Factors influencing the recession rate of Niagara Falls since the 19th century
Yuichi S. Hayakawa , Yukinori Matsukura
Geoenvironmental Sciences, Life and Environmental Sciences, University of Tsukuba, 1-1-1 Ten-nodai, Tsukuba, Ibaraki 305-8572, Japan
a b s t r a c ta r t i c l e i n f o
Article history:
Received 3 February 2009
Received in revised form 10 April 2009
Accepted 14 April 2009
Available online 17 April 2009
Keywords:
Waterfall recession
Knickpoint
Bedrock river
Erosion rate
The rate of recession of Niagara Falls (Horseshoe and American Falls) in northeastern North America has
been documented since the 19th century; it shows a decreasing trend from ca. 1 m y-1
a century ago to
ca. 0.1 m y- 1
at present. Reduction of the flow volume in the Niagara River due to diversion into bypassing
hydroelectric schemes has often been taken to be the factor responsible, but other factors such as changes in
the waterfall shape could play a role and call for a quantitative study. Here, we examine the effect of physical
factors on the historically varying recession rates of Niagara Falls, using an empirical equation which has
previously been proposed based on a non-dimensional multiparametric model which incorporates flow
volume, waterfall shape and bedrock strength. The changes in recession rates of Niagara Falls in the last
century are successfully modeled by this empirical equation; these changes are caused by variations in flow
volume and lip length. This result supports the validity of the empirical equation for waterfalls in rivers
carrying little transported sediment. Our analysis also suggests that the decrease in the recession rate of
Horseshoe Falls is related to both artificial reduction in river discharge and natural increase in waterfall lip
length, whereas that of American Falls is solely due to the reduction in flow volume.
 2009 Elsevier B.V. All rights reserved.
1. Introduction
Waterfalls, a typical form of knickpoint or knickzone, are commonly
observed in bedrock rivers across the world, including active
orogens, volcanic areas and glacially- or tectonically-deformed areas
(Gilbert, 1907; von Engeln, 1940; Derricourt, 1976; Bishop et al., 2005;
Crosby and Whipple, 2005; Hayakawa and Oguchi, 2006, in press;
Lamb et al., 2007). Headward erosion of waterfalls, often referred to as
recession, is one of the central themes in fluvial geomorphology
because the rates of waterfall recession are usually much greater than
the other erosional processes involving bedrock rivers, such as
incision of the entire riverbed (Begin et al., 1980; Kukal, 1990; Wohl,
1998, 2000; Hayakawa and Matsukura, 2002). Bedrock erosion at
waterfalls can be significant even if rivers lack transported sediment to
act as a tool of abrasion, probably due to the frequent occurrence
of plucking and/or cavitation by rapid stream flows including jet
flows (Barnes,1956; Young,1985; Bishop and Goldrick,1992; Whipple
et al., 2000; Pasternack et al., 2006). Niagara Falls is a famous waterfall
in northeastern North America that has long been studied by
geologists and geomorphologists (Gilbert, 1907; Spencer, 1907;
Tinkler, 1987). Historical changes of the waterfall shape are well
documented, especially in the last 200 years, and the actual recession
rates in that period are well known. The long-term average recession
rate of Niagara Falls is estimated to have been ca. 1 m y-1
during the
Holocene (e.g., Gilbert, 1907; Ford, 1968), but it has decreased to the
order of 0.1­0.01 m y-1
in recent decades (Tinkler,1993). The slowing
of the waterfall recession rate is commonly attributed to the reduction
in water flow over the waterfall as a result of the construction of
several large power plants in the river (e.g., Tinkler,1993,1994). These
plants divert water well above the falls and return it to the Niagara
River well below the falls. The shape of the falls in plan has also
changed during the recession process, and the curved shape of the
waterfall crest (the `horseshoe' shape) is supposed to be more stable
in its stress distribution and thereby decreases the recession rate
(Philbrick, 1970). The greater length of the curved lip may also cause
the stream to be shallower with less tractive force at the lip, also
reducing the recession rate, although the effects of lip length on the
recession of Niagara Falls are yet to be quantified. Here we quantitatively
examine the recession rate of Niagara Falls based on a
previously proposed model, which estimates the waterfall recession
rate from relevant physical parameters including the stream discharge
(total flow) and waterfall shape.
