erskine, W.D., Warner, R.F., 1988. Geomorphic effects of alternating
flood-and drought-dominated regimes on NSW coastal rivers. In:
Warner, R.F. (Ed.), Fluvial Geomorphology of Australia. Academic
Press, Sydney, pp. 223­244.
Ferguson, R.I., Werritty, A., 1983. Bar development and channel changes
in the gravely River Feshie. In: Collinson, J.D., Lewin, J. (Eds.),
Modern and Ancient Fluvial Systems. International Association of
Sedimentologists Special Publication, vol. 6, pp. 133­143.
Fuller, I.C., in press. Geomorphic work during a "150-year" storm:
contrasting behaviour of river channels in a New Zealand catchment.
Annals Association of American Geographers.
Fuller, I.C., Heerdegen, R.G., 2005. The February 2004 floods in the
Manawatu, New Zealand: hydrological significance and impact on
channel morphology.Journal of Hydrology (New Zealand) 44, 75­90.
Fuller, I.C., Passmore, D.G., Heritage, G.L., Large, A.R.G., Milan, D.J.,
Brewer,P.A., 2002. Annual sediment budgets in an unstable gravelbed
river: the River Coquet, northern England. In: Jones, S., Frostick, L.E.
(Eds.), Sediment Flux to Basins: Causes, Controls and Consequences.
Geological Society Special Publication, vol. 191, pp. 115­131.
Fuller, I.C., Large, A.R.G., Milan, D.J., 2003a. Quantifying channel
development and sediment transfer following chute cutoff in a
wandering gravel-bed river. Geomorphology 54, 307­323.
Fuller, I.C., Large, A.R.G., Charlton, M.E., Heritage, G.L., Milan, D.J.,
2003b. Reach-scale sediment transfers: an evaluation of two
morphological budgeting approaches. Earth Surface Processes and
Landforms 28, 889­903.
Fuller, I.C., Large, A.R.G., Heritage, G.L., Milan, D.J., Charlton, M.E.,
2005. Derivation of reach-scale sediment transfers in the River
Coquet, Northumberland, UK. In: Blum, M., Marriott, S., Leclair, S.
(Eds.), Fluvial Sedimentology VII. International Association of
Sedimentologists Special Publication, vol. 35, pp. 61­74.
Gardner, J.S., 1977. Some geomorphic effects of a catastrophic flood
on the Grand River, Ontario. Canadian Journal of Earth Sciences
14, 2294­2300.
Graf, W.L., 1983. Downstream changes in stream power in the Henry
Mountains, Utah. Annals Association American Geographers 73,
373­387.
Grant, G.E., Swanson, F.J., 1995. Morphology and processes of valley
floors in Mountain Streams, Western Cascades, Oregon. In: Costa,
J.E., Miller, A.J., Potter, K.W., Wilcock, P.R. (Eds.), Natural and
Anthropogenic Influences in Fluvial Geomorphology. Geophysical
Monograph, vol. 89. American Geophysical Union, pp. 83­102.
Gupta, A., 1983. High-magnitude floods and stream channel response.
In: Collinson, J.D., Lewin, J. (Eds.), Modern and Ancient Fluvial
Systems. International Association of Sedimentologists Special
Publication, vol. 6, pp. 219­227.
Hack, J.T., Goodlett, J.C., 1960. Geomorphology and forest ecology of
a mountain stream in the Central Appalachians. US Geological
Survey Professional Paper 347.
Harvey, A.M., Hitchcock, D.H., Hughes, D.J., 1979. Event frequency
and morphological adjustment of fluvial systems in upland Britain.
In: Rhodes, D.D., Williams, G.P. (Eds.), Adjustments of the Fluvial
System. Kendall Hunt, Dubuque, Iowa, pp. 139­167.
Heerdegen, R.G., Shepherd, M.J., 1992. Manawatu Landforms--
Product of tectonism, climate change and process, In: Soons, J.M.,
Selby, M.J. (Eds.), Landforms of New Zealand, 2nd edition.
Longman, Auckland, pp. 308­333.
Hickin, E.J., 1983. River channel changes: retrospect and prospect. In:
Collinson, J.D., Lewin, J. (Eds.), Modern and Ancient Fluvial
Systems. International Association of Sedimentologists Special
Publication, vol. 6, pp. 61­83.
Hicks, D.M., Shankar, U., 2003. Sediment from New Zealand Rivers.
NIWA Chart Miscellaneous Series, vol. 79.
Horizons, 2002. Pohangina-Oroua Catchment Control Scheme
Review. Report No. 2002/EXT/523.
Kochel, R.C., 1988. Geomorphic impacts of large floods: review and
new perspectives on magnitude and frequency. In: Baker, V.R.,
Kochel, R.C., Patton, P.C. (Eds.), Flood Geomorphology. John
Wiley & Sons, Chichester, pp. 169­187.
Lapointe, M.F., Secretan, Y., Driscoll, S.N., Bergeron, N., Leclerc, M.,
1998. Response of the Ha! Ha! River to the flood of July 1996 in
the Saguenay Region of Quebec: large scale avulsion in a glaciated
valley. Water Resources Research 34, 2383­2392.