2. Existing record of recession rates of Niagara Falls
Niagara Falls is located in the middle portion of the Niagara River,
draining northward from Lake Erie to Lake Ontario (Fig. 1). The water
of the Niagara River is supplied from the upstream Great Lakes, of
glacial origin, and the stream discharge is controlled mainly by the
water level in the lakes, not by immediate precipitation. The climate in
the area is relatively cold, with active frost weathering of bare rocks
exposed along the river from winter to spring.
The waterfall originally formed at the Niagara Escarpment running
west to east between the lakes, when the Laurentide ice sheet had
Geomorphology 110 (2009) 212­216
 Corresponding author. Tel.: +81 29 853 5691; fax: +81 29 853 4460.
E-mail address: hayakawa@geoenv.tsukuba.ac.jp (Y.S. Hayakawa).
0169-555X/$ ­ see front matter  2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.geomorph.2009.04.011
Contents lists available at ScienceDirect
Geomorphology
journal homepage: www.elsevier.com/locate/geomorph
receded and the river started draining over the escarpment approximately
12,500 years ago (Tinkler et al., 1994). Since then the waterfall
has receded for ca. 11 km, leaving a deep narrow gorge downstream,
called the Great Gorge. The long-term mean recession rate of the
waterfall is therefore approximately 1 m y-1
, although the rate was
considerably lower for about 5000 years during 10,500­5500 yBP (ca.
0.1 m y-1
at minimum), due to the abrupt decrease of flow caused by
water level lowering and changes in stream courses in the upstream
lakes (Lewis and Anderson, 1989; Tinkler et al., 1994). The mean
recession rate prior to that interval was similar to that after 5500 yBP
(ca. 1.6 m y-1
).
Niagara Falls at present consists of several falls, due to diversion of
the stream slightly upstream of the site. The largest fall in the western
Canadian side, named Canadian or Horseshoe Falls and henceforth
referred to as the Horseshoe, has a horizontal curved 762-m-long rim,
slightly overhanging face, and a single uninterrupted fall of water with
a 51-m height. Another fall at the eastern side, named American Falls,
has a relatively straight rim with a length of 335 m. The face of
American Falls is nearly vertical but there is no overhang and the
lower portion is filled with fallen rock blocks, like talus deposit, for
nearly half of its total height of 54 m.
Changes to Niagara Falls during historical times have been
recorded in various documents since the 1600s. The early records
were provided simply as artists' pictures, but since the 19th century
detailed topographic measurements have been conducted by geologists.
Philbrick (1970) gave a detailed summary of the changes in
horizontal morphology of Niagara Falls based on the literature, so that
we can see the dimensions of the waterfalls in the past (Fig. 2). The lip
length of the Horseshoe has increased dramatically as its horizontal
shape became progressively more curved over recent centuries,
whereas that of the American Falls has merely changed. The height
Fig. 1. Study site. Niagara Falls at present comprises two major falls, Horseshoe Falls and American Falls.
Fig. 2. Historical changes of the plan shape of Horseshoe Falls (after Philbrick, 1970).
213Y.S. Hayakawa, Y. Matsukura / Geomorphology 110 (2009) 212­216
of Niagara Falls has scarcely changed since its formation in the
Holocene, as suggested by the constant height of the Great Gorge.
Since the late 19th century, diversion of water to hydroelectric
power plants has caused a severe reduction in water flowing over
Niagara Falls (Tiplin, 1988). The flow is now regularized to prevent
shortages of flowing water for tourists. The changing flow in the river
has been documented in the literature (e.g., International Joint
Commission, 1953). The presence of Goat Island between Horseshoe
and American Falls causes 90% of the river water to pass over the
Horseshoe and 10% over American Falls.
The recession rate of Niagara Falls, the Horseshoe Falls in particular,
has been precisely recorded in the last century (Gilbert, 1907;
Philbrick, 1970; Tinkler, 1987). The largest recession rate of the
Horseshoe was reported by Gilbert (1907) to be 2.0 m y-1
in 1875­
1905, whereas the average rate in this period was about 1.3 m y-1
(International Joint Commission, 1953). The rate has progressively
decreased due to the building of power plants during the past century
(International Joint Commission,1953; Table 1), and the current rate is
supposed to be less than 0.1 m y-1
although no certain record is
available for the recent recession (Tinkler, 1993, 1994).
The rate of recession of American Falls is much less than the
Horseshoe. Gilbert (1907) inferred that the recession rate of American
Falls had been about 0.1 m y-1
for hundreds of years. The modern rate
is assumed to be of the order of 0.01 m y-1
(Tinkler, 1993, 1994).
3. Estimating waterfall recession rates from empirical equations
3.1. An empirical equation for estimating waterfall recession rates
To quantify the effects of flow and lip length on the recession rate
of Niagara Falls, we use a previously proposed model which estimates
the relationship between the waterfall recession rate and relevant
physical parameters, namely erosive force and bedrock strength
(Hayakawa and Matsukura, 2003a). Supposing that the rate of
waterfall recession depends on the erosional force of the stream and
the strength of the resisting bedrock, dimensional analysis finds a
dimensionless index, FR, based on these variables:
FR =
AP
WH
ffiffiffiffiffi

Sc
r
1
where A (L2
) is the upstream drainage area of a waterfall; P (L T-1
) is
the mean annual precipitation in the drainage basin, so that the
product of A and P accounts for the annual stream flow over
the waterfall; W (L) and H (L) are the width (lip length) and height
of the waterfall,  (M L-3
) is the water density, and Sc (M L-1
T-2
) is
the unconfined compressive strength of bedrock. The dimensionless
index FR represents the balance between the erosional force and
bedrock resistance as a whole, where all these parameters are given in
the SI unit.
Examination of the relationship between the FR index and the
waterfall recession rate, E, using data for waterfalls in the Boso
Peninsula of eastern Japan gives the following equation (Hayakawa
and Matsukura, 2003a):
E = 99:7FR
0:73
= 99:7
AP
WH
ffiffiffiffiffi

Sc
r 0:73
2
This equation has been found to give good order-of-magnitude
estimates of waterfall recession rates in many areas (Hayakawa and
Matsukura, 2003b; Hayakawa, 2005; Hayakawa and Wohl, 2005;
Hayakawa et al., 2005, 2008a), with the exception of rivers carrying
abundant transported sediments (Hayakawa et al., 2008b, 2009).
As the proxy of discharge in the FR index, A and P have been
individually measured in previous studies, because direct measurement
data of the discharge is not available for most waterfalls.
Fortunately, the stream discharge has long been measured in the
Niagara River. The flow volume of the Niagara is stable in normal
conditions, because it depends on the water budget in the Great Lakes,
which is too large to be significantly affected by individual storms and
other local and short-term climatic events (Tinkler, 1993). For the
same reason, estimation of the discharge of the Niagara River from
drainage area and precipitation is inappropriate. We therefore use
stream discharge, Q (L3
T-1
), for the FR index, instead of the product
of A and P. Eqs. (1) and (2) become:
FR =
Q
WH
ffiffiffiffiffi

Sc
r
3
and
E = 99:7FR
0:73
= 99:7
Q
WH
ffiffiffiffiffi

Sc
r 0:73
4
3.2. Data acquisition and model applications
The present water discharge over Niagara Falls is controlled
artificially, based on an international agreement between Canada
and the US made in 1953. The regulated flow discharge is 2832 m3
s-1
in summer daytime (0800 to 2200 h from April 1st to September 15th
and 0800 to 2000 h from September 16th to October 31st), and half of
that value (1416 m3
s-1
) at all other times. The annual-mean flow
discharge at present is therefore 1770 m3
s-1
, which is 30% of the
natural flow level of ca. 5760 m3
s-1
until a century ago (Tinkler,
1993). The present annual-mean discharge over the Horseshoe is
1593 m3
s-1
(90% of the Niagara River discharge) and that over
American Falls is 177 m3
s-1
(10%).
Since water was first diverted through hydroelectric power plants
in the late 19th century, the water discharge over the waterfalls was
progressively reduced, until the Canada­USA treaty was agreed in
1953. The discharge was reported to be 4147 m3
s-1
(72% of natural
flow) in 1905­1927, and 3456 m3
s-1
(60%) in 1927­1950 (International
Joint Commission, 1953). Table 2 summarizes the historical
changes of flow volume over Horseshoe and American Falls.
Table 1
Actual recession rates of Horseshoe Falls and American Falls over differing time periods.