Laronne, J.B., Duncan, M.J., 1992. Bedload transport paths and gravelbar
formation. In: Billi, P., Hey, R.D., Thorne, C.R., Tacconi, P. (Eds.),
Dynamics of Gravel-Bed Rivers. Wiley, Chichester, pp. 177­202.
94 I.C. Fuller / Geomorphology 98 (2008) 84­95
Lewin, J., 1989. Floods in fluvial geomorphology. In: Beven, K.J.,
Carling, P.A. (Eds.), Floods: Hydrological, Sedimentological and
Geomorphological Implications. Wiley, Chichester, pp. 265­284.
LINZ., 2005. http://www.linz.govt.nz/rcs/linz/pub/web/root/core/
Topography/aerialandorthophotos/index.jsp. Accessed 26 September
2005.
Lisle, T.E., 1981. The recovery of aggraded stream channels at gauging
stations in northern California and southern Oregon. Proceedings,
Symposium on Erosion and Sediment Transport in Pacific Rim
Steeplands, Christchurch, New Zealand, 1981. International
Association of Hydrological Sciences IAHS-AISH Publication,
132, pp. 189­211.
Magilligan, F.J., 1992. Thresholds and the spatial variability of flood
power during extreme floods. Geomorphology 5, 373­390.
Magilligan, F.J., Phillips, J.D., James, L.A., Gomez, B., 1998.
Geomorphic and sedimentological controls on the effectiveness
of an extreme flood. Journal of Geology 106, 87­95.
Meteorological Society, 2004. A very stormy period, with severe
flooding and gales over southern North Island 14­17 February
2004. Newsletter, March 2004, 96, Meteorological Society of New
Zealand, pp. 21­23.
Meyer, G.A., 2001. Recent large-magnitude floods and their impact on
valley floor environments of northeastern Yellowstone. Geomorphology
40, 271­290.
Miller, A.J., 1990. Flood hydrology and geomorphic effectiveness in
the central Appalachians. Earth Surface Processes and Landforms
15, 119­134.
Miller, A.J., 1995. Valley morphology and boundary conditions
influencing spatial patterns of flood flow. In: Costa, J.E., Miller,
A.J., Potter, K.W., Wilcock, P.R. (Eds.), Natural and Anthropogenic
Influences in Fluvial Geomorphology. Geophysical Monograph,
vol. 89. American Geophysical Union, pp. 57­81.
Nanson, G.C., 1986. Episodes of vertical accretion and catastrophic
stripping: a model of disequilibrium flood-plain development.
Bulletin of the Geological Society of America 97, 1467­1475.
Neill,C.R., 1973. Hydraulicandmorphologic characteristics of Athabasca
River near Forth Assiniboine--the anatomy of a wandering gravel
bed river. Report REH/73/8. Alberta Research Council, Highways and
River Engineering Division, Edmonton.
Nolan, K.M., Marron, D.C., 1985. Contrast in stream-channel response
to major storms in two mountainous areas of California. Geology
13, 135­138.
Philpott, J., 2005. The Lower Kiwitea Scheme--Draft Review.
Horizons Regional Council Report No. 2005/EXT/632.
Pickup, G., Warner, R.F., 1976. Effects of hydrologic regime on magnitude
and frequency of dominant discharge. Journal of Hydrology
29, 51­75.
Pitlick, J., 1993. Response and recovery of a subalpine stream following
a catastrophic flood. Bulletin of the Geological Society of
America 105, 657­670.
Reid, I., Frostick, L., 1994. Fluvial sediment transport and deposition.
In: Pye, K. (Ed.), Sediment Transport and Depositional Processes.
Blackwell, Oxford, pp. 89­155.
Richards, K., 1999. The magnitude-frequency concept in fluvial geomorphology:
a component of a degenerating research programme?
Zeitschrift fur Geomorphologie. Supplementband 115, 1­18.
Schumm, S.A., Lichty, R.W., 1963. Channel widening and floodplain
construction along Cimarron River in southwestern Kansas. US
Geological Survey Professional Paper 352-D, 71­88.
Sloan, J., Miller, J.R., Lancaster, N., 2001. Response and recovery of
the Eel River, California, and its tributaries to floods in 1955, 1964
and 1997. Geomorphology 36, 129­154.
Whitehouse, I.E., Pearce, A.J., 1992. Shaping the Mountains of New
Zealand, In: Soons, J.M., Selby, M.J. (Eds.), Landforms of New
Zealand, 2nd edition. Longman, Auckland, pp. 144­160.
Wolman, M.G., Gerson, R., 1978. Relative scales of time and effectiveness
of climate in watershed geomorphology. Earth Surface
Processes 3, 189­208.
Wolman, M.G., Miller, J.P., 1960. Magnitude and frequency of forces
in geomorphic processes. Journal of Geology 68, 54­74.
95I.C. Fuller / Geomorphology 98 (2008) 84­95