Duration Recession rate
E (m y-1
)
Source
Horseshoe Falls 1842­1875 (H1) 1.2­1.3 Gilbert, 1907; International
Joint Commission, 1953
1875­1905 (H2) 1.3­2.0 Gilbert, 1907; International
Joint Commission, 1953
1905­1927 (H3) 0.98 International Joint
Commission, 1953
1927­1950 (H4) 0.67 International Joint
Commission, 1953
1950­2000 (H5) 0.1 Tinkler, 1993, 1994
Modern (H6) 0.1 Tinkler, 1993, 1994
American Falls 500 y (­1905) (A1) 0.098 Gilbert, 1907
Modern (A2) 0.01 Tinkler, 1993, 1994
Table 2
Historical changes in flow discharge over Niagara Falls.
Duration Discharge over
Niagara Falls
(m3
s- 1
)
Discharge over
Horseshoe Falls
(m3
s-1
)
Discharge over
American Falls
(m3
s-1
)
(100%) (90%) (10%)
1842­1905 5760 5184 576
1905­1927 4147 3732 415
1927­1950 3456 3110 346
1950­present 1770 1593 177
214 Y.S. Hayakawa, Y. Matsukura / Geomorphology 110 (2009) 212­216
Lip lengths of the Horseshoe in differing years have been obtained
by measuring the lengths of its past crest lines on the map by Philbrick
(1970), on which the development of the shape of Niagara Falls is well
documented (Fig. 2). Lines based on topographic measurement
surveys have been available since 1678, but only the lines after 1819
are used because the left end of the lines in 1678 and 1764 are unclear.
The lip length of the Horseshoe has increased in recent times from
420 m in 1842 to 762 m in 2000 as the profile becomes more curved
(Table 3).
To estimate the intact rock strength of the waterfalls, Schmidt
hammer (N-type) measurements were made using the repeated
impact method, which involves averaging over 20 impacts at each
point after eliminating outliers (Hucka, 1965; Matsukura and Tanaka,
2000; Matsukura and Aoki, 2004). Since the waterfall consists of two
major different rock layers, dolomite and shale, the Schmidt hammer
rebound values, RN (%), of both rocks are measured at several
locations around the waterfall. The rocks close to the waterfall are
generally fresh, but those exposed along the walls of the Great Gorge
are commonly weathered, probably by freeze-thaw processes. The
measured Schmidt hammer rebound values are summarized for each
rock type and weathering conditions in Table 4. Although the dolomite
layer along the Gorge, forming the caprock structure of the waterfall,
has been considered to be much harder than the shale layer, no
significant differences were found between dolomite and shale,
whereas differences were much clearer between fresh and weathered
conditions (Table 4). The Schmidt hammer rebound values may
include the effects of joints or anisotropy of the bedding planes in the
shale at least to some extent, because the bedding spacing is small
enough (several centimeters) to be reflected in the rebound values.
We then use average values for the fresh rocks in calculating the
unconfined compressive strength for the FR index, because rocks
under flowing water should be fresh and their strength is directly
related to the erosion of the waterfall. The Schmidt hammer rebound
values are then converted to the unconfined compressive strength Sc
(MPa) using the Schmidt hammer equipment conversion equation
(log (Sc /0.098)=0.0307 RN +1.4016).
By substituting the parameter values for the Horseshoe into Eq. (4),
the present recession rate is estimated to be 0.15 m y-1
. The past
recession rates of the Horseshoe were also computed for the following
six intervals: 1842­1875 (H1), 1875­1905 (H2), 1905­1927 (H3),
1927­1950 (H4), 1950­2000 (H5) and 2000­ (H6). These have
differing discharge and lip length values, and the recession rates
varied from 0.15 to 0.55 m y-1
(Table 5).
The present recession rate of American Falls is estimated to be
0.05 m y-1
with the regulated flow (A2, Table 5). Assuming that the
lip length, height and rock strength have not changed, the recession
rate of American Falls in the past, i.e., under the natural conditions
prevailing until a century ago (­1905), is estimated to be 0.13 m y-1
(A1, Table 5).
4. Discussion and conclusions
The estimated recession rates of the Horseshoe during the last
century (0.15­0.53 m y-1
) are of the same order as the actual
recession rates (0.1­1.2 m y-1
). The estimated recession rates of
American Falls (0.05­0.12 m y-1
) are also in good agreement with the
actual recession rates (0.01­0.1 m y-1
). Agreement is also found in
the relationships between the temporally shifting FR and the actual
recession rates of the Horseshoe and American Falls, in which all the
data for each period are plotted around the line of Eq. (4) (Fig. 3). The
Niagara River transports almost no sediment, because it is trapped
upstream in the Great Lakes. Therefore, the results obtained for
Niagara Falls support the validity of Eq. (4), which has been applied to
Table 3
Historical changes in lip length of Horseshoe Falls.
Year Lip length (m)
1819 473
1842 420
1875 501
1886 586
1890 614
1927 626
1964 723
2000 762
Table 4
Schmidt hammer rebound values (RN) averaged for each rock type and weathering
condition.
Condition
Fresh Weathered
Dolomite 53.4 (4.7) 31.7 (4.0)
Shale/mudstone 51.3 (3.4) 29.7 (1.4)
Average 52.3 30.7
(%)
Numbers in parentheses show the standard deviation.
Table 5
Summary of data for parameters giving the FR index and FR-based waterfall recession
rates, where Q is discharge, W is lip length of waterfall, H is height of waterfall, RN is
Schmidt hammer rebound value of bedrock, and Sc is unconfined compressive strength
of bedrock converted from RN.
Waterfall Duration Q
(m3
s-1
)
W
(m)
H
(m)
RN
(%)
Sc
(MPa)
FR×103
(-)
Computed
rate (m y-1
)
Horseshoe
Falls
1842­1875
(H1)
5184 420 51 52.3 99.7 0.766 0.53
1875­1905
(H2)
5184 501 51 52.3 99.7 0.643 0.47
1905­1927
(H3)
3732 614 51 52.3 99.7 0.377 0.32
1927­1950
(H4)
3110 626 51 52.3 99.7 0.309 0.27
1950­2000
(H5)
1593 723 51 52.3 99.7 0.137 0.15
2000­ (H6) 1593 762 51 52.3 99.7 0.130 0.15
American
Falls
­1905 (A1) 576 335 54 52.3 99.7 0.101 0.12
Modern (A2) 177 335 54 52.3 99.7 0.031 0.05
The modern lip length and height of the waterfalls are given by Niagara Falls State Park, NY.
Fig. 3. Relationships between the calculated FR values and actual recession rates for the
Horseshoe (solid circle) and American Falls (solid triangle). The suffix for each plot
(H1­H6 and A1­A2) corresponds to each durations shown in Tables 1 and 5.
Translucent arrows indicate time direction.
215Y.S. Hayakawa, Y. Matsukura / Geomorphology 110 (2009) 212­216
waterfalls in detachment-limited rivers with less sediment (e.g.,
Hayakawa and Matsukura, 2003a,b; Hayakawa et al., 2008b).
The present model gives the best results when changes in both water
discharge and waterfall width are incorporated into the equation. If
discharge is recovered to the past natural level, the model equation
predicts the recession rate at the Horseshoe with the current shape to
increase to 0.4 m y-1
; whereas, if the discharge had not been reduced
for the past hundred years and only the lip length changed, the present
recession rate at the Horseshoe would be 0.3 m y-1
. These expected
rates with changes in either discharge or width alone are relatively apart
from the actual recession rate. This implies that the reduction of flow
discharge due to water diversion though hydroelectric power plant is
not the only reason for the reduced recession rate at the Horseshoe; the
natural change of its shape is also important. The lip length of American
Falls, in contrast, has scarcely changed, and the reduction of discharge is
sufficient to explain the decreased recession rate of the waterfall. This
observation supports the qualitative argument by Philbrick (1970) that
the decreased Horseshoe recession rate is due to both the flow reduction
and changes in its horizontal curved shape causing stress dispersion.
Philbrick (1970) emphasizes that the presence of notches in the plan lip
form is a major factor causing instability enhancing erosion. The effect of
such plane shape is not incorporated in the FR index. A distinct notch
along the lip of the Horseshoe existed earlier (H1 to H4) but disappeared
recently (H5 and H6). This may explain the overestimation of the
recession rate for H5 and H6 and underestimation for H1 to H4 (Fig. 3).
In contrast, the data for American Falls are less apart from the line in
Fig. 3, probably because the shape of American Falls has not changed
abruptly throughout the last centuries without distinct notches nor
curved lip.
Further studies on recession rates of Niagara Falls in the Holocene
and the palaeoenvironments of the surrounding area are necessary in
relation to the long-term changes in the parameters incorporated in
the FR index as well as the other factors such as waterfall plan shape.
The mechanisms of erosion of Niagara Falls also remain uncertain, and
the classical caprock model for waterfall erosion (Gilbert, 1907)
should be reassessed because the dolomite and shale layers have
similar rock strengths, and undercutting of waterfall face is not
common to many waterfalls (Young, 1985; Lamb et al., 2007).
Acknowledgements
This research is supported in part by a Grant-In-Aid for JSPS
Fellows from the Ministry of Education, Science, Sports and Culture of
the Japanese Government (19-2271). Ellen Wohl and an anonymous
reviewer provided valuable comments on a previous version of the
manuscript. We also thank Takashi Oguchi for his help in improving
the paper.
References
Barnes, H.L., 1956. Cavitation as a geological agent. American Journal of Science 254,
493­505.
Begin, Z.B., Meyer, D.F., Schumm, S.A., 1980. Knickpoint migration due to baselevel
lowering. Journal of Waterway Port Coastal and Ocean Division, ASCE 106, 369­388.
Bishop, P., Goldrick, G., 1992. Morphology, processes and evolution of two waterfalls
near Cowra, New South Wales. Australian Geographer 23, 116­121.
Bishop, P., Hoey, T.B., Jansen, J.D., Artza, I.L., 2005. Knickpoint recession rate and
catchment area: the case of uplifted rivers in eastern Scotland. Earth Surface
Processes and Landforms 30, 767­778.
Crosby, B.T., Whipple, K.X., 2005. Knickpoint initiation and distribution within
fluvial networks: 236 waterfalls in the Waipaoa River, North Island, New Zealand.
Geomorphology 82, 16­38.
Derricourt, R.M., 1976. Retrogression rate of the Victoria Falls and the Batoka Gorge.
Nature 264, 23­25.
Ford, D.C., 1968. Waterfalls. In: Fairbridge, R.W. (Ed.), The Encyclopedia of Geomorphology.
Reinhold Book Corporation, New York, pp. 1219­1220.
Gilbert, G.K.,1907. Rate of recession of Niagara Falls. U.S. Geological Survey Bulletin 306,
1­31.
Hayakawa, Y., 2005. Reexamination of a predictive equation of waterfall recession rates
in Boso Peninsula, Chiba Prefecture, Japan. Geographical Review of Japan 78,
265­275.
Hayakawa, Y., Matsukura, Y., 2002. Recession rates of waterfalls: a brief review. Annual
Report of Institute of Geoscience, University of Tsukuba, 28, 1­4.
Hayakawa, Y., Matsukura, Y., 2003a. Recession rates of waterfalls in Boso Peninsula,
Japan, and a predictive equation. Earth Surface Processes and Landforms 28,
675­684.
Hayakawa, Y., Matsukura, Y., 2003b. Recession rates of Kegon Falls in Nikko, Tochigi
Prefecture, Japan. Journal of Geography (Tokyo) 112, 521­530 (in Japanese with
English abstract).
Hayakawa, Y.S., Oguchi, T., 2006. DEM-based identification of fluvial knickzones and its
application to Japanese mountain rivers. Geomorphology 78, 90­106.
Hayakawa, Y.S., Oguchi, T., in press. GIS analysis of fluvial knickzone distribution in
Japanese mountainwatersheds. Geomorphology. doi:10.1016/j.geomorph.2007.11.016.
Hayakawa, Y.S., Wohl, E.E., 2005. Recession rate of Poudre Falls in Rocky Mountain Front
Range, Colorado, USA. Geographical Review of Japan 78, 853­858.
Hayakawa, Y.S., Yokoyama, S., Matsukura, Y., 2005. Recession rates of waterfalls in and
upstream of the Tateno Canyon, Aso Volcano. Transactions, Japanese Geomorphological
Union 26, 439­449 (in Japanese with English abstract).
Hayakawa, Y.S., Yokoyama, S., Matsukura, Y., 2008a. Erosion rates of waterfalls in postvolcanic
fluvial systems around Aso volcano, southwestern Japan. Earth Surface
Processes and Landforms 33, 801­812.
Hayakawa, Y.S., Obanawa, H., Matsukura, Y., 2008b. Post-volcanic erosion rates of
Shomyo Falls in Tateyama, central Japan. Geografiska Annaler 90A, 65­74.
Hayakawa, Y.S., Matsuta, N., Matsukura, Y., 2009. Rapid recession of fault-scarp waterfalls:
Six-year changes following 921 Chi-Chi Earthquake in Taiwan. Transactions, Japanese
Geomorphological Union 30, 1­13.
Hucka, V.,1965. A rapid method of determining the strength of rocks in situ. International
Journal of Rock Mechanics and Mining Sciences 2, 127­134.
International Joint Commission, 1953. Report on the Remedial Works Necessary to
Preserve and Enhance the Scenic Beauty of the Niagara Falls and River. International
Joint Commision, Washington, DC.
Kukal, Z., 1990. The rate of geological processes. Earth Science Reviews 28, 1­258.
Lamb, M.P., Howard, A.D., Dietrich, W.E., Perron, J.T., 2007. Formation of amphitheaterheaded
valleys by waterfall erosion after large-scale slumping on Hawai i.
Geologcial Society of America Bulletin, 119, 805­822.
Lewis, C.F., Anderson, T.W., 1989. Oscillations of levels and cool phases of the Laurentian
Great Lakes caused by inflows from glacial Lakes Aggasiz and Barlow-Ojibway.
Journal of Paleolimnology 2, 99­146.
Matsukura, Y., Aoki, H., 2004. The Schmidt hammer: a brief review and some problems
in geomorphology. Transactions of the Japanese Geomorphological Union 25,
175­196 (in Japanese with English abstract).
Matsukura, Y., Tanaka, Y., 2000. Effect of rock hardness and moisture content on tafoni
weathering in the granite of Mount Doeng-Sung, Korea. Geografiska Annaler 82A,
59­67.
Pasternack, G.B., Ellis, C.R., Leier, K.A., Vallé, B.L., Marr, J.D., 2006. Convergent hydraulics
at horseshoe steps in bedrock rivers. Geomorphology 82, 126­145.
Philbrick, S.S., 1970. Horizontal configuration and the rate of erosion of Niagara Falls.
Geological Society of America Bulletin 81, 3723­3732.
Spencer, J.W.W., 1907. The Falls of Niagara, Their Evolution and Varying Relation to the
Great Lakes. Department of Mines and Surveys, Geological Branch, Ottawa.
Tinkler, K.,1987. Niagara Falls 1750­1845: the idea of a history and the history of an idea.
Geomorphology 1, 69­85.
Tinkler, K.J., 1993. Field guide Niagara Peninsula and Niagara Gorge. Third International
Geomorphology Conference, McMaster University, Hamilton, Ontario, Canada. 24p.
Tinkler, K.J., 1994. Entre Lacs: a postglacial peninsula physiography. In: Gayler, H.J.
(Ed.), Niagara's Changing Landscapes. Carleton University Press, Ottawa, Canada,
pp. 13­51.
Tinkler, K.J., Pengelly, J.W., Parkins, W.G., Asselin, G., 1994. Postglacial recession of
Niagara Falls in relation to the Great Lakes. Quaternary Research 42, 20­29.
Tiplin, A.H., 1988. Our romantic Niagara: a geological history of the river and the falls.
The Niagara Falls Heritage Foundation, Niagara Falls, Ontario, Canada. 216 p.
von Engeln, O.D., 1940. A particular case of Knickpunkte. Annals of American
Geographers 30, 268­271; 281­284.
Whipple, K.X., Hancock, G.S., Anderson, R.S., 2000. River incision into bedrock:
mechanics and relative efficacy of plucking, abrasion, and cavitation. Geological
Society of America Bulletin 112, 490­503.
Wohl, E.E., 1998. Bedrock channel morphology in relation to erosional processes. In:
Tinkler, K.J., Wohl, E.E. (Eds.), Rivers over Rock. American Geophysical Union,
Washington, DC, pp. 133­151.
Wohl, E.E., 2000. Mountain Rivers. American Geophysical Union, Washington, DC.
Young, R.W., 1985. Waterfalls: form and process. Zeitschrift für Geomorphologie
Supplementband 55, 81­95.
216 Y.S. Hayakawa, Y. Matsukura / Geomorphology 110 (2009) 212­